Figuring out how to brew keptinis

Mash baking in my kitchen oven

I visited Vikonys in Lithuania and saw how the Lithuanians there brew keptinis. The basic idea is straightforward enough: do a normal mash, then bake the mash in a huge Lithuanian duonkepis oven to get caramel flavours by toasting the sugars in the mash. This is an important idea, because it’s a completely “new” type of brewing process that creates flavours you cannot make with normal techniques.

My problem was that although I’d seen it done, I didn’t know what temperature to bake the mash at, because I didn’t dare stick my thermometer into that very hot oven. It was no use asking the brewers, because they determine the temperature by the amount of wood they burn in the oven. So how to put together a recipe that other people could use? I decided the only way to do it was to brew keptinis myself at home at different temperatures to find out what worked and didn’t work.

Vytautas Jančys putting the mash into an oven so hot he’s wearing huge gloves and a plexiglass visor.

My goal was to recreate that huge, massive earthy and caramelly flavour I remembered from the keptinis I’d tasted in Vikonys.

I guess I must be the only non-Lithuanian whose first solo brew was a keptinis.


I did a little reading on caramelization and found that it’s actually chemically very complex, to the point that it’s not yet fully understood. By looking at the most common sugar types in wort and their caramelization temperatures I found I could get a tiny amount of caramelization at 110C (230F), but for most of it I’d have to get to 160-180C (320-356F).

I also learned that at too high temperatures all that lovely caramel is burnt into sooty black stuff, so I wanted to avoid that.

Simonas’s estimate of the temperature of the duonkepis oven at 300C (572F) seemed reasonable to me, but since the temperature clearly fell during the baking it seemed reasonable to keep my electric oven a little lower. I decided to try 250C (482F).

First attempt

The starting point was figuring out the batch size, because the limiting factor was our kitchen oven. Or, to be precise, our oven pans. So I took out my wife’s various oven pans and computed the total volume. 9 liters. Okay, so that would be my total mash volume. I thought malts and water would fill equal amounts, meaning I could have 4.5 liters of malts and 4.5 liters of water. So I weighed 4.5 liters of malts and found that was 2.2 kilos (4.85 lb). From that, I could work out that to follow the Lithuanian recipe my batch size would be 5.6 liters with 33 grams (1.16 oz) of hops.

The mash.

I heated water, put the malts in a kettle and added water until I had the mash at 65C (149F). Quite soon I realized having this in a kettle was dumb, because the temperature kept dropping. So I had to add more hot water periodically.

While the mashing was going on I heated the kitchen oven to 250C (482F), and once I had mashed for an hour I scooped the mash over into the oven pans. And now I found to my surprise that 4.5 liters of malts plus the 3.2 liters of water I had added in total filled … 4.5 liters. So I could have doubled the batch size and still had room for
everything. Ah well. Small batch size for the first try was probably a good idea anyway.

The mash ready in oven pans.

Then the boxes went into the oven and I settled down to watch the show.

After 10 minutes the water in the boxes started boiling, so clearly the boxes had reached 100C (212F) already. I decided 250C (482F) for three hours might be a bit over the top and reduced the temperature to 180C (356F).

25 minutes. The visible water was mostly gone, but bubbles kept popping up, so there was still boil going on.

At 45 minutes it was still boiling, and the tops started going brown. A caramelly smell was coming out of the oven, so clearly I was on the right track.

65 minutes: the top was drying out and getting browner, but still boiling. Some smoke was beginning to appear.

The first box dried out at 2 hours, with the other two still boiling. Colour had kept darkening the whole time.

Getting pretty dark and fairly dry.

At 2.5 hours only one box was still boiling and now they were
really getting dark.

After 3 hours the last one was still boiling when I took them out. I used my thermometer to check the center temperature and was astonished to get readings of 95-98C (203-208.4F) for all of them. 3 hours at 180C
and they still haven’t gone above 100C (212F)?!? What the …

The mash after baking.

Poking at the mash I found there was a hard crust on top, with soft gooey caramelly stuff right underneath (promising!). The dark layer was very thin and underneath the mass of the mash was roughly as pale as when I started.

Soft, gooey, caramelly stuff.

Anyway, I had work to do. I used a large wooden fork to break the crust on the boxes and dump the contents into the lauter tun. Then I added hop tea and hot water. Finally I ran off about 4 liters of wort. The idea was to get 5.6 liters, but my lauter tun is quite big, so I ended up losing 1-2 liters at the bottom that I couldn’t get out. An unexpected downside of the small batch size.

The wort was at OG 1.072, and I pitched a dry flake of #20 Espe that I had taken from my freezer that morning. The wort was around 30C (86F). I pitched at 19:00, but by the time I went to bed, there was still no activity, although foam started appearing during the night. It took roughly 24 hours for fermentation to really take off, so probably the viability of the yeast I’d taken out was not the best.

24 hours after that the beer was mostly fermented and I bottled it.

The mash opened up.

I waited a few days, and then it was time for the big moment: the tasting. And my keptinis tasted like… vossaøl!??! Instantly I knew what had happened. Since the temperature had never been above 100C (212F) I had basically done a 3-hour boil and got the same flavour that the brewers in Voss get from their long wort boils. It’s caramelly, sure, but not with the same depth and intensity.

The reason I didn’t get above 100C (212F) was so obvious I’m embarrassed I hadn’t thought of it beforehand: liquid water under normal conditions never goes above 100C (212F). If it does, it becomes steam and escapes. So as long as there was visible boiling there was still liquid water, which meant the temperature couldn’t be above 100C (212F).

How was that possible at 180C (356F) for 3 hours? Well, evaporation of water actually consumes a lot of energy. A water molecule evaporates when accidental collisions with other molecules give it high enough speed to escape the liquid, but all of that kinetic energy is then lost from the liquid. Other molecules banged into the one that escaped, giving their kinetic energy to the escaping molecule so that it’s knocked loose, taking all that kinetic energy with it. The sum of the kinetic energy of all the molecules is the heat of the liquid.

I concluded that the reason this attempt didn’t work out was that I had too much water, so that it took too long to evaporate it all, and that the temperature wasn’t high enough. So I would just have to try again.

Second attempt

I decided this time I would mash in an insulated coolbox, which would keep the temperature better, so I wouldn’t have to add so much water. And I would bake at 250C (482F), because although that clearly was very hot, the previous attempt showed that the darkening was only on top, anyway.

And I realized I could use my wife’s meat thermometer to constantly measure the center temperature, so that I could see what was going on all the time.

Thus prepared I tried again. This time I needed only about 2-3 liters of water to mash initially, despite having 3.3 kilos (7.28 lb) of malts (that’s what was left). I ended up adding 4 liters in the end, but that still better than last time, when I had 1.45 liters per kilo, since I now had 1.2 liters. Still, this was less of a difference than I’d hoped.

Mash in the oven, with meat thermometer.

Anyway, after an hour I scooped the boxes full of mash, and put them in the oven, which had been pre-heated to 250C (482F). Then I stuck in the meat thermometer and settled down on the floor in front of the oven.

It took 14 minutes before it started boiling at the edges, but at that point, the center temperature was still only 59C (138F). Interesting! So there was clearly quite a temperature gradient from the edge to the center.

After 30 minutes the center temperature was still only 70C (158F), but the mash started turning brown on top, and it also swelled up, overflowing the boxes. I put a larger pan underneath to catch the spill, so there wouldn’t be so much to clean up afterward. I also made a note for later: don’t fill the boxes completely — leave some room for this swelling.

After 36 minutes some burnt smell started coming off, despite the center temperature being only 75C (167F).

After 45 minutes we were at 94C (201F) and the crust on top was going darker brown.

After 51 minutes the center temperature reached 100C (212F). Now I needed to wait for the water to evaporate out so we could go above and get to the interesting caramelization temperatures.

Mash after 75 minutes.

So there I sat on the floor in front of the oven, watching the thermometer, waiting. And waiting. And waiting. And waiting. Three hours after I put in the mash, the center temperature was still stuck at 100C (212F). Or actually 101C (214F), according to the thermometer. (I don’t know why.)

I took out the boxes, and this time the crust was black as soot. But still we hadn’t gone above 100C (212F)! How on earth was I going to get this to work without burning the whole thing to cinders? The Lithuanians had gotten a beautiful brown caramelly crust. Why couldn’t I?

It’s pretty black.

Just like last time, the black crust on top turned out to be thin, with most of the mash much paler. Although it had definitely darkened a bit from where we started.

The graph below shows the temperature development.

Center temperature graph

Again the temperature remained stuck at 100C (212F) the whole time.

One interesting thing to note is that the mash went through a kind of second mashing for about 45 minutes. So fermentability should be very good, but somewhat reduced again by the caramelization.

The black layer is not so thick, though.

I did the hop tea, lautered the wort, and pitched the yeast, and started on the cleaning in between. That turned out to be a nightmare. The high temperatures had burned the sugars to a dark, hard substance
that was very difficult to get off. The non-stick pans weren’t too bad, but the older ones took hours to clean. One of them I was just barely able to clean at all, despite using so much brewer’s cleaning agent that I scraped off some of the surfaces.

Black sticky mess.

And then there was the pan I’d used to catch the overspill. Sugary wort had dripped down onto this metal heated to 250C (482F) where it had boiled furiously, eventually forming bone dry and rock hard shapes looking more like congealed lava than anything edible. This stuff was so hard I seem to have unintentionally invented a new building material. The process for producing it is perhaps not cost-effective, but you do get beer as a side-effect, so …

Sugar “art”.

I didn’t even try to clean up this pan. I told my wife I’d buy a new one and just threw it away. (I want to record my gratitude for her being understanding about this.)

Eventually, the beer was finished fermenting and I could taste it, gingerly. I was fully expecting to find that I had brewed undrinkable sooty water, but the beer was actually quite nice. It didn’t really have the caramel flavour, but an odd, toasty, minerally flavour. It wasn’t what I’d tried to brew, but not bad at all. That was reassuring, because it helped convince me that brewing this type of beer isn’t that difficult. It may be tricky to control the flavours
precisely, but you do get something interesting and drinkable even if you overdo the baking.

Third attempt

After having gone too low and then too high without ever getting the center temperature above 100C (212F), I had some thinking to do to figure out what was missing. In the end, I decided the problem might be that the duonkepis oven has an outlet for the smoke and hot air through the chimney, while our oven is not that well ventilated by default. So perhaps a lower temperature with the fan turned on, in order to vent the humidity out?

I decided to try 190C (374F) with the fan on. This time mashing and all of that went pretty quickly, since I was getting used to the drill. Into the oven it went, and:

Center temperature graph for attempts two and three

Still no luck getting above 100C (212F). (And why did the thermometer show 100C this time, but 101C (214F) last time? No idea, really.)

The inside of the oven was clearly quite wet the whole time. My glasses fogged up when I opened the oven, there was visible condensation on the lamp inside the oven and visible boiling went on for the first 1.5 hours. In fact, this time it boiled so furiously it spattered the door and walls in sugary liquid. I was glad I’d put some extra large pans underneath to catch any spill, because although the swelling mash did not flow over this time, the boil was so strong that some of the wort did.

Very dark brown mash with caramelized sugar on top.

In fact, I seemed to be getting caramelization in this sugary, sticky layer on top of the mash. Once the top layer started turning really dark brown, I stopped, figuring it would be a bad idea to burn away any caramel I might have created.

Once I broke the crust I found that I also seemed to be getting caramelization along the edges of the boxes. The cleaning was also easier this time. The boxes were sticky and gooey, but this was the kind of sugars that easily give in to hot water and not the basalt-like black stuff from the previous attempt. So this time it was looking promising.

Sticky sugary stuff

And after fermentation, I did indeed find that this attempt was my closest yet. There was a clear caramelly flavour, distinctly different from the caramel flavour of a vossaøl, although without the deep intensity of the original.

I’m beginning to suspect that another aspect of the traditional oven may be important: the brickwork, which stores the heat for a long time. The mash boxes are placed directly on the very hot bricks, and this may also contribute to the flavour. Anecdotal reports from other brewers indicate the same. So if you have a pizza oven that may be better suited for keptinis than an ordinary electric oven. And if you do have an electric oven, a pizza stone, quarry tile, or something similar to place the boxes on may also be a good idea.

However, one thing I definitely learned was that you get an interesting beer pretty much no matter what you do.

Fermenting keptinis

For all of these beers I ended up with a pretty high final gravity, but this could be because the #20 Espe I was using was not really as vigorous as it should have been. I’m not sure. I reused the same yeast from batch to batch, in order to get experience with that as well, which means that whatever it was that made these fermentations so sluggish may simply have been a problem with the frozen yeast I started with. Or it may have been something about keptinis wort. I simply don’t know.

Summing up

The actual recipe can be found at the bottom of the original blog post, but I think it’s useful to summarize the main things I learned:

  • If your partner has strong feelings about the oven pans, maybe start by discussing the project with her/him.
  • The best temperature seems to be 190C with the fan on, until the top starts heading toward black.
  • Use small pans: the reactions you’re looking for happen around the edges.
  • Try to keep the amount of water in the mash to a minimum.
  • Use a pizza stone/quarry tile if you can.
  • Leave room for the mash to swell.
  • Put something underneath the pans to catch stuff that boils
  • Stop once the top starts going really dark brown. (Although you don’t have to. Some keptinis has traditionally been baked totally black.)

And, last but not least, don’t worry too much about burning the mash. You get interesting flavours anyway.

Source & comments


Posted in beer | Tagged , , | Leave a comment

Tools to Convert WordPress Site to HTML

WordPress is one of the leading CMSs (Content Management Systems) today and powers over 30% of the websites all across the globe. It is both flexible and easy to use. This becomes simplified when you even choose WordPress Hosting for your WordPress website. However, there are times when you as a website owner would want to return to the basic HTML version of your website. There can be many reasons for wanting to switch to the static HTML version such as slower website loading speed, or security concerns. Converting WordPress site to HTML can be quite a hassle. Nevertheless, no task is impossible. All it needs is time and patience to carry it out.

In this post, we’ll cover how to convert WordPress site to HTML without losing data in simple steps.

Need for WordPress site to HTML Conversion

One of the prime reasons for converting from WordPress to a static HTML site is the page load speed. With WordPress, the pages are dynamic in nature and every time a user loads the website, it has to generate pages dynamically. However, with a static website, this problem is eliminated because there is no dynamic content. This improves the site’s speed and stability.

Another aspect is that of security. With no dynamic data storage, there is nothing that an intruder can exploit and hack into. Also, a static website eliminates the use of a database to store its content and can be loaded faster.

Tools to Migrate from WordPress to HTML

One of the most common questions I have come across is, how can one convert WordPress websites to HTML. I do agree that to convert WordPress to HTML is a tricky business. Nevertheless, plugins make the work easier.

Let us cover some tools used to convert WP to HTML.

  1. Simply Static
    Simply Static is a static site generator developed by Code of Conduct LLC and is an open-source software. The main purpose of this plugin is to create a static site from your WordPress code. At the core, what this tool does is generate HTML copies of your WordPress website pages, thus your WordPress installation remains an entirely different entity from the static site.
    convert wordpress website to html
    • Includes images, CSS, JS and any other files of which there are links
    • Replaces URLs for WordPress installation with absolute URLs, relative URLs or URLs created for offline use
  2. WP Static HTML Output
    WP Static HTML Output is a simple and powerful tool for converting a WordPress site into HTML. It is an open source software that generates an HTML copy of your entire site. It allows unlimited export targets for a single website.
    • Deploys the static website to a local folder, Amazon S3, FTP, Dropbox and Github automatically
    • It offers bilingual support (English and Japanese)
  3. Jekyll
    Jekyll is a simple, blog-aware and popular static site generator supported by GitHub. Typically Jekyll is used to convert blogs and small projects. It takes a template directory containing text files in any format, follows steps like converting & rendering, and generates static HTML websites. To learn how to build a static website using Jekyll, check out this post.
    • A Jekyll website is a collection of files in the form of a directory. Also, since Github pages are deeply integrated with Jekyll, you can store your entire website in a Git repository
    • Since Jekyll is a large community, they have numerous documentation and installation guides. This makes it easy to import content from WordPress and convert to HTML
  4. Pelican
    Pelican is another popular static site generator that allows users to write content in Markdown, reStructuredText or AsciiDoc format and import from WordPress, Dotclear or RSS feeds. Also, it has several tools that help with the generation of sitemap files.
    Pelican helps convert wordpress website to html
    • Supports features like Atom and RSS feed, social media integration, commenting tools like Disqus, Google Analytics and code syntax highlighting
    • Helps in customizing themes using Jinja templates and publishing the content in several languages

Selecting a tool to convert WordPress code to HTML is purely based on the tool that you find simple and easy to work with. I hope this post helped you understand how to convert WordPress site to HTML. At the end, the main motive of converting WP to HTML is to protect it from potential hackers and increase the speed and performance of your website.

Source: recellerClub

Posted in Dev | Tagged , | Leave a comment

Lithuanian Beer Yeast


Borne from an exclusive partnership with the famed Jovaru® Brewery’s “queen of Lithuanian farmhouse beer,” this unique yeast complements farmhouse beers with citrusy esters and restrained phenols. The strain produces a character of lemon pith, black pepper, and a soft mouthfeel. Į sveikata!

This strain tests positive for the STA1 gene, an indicator of Saccharomyces cerevisiae var. diastaticus. This strain may have the ability to metabolize dextrins over time, resulting in higher than expected attenuation.

JOVARU® is an ALDONA UDRIENĖ® yeast strain that originates from the famous JOVARU® Brewery that has been in operation since the mid-19th century.

Temperature: 70-95 °F (21-35 °C) || Attenuation: 80-85%

Flocculation: Medium-Low || Alcohol Tolerance: 10% ABV


This culture is a single strain of Saccharomyces cerevisiae, isolated from a Lithuanian Farmhouse mixed culture kindly provided by Julius Simonaitis via Lars Garshol. Across a wide temperature range, this culture will throw a potent mix of orange, tropical fruit and stone fruit esters that is reminiscent of POG Juice® (passionfruit, orange, guava) accompanied by distinct spicy, earthy and herbal undertones.

Temperature: 75-95 °F (24-35 °C) || Attenuation: 76-82%

Flocculation: High || Alcohol Tolerance: High


Pakruojis Lithuanian Farmhouse is a single strain of STA1+ Saccharomyces cerevisiae, isolated from a Lithuanian brewery. This yeast produces beer with a dry, crisp and silky mouthfeel, an ester profile of complex citrus fruit, and a balanced rustic earthiness with undertones of white peppercorn. This strain exhibits high diastatic activity.

Temperature: 75-95 °F (27-35 °C) || Attenuation: 90-100%

Flocculation: Low || Alcohol Tolerance: High

Posted in beer | Tagged , , | Leave a comment

Notepad++: A guide to using regular expressions and extended search mode

The information in this post details how to clean up DMDX .zil files, allowing for easy importing into Excel. However, the explanations following each Find/Replace term will benefit anyone looking to understand how to use Notepad++ extended search mode and regular expressions.

If you are specifically looking for multiline regular expressions, look at this post.

You may already know that I am a big fan of Notepad++. Apparently, a lot of other people are interested in Notepad++ too. My introductory post on Notepad++ is the most popular post on my speechblog. I have a feeling that that is about to change.

Since the release of version 4.9, the Notepad++ Find and Replace commands have been updated. There is now a new Extended search mode that allows you to search for tabs(\t), newline(\r\n), and a character by its value (\o, \x, \b, \d, \t, \n, \r and \\). Unfortunately, the Notepad++ documentation is lacking in its description of these new capabilities. I found Anjesh Tuladhar’s excellent slides on regular expressions in Notepad++ useful. After six hours of trial and error, I managed to bend Notepad++ to my will. And so I decided to post what I think is the most detailed step-by-step guide to Search and Replace in Notepad++, and certainly the most detailed guide to cleaning up DMDX .zil output files on the internet.

What’s so good about Extended search mode?

One of the major disadvantages of using regular expressions in Notepad++ was that it did not handle the newline character well—especially in Replace. Now, we can use Extended search mode to make up for this shortcoming. Together, Extended and Regular Expression search modes give you the power to search, replace and reorder your text in ways that were not previously possible in Notepad++.

Search modes in the Find/Replace interface

In the Find (Ctrl+F) and Replace (Ctrl+H) dialogs, the three available search modes are specified in the bottom right corner. To use a search mode, click on the radio button before clicking the Find Next or Replace buttons.

Cleaning up a DMDX .zil file

DMDX allows you to run experiments where the user responds by using the mouse or some other input device. Depending on the number of choices/responses (and of course the kind of task), DMDX will output a .zil file containing the results (instead of the traditional .azk file). This is specified in the header along with the various response options available to the participant. For some reason, DMDX outputs the reaction time twice—and on separate lines—in .zil files. Here’s a guide for cleaning up these messy .zil files with Notepad++. Explanations of the Notepad++ search terms are provided in bullet points at the end of each step.

Step 1: Backup your original result file (e.g. yourexperiment.zil) and create a copy of that file (yourexperiment_copy.zil) that we will edit and clean up.

Step 2: Open yourexperiment_copy.zil in Notepad++ (version 4.9 or later).

Step 3: Remove all error messages.All lines containing DMDX error messages begin with an exclamation mark. Let’s get rid of them.

Bring up the Replace dialog box (Ctrl+H) and select the Regular Expression search mode.

Find what: [!].*

Replace with: (leave this blank)

Press Replace All. All the error messages are gone.

    • [!] finds the exclamation character.
  • .* selects the rest of the line.

Step 4: Get rid of all these blank lines.

Switch to Extended search mode in the Replace dialog.

Find what: \r\n\r\n

Replace with: (leave this blank)

Press Replace All. All the blank lines are gone.

    • \r\n is a newline character (in Windows).
    • \r\n\r\n finds two newline characters (what you get from pressing Enter twice).

Step 5: Put each Item (DMDXspeak for trial) on a new line.

Switch to Regular Expression search mode.

Find what: (\+.*)(Item)

Replace with: \1\r\n\2

Press Replace All. “Item”s have been placed on new lines.

    • \+ finds the + character.
    • .* selects the text after the + up until the word “Item”.
    • Item finds the string “Item”.
    • () allow us to access whatever is inside the parentheses. The first set of parentheses may be accessed with \1 and the second set with \2.
  • \1\r\n\2 will take + and whatever text comes after it, will then add a new line, and place the string “Item” on the new line.

So far so good. Our aim now is to delete duplicate or redundant information (reaction time data).

Step 6: Remove all newline characters using Extended search mode, replacing them with a unique string of text that we will use as a signpost for redundant data later in RegEx. Choose a string of text that does not appear in you .zil file—I have chosen mork.

Switch to Extended search mode in the Replace dialog.

Find what: \r\n

Replace with: mork

Press Replace All. All the newline characters are gone. Your entire DMDX .zil file is now one very long line of (in my case word-wrapped) text.

Step 7: We’re nearly there. Using our mork signpost keyword, let’s separate the different RT values.

Stay in Extended search mode.

Find what: ,

Replace with: ,mork

Press Replace All. Now, mork appears after every comma.

Step 8: Let’s put the remaining Items on new lines.

Switch to and stay in Regular Expression search mode for the remaining steps.

Find what: mork(Item)

Replace with: \r\n\1

Press Replace All. All “Item”s should now be on new lines.

Step 9: Let’s get rid of those duplicate RTs.

Find what: mork ([^A-Za-z]*)mork [^A-Za-z]*\,mork

Replace with: \1,

Press Replace All. Duplicate reaction times are gone. It’s starting to look like a result file 🙂

    • A-Z finds all letters of the alphabet in upper case.
    • a-z finds all lower case letters.
    • A-Za-z will find all alphabetic characters.
    • [^…] is the inverse. So, if we put these three together: [^A-Za-z] finds any character except an alphabetic character.
    • Notice that only one of the [^A-Za-z] is in parentheses (). This is recalled by \1 in the Replace with field. The characters outside of the parentheses are discarded.

Step 10: Let’s get rid of all those morks.

Find what: mork

Replace with: (leave blank)

Press Replace All. The morks are gone.

Step 11: Separate each participant’s data from the next.

Find what: (\**\*)

Replace with: \r\n\r\n\1\r\n\r\n

Press Replace All. The final product is a beautiful, comma-delimited .zil result file that is ready to be imported into Excel for further analysis.

Notepad++, is there anything it can’t do?

Please post your questions in the comments below, rather than emailing me. This way, others can refer to my answers here, saving me many hours of responding to similar emails over and over.

Update 20/2/2009: Having trouble understanding regexp? I have created a new Guide for regular expressions. Check it out.

Source and comments

Posted in Dev | Leave a comment

What is the ideal HERMS coil length? Theory and experiments


When we started selling HERMS coils 2 years ago, we made our 40cm coil 15 meters (50 feet) long. The 30cm coil was 11.5m long. Why? That’s what most home brewers were doing.

But we recently started to suspect that such a long coil is not optimal. Sure, a longer coil will give you a higher output temperature, but it will also have more friction, reducing your flow. Each extra loop of the coil will cause a small temperature rise, but with diminishing returns for each extra loop. But each extra loop will also add a additional amount of friction.
The amount of heat transferred to the mash tun depends on both the achieved input-to-output temperature difference and the flow rate, so there is a length that results in the maximum heat transfer. We set up an experiment to estimate what this length would be.

Below, you see a photo of the HLT we used for testing. It has a 40cm HERMS coil, made of stainless steel tubing (0.5mm wall thickness and 11mm inner diameter). The element is a 3500W element. Elco dry fired it once, which is why it turned black. It still works fine though. On the bottom right, you can see the temperature sensor and the dry fire protection which was installed after ruining the element.


Based on our experiments, an 8 meter coil maximizes the heat transfer. This of course depends on many factors, like the coil material, your pump pressure and other losses in the system. We also found some other simple optimizations you can apply to your brewing setup and came up with a complete new coil design for big kettle setups.

HLT with BrewPi HERMS coil

Hot Liquor Tun setup with whirlpool, heating element and HERMS coil.

Heat transfer as function of flow and temperature difference

The amount of heat that is transferred every second (power in Watts) can be calculated by multiplying the volume that is transferred, the temperature difference between coil input and output and the specific heat capacity of water. In formula form:

\[ P = V \cdot \Delta T \cdot c_p\]

So to heat up your mash as quickly as possible, you want to maximize this equation.

Experiment setup

To find the ideal length, we needed to determine 2 things:

1. How much does each loop of the coil contribute to temperature rise?

It costs a lot of work and money to make multiple coils of different lengths, so we came up with an alternative: we decreased the amount of water in the HLT, submerging only a fraction of the coil. We assumed that only the submerged loops are contributing to the output temperature significantly.

Timelapse of HLT water levels

Hot Liquor Tun water levels for each successive test.

2. How much friction does each loop add?

This one is a bit harder to test, so we took a more theoretical approach. We measured the flow rate twice, once bypassing the entire coil and once pumping through the coil. From the difference in flow rate, we calculated the friction factor of the entire coil. We used that value to calculate the friction that shorter coils would have.

The experiments

Flow rate

We measured the flow simply by filling a bucket with the return tube of the HLT, timing it and weighing the amount of water. The results were:

  • 5.8 liters per minute with the coil
  • 9.8 liters per minute without the coil.

Switching between with and without coil could be done by simply toggling a valve. Because it is fully automated Elco’s system has a lot of tubing, meaning there will be a lot more friction. In simpler setups with less tubing you will get higher flow. We correct for that later.

Flow calculations

There’s a lot of math in this section, feel free to skip it. What we did is measure the flow with and without the coil and from the difference we calculated the friction of the coil.


To calculate  (flow velocity in meters per second) we used the Bernoulli formula.

Bernoulli formula:

\[(p_{1} + \frac{1}{2}\rho v_{1}^{2} + \rho g z_{1})-(p_{2} + \frac{1}{2}\rho v_{2}^{2} + \rho g z_{2})=f \frac{1}{2} \frac{L}{d} \rho v^{2} + K \frac{1}{2} \rho v^{2}\]

Without the coil, this formula simplifies to:

\[p_{pump}= K \frac{1}{2} \rho v^{2}\]

This allows us to calculate K, our dimensionless loss factor for everything except the coil.

\[K = \frac{2 \cdot p_{pump}}{\rho v^{2}}\]
\Big[ \frac{kg}{m^{3}} \Big]

Our pump pressure is known from manufacturer (4.5m head), and the density of water  at 60°C is 980 


Now that we have a good estimation for the loss factor K, the friction factor  of the the coil can be determined from the difference in flow velocity. The pressure of the pump does not change, so we can solve the following formula.

\[p_{pump} = f \frac{1}{2} \frac{L}{d} \rho v_{1}^{2} + K \frac{1}{2} \rho v_{1}^{2} = K \frac{1}{2} \rho v_{2}^{2}\]
\[f=K \cdot \frac{d}{L} \frac{v_{2}^{2}-v_{1}^{2}}{v_{1}^{2}}\]

Now that the friction factor of the coil is calculated it can be used to estimate the speed of the wort for different coil lengths.

\[v = \sqrt{\frac{2 \cdot p_{pump}}{\rho (f \frac{L}{d}+K)}}\]

We got a resulting friction factor of 0.0407, giving the following flow curves for a K of 30.

Pump flow and speed for Elco's system

Volume flow and velocity for Elco’s system for various coil lengths

Elco’s system has a lot of tubing and valves, because it is fully automated. You can see the schematic of the entire system here:

Elco’s brewery with automated valves, interactive demo

To mimic our experiment in this setup, select the mash recirculation step. By toggling valve v6 we could bypass the coil.

And here are some photos of the setup:

Temperature sensors on coil

With the temperature sensors inline with the coil, we could measure the coil in and output temperature.

Elco's setup water pump cluster

This cluster pumps water from the HLT or water input to the HLT (top valve v1) or router cluster (bottom valve v2)

Elco's setup wort pump cluster

This cluster gets water from the MT (top left v9) or BK (bottom left v10) and pumps it to the router cluster via the CFC (bottom right v12) and bypassing the CFC (top right v11) or drain (middle v13)

Elco's setup router cluster

This router cluster gets clean water from the left or wort from the right. The connections at the bottom are MT in (v3) coil out (v5) coil in (v7) or BK in (v8). By toggling v6, in between v5 and v7, we could bypass the HERMS coil.

This shows the water input on the right, with a valve for the water pump cluster and the CFC and a ‘to fermenter’ output on the left.

As you can see from the photos above, Elco’s setup has a lot of tubing and valves. This results in a lot of friction in the system, and therefore a high K factor. We think a K of 15 will be more representative of a typical homebrew setup.

A loss factor K of 15 will result in the following flow charts.

Pump flow curve using f value of 0.0407 and K value of 15

Volume flow and velocity for a more typical system with less tubing and valves

Coil efficiency vs. length

To measure the temperature distribution through the coil, we measured the output temperature of the coil for different water levels in the HLT. We measured:

  • 1.5 loops submerged
  • 3.5 loops submerged
  • 5.5 loops submerged
  • 7.5 loops submerged
  • 9.5 loops submerged
  • 11.5 loops submerged

For each water level in the HLT we recirculated the mash tun water through the HERMS coil to go from 55°C to 65°C, while trying to keep the HLT at 70 degrees. We measured:

  • Coil in temperature
  • Coil out temperature
  • Mash tun in temperature
  • Mash tun out temperature
  • HLT temperature

Here you can see the measurements form having 11.5 loops submerged:

brewpi data test 2 using 11.5 meter submerged coil

Raw data as recorded by the BrewPi brewing controller

For each measurement we started the HLT at 70 degrees and the MT at 55. We then closed valve v6 to switch from bypassing the coil to pumping through the coil.
After each run, we would discard some water from the MT, add some fresh cold water and make sure it was back at 55 degrees with the same volume for the next test.


We measured the input temperature difference (HLT – coil in) and the output temperature difference (coil out – coil in), so for each point in time we can calculate the efficiency  of the submerged section:

\[\eta = 1 - \frac{\Delta T_{out}}{\Delta T_{in}}= 1- \frac{coil_{out}-coil_{in}}{HLT - coil_{in}}\]

Each experiment would give us a set of points for the efficiency for a certain coil length. You can see these as a set of dots in the chart below. We only used the points for which the input temperature difference was between 4 and 8 degrees, which is the typical use case.

efficiency plot with temperature difference legend

Efficiency of a HERMS coil for different coil lengths

\eta = \eta_{max} \cdot (1 - e^\frac{-t}{\tau})

We fitted a curve through these points, with the formula  where  is a time constant based on the lumped thermal capacity model. Our fitted curve has a time constant of 4.5,  of 0.79, and  is the time spent in the submerged part of the coil, which can be calculated form the flow velocity.


Now we can use this formula to calculate the coil output temperature for different lengths by calculating the time spent in coil () first and entering it in this formula.

Heat transfer analysis based on our experiments

Now that we have determined how coil length changes flow rate and efficiency, we can use the formulas to calculate how much heat will be transferred for each coil length and temperature difference. As you can see in the chart below, the curves are as expected: making the coil longer increases the efficiency, but after a certain point the loss in flow rate cancels out the higher output temperature.

Power curve with respect to coil length

Power transferred to mash with different temperature differences between HLT and mash temperature

We have found 8.2 meters (dotted line) to be the most efficient length. The graph above is calculated with K = 15.

From the graph, you can also see that if your temperature difference is small, the power with which you heat your mash tun is low.

Recommendations for your own HERMS setup

During our experiments, we also found some quick improvements you can use to make your own mash go faster.

Exit the coil at the bottom

We did our experiments twice, once with the wort entering the coil at the top and leaving on the bottom and once the other way around. It turns out that having the wort in the HLT coil flow from top to bottom is more efficient that from bottom to top. This is because the water in the HLT is hotter closer to the element, even when there’s a whirlpool in the HLT. Our experiments found a 12% increase in efficiency when wort flowed from top to bottom.

coil efficiency comparison flow direction HERMS coil

Two efficiency curves with changing flow direction within HERMS coil

Keep the temperature difference high

V \cdot \Delta T \cdot C

Because the heat transferred per second is , to reach your mash temperature quickly, you will want to keep the temperature difference high. However, you don’t want to overshoot your target temperature. If you have a lot of water in the HLT, your HLT will continue to heat your mash if you keep pumping, even after you stop heating your HLT. All the heat that you put into your HLT will eventually distribute between your HLT and MT. A small example:

Let’s say you have 40L in your HLT and 40L in your mash tun. Your target temperature is 65 degrees and your mash is currently at 63 degrees. How high can you set your HLT to not overshoot? If you assume there are no temperature losses, the answer is 67 degrees. The equilibrium situation is that the HLT loses 2 degrees to the MT and both end up at 65:

\[ \frac{63 \cdot 40 + 67 \cdot 40}{40 + 40} = 65\]

Now let’s take another look with only 10L in the HLT, still 40L in the mash tun. Now you can heat up the HLT to 73:

\[ \frac{63 \cdot 40 + 73 \cdot 10}{40 + 10} = 65\]
\frac{\textrm{HLT volume}}{\textrm{mash volume}}

With a 4 times smaller HLT, you can heat your mash 4 times quicker without overshooting. If you use BrewPi for mashing, you would set your Kp for the PID that manages the HLT setpoint to the  to automatically manage the HLT setpoint.

So our recommendation is:

  • You can use a big kettle for your HLT, but don’t fill it up fully until you are heating your sparge water.
  • Mount your HLT coil as close to the bottom as possible
  • Use a low and wide HERMS coil if possible.

Whirlpool or stir your HLT

We measured the temperature response with and without recirculating our HLT water. This makes a big difference in the temperature exchange, as you can see from the graph below. Without stirring your HLT, the water around the coil will be a bit cooler, decreasing the output temperature.

Whirlpool vs no whirlpool in the HLT

Whirlpool vs. no whirlpool in the HLT

Our new HERMS coils

We changed the HERMS coils we sell based on these experiments.

Our 30cm HERMS coil has been shorted from 12 windings to 8 windings. It is now 7.5m long instead of 11.3m.

Our new 30cm HERMS coil

We are going to discontinue our 40cm HERMS coil, which was 15m long. The longer coil does not help with heat transfer. We have come up with an alternative though!
For big kettles, we have made a prototype of an entirely new design, suitable for big kettles and high power elements.

To maximize the heat transfer, we want to:

  • Maximize the time the wort is in the coil, to increase the efficiency
  • Reduce the friction of the coil

The only way to do both at once, is to increase the diameter of the coil. We chose to do this by having 3 coils in parallel, ending in one 1 inch fitting.

This design will have:

  • A lower velocity and therefore less friction, because the flow velocity per coil is lower and the coil is shorter
  • A lot of surface area
  • A very low height, making it possible to have only very little water in the HLT to be able to work with bigger temperature differences.

The 30 cm HERMS coils are back in stock now. For the new 3-way parallel super awesome high flow coil, we only made a few prototypes. They’ll be available in about 4 weeks.

Source and comments

Posted in beer | Leave a comment

Can you convert html, css and javascript into a exe?

  • Electron, by GitHub. Lets you build cross-platform Windows, MacOS, and Ubuntu applications from HTML/JavaScript/CSS. It works by embedding them into a Chromium build alongside with Node.js. Many applications such as Slack and Visual Studio code are built with Electron.
  • Cordova, by the Apache Foundation. Targets Android, iOS, and Windows from HTML/JavaScript/CSS. It works by embedding them into a WebView (or whatever is the equivalent on the target platform).
  • Proton Native. Works with React Native components and does not use Electron (see above), which makes generated apps therefore more “lightweight” than the ones generated by Electron. Uses libui under the hood.
  • Yes, there is a program HTML Compiler that allows you to convert an entire HTML application (using CSS, JavaScript, Images and more) into a standalone Windows application.
  • There is also ExeOutput for PHP that lets you create applications in native format for Windows with PHP, JavaScript and HTML. Convert PHP scripts, websites, JavaScript, HTML, databases into stand-alone apps for Windows (single EXE files) that do not require a Web server nor PHP distribution.

For window apps

If you need something to present you a window environment and a webserver emulation, choose ZZEE or ExeOutput, these are decent tools, maybe ZZEE is a bit outdated regarding the PHP version. So for “HTML, CSS and Javascript”, the answer is clearly something like this.

For console apps

If you only need a tool to run in the background or from command line, here’s my pet project called RapidEXE – it’s quite simple to use, somewhat extendable and features PHP 7 out of the box. Also, it’s free and always will be. But since there’s no browser involved, technically this is not solution for HTML/CSS/JS; just something very similar, like using web-related languages to create EXE files.

A real compiler: JSC

If you want (something-like-)javascript to be compiled to a standalone EXE, give a chance to JSC, it’s there in your Windows, you just probably never met before. JSC gives you extremely small executables (because of the .NET already built into Windows itself) and they have full power of .NET so you can do pretty much anything you want. Flipside: apart from the syntax being familiar, not a lot will remind you of Javascript, you have to learn a lot of new practices to get along.

A bonus track: “chrome –app” and RAR

If you can live with a dependency such as Chrome, you can do a nasty but working solution: RAR can create a self-extracting EXE for you and run a command after extraction (!!), so you can pack up a little micro-universe of HTML/CSS/JS files; then you can call Chrome with “–app=something.html” that will give you a plain window with a browser environment and your HTML file in it. I know this is not a compiler; however, it’s a way to create an EXE that is self-contained and gives the user a web-liek experience in an application-like window. Sure, Javascript will be a bit limited because you’re running in a local file, not a proper website, but it’s not that painful. One lovely thing about this technique is that Chrome will remember the position and size of your window so it will pretty much feel like an application.

Source and comments

Posted in Dev | Leave a comment

How to Prevent DMS in Beer

A  Look at Studies and Experimental Brew with Tested Results

The past few weeks I have casually been combing through over 20 academic studies on DMS, a topic in brewing that has benefited from a surprisingly vast amount of research and experimentation. Of most interest to me, other than understanding the various aspects of the brewing process that can prevent or lead to DMS in beer, is to determine if DMS is even something brewers need to worry about anymore. If so, how specifically can DMS shape or alter hop aromatics and flavor. The move to higher malt modification has generally created malts that are less susceptible to creating DMS and likely a reason for DMS weighing less on brewers minds. However, as one of the studies detailed below shows, in tested commercial example of ales, DMS was found in amounts above taste thresholds and in comparable amounts as the lagers tested. Could it be that DMS even at low but detectable levels is playing a greater role in shaping the overall flavor profile of beer?

In this post, I first outline the relevant tested information on DMS creation and prevention throughout the brewing process outlined in the academic literature reviewed.  I then applied the information to brew an experimental beer doing nearly everything wrong to see if any DMS was actually created. After all, if the beer ends up being fine, despite going against traditional anti-DMS protocol, then this would all be a non-issue anyways. To go further than just taste the poorly brewed beer, I sent three samples to Oregon State University where levels of DMS were tested. The results of the test and tasting notes, including the effect DMS had on hop aroma and flavor are outlined below the research.

What is DMS?

Dimethyl sulfide (DMS) is a sulfur compound that is typically considered an off-flavor in beer at high concentrations and is introduced into beer from the thermal decomposition (wort heating) of S-methylmethionine (SMM) produced in the embryo of barley during germination.1 SMM finds its way into wort because it is readily dissolved from malt at all mashing temperatures. During fermentation, yeast can also reduce another precursor called dimethyl sulfoxide (DMSO), that can turn into DMS in the finished beer. This reaction is by a yeast enzyme called DMSO reductase.2 The extent of this conversion of DMSO is dependent upon the particular year strain employed.3DMS can also arise from wort infected with spoilage organisms.4


DMS is typically described as having an aroma of cooked or creamed corn; however, the literature goes further with descriptors such as sauerkraut-like, black olive, canned corn, cabbage, rotten onions, and blackcurrant buds. DMS is a major contributor to the aroma other beverages and foods including many vegetables like asparagus where DMS is the “dominating character impact compound” and is generally described as having odor sensations of germ-like, mushroom-like, and nutty. 5 White wines, milk, rum, and numerous cooked vegetables are also documented for having a DMS aroma and flavor component.6

Prevalence in Commercial Beers

Most brewers probably don’t think their beer (especially ales) contains detectable levels of DMS, however, an interesting study of thirteen commercially available beers were tested with surprising results. Seven domestic ales, two domestic lagers, two imported lagers, and two non-alcoholic lagers were among the beers tested and found that DMS was at or above flavor thresholds (30-60 µg/L)7 in all but one of the beers, which shows that DMS was contributing to the flavor profile of the beers. Surprisingly, DMS levels in the ales were comparable with the lagers, which are known for producing higher amounts.8 The flavor threshold for DMS is relatively low at 30 µg/L, so it doesn’t take much to contribute to the flavor of a beer.9


Most of the DMS in beer produced from the SMM precursor is lost (evaporated) during the boiling process. Specifically, it was determined that the half-life of SMM at boiling point is 37 minutes, which means that half of the SMM is evaporated from the boil pot at 37 minutes. A boil of 90 minutes removes approximately three quarters of the DMS and most of the SMM is lost during a 120 minute boil. This half-life of SMM is increased when the temperature is decreased. For example, if wort is heated but not boiled (like a no-boil berliner weisse) to around 190°F, it would have to be held at this temperature for approximately 130 minutes to evaporate half of the SMM precursor.10 The physical location of the brewery can alter the evaporation of SMM because wort boils at lower temperatures at increased elevations, which would result in much less conversion of SMM to DMS in the boil.11

The vigor of a boil also contributes to final DMS levels in beer. One study looked at the power input of a boil and the impact on DMS. The authors found that the when the power input used for boiling increased, the levels of DMS decreased (in a DMS water solution). During a 60 minute boil almost no DMS was detected at 1,500 watts where at 1,000 watts approximately 175 ppb of DMS was detected. At 500 watts, about 300 ppb of DMS was detected. 12 A small electric stove uses about 1,500 watts per hour while it heats on medium or high.13


During hot stands where the wort remains at hot temperatures after the boil (now typically called hopstands) any remaining SMM in the wort that wasn’t evaporated during the boil will continue to hydrolyse and more free DMS will be formed. One study extended a hot wort stand by one hour and found higher DMS concentrations from an average of around 49 µg/L to 104 µg/L in the extended hot stands.14 This means if a short boil time is combined with an extended hop stand, more DMS in the final beer could be formed. Or in the commercial realm, if a long whirlpool is conducted in a well-insulated vessel or long transfer times occur at hot temperatures, this could lead to a continued breakdown of SMM that survived the boil.

Foam During Boil

One of the most recent studies on DMS looked at the role of beer foam during boiling and the impact of the vaporization of DMS from wort. The authors found that the vaporization of DMS was strongly enhanced in the presence of beer foam versus trials where an antifoam agent was used. Although the direct cause of this was not the point of the study, they speculated that one explanation could be that DMS might actually concentrate in the foam and is then stripped from the wort by the rising vapor of the boil.15

Role of Carbon Dioxide

During fermentation, measured levels of DMS are purged via C02 escaping. Specifically, one study found that there is a significant decrease in DMS levels over the course of the first five days or so of vigorous fermentation, with days 1-2 seeing the most dramatic decreases. However, after day five there was a measured increase in DMS in the fermenting beer, which was produced by the yeast from the DMSO precursor. 16 The Miracle study found an increase in overall sulfur volatiles by the yeast after fermentation was completed with just a short contact time with yeast present, finding that prolonged post-fermentation conditioning in primary can lead to more sulfur-containing volatiles. Specifically, DMS increased to 43.4 μg/L with 30-day residence in a keg vs. 32.6 μg/L with only 14-day residence in the keg.

Why would DMS be created after primary fermentation is complete? Gibson found that when nitrogen is present in the wort in excess a constant and low level of DMSO reductase is detectable. Most typical worts have enough assimilable nitrogen for the yeast, however after primary fermentation and attenuation has been achieved the nitrogen levels in the yeast cells become depleted, which can lead to an increase of DMSO and thus DMS.17 To avoid DMS creation after primary fermentation is over, I’m curious if adding a very small amount of yeast nutrient containing nitrogen around day five would be beneficial.

Open/Closed Fermentation

One study looked at the difference of DMS in beers fermented in an open fermentation system (fermenter with a partially opened top) vs. an enclosed conical vessel. They found that DMS in the open topped fermenter contained the amount of DMS anticipated from the potential of the wort, however, the beer fermented in the conical vessel contained more DMS than could be accounted for from the worts potential. The difference was stark with an ale fermentation resulting in only 9 µg/ in the open fermenter and the conical resulted in 57 µg/ at the time of racking.18 Homebrewers could easily mimic this open fermentation by covering the top of the carboy with sanitized tin foil for the first five days or so of vigorous fermentation then replacing the foil with an airlock for the remainder of time in primary.

Fermentation Temperature

A 1980 Journal of the Institute of Brewing looked at both original gravities of wort and the temperature of fermentation and the effects on DMS in beer. The authors found that worts of high gravities create substantially more DMS than worts of low gravities. Specifically, fermentations of a 1.060 wort had over three times more DMS than a 1.033 wort (adjusted with either glucose or fructose to achieve the higher gravity). With respect to fermentation temperature, it was found that as the temperature increased the level of DMS decreased. The maximum DMS level achieved with a ferment at 46°F was five times greater than a ferment conducted at 77°F.19 This is because DMS production by yeast is greater at lower temperatures.

Gravity of Wort

High gravity worts that are increased in fermentables by the use of sucrose (table sugar) may actually see lower levels of DMS in the finished beer. This is because sucrose can lead to increased DMS volatility. A recent study found that a solution consisting of 10% sucrose (in water) saw a remarkable increase in DMS evaporation.20 However, worts are low in sucrose with typical fermentable sugars profiles of maltose (42-47%), maltotriose (11-13%), hexose (7-9%), and sucrose (2-3%).

Unmalted Cereals 

Because barley malt is the main source of SMM, the DMS precursor, the less used in the grist the lower potential for DMS in the finished beer. One way of using less barley is by substituting some of the grist with unmalted cereals, which can also contain DMS precursors, but likely will boil out completely as corn has been shown to do.21 Interestingly, cereals with high gelatinization temperatures, which require boiling prior to usage, may actually have less DMS precursors because this processes breaks down and evaporates DMS, leading to less total DMS precursors in the boil compared to an all malt grist.22

Yeast Storage/Health

The storage of harvested yeast may also play a role in the levels of DMS in beer. A 1985 study examined the DMS levels of beer with yeast stored for 5 days at 34°F under beer (like most homebrewers save yeast) and one stored in water and agitated twice a day and another stored under water receiving constant agitation (via stir bar). Levels of DMS were lower in beers stored unagitated under beer than with agitation. Specifically, the agitated yeast had 25% more DMS in the finished beer which was likely caused by yeast damaged (oxygen toxicity) that may have occurred from stirring resulting in a less vigorous ferment stemming from less healthy yeast pitched.23 This finding is a good reminder to ensure pitching an appropriate amount of healthy yeast, because a healthy vigorous fermentation can aide in controlling DMS Levels.


DMS can also be introduced into beer through hops. At the end of a typical 60 minute boil all of the sulphur volatiles have evaporated, however during late hoping and whirlpool additions three sulfides peak, including DMS, which is then detectable in the wort. Some of this DMS will be reduced during fermentation via C02, but some could survive and stay in the beer until packaging.24

pH Levels

It was found that worts with a high starting pH can lead to greater amounts of DMS in the final beer. The following chart shows test results of wort fermented with with S. cerevisiae NCYC240 at different starting pH levels.25

Initial Wort pHFinal Beer pHFinal DMS (µg/L)

Experimental Brew & Results

I decided to brew a NEIPA to test out the research and to see how potential DMS could influence hop aromatics and flavors. Based on the literature, I took the following actions to potentially encourage DMS production:

I aimed for a mash pH over 5.4 which is slightly higher than most typically try to achieve in pale hoppy beers. I tend to like NEIPAs finishing with a slightly higher pH anyways because it seems to smooth out the finish of the beer (less of a bite) by also requiring less direct bittering by hops because beer can actually taste more bitter as the pH increases even when the iso-alpha-content remains the same.26 To help position myself for a higher finishing pH I generally aim for a higher mash pH of around 5.4-5.5, which according to the pH study above, could actually be slightly increase the final DMS content in the beer, which seems to be strain dependent.

At boiling temperatures the half-life of SMM is 37 minutes, so it seemed like a bad idea (or in this case a good idea) to heat the wort up 212℉ and immediately cut the heat. I proceeded to scoop out any foam that was created from the top of the wort as it heated to discourage DMS evaporation (again based on the study above). Because the beer didn’t reach a roiling boil, the foam creation was considerably less than normal, so there wasn’t much to remove. Rather than chilling the batch at this point, I opted to do a 60 minute hopstand, which should give the SMM remaining in the wort from not being boiled out plenty of time to hydrolyse into DMS.

On day five of fermentation, I dry hopped the beer and then capped the fermenter by sealing the 10 gallon keg the wort was in. Learning that after primary fermentation is largely over the nitrogen levels in the yeast cells can become depleted leading an increase in DMSO by the yeast, should this occur, the sealed keg should trap any DMS trying to escape via C02 (as well as potentially retain the dry hop aromatics).

Click to Expand Recipe Details
DMS Results

As you can see from the results above, the testing showed a total of 177.2 µg/L of DMS. To put the results into context, ales will generally have less than 30 µg/L, but below 100 µg/L, and anything above 100µg/L is when the flavors associated with DMS start to really come through. When I transferred the beer from the fermenting keg (with C02) to the serving keg with the hops, the aroma coming out of the receiving keg lid immediately reminded me of a hoppy no-boil berliner weisse I made a few years ago. During the first couple weeks in the fermenter the berliner had a very strong musty and briney aroma that made me think I screwed something up. Eventually, with extended time in the fermenter (about 30 days total) and further souring, the aroma did seem to fade quite a bit, but not completely make it’s way out of the beer (I actually enjoyed it, thanks to dry hopping). Now being fairly confident after smelling this experimental beer that what I was likely picking up in the no-boil berliner was DMS, I’m not sure if it cleared up a little from the extended time in the fermenter allowing C02 to push some DMS out or if the flavors and aromas from souring eventually dominated. This experimental beer on the other hand didn’t have additional time in the fermenter (plus it was sealed after day five) and the aroma definitely didn’t fade!

At first pour it’s got a lot of what I would cautiously describe as hop aroma, but with a healthy dose of aromatic hops (Citra, Galaxy, and Amarillo) in two separate large additions, it’s not what you would expect at all. The hop aroma is difficult to describe because it’s being altered so much from the DMS. I’ve settled on calling it an overripe and borderline rotten fruit smell, which sounds about as good as it tastes. This fairly strong “hop” flavor quickly fades though and the briney vegetative aroma comes through even more. Having gotten the results back and knowing that DMS is very prevalent the beer, I wouldn’t say the smell reminds me of creamed corn, maybe this is largely from the dry hopping. To me, the smell is much closer to cabbage and sauerkraut, I can even see how you could describe it as asparagus (based on the research) as opposed to corn.

The overall beer itself is questionably drinkable, it still has a great mouthfeel and all of that, but it’s just so strange. There is no question (and this isn’t really a surprise) that DMS in a high concentration is enough to dominate the hop flavor and aroma. I’m now curious how much DMS might influence hop flavor and aroma in smaller concentrations. For example, if a NEIPA is tested for lets say 50 µg/L, how might even this smaller amount alter the aroma of fresh and fruity hops? I’m particularly interested in this question because I have often thought that as hops fade with time, especially in commercial bottles a few months from packaging, they seem to drift from fresh fruity aromatics to more earthy and vegetative aromas (often I think green onion). I’m not saying for sure this is DMS eventually coming through as the hops fade, but this experiment does seem to convince me that it makes sense to try to keep DMS levels as low as possible in hop forward styles to prevent any sort of aroma competition.

I will likely make a few changes to my process after researching for this article. For example, when it comes to no-boil beers, I think I’m going to start boiling them. I’m already heating the wort for a no-boil beer anyways so why not just let it come to a boil? My memories of that previously un-boiled berliner I talked about really stick out to me as a good example of trying to purge that DMS out of the beer early on.

I’m also going to start experimenting with putting a piece of tinfoil over the small carboy opening on my Big Mouth Bubblers during the aggressive stage of fermentation and replace it with an airlock after around days 3-5. When I ferment in a keg, I might consider leaving the pressure release valve open to mimic an open fermentation to encourage more DMS removal by C02.

The next time I try capping the fermenter when adding dry hops in an attempt to retain hop aromatics, I’m going to think about waiting a few extra days than I typically would to make sure that first big drop in DMS has occurred. I might even try adding a tiny bit of yeast nutrient to boost the nitrogen content again to potentially prevent additional DMS from forming via the yeast.

Source and comments

Posted in beer | Leave a comment

Effects of mash parameters on fermentability and efficiency in single infusion mashing

A PDF version of this article can be found here: Effects_of_mash_parameters_on_attenuation_and_efficiency.pdf

It is commonly known that there are many factors that effect the fermentability (limit of attenuation) of brewing wort. The series of experiments conducted here are aimed at understanding not only the qualitative impact but also quantify the changes of fermentability depending on the parameters that were considered for evaluation.

In Understanding Attenuation it is mentioned that final attenuation of a beer mainly depends on 2 factors: limit of attenuation (i.e. attenuation potential) and the yeast’s ability to come close to that limit of attenuation. The limit of attenuation is only effected by mashing and the Fast Ferment Test has been introduced to determine it. For simplicity sake, these experiments only focus on single infusion mashes and explore the effects of the following mash parameters:

  • saccharification rest temperature: This is the first factor that comes to mind for all grain brewers. For a single step saccrification rest, the mash temperature has a great effect on the fermentability of the resulting wort. The lower the temperature (within a given range of course) the longer the beta-amylase will be able to work and produce maltose. See The Theory of Mashing.
  • time: Along with temperature this is one of the more important parameters in mashing. It determines how much time the enzymes will have to work on the mash.
  • mash pH: the beta and alpha amylase enzymes have different optimal pH ranges and therefore the mash pH can effect the activity balance between these enzymes. See The Theory of Mashing.
  • grind: larger grits of endosperm make it harder for the mash water to fully hydrate them and make the starches accessible to the enzymes. As a result lots of starch is released later when the beta amylase activity has already diminished. The result of a coarser grind is a lower limit of attenuation. See The Theory of Mashing.
  • water to grist ratio: the enzymatic activity of the amylases is effected by the thickness of the mash. Thinner mashes enhance the maltose production and therefore increase the fermentability. See The Theory of Mashing.
  • grain bill composition (base malt): mashes with high diastatic power (Pilsner, Pale) will produce more fermentable worts since they contain a lager amount of beta-amylase which can produce more maltose compared to mashes with lower diastatic power (Munich or large amounts of unmalted grains) assuming the same saccrification rest temperature.



Materials and Methods

Figure 1 – The samples were mashed in a stainless steel thermos bottle.

Figure 2 – The mash was strained batch-sparge style though a wire strainer

Figure 3 – The wort was then boiled for 15 min over medium heat

Figure 4 – After the boil was complete, the hot wort was filtered through a paper towel set in a funnel

Figure 5 – Fermentation of the sample with bread yeast

All experiments were conducted at a relatively small scale with minimal overhead. The strike water was heated in the micro wave. The time spent in the micro wave was noted for each experiment and provided a guidance for future experiments. After a number of experiments it was possible to come close to a desired mash temp by estimating the necessary heating time. The water was then added to a small steel thermos bottle where it was left for about 5 min to settle and if necessary adjustments were made by heating it some more or removing the top to let it cool quicker. A Styrofoam stopper was fitted for the bottle which also held an alcohol filled thermometer. The tip of the thermometer reached into the water/mash and it was possible to read its temperature without opening the bottle. The settled temperature was recorded and the milled grain was sirred into the water. 5 min into the mash the mash temperature was recorded. It was also recorded 30 min into the mash. Then the mash was stirred and its temperature recorded 5 min later. The last time the temperature was recorded was at the end of the mash.

Once the 60 min mash was complete a sample of the wort was tested for starch conversion with iodine on chalk. This was not done for all mashes because sometime I simply forgot. The mash was then lautered through a stainer set over a pot. In all cases it took less than 2 min to bring the first runnings to a boil. In the meantime sparge water was heated in the microwave and the grain was well mixed with the water before it was lautered again through the strainer. This sparging technique resembled a batch sparge.

After the wort was boiled for 15 min it was strained though a paper towel set in a funnel. The funnel was placed into a clean 12 oz bottle. Initial experiments sanitized the bottles and the funnel in boiling water. But that became to cumbersome and due to the high pitching rate used the effect that infections could have on the attenuation measurement was deemed insignificant. As a result the bottles were only cleaned and somewhat sanitized in the dishwasher. The wort in the bottle was then topped of with reverse osmosis water when necessary to reach the same amount of final wort volume for all experiments. Initial experiments didn’t care about the precise amount of wort in the bottle as long as there would be enough for 2 hydrometer readings, but later it was decided to keep the volume the same to get a measure of the efficiency along with the attenuation.

After capping the bottle with aluminum foil, the wort was left to cool at room temperature. Once cooled the original extract was measured with a hydometer (range 0 – 40 Plato) and then pitched with 1/2 teaspoon of Fleischman instant bread yeast and later experiments used Shaw’s brand instant bread yeast. Dry bread yeast was chosen for its low cost and its consistency. In previous fast ferment tests, where it fermented along side other yeasts, it has been determined that it attenuates similar to other ale yeasts.

The samples were fermented at about 20 C (70 F) for 4-6 days. They were occasionally shaken to rouse the yeast. After the 4-6 days of fermentation the yeast started to settle and no visible signs of fermentation were left. The final extract (=final gravity) of the sample was measured with a more precise hydrometer (range 0.990 – 1.020). Both hydrometers were calibrated with various sugar solutions (20, 10, 5, 2.5 and 0 Plato) and the readings were also corrected for temperature.

Time experiments

For the time experiments, the following parameters were kept constant:

  • grain type: Weyermann Pilsner Malt
  • grain weight: 70g
  • mill gap spacing: 0.55 mm
  • reverse osmosis water and no pH adjustment of the mash. The resulting mash pH was measured as 5.4
  • strike water volume: 240 ml
  • mash starting temperature: 67 C and 72 C
  • sparge water volume: 250 ml
  • lauter efficiency was estimated to be about 90%
  • boil time: 15 min

This series of experiments was actually done last, but it best illustrates how different temperatures effect mashing and as a result its data is presented first. To reduce the temperature drop during mashing, the thermos bottle was wrapped in sheets of foam generally used for packaging. But despite these efforts a temperature drop of ~ 3C/hr was still observed.

Temperature experiments

For the temperature experiments, the following parameters were kept constant:

  • grain type: Weyermann Bohemian Pilsner Malt
  • grain weight: 80g
  • mill gap spacing: 0.55 mm
  • reverse osmosis water and no pH adjustment of the mash. The resulting mash pH was measured as 5.4
  • strike water volume: 240 ml
  • mash time: 60 min
  • sparge water volume: 250 ml
  • lauter efficiency was estimated to be about 90%
  • boil time: 15 min

pH experiments

For the pH series experiments, the following parameters were kept constant:

  • grain type: Weyermann Bohemian Pilsner Malt
  • grain weight: 70g
  • mill gap spacing: 0.55 mm
  • reverse osmosis water
  • strike water volume: 240 ml
  • starting mash temperature: 73 – 74.2 C
  • mash time: 60 min
  • sparge water volume: 250 ml
  • lauter efficiency was estimated to be about 90%
  • boil time: 15 min
  • final wort volume: 15 min

The mash pH was ajusted with either white distilled vinegar (5% acetic acid) or baking soda (5% w/w NaHCO3 solution) which was added by volume with a small syringe.

The pH of the samples was measured at the end of the mash. Because the probe of the pH meter is getting old and its calibration function wasn’t working anymore, the correction of the measured value was done externally by measuring the sample, a 4.01 reference buffer and a 7.01 reference buffer. The following equation was used to determine the actual pH of the sample:

corrected pH = 4 + 3 * ((sample pH – 4.01 buffer pH) / (7.01 buffer pH – 4.01 buffer pH))

In addition to that, the samples were also tested with EMD’s colorpHast strips, which were read against the color scale in tungsten light. Reading them in fluorescent light actually changes their color. All samples were cooled to room temperature before measuring their pH. The pH meter that was used is a Milwaukee pH53.

Mill gap experiments

For the Mill gap experiments, the following parameters were kept constant:

  • grain type: Weyermann Bohemian Pilsner Malt
  • grain weight: 70g
  • mash average temperature: 68 C
  • reverse osmosis water and no pH adjustment of the mash. The resulting mash pH was measured as 5.4
  • strike water volume: 240 ml
  • mash time: 60 min
  • sparge water volume: 250 ml
  • lauter efficiency was estimated to be about 90%
  • boil time: 15 min

The grain was milled with an adjustable 2 roller mill (Schmidling Maltmill). The gap of that mill was measured with a feeler gauge with an estimated precision of +/- 0.02 mm.

Mash thickness experiments

For the mash thickness experiments the following parameters were kept constant

  • grain type: Weyermann Bohemian Pilsner Malt
  • grain weight: 70g (for the 2.57l/kg and 5 l/kg series) and 120g (for the second 2.57 l/kg series)
  • mill gap spacing: 0.55 mm
  • reverse osmosis water and no pH adjustment of the mash. The resulting mash pH was measured as 5.4
  • the total amount of water used was 5.6 l/kg for all experiments. (70g grist used a total of 390ml and 120g grist used 670ml water)
  • lauter efficiency was estimated to be about 80%
  • mash time: 60 min
  • boil time: 15 min

The strike water was calculated to reach a mash thickness of 2.57 l/kg and 5 l/kg respectively. Because of the larger temperature drop that was encountered for the 2.57 l/kg mash with only 70g of malt, another series for that mash thickness was recorded with more malt and strike water to keep the actual mash volumes and temperature drops the same between the two series. After mashing additional water was added to adjust the total water use per grain to 5.6 l/kg. In order to keep the lauter efficiency the same for all these mashed this series of experiment was lautered using the no-sparge method in which all the malt and water were combined before the mash was strained though a strainer into the pot used for boiling.

Malt type experiments

For the malt type experiments the following parameters were kept constant

  • grain weight: 70g
  • mash average temperature: 70 C
  • mill gap spacing: 0.55 mm
  • reverse osmosis water with treatment as necessary to keep the pH values comparable
  • strike water volume: 240 ml
  • mash time: 60 min
  • sparge water volume: 250 ml
  • lauter efficiency was estimated to be about 90%
  • boil time: 15 min

Since the color of the malt has an effect on the mash pH, one mash w/o adjusting the pH was done and based on the resulting mash pH the strike water was treated with vinegar or NaHCO3 solution such that the mash pH of the 2nd mash would be close to 5.3-5.5.

Calcium content experiments

For the Calcium content experiments the following parameters were kept constant

  • grain weight: 70g
  • grain type: Weyermann Pilsner
  • mash average temperature: 69 C
  • mill gap spacing: 0.55 mm
  • strike water volume: 240 ml
  • mash time: 60 min
  • sparge water volume: 250 ml
  • lauter efficiency was estimated to be about 90%
  • boil time: 15 min

To create mash water with varying levels of calcium ions but constant residual aklaklinity (which would change the pH of the mash) 1.05g calcium chloride and 1.00g chalk were added to 2l distilled water. This resulted in a water with ~343 mg/l (ppm) Ca and a residual alkalinity of 0. To create a strike water with different Ca concentrations, the treated water was blended with distilled water at varying rated. But In order to keep the Ca concentration in the final wort the same, the sparge water was blended by using the reverse ratio. E.g. if an experiment used 20% 343 mg/l Ca water and 80% distilled water, the sparge water would consist of 80% 343 mg/l Ca water and 20% distilled water. Since the chalk does not readily dissolve in water, the treated water was shaken before use to resuspend the chalk.

Results and Discussion


Figure 6 – The mash time dependency of the limit of attenuation and brewhouse efficiency shown for 2 different mash rest starting temperatures

Figure 7 – The iodine reactions of the mash liquid as they were recorded after the mash time was over. The iodine reaction was measured by dropping a drop of mash liquid onto a piece of chalk and adding a drop of iodine solution

When the relationship between mash time, fermentability (limit of attenuation) and extract efficiency was examined, 2 trends were observed. The lower of the two mash temperatures (67C / 152F) resulted in a slower initial increase in extraction efficiency and fermentability compared to the higher mash temperature (72C / 162F). This is explained by a lower activity of both the alpha and the beta amylase enzymes at the lower temperature. And in the long run the mash with the lower mash temperature was able to create a more fermentable wort. This is a result of the greater stability of the beta amylase enzymes at the lower temperature. While they are initially not as active as in the higher temperature mash, they are active for a longer time and can therefore produce more maltose which increases the fermentability.

At temperatures at which the denaturation of the beta amylase dominates (70C/160F and above) the beta amylase becomes more of a “sprinter”. It is able to create maltose at a higher rate due to the higher temperature which speeds up the reaction, but it denatures much quicker at the higher temperatures and as a result is not able to “cover as much ground” as it is able to at lower temperatures.

The higher temperature mash outperformed when it comes to extraction efficiency. This is a result of the stronger alpha amylase activity and possibly better gelatinization of the starches. Alpha amylase is the main enzyme for converting starch into water soluble glucose chains.

Unfortunately the temperature drop of the mash during the experiments was substantial ( ~3 C/hr, see Figure 8) which means that the beta amylase suffered less damage than it would have if the temperature for the 72 C mash would have remained at that temperature for a longer time. This can be seen in the continued increase of fermentability. While this could also be a result of alpha amylase activity it is assumed to be the result of lingering beta amylase activity. While alpha amylase is able to produce fermentable sugars as well it is not very effective at it. The effects of the increased denaturation of beta amylase can be seen in the steep rise of the fermentability that quickly falls behind the fermetability curve for the 67C mash temperature. This temperature dependent denaturation rate of the beta amylase is the main factor in the mash rest temperature dependence of the limit of attenuation.

Another observation (Figure 7) that was made is that the hotter mash was able to reduce the iodine reaction (indication for free starch in the mash liquid) quicker than the mash at 67 C. This is another result of the increased alpha amylase activity in the mash done at 72 C.

Figure 8 – Temperature profile for the various mashes done in this series of experiments. The average temperature drop was about 3 C per hour


The next series of experiments discussed here evaluated the effect of the mash temperature on the attenuation. Unfortunately the small mashing vessel that was used (stainless steel thermos bottle) and the stirring at 30 min caused a significant mash temperature loss. Because of that the average mash temperature was used as the x-axis in Figures 10 and 11. The approximate mash profiles for all the sample mashes that were done are shown in Figure 9. Dough-in happened at time 0 and the mash was quickly lautered after 60 min. The average mash temperature drop was 4C (8F). The wort pH was measured at 5.5 (at room temperature)

Figure 9 – The mash profile for the different mashes that were done for determining the attenuation dependency on the mash temperature

Figure 10 shows the limit of attenuation numbers that were measured for the various mash experiments in this series. Two things were surprising. First, all data points track along a curve very nicely and there is not much difference between points at similar temperatures. This is an indication for the high repeatability and low statistical error of this experiment. Second, the curve matches the few data points that are listed in one of the literature references [Narziss 2005] fairly closely.

Figure 10 – The attenuation, and estimated efficiency for the mash temperature experiment

Another remarkable observation is the almost perfectly linear shape of the attenuation slope after its peak. The linear function that has been fitted to match the data points has a slope of ~ 4 %/C. This means that increasing the temperature by one degree Celsius (1.8 F) will lower the limit of attenuation by 4%. This was found to be true over a fairly wide temperature range (12 C / 24 F). However, the slope leading up to the peak of attenuation is not as steep. This is assumed to be a result of stronger beta amylase activity which causes the production of maltose as soon as starch gelatenizes and alpha amylase is breaking down the amylose and amylopectin molecules into shorter dextrines. Literature sources that show data for the fermentability of single infusion mashes at various rest temperatures don’t show such a linear relationship past the peak of fermentability [Briggs, 2004]. It is assumed that the different shape of the temperature vs. attenuation curve shown in this experiment is effected by the relatively large temperature drop over the rest time.

In addition to the attenuation, an iodine test on chalk was performed at the end of the mash and the beginning of the boil. The results can be seen at the top of Figure 10. Note the mahogany color, witch is an indication of existing dextrines. This was expected to be seen for samples with a low limit of attenuation, but even sample number 10, which achieved 87%, shows a significant mahogany iodine reaction. More iodine tests for samples past the attenuation peak need to be added to confirm a trend. Surprisingly little of the iodine reaction carried over into the boil even though care was taken to heat to wort quickly to minimize the time spend in a temperature range that favors the alpha amylase activity (68 – 78 C/ 155 – 172 F). Note that this quick heating is not realistic in brewing practice where even without performing a mash-out, the collected wort spends some time in this range while it is heated to boiling.

Later experiments (number 12 and up) were done in a way that allows for a comparison of the brewhouse yield by recording the post boil volume. Because the lauter efficiency can be assumed to be the same for all experiments (all mashes were lautered batch sparge style with 1 sparge), the differences in brewhouse yield (also known as brewhouse efficiency) must be largely due to changes in the conversion efficiency (a measure of how well the mash converted the starches available in the grist). From the few data points that exist so far it is evident that lower temperature mash rests lead to a lower mash efficiency. Which is expected since not as many starches have gelatinized yet and the alpha amylase which is the major enzyme responsible for liquefying starches is less active. Note that all experiments were done as 60 min mashes and that the lower efficiency at the lower mash temps can be accounted for by mashing for a longer time.

Figure 11 – The expected limit for the final extract of a beer brewed from a 12 Plato (1.048 SG) wort in relation to the mash temperature

Figure 11 shows another twist on representing the attenuation data. For this chart, all the brewed worts were assumed to have an extract content of 12 Plato (equivalent to 1.048 SG). Instead of showing the limit of attenuation, the attenuation was used to calculate the lower limit of the final extract (=gravity) of the beer. This is the extract reading you would expect to get from a fast ferment test or a very well attenuating yeast which left the beer with very little or no fermentable extract.


pH meter vs. colorpHast strips

Table-1 pH adjustments and the measured pH values

Table-1 lists the pH corrections that were done to the mashes and the pH that was measured with a pH meter and the pH determined with colorpHast strips. The first column lists the experiment number and the second column the amount of acid or baking soda solution that was added. It is interesting to see that the pH meter measurements for experiments 19 and 23, which both received no pH treatment, are fairly close while the strips must have been interpreted differently between the 2 experiments. Another remarkable observation are experiments 24, 25 and 26, which received different amounts of acid, caused different pH meter readings, but all read 4.7 on the color scale for the pH strips. This effect was also observed when I tested the pH meter against the strips a few months back. The results of that test are published here. The next section will show that 2 of these 3 samples (25 and 26) show very different limit of attenuation values which leads to the conclusion that there was indeed a pH difference between the two.

Figure 12 – A plot of the colorpHast reading vs. the corrected value from the pH meter

Figure 12 shows a plot of the colorpHast reading vs. the corrected pH meter value. The error was at most 0.5 (sample 24), which is significant enough to matter in brewing, but it happened at a pH level which is generally not achieved in mashing.

mash results

The mashing experiments for the pH series were performed with a single infusion mash that started at or around 73.5 C (164 F). This temperature is well on the slope that was determined earlier in the temperature series of experiments. The results of the pH experiments are shown in Figure 13.

Figure 13 – iodine reaction, limit of attenuation and brew house yield as a function of the mash pH

Figure 14 – A comparison of the wort colors after a 15 min boil for experiments 23 (pH 5.5) and 22 (pH 6.5). The darker color of the higher pH wort is remarkable. It is a result of the stronger maillard reactions at the higher pH.

Figure 15 – For experiment 28 (mash pH 4.55) the wort at the end of the mash and the beginning of the boil was tested for starch and both tested positive

The first observation was that the optimum for the limit of attenuation (fermentability) seems to be between pH 5.4 and 5.6. This matches the pH range that was given for the beta amylase enzyme in Narziss [Narziss 2005]. In addition to that the shape of the curve matches the attenuation data that Narziss lists for a few pH values. But he gave no information about the mash temp, mash thickness (though ~ 4 l/kg is likely) or other parameters that can effect the limit of attenuation. Because of that it can be assumed that all pH values given in that book were measured at room temp.

Outside of the 5.4 – 5.6 range the limit of attenuation declines. The slope of decline is steeper towards lower mash pH and a little less steep towards higher mash pH. This might be the result of a different decline of beta amylase activity. Number 26 is an outlier, but can be explained by being mashed at a lower than average mash temp. Though the termperature series showed an attenuation decline of 4% for every degree Celsius, it is unknown if this relationship also exists for low mash pH conditions.

The brewhouse yield (also known as brewhouse efficiency) also shows a dependency on the mash pH. This is not new as many home brewers report an increase in their brewhouse yield once they correct the mash pH to be in the optimal range for mashing. The optimum for the brewhouse yield seems to be near a pH of 5.2-5.3, which is a little lower than the optimum for the fermentability. This might be the result of of better alpha amylase activity in that range, although Narziss reports the alpha amylase pH optimum as being between 5.6 and 5.8 [Narziss 2005]. At the extreme end of the pH range that was tested are 2 outliers for brewhouse efficiency. Experiment 28 was done close to a pH of 4.5 and its brewhouse efficiency was greater than the efficiency for the experiment done at 4.8 pH. The same is true for experiment 22. But when that experiment was repeated as 27, the resulting brewhouse efficiency followed the established trend. It is assumed that the measurements were incorrect rather than being unexpected peaks of efficiency.

The iodine reaction at the end of the mash followed the trend that was already seen for the limit of attenuation and the brewhouse yield. The least reaction remained in the optimum mash range of 5.2 – 5.6 while at the extremes to either side native starch was still present in the mash liquid after a 60 min mash (purple color of the iodine test). Unfortunately no iodine test results are available for either of the two experiments that didn’t receive any mash pH treatment. For experiment 28 (mash pH 4.55) the boiled wort was tested in addition to the mash and starch was still present. This shows that a pH well below 5.0 does not provide sufficient enzymatic activity to convert the mash. As a result of that the attenuation and brewhouse yield suffer and native starch will be present in the wort and subsequently in the beer.

Having a lower brewhouse yield and stronger iodine reaction also contradicts the pH and temperature ranges shown for the alpha amylase in John Palmer’s How To Brew [Palmer 2006]. There it is shown that the alpha amylase is active well towards 4.5 pH and below.

Mill gap spacing

Figure 16 – limit of attenuation and brewhouse yield as a function of the mill gap spacing.

Figure 16 shows the results of the mill gap spacing series of experiments. The limit of attenuation shows little dependency on the tightness of the grind. This seems to indicate that either only little new starch is released from the coarser grind during mashing or that the beta amlylase shows activity throughout the duration of the mash. Given that the mash temperature was 68 C, similar to one of the temperatures in the mash time experiment it is assumed that the relatively small change of attenuation is a combination of both.

The brewhouse yield shows more dependency on the tightness of the crush. The tighter the crush, the higher the brewhouse yield. But the difference between a 0.35 mm (13 mil) crush and a 0.95 mm (37 mil) crush was only about 10%. Reports of significantly increased brewhouse efficiencies by home brewers after adjusting the mills to crush more tightly led me to expect a larger difference. But some of the default malt mill settings for home brewers and home brew stores are also greater than 0.9mm (37 mil) which was not taken into account when the decision was made to only evaluate mill gap spacings between 0.35 and 0.95 mm. The brewhouse yield dependence on the mill gap spacing can be explained by the larger amount of flour and smaller grits that are produced with a tight crush. These make it easier for the mash water to access the starches which is necessary for the enzymes to convert them into soluble carbohydrates (i.e. dextrines and sugars).

The literature reports that finer grists result in higher levels of fermentability. This behavior was not confirmed in this experiment but given the higher conversion efficiency that is achieved with the finer grists and the relatively unchanged fermentability it can be said that the finer grists produce more fermentable extract compared to coarser grists. If this was also the case for mashes that show less conversion efficiency sensitivity towards the mill gap spacing (e.g. if the close to 100% conversion efficiency is achieved with any mill gap spacing within a given range) such mashes would likely show a higher fermentability for finer grists.

Figure 17 – iodine test on a piece of spent grain from experiment 33 (0.95 mm) (circled in red)

For experiment 33 (0.95 mm mill gap) a piece of spent grain was taken from the spent grain, crushed on chalk and tested with iodine (Figure 17). The result was a deep black-purple reaction that is an indication of the presence of native starch. This shows where the lost efficiency is it is still present as native starch in the grain. But this was also the case for the spent grain from mash experiments with tighter mill settings. For some reason the mash experiments don’t achieve the level of conversion efficiency which I commonly see in full size 60 min single infusion mashes. An indication of that is already present in the iodine tests which oftentimes showed a residual reaction after 60 min, something I don’t see in full sized mashes either. One reason for that could be temperature loss and lower level of agitation (stirring) that happens for the experimental micro mashes.

Mash thickness

Figure 18 – the effect of mash thickness on attenuation and efficiency

The results for mash thickness were somewhat surprising. Contrary to common believe no attenuation difference was seen between a thick mash (2.57 l/kg or 1.21 qt/lb) and a thin mash (5 l/kg or 2.37 qt/lb). Home brewing literature suggests that thin mashes lead to more fermentable worts, but technical brewing literature suggests that the mash concentration doesn’t have much effect in well modified malts [Narziss, 2005]. Briggs cites data that doesn’t show a change in fermentability when the mash thickness is changed [Briggs, 2004]. This was confirmed by these eperiments where all the data points were on the same curve that had already been established in the temperature experiment.

Note, that the experiments for the 2.57 l/kg mash were run twice because the initial experiment resulted in a small mash volume that lost 5 degree Celsius over the duration of the mash. To keep the temperature drop between the experiments the same the mash volume was increased and the result was a 2 degree Celsius temperature drop which matched the temperature drop for the 5 l/kg mash. But in the end that didn’t make a difference.

A significant difference was however found in the efficiency. The brewhouse efficiency of the tick mashes remained almost constant between 58 and 60% over the temperature range of the experiments, but the brewhouse efficiency for the thinner mash showed a strong dependency on the temperature and was always better than the efficiency of the tick mash. That leads to the conclusion that thinner mashes perform better and allow for better extraction of the grain. Briggs also reports that thinner mashes can convert more starch but that most of the conversion potential is reached at a water to grist ratio of 2.5 l/kg [Briggs, 2004]

Malt type

Figure 19 – The limit of attenuation of mashes with 3 different malts at a mash temperature of 69C

Figure 20 – The brewhouse efficiency that was achieved with 3 different malts at a mash temperature of 69C

Different malts, especially malts kilned at different temperatures provide different amounts of enzymes to the mash. This change in enzymatic strength leads to differences in the limit of attenuation and extraction efficiency.

Figure 19 shows the attenuation results for 2 different pH values for each malt. There is not much difference between the light Munich (Weyermann Munich I) and the Pilsner malt. But the dark Munich malt shows a significant drop in the limit of attenuation when mashed at about the same rest temperature. This is a result of lower beta amylase activity due to the significantly reduced beta (and to some extend also alpha) amylase amount. Note that the Dark Munich experiment with the lowest attenuation was also mashed at a rest temperature that was about 2C higher than the average for the other malts at this pH. If we assume that there is a 4%/C drop of attenuation (as shown in Temperature) the limit of attenuation at a comparable mash rest should have been a little higher (about 63%).

Note that the limit of attenuation of a mix of these base malts would not be the weighted average of the individual malt’s fermentabilities. This is because the pilsner malt has much more enzymatic strength than the dark Munich and will raise the fermentability by more than what its actual weight portion of the grist is.

The brewhouse efficiency (Figure 20) was not as strongly effected even though the higher kilned malts showed a slightly lower brewhouse efficiency. This is somewhat surprising as recent full size (5 gal batch size) mashes with the Best Malz Dark Munich showed a much slower conversion than mashes that contain all or larger amounts of low kilned malts.


Figure 21 – The limit of attenuation and efficiency for 4 mashes with different calcium ion concentrations in the mash water. The numbers above the data points show the pH that was measured for room temperature sample of the mash

As mentioned earlier, none of the mashes so far has been able to get to a conversion efficiency where 100% of the starch in the grist is converted to soluble extract (mainly dextrins and sugars). This rarely happens in the full size brewing batches (5 gal) even if they are single infusion mashes. One difference between the micro and the full size mashes is that water that is used. The micro mashes used reverse osmosis water with an approximate dissolved solid content of 30-50 mg/l (=ppm) while the full size mashed use reverse osmosis water and brewing salts. Based on this it was suspected that the mineral composition of the water has more impact than simply though the residual alkalinity and subsequently the mash pH.

One ion of particular interest is Calcium. Briggs notes that calcium stabilizes the alpha amylase [Briggs, 2004]. But since alpha amylase is still fairly stable at common saccharification rest temperatures (below 70C / 160F) the stabilizing effect of calcium shouldn’t matter much.

The experiments showed a slight downward trend for the fermentability if the amount of calcium in the mash is increased. But the difference between the highest and lowest attenuation is only about 1% which is well within the error range for these experiments. Based on that I would say that the calcium content of the mash has no impact on the fermentability.

A slightly stronger effect was seen for the efficiency. Here an almost 3% span was observed but since this is based on only a few samples that didn’t match in their mash pH either the result may easily be called conclusive although the trend, the higher the calcium content the better the efficiency, matches the expectations.


Attenuation and efficiency of the mash are effected by many mash parameters. Some have more impact others have less. When using a single infusion mash, temperature and time are the best parameters that a brewer can work with to target a specific fermentability of the wort. The time should be long enough to allow for complete conversion of the mash or at least a wort that doesn’t contain any starch (negative starch test). This might be achieved after 15-30 min, but a longer mash rest may be needed to achieve the desired fermentability. The mash pH should always be controlled and kept between 5.4 and 5.7 when measured at room temperature (5.05 – 5.35 when measured at mash temperature). This pH control can be done through the brew water design (residual alkalinity) and/or acid/salt addition to the mash.

Brewers that don’t mill their own grain will not be able to effect the tightness of the crush and will have to accept lower conversion efficiencies or ensure that the mash has enough time and “strength” to achieve an acceptable conversion of the starches. If the mill gap spacing can be controlled the conversion efficiency can be improved through a tighter crush. But at some point the crush might be to tight for a resonable run-off speed.

The thickness of the mash doesn’t seem to effect the fermentability of the wort that is produced but thinner mashes can significantly improve the conversion efficiency. As a result brewers who see low efficiency from their mashing may try to use a thinner mash (3-4 l/kg or 1.5 – 2 qt/lb) as they were shown to convert more starches.

When working with large amounts of highly kilned malts attenuation and efficiency problems can arise due to the lower diastatic power (enzymatic strength) of these malts. This can be counteracted by lower mash temperature and longer mashes or the addition of a diastatic stronger malt to the grist (e.g. 10-20% of Pale/Pilsner malt)

While the water composition may also have an impact on attenuation and efficiency besides the change in mash pH through the residual alkalinity, its impact is considered small and secondary.



Time series

Time series data.gif

Temperature series

Temperatur series data.gif

pH series

PH series data.gif

mill gap series

Mill gap and attenuation table.gif

mash thickness series

Mash thickness data.gif

malt type series

Malt series data.gif

Calcium series

Calcium series data.gif


  • [Narziss, 2005] Prof. Dr. agr. Ludwig Narziss, Prof. Dr.-Ing. habil. Werner Back, Technische Universitaet Muenchen (Fakultaet fuer Brauwesen, Weihenstephan), Abriss der Bierbrauerei. WILEY-VCH Verlags GmbH Weinheim Germany, 2005
  • [Palmer, 2006] John J. Palmer, How to Brew, Brewers Publications, Boulder CO, 2006
Posted in beer | Leave a comment


Короткого хранения

  • 700г – зелёный перец
  • большой пучёк – свежая кинза (кориандр) и
  • одна большая головка – чеснок
  • 70г – соль
  • 2 дня сушить помешивая

Долгого хранения

  • 1кг – красный перец
  • 0.5ч.л. – Уцхо-сунели (fenugreek blue, Trigonella caerulea)
  • 1ч.л. – хмели-цунели
  • 1ч.л. – кориандр
  • 100г – соль
  • 100г – чеснок (одна большая головка)
  • 2 дня сушить помешивая
  • 500г перец
  • 150г чеснок
  • 75г соли
  • 2с.л. хмели-цунели
  • 2с.л. кориандр
  • 3 дня

Posted in Food and drink, Recipes | Leave a comment

Improve mash efficiency in your stainless steel mash tun

Mash efficiency refers to the extent to which you are able to pull the sugars out of the malted grain and into the water. A high percentage of sugar recovery (75-90%) is called ‘high mash efficiency’. A low percentage (60-75%) is called ‘low mash efficiency’. Since the amount of sugar in the water correlates quite closely to the amount of alcohol you will have in the beer, it is important to understand mash efficiency in order to produce consistent results.

High quality beer brewing equipment

A low efficiency can be compensated for by increasing your grain bill (the amount of malted grain you start with). For example, if your mash efficiency is at 70% resulting in an alcohol content of 4% in your beer, you can increase your grain bill by 20% to bring the sugar and alcohol content higher. But this also adds cost to your brewing as the extra grain costs money and with more grain, you need a larger mash tun which also adds cost. The best way to increase mash efficiency is through proper technique.

First, it is important to ensure is that your ingredients are fresh. The yield will be lower with oxidized/stale grain.

Second, it is important to have a proper crush size and properly crush the grain. For most systems, a crush size of 0.04″ (1mm) is the best (if routinely seeing a stuck mash where wort will not flow through the grain bed, one solution can be to increase the crush size slightly). It is also very important that grain be ‘crushed’ and not ‘ground’. Grinding involves shearing of grain into fine particles (that gum up), whereas crushing sends cracks throughout the grain that open up to insides of the grain to water and enzymatic digestion. A 3 roller mill generally provides a better crush than a 2 roller mill, and the speed of the mill is also very important to maintain. The mill rollers should rotate at about 150rpm, whereas most electric drills operate at 600-1800rpm. Pulleys can be used, or a simpler solution is to purchase a variable speed drill and use it on a low setting. If the variable speed drill is a manual type (where depressing a button manually will be tiresome and inconsistent) it can be helpful to use a ‘120V fan speed controller’ to control the drill speed to 150rpm (if it is hard to calculate the speed, put a piece of visible tape on the drill chuck and count the number of times it passes each second—it should be close to 3/second). If the drill is not strong enough to turn the mill on low speed, the grain intake inside the hopper can be restricted to let less grain through at one time—it will take slightly longer but your crush will be more effective.

A third step is to ensure proper water chemistry. Water chemistry is a complex topic which we won’t get into here, but if you suspect water issues, try brewing with bottled (spring or RO) water that you know is fairly neutral, add the minerals you need, and compare results.

Fourth, ensure mash pH is around 5.2-5.3.

Fifth, ensure the grain is thoroughly mixed with the water. Clumps of grain and air can easily form when mixing the grain with the water. These clumps prevent water from accessing the crushed grain inside the clump, preventing the carbohydrates from being dissolved into the water and preventing enzymes from breaking the carbohydrates up into sugars that the yeast can digest. Mash in using the following steps:

  1. Ensure the pump is turned off as the grain should be given time to settle slowly.
  2. Add grain (mixed with rice hulls if using rye or a high percentage of high protein and/or no husk grain like wheat) slowly in a circular motion so it doesn’t ‘pile up’ on the surface, and allow time for the grain to sink naturally (for commercial customers our Grain Bag has a chute that is controllable and can make this easy).
  3. Dark malt should be crushed separately and added last as it tends to crush finer which impedes the flow of wort; keeping it near the surface ensures it does not form a dense ‘mat’ near the bottom.
  4. Ensure any grain/air clumps are broken up and a proper grain bed forms. The BREWHA Mash Mixer is an easy and effective way to mix the mash but do not overmix (as this can pulverize the grain increasing the chance of a stuck mash) and DON’T go near the bottom of the Colander to the disrupt the grain bed that has naturally formed by the larger grains settling faster through the water.

Sixth, closely regulate mash temperature. The enzymes that work to cut up carbohydrates into smaller sugars that yeast can digest (in the process called ‘saccharification’), producing CO2 and ethanol (alcohol) in the process, work best at specific temperatures. If the mash is not kept at the precise temperature the recipe calls for, optimal results and consistent efficiency will not be achieved. The first step to ensure the proper temperature is to ensure the strike temperature is sufficiently high. Since the grain addition will cool the water, the temperature of the water before adding grain should be above the mash temperature. Grain will generally cool water by 5-8C/10-15F so the strike temperature should be that much higher. After mashing in, check the temperature IN the grain bed with a thermometer to ensure the strike temperature was high enough. The mash temperature can be regulated with BREWHA vessels since the mash vessels (mash tun or fermenters) have built in ports for heating elements and temperature sensors to precisely control the temperature of the mash. If a large temperature rise is needed, the temperature in the cone can be set substantially higher than the mash temperature to reduce the time needed to raise the temperature in the grain bed; this is similar to a decoction type method which is explained more fully in the BIAC manual mashing section—see pg. 47 for details.

The seventh tip for increasing mash efficiency is by stirring or using a recirculating pump. The pump pulls wort (sugary water) out of the bottom of the mash vessel (past the temperature sensor and heating element under the mash screen) and puts it back into the top of the mash vessel above the grain bed. This circulating action provides movement of the enzymes and water past the crushed grain resulting in a more even temperature throughout the gain bed and the greatest interaction of the enzymes and grain carbohydrates, leading to greatest efficiency. When circulating with the pump, the return flow back into the mash vessel should be throttled with a valve to ensure the grain bed is not compacted as this will impede saccharification (enzymes converting the complex sugars/starches into simple sugars the yeast can digest) and could damage the heating element. mash recirculation ratesLet the grain sit for a minimum of 10 minutes before turning the pump on to allow the grain to absorb water and expand; start the pump with the return valve (on Mash Colander or pump exit) closed and open slowly until the proper recirculation rate is achieved (see the table to the right for recommended recirculation rates). Check the temperature of the returning wort to ensure it is not cooling the grain bed—particularly with a Small Colander, the return rate needs to be very slow, and cooling can occur in the hose; if this occurs, the temperature of the wort in the cone should be raised by up to 2C/4F above the mash temperature (a higher temperature is used with very low flow rates like are needed in the Small BIAC as more cooling occurs the longer the wort remains in the hose). If the wort level is rising at the top of the mash vessel, it indicates that too much volume is being moved by the pump and the flow should be throttled with the valve (install the valve ONLY on the downstream side of the pump either on the pump exit or on the Mash Colander port, or the pump will not work properly and could be damaged). If the wort stops moving down through the grain bed (or moves only very slowly) it likely means the grain bed is compacted and may need to be stirred up and reformed. If the pump plugs frequently when using the BIAC (due to grain falling to the bottom of the fermenter and entering the pump hose), the hose from the fermenter to the pump can be attached to the side racking port (instead of the bottom) to prevent grain from being drawn into the pump. With a very thick mash, channeling of the wort down through the mash may occur so use of the Mash Mixer at intervals to disrupt the top 4/5 of the mash bed can be helpful.

The eighth action is to slowly sparge the grain while lautering. The tendency at the end of mashing is to want to hurry the lautering and rush into boil. This is a mistake as far as increasing efficiency is concerned. Residual sugars inside the grain do not immediately flush out with fresh water but slowly diffuse out. If not given sufficient time (around 60 minutes to sparge and lauter is not too long) they will remain in the grain and recovery of sugar will be decreased. Hot sparge water (75C/167F) will generally give better results than cold sparge water as sugars are more soluble at higher temperatures.

To confirm that saccharification is complete, a starch/iodine test can be completed. If starch is still present, a drop of iodine will turn black. If starch is not present (ie. the enzymes have converted all the available starch into smaller sugars that the yeast can digest) the iodine will stay a yellow color. In most situations, wort will convert quickly (within 15-30 minutes) so the best way to check for starch (this tip was given to me by Ken Grossman of Sierra Nevada) is by removing a piece of grain and dropping the iodine directly on the grain; ideally the entire grain should stay yellow, but since this is rarely achieved at least all but the tip (where it is hardest for water/enzymes to get to) should stay yellow. If most of the grain turns black, let the mash run a while longer. (It goes without saying—don’t add the iodine back into the mash; throw it out or wash it down the drain.)

With an appropriate mash period (60-90 minutes is typical), proper crush size and water chemistry and by observing the above actions, consistently high mash efficiency can be achieved.

Proper crush size and tips for preventing a stuck mash

Posted in beer | Leave a comment