Forgot your password?
typodupeerror
Beer Hardware Hacking Build

BrewPi: Raspberry Pi and Arduino Powered Fermentation Chamber 96

Posted by Unknown Lamer
from the includes-kitchen-sink dept.
For the homebrew hardware nerds out there who also homebrew beer: "BrewPi is an open source fermentation controller that runs on an Arduino (for now) and a Raspberry Pi. It can control your beer temperature with 0.1 degree precision, log temperature data in nice graphs and is fully configurable from a web interface." Source code. The article has lots of photos and screenshots. The project involves rewiring the compressor's electrical connection through a PID controller, and includes both a fancy OLED display on the fridge and support for logging statistics and control over the web. If you've ever had the joy of gradually crash-cooling a lager (not too fast, not too slow), the software includes settings to effect gradual temperatures changes in the fermenting wort. Certainly fancier than a Johnson controller and a probe attached to a fermenter with a strip of insulating tape.
This discussion has been archived. No new comments can be posted.

BrewPi: Raspberry Pi and Arduino Powered Fermentation Chamber

Comments Filter:
  • Why does it have to be a Raspberry? Why do you have to control it in C? Is it because people don't teach or know how to do it the old school way anymore? Is the effect somehow less if you DONT have a fancy OLED display?

    THIS is why I cancelled my "MAKE" subscription - too much sizzle and not enough steak.

    • by santax (1541065)
      Of course it has to have a oled display. That way you can program tetris on it and tell the wife the kettle is the computer that you build yourself.
    • by afaik_ianal (918433) on Monday October 01, 2012 @09:29PM (#41521133)

      You reckon you can do temperature control in an uncontrolled environment with a handful of transistors? Keep in mind the external temperature is uncontrolled. The yeast itself generates heat within the ferment at varying (and often unpredictable) rates.

      The simplest approach you can reasonable consider for the level of control they're looking for in their environment would be a PID controller, which if memory serves me correctly will have more than a handful of transistors in it. You're going to need a pretty decent PID to handle temperature profiling, which their solution

      They're using cheap, off-the-shelf parts to solve their problem, which in turn allow them to put extra features in there like web control, and it's now much easier for anyone to do the same thing. The parts may not have existed in 1964, but you need to keep in mind The Doors aren't the only good thing to happen since then.

      • Re: (Score:3, Funny)

        by Anonymous Coward

        Dude do you know how many transistors you can fit in your hand nowadays?

    • by OzPeter (195038) on Monday October 01, 2012 @09:45PM (#41521239)

      Why does it have to be a Raspberry? Why do you have to control it in C? Is it because people don't teach or know how to do it the old school way anymore?

      Do you want to explain how to make a tunable PID loop out of transistors? Yes it can be done, but the commercial loop controllers I used 30 years ago certainly had their foibles. And there is a hell of a lot of useful fundamental old school control theory that can be learned in doing it the way they have. Proper manual loop tuning is just as much an education as building a circuit out of individual parts.

      • by Formalin (1945560)

        It's a lot easier to make them with comparators or opamps as opposed to discrete transistors.

        They're pretty good at P, I, and D. Kinda what they were devised for... op[erational] amplifier. Analog math / computing.

        Can be fairly simple circuits, but tuning takes some work (not that it doesn't in digital, either). Digital has the advantage of fitting a wider range of applications without needing physical component changes and so on.

        I kind of feel sad when I see people replacing a mechanical thermostat, or sim

        • by Formalin (1945560)

          Having read the thing now, digital is the way to go in this instance.

          I'd have done it cheaper though :-p

          I still see things all the time that really ought to be analog, which is what I was talking about anyway.

        • by Sqweegee (968985)

          I built an analog PID controller years ago in an industrial controls course exactly that way using opamps. We only used it to keep a couple of servos aligned but it could have been adapted to anything.

          Tuning was a pain, as it was part of the whole point of the exercise we needed to swap out the individual resistors, caps, and inductors to achieve the tune values we had calculated for a proper response as well as specific improper responses each stage could cause. Even the small leakage in the capacitors cau

    • Glad you're gone.

      Look, you can do it the easy way or the hard way. When tech is cheap, why not do it the easy way? And when you've got the extra power, why not go for some bells and whistles?
      And did you read the goddamn article? The guy wants to glance at his brewer and make sure it's still doing what it should. Should he make it beep morse just because he can?! Seriously, what kind of idiot would build something like this without a display?
    • Re: (Score:3, Informative)

      by Anonymous Coward

      According to the comments here: http://hackaday.com/2012/10/01/brewpi-is-a-raspberry-pi-in-charge-of-beer-fermentation/

      >The Pi is a bit too unstable to be trusted with direct control over the fermentation. The slave controller runs all the control algorithms, the Pi is for interfacing and data logging.

      • by ncc74656 (45571) * <scott@alfter.us> on Tuesday October 02, 2012 @12:28AM (#41521971) Homepage Journal
        That sounds a bit fishy. A Raspberry Pi can't keep up with switching a fridge on and off to keep temperature in a range? That beggars belief. I have an Apple II bit-banging a 1-Wire temperature sensor and clock and toggling a relay on and off for the same purpose. If a 1-MHz 6502 can run a fermentation fridge, I'd think a Raspberry Pi would be more than up to the task. I'm pretty sure there's a 1-Wire USB interface available that would let the RPi use the same sensor I'm using, with less overhead.
        • Re: (Score:2, Informative)

          by Anonymous Coward

          Elco here. The raspberry pi is easily fast enough, but it would crash on me sometimes. Just hang and require a reboot. A microcontroller just doesn't crash.

    • by jcr (53032) <jcrNO@SPAMmac.com> on Monday October 01, 2012 @10:53PM (#41521559) Journal

      Is it because people don't teach or know how to do it the old school way anymore?

      Got it on the first guess. Hell, I'd be surprised if 2% of /. readers can even read resistor color codes.

      Of course, that being said, it's probably less work to calibrate a computer controlled sensor and cooler than to achieve the same precision with thermistors.

      -jcr

      • Hell, I'd be surprised if 2% of /. readers can even read resistor color codes.

        BBROYGBVGW - For political correctness, I won't share the mnemonic I used 30 years ago. Now get off my lawn.

    • by honkycat (249849) on Monday October 01, 2012 @11:06PM (#41521621) Homepage Journal

      Why does it have to be a Raspberry? Why do you have to control it in C? Is it because people don't teach or know how to do it the old school way anymore? Is the effect somehow less if you DONT have a fancy OLED display?

      Why? Because these methods are objectively better than the old way. And I say this as someone reasonably experienced in the old school way.

      The controller is far more sophisticated than a simple PID loop, even if you ignore the ability to log the temperature history, store profiles, query the status remotely, etc.

      The OLED display isn't necessary, but it's actually a helpful, useful feature to be able to display the status of the system clearly and unambiguously.

      And good luck implementing this "the old school way" in the same period of time as someone using these tools could do it.

      It doesn't have to be a Raspberry Pi or Arduino with a spiffy display, but there's no glory in doing things the hard way. There's educational value in it, and in knowing the basics, but a system that works doesn't benefit from having been hard to build.

    • by anubi (640541) on Monday October 01, 2012 @11:37PM (#41521761) Journal
      I guess if you consider the problems you would have calibrating an analog version, along with the display interface, the digital stuff has become more practical than the old analog stuff we grew up with. Did you forget all those trimpots we always had to design in because we never knew the exact parameters of our parts?

      Those old analog panel meters are now quite expensive. As well as being fragile, and never were much better than 2% accurate. The cheapest 3 1/2 digit display you get today blows them out of the water.

      Personally, I enjoyed this article - as I build a lot of heat transfer controllers - and I find it very interesting what others are doing.... Not to say I would do it the same way, but others often find some way around a situation that has stymied me in the past.

      I am very interested in when people start controlling compressor speed via variable-speed three-phase drives. Some washing machines already use these pancake-like motors to completely eliminate the transmission, controlling the motor speed and direction directly from a microcontroller so as to seamlessly shift the motor function between agitate and spin. I have been waiting with baited breath for a refrigeration compressor made in that manner so I can optimize the refrigeration process for whatever heat transfer I can get into the evaporator coil and out of the condenser coil. The only control point I now have is either flat all-or-nothing from the compressor and limited control of the thermal expansion valve. With finer control of heat exchanger temperatures, I can control not only air temp but also the humidity.

      Articles like this constitute a design snippet to show a usable part of a larger system.
    • by AC-x (735297) on Tuesday October 02, 2012 @04:14AM (#41522799)

      The Arduino does the temperature control (as that's what microcontrollers are good for) and the Pi is there to add a nice display and web server (as that's what a mini SBC is good for), seems like using the right tools for the right jobs to me. If you don't want a fancy display you could just use the Arduino part and skip the Pi.

      • The Arduino does the temperature control (as that's what microcontrollers are good for) and the Pi is there to add a nice display and web server (as that's what a mini SBC is good for), seems like using the right tools for the right jobs to me. If you don't want a fancy display you could just use the Arduino part and skip the Pi.

        In that vein, it's worth noting that an entire RPi sells for less(sometimes about the same, depending on distributor) than the most common W5100-based Arduino ethernet shields, while being markedly more powerful and having convenient features like USB host support and some I/O pins of its own. This probably helps explain some of the recent popularity of 'Arduino+RPi' projects where the RPi is, perhaps, a bit on the overkill side from a purely technical perspective...

    • by EmagGeek (574360)

      Agreed. This same thing can be accomplished with $2 worth of op amps and passives, with no software to be buggy.

      When the only tool you have is a hammer, everything looks like a nail.

    • by Tassach (137772)

      Why do you need electronics at all?

      All you need to brew alcohol is watertight container of moderate size, a sugar-rich solution, and a teaspoon or two of yeast. People were doing it for thousands of years. There's even some evidence to suggest that the neolithic transition to sedentary farming was motivated by the desire to have a large surplus of grains available to ferment.

      • by bonehead (6382)

        You don't "need" electronics. Particularly if all you care about is that the final product have some alcohol content.

        Something like this is handy for when you want to produce the best product you can, and even more importantly, when you want to be able to consistently produce a beer that tastes the same from batch to batch.

        Fermentation temperatures can have a large effect on the taste of a finished beer. Even from the exact same recipe you can have very different tasting products from summer to winter if

      • For lagers, as opposed to ales, you generally need slightly lower temperatures than your ordinary household ambient temperature. It's difficult these days to find a friendly, local cave with the proper temperature.

        The common home brewer solution to this problem, absent a sufficiently cool basement, is to use a chest freezer with an aftermarket temperature controller. The temp controller is plugged into the wall, the chest freezer plugs into that, and a probe goes inside. It's a reasonably cheap solution

  • by exabrial (818005) on Monday October 01, 2012 @09:12PM (#41521013)
    I'd be interested in multiple control points... I've started to do pressurized fermentation once the gravity of the beer of approaches the final gravity. This carbonates the beer without a secondary fermentation, and reduces risk of introducing oxygen and contamination into the beer. If the system could electronically monitor specific gravity then seal the airlock, but blow off any excess over 10psi... would be awesome!
    • Find a sensor that does that, plus a suitable valve, and it can't be very hard to add that.

      • The valve is the easy bit (if you don't mind manually switching a tap on in response to an alert).

        The sensor is not so simple. Even if digital hydrometers were cheap and easily available, you'd need to find a way of mounting it inside the fermenter such that you're not going to create a haven for bacteria (hint: any kind of non-smooth surface inside the fermentation vessel is a haven for bacteria).

        I've heard of people using ultrasonic transducers for measuring the gravity of fluid in a pipe, but it sounds

        • by norpy (1277318)

          Why open a valve electrically?

          There are valves you can get from 0.25psi upwards that will do it mechanically, and i'm sure there are digital probes that will measure specific gravity. If not you could make one with a float, a webcam and some crude image processing.

  • by tpstigers (1075021) on Monday October 01, 2012 @09:14PM (#41521033)
    is what technology is for.
    • by Anonymous Coward

      Bah, my girlfriend has had a yeast fermentation chamber in her Pi for many years now, without using any fancy technology.

      -- Ethanol-fueled

    • by bondsbw (888959)

      And completely free-as-in-beer.

    • I just don't like the term "BrewPi".

      Most beer drinkers will read it as "BrewPee"

      No thanks.

  • Still waiting for my Gertboard [raspberrypi.org]
  • For something like a fridge that has a lot of thermal mass I'd be tempted to ditch a PID approach and go for a PD controller. Using the PD configuration I've achieved great results with ovens heating large steel canisters and having them closely follow ramping and soaking profiles.

  • I'm looking in my local RadioShack and I noticed the newer Android-powered Arduino.

    Since I'm not seeing Android on the Raspberry Pi, what would be the advantage of the Android Arduino over the Pi besides the much higher price and Android operating system? The Pi already has a fairly useful Debian with X Windows which I didn't see on the Arduino.

    • by drinkypoo (153816)

      I'm looking in my local RadioShack and I noticed the newer Android-powered Arduino.

      I haven't noticed this. What are you talking about?

      Since I'm not seeing Android on the Raspberry Pi, what would be the advantage of the Android Arduino over the Pi besides the much higher price and Android operating system?

      No idea. The Pi is supposed to have Android, but they refuse to release it! They claimed the only problem with ICS was sound, but it's been months since that claim and there's been no release, no source, no howto, no nothing. This is badly, sadly disappointing. I thought Raspi was supposed to be an open platform (binary blobs aside) but there's been sadly little openness so far. It's very much been "when we're ready" and no status updates whatsoever. Tons o

      • he is most likely talking about the 276-0129 (retailing at $84.99) i would say very few RadioShacks have this IN STOCK but all of them should be able to order it Direct2U.

        (speaking as a former Tandy/Radio Shack employee)

        • by drinkypoo (153816)

          Oh, you mean the ADK. Yawn. Arduino with Android connectivity and an extra-big price tag. I thought it might have been something relevant, but it's certainly not an Android-powered Arduino.

  • Can this ferment viili [wikipedia.org] or filmj:olk [wikipedia.org]?
  • I know about PID and fuzzy, but what the hell is a Johnson controller?
  • A friend and I make fermented apple cider, but very primitively. Basically, we buy a glass gallon of unfiltered apple-juice, an airlock, a packet of champaign yeast, and some suitable sugar. We then take a few or so cups of the apple-juice and blend it with the sugar, bring it to a simmer, add the yeast, stir it vigorously, and pour it back into the original glass bottle, then finally attach the airlock/CO2-indicator. After about one week at approximately 76 degrees-or-less, we have a pretty good preservati
    • You can definitely make great cider that way (I've done it!). Instead of Champagne yeast, try Cote Des Blanc or an English ale yeast to get a bit of extra complexity (Champagne yeast is really neutral... an estery yeast really helps in a cider), and bump the gravity up to around 1.070-80 (maybe even 1.100 if you want to make an Applewine instead). You have to let it ferment for around 3 months (and rack it at least once! Sulfites are your friend), but the end result is pretty great with not much more effor

      • Excellent advice; I will definitely apply it. But apple honey mead, absolutely wicked! No, I couldn't wait for that. I would be counting bubbles until I went insane. Maybe if I could leave it in a different State while it aged.
        • by cayenne8 (626475)

          Excellent advice; I will definitely apply it. But apple honey mead, absolutely wicked! No, I couldn't wait for that. I would be counting bubbles until I went insane. Maybe if I could leave it in a different State while it aged.

          Nah...just make 2-4 batches each time you make it....drink 2 batches....put two batches back.

          When each batch is about ready...brew 2-4 more batches...that way, you are drinking the young stuff (to satisfy your thirst), but you are also putting stuff back...and if brewing every coupl

      • by mug funky (910186)

        wifey has a sulphite sensitivity. if i don't want to spend the night scratching her back, preservatives are out.

        makes things tricky - gotta be much more sanitary and also not leave things too long. fast ferments are good though (i'm trying my hand at a Caribbean rum now, as those ferment out in 3 days), but tend to limit the flavour profiles you can get.

        • by cayenne8 (626475)

          (i'm trying my hand at a Caribbean rum now, as those ferment out in 3 days),

          Can you go more into the 'caribbean' rum thing? Are you saying you're fermenting...the distilling something into rum?

          Where are you living....I don't know of any countries or US states really....where distilling, even for personal use, is legal.

          Not that I'm not wanting to try it myself...hope to soon, guessing that a small batch set up won't draw attention...but just curious as to what you're doing...?

    • by mug funky (910186)

      aussie excise tax on alcohol is 75 bucks per ethanol litre.

      putting that in perspective, if you're making a wine in a 30 litre fermenter, you'll be using about 5 kg of fermentables. this will yield about 4.5kg of ethanol, which will run you about 428 bucks of excise for 6 bucks worth of ingredients.

      • by cayenne8 (626475)

        putting that in perspective, if you're making a wine in a 30 litre fermenter, you'll be using about 5 kg of fermentables. this will yield about 4.5kg of ethanol, which will run you about 428 bucks of excise for 6 bucks worth of ingredients.

        Is this for the home brewer?

        Are you saying over there..they charge you all this tax for alcohol, on raw ingredients that 'could' be used to make DIY home brewed alcohol?

        If so...wow....

        Here in the US...the ingredients you buy are only subject to the usual sales tax any

    • take a few or so cups of the apple-juice and blend it with the sugar, bring it to a simmer, add the yeast, stir it vigorously, and pour it back into the original glass bottle

      Wait, you're adding yeast to hot cider? When brewing beer, you have to get your wort temperature down below 80 before you add your yeast. Sounds like it's working for you, though, maybe champagne yeast is tougher stuff than regular brewer's yeast.

  • I ran across the BrewBit [brewbit.com]. By the looks of the blog, it's pretty far along (photos of the circuit board) but I have no idea what it's going to sell for.

  • This seems like overkill. If you just want a web interface, there is an Arduino library that will allow you to establish a SLIP connection over USB to connect to the Arduino via a webbrowser, no ethernet shield required: SerialIP [arduino.cc]. It does take up a bit of memory, though, which leads to the next suggestion...

    How did you max out the program memory on your Arduino? Why on earth would you use 16 bits (0.001953125 dC precision) to store temperatures, when your thermometer only has +/-0.5 dC resolution? Does

    • Re: (Score:3, Informative)

      by Anonymous Coward

      Elco here,

      The Arduino could by itself send some simple values to a server. It could not provide a fancy interface. The thermometer has ha 0.0625 resolution, a 0.5 deg absolute accuracy. I am filtering the temperature data and need the higher resolution for the control algorithm. If you don't filter, and your sensor jumps from 20.000 to 20.0625, what's the slope?

      The program memory is full because of the 16 bit variables and that they are all configurable from the web interface. And no for brewing you don't N

      • Okay, take this with a grain of salt since I'm not really an expert on web design. But it should be possible to write your web interface in Javascript, store it locally, and use it to communicate via JSON with the Arduino over USB.

        If you're considering producing an Arduino shield, it's not too costly to add a bit of memory as well for data logging and storage.

        Just something to consider. It seems like it would make your design cheaper and more robust, for the cost of a bit of programming.

  • Also, I think you should generalize this project to include maintaining a bread yeast culture.

  • If they used a pentium they could warm the fermentation vats too ..... now I wonder if I could patent the first beer-cooled data center.
  • I love it, really, but has anyone else noticed that using a Raspberry Pi (instead of any Linux platform at all) is the greatest way to get publicity for your computer controlled hardware project?

    Brewing controlled by an HP netbook would never have made the headlines.

    • Meh, it is kind of a cool platform, in theory. For $25, being able to control some GPIO pins, connect via ethernet and USB, and display info on a TV are nice features.

  • by jkinney3 (535278) on Tuesday October 02, 2012 @06:56AM (#41523781)
    BEER is good. BEER should be studied in school. BEER should be required for graduation. BEER should be in the home. BEER should be promoted for the betterment of society. BEER is good for everyone of all ages.
  • Accuracy has nothing to do with which processor you choose or which architecture runs your control software.

    It has everything to do with the design of your plant and the tuning of your PID loops.

    You could control a well-designed brew plant to 0.1C accuracy using a bunch of op-amps if you wanted to. A brew plant function just isn't going to be so fast that you need any kind of horsepower or complexity to run it.

  • by Hillgiant (916436) on Tuesday October 02, 2012 @07:11AM (#41523903)

    Imagine a beowulf cluster of these.

"Well, social relevance is a schtick, like mysteries, social relevance, science fiction..." -- Art Spiegelman

Working...