Peltier TEC Carboy Cooler

[nayr14]

Before I started brewing my first beer, I wanted to setup a way to cool my carboy down to a desired temperature. I looked into purchasing a refrigerator but I don't really have room in my apartment for an appliance like that. So instead I built this!

I purchased two used CPU heatsinks off Amazon (http://www.amazon.com/gp/product/B002G1YPH0/?tag=skimlinks_replacement-20) placing my water block (http://www.amazon.com/gp/product/B00DHC2GJ6/?tag=skimlinks_replacement-20) and two peltiers between each side of the water block and the heatsink, thermal paste on each side. I ordered a 12V 30A power supply to run all this. One pelteir is a TEC-12706 which uses 60W @12V. The other is a TEC-12715 which uses 140W @12V; both pelteirs can run @ 15.4V. I also purchased a "buck step down voltage regulator" (http://www.amazon.com/dp/B00C4QVTNU/?tag=skimlinks_replacement-20) to drop the voltage to control temperature.


Currently I'm able to maintain the water temperature surrounding the carboy at 62F (75F ambient) with only the more powerful cooler running @ 4.6V through the step down converter. With the two fans running (one in the power supply, the other on the heatsink) I'm drawing 40W according to my Kill-A-Watt. The yeast temperature range was 60-72, so I figured 60-62F water surrounding the carboy would be sufficient.

I'm monitoring the temperature via my Raspberry Pi which writes to a file that I can view via http.

After I ordered everything I found this article which helped me (attempt) to make calculations: http://www.shine7.com/aquarium/chiller.htm The digital temperature controller is a POS, it reverses the polarity immediately when the temperature drops out of range and heats the water with the same current!!

The next step might be staging different voltages with a relay board for the Pi so I can maintain temperature in a varying environment.

 

[Porchfddler]

I don't have much to comment on, but I found your thread interesting! I have a cask ( pin) which I had considered making a peltier cooler for as in this posting http://thedatastream.4hv.org/beer_cooler.htm. I got a little bogged down with all the specs so I never made it. You seem do have done a great job.

 

[mattd2]

Are those CPU heatsinks using heattubes? That is pretty sweet. What is the most efficient voltage to run the TEC at, and would you consider changing the control methodology to having the pump/TEC run at the best efficiency point for short bursts to keep the temp in check (like a fridge/freezer works)?

 

[jakenbacon]

Hell of a first post, you have been officially welcomed fellow HBTalker!

 

[MindenMan]

Amazing! I had thought of using Peltier heat pumps, in a similar way. As I remember peltiers are only able to raise or lower a temperature a finite amount. It appears you have worked around that problem.

 

[Unknown303]

Nice work! Have you considered using diodes to control current flow on the POS temp controller? Then you wouldn't get the effects of the reverse polarity.

 

[nebulous]

i looked at doing something TEC based a while ago because i had a bunch of water cooling parts laying around but it didnt develop very far. i guess what im most interested in what kind of temperature fluctuation range do you have and does the system run thw entire time by just adjusting the volatage to the TEC or does the whole thing ever fully turn off? i would also assume the more insulated the fermenter and its water bath are the more efficient this set up would be? but as a bit of a noob on hbt i must agree; one hell of a first post

 

[nayr14]

Are those CPU heatsinks using heattubes? That is pretty sweet. What is the most efficient voltage to run the TEC at, and would you consider changing the control methodology to having the pump/TEC run at the best efficiency point for short bursts to keep the temp in check (like a fridge/freezer works)?

Yes they are used from Amazon. They almost fit the peltier perfectly. I worry more about using the peltier at high power and low efficiency then at low power and low efficiency. Currently it almost maintains temperature on it's own. I've been running it around 3.2V to maintain ~65F. I was thinking of staging two different voltage regulators, one at 3v and another at 5v and write a program to switch between the voltage regulators using a relay board for the Pi.

The most efficient voltage can be calculated by using this: http://www.overclock.net/t/249758/ultrasonic2s-tec-calculator. If you check the chart

TEC's take some research to understand everything. I jumped in without doing much technical research and wasted a few bucks.

Nice work! Have you considered using diodes to control current flow on the POS temp controller? Then you wouldn't get the effects of the reverse polarity.

I have, originally I though the amount of power needed would be too much due to a bad peltier I was using. I opened it up to see if I could wire around the relay that reverses the current but I gave up. I rather use the Raspberry Pi and a relay board with multiple voltage regulators to switch between (or switch on the second peltier) to cool the carboy.

i looked at doing something TEC based a while ago because i had a bunch of water cooling parts laying around but it didnt develop very far. i guess what im most interested in what kind of temperature fluctuation range do you have and does the system run thw entire time by just adjusting the volatage to the TEC or does the whole thing ever fully turn off? i would also assume the more insulated the fermenter and its water bath are the more efficient this set up would be? but as a bit of a noob on hbt i must agree; one hell of a first post

Here are the temps over the last 16 days taken every minute(ignore the drops):

In the last week I've been able to control the temperature more accurately. The large spike was when I transferred it to a secondary. Minor changes to the voltage caused temperature changes once I had the temperature under control, but most of what you see is room temperature fluctuations. I might insulate the water block.. one issue is that when the other peltier isn't on the other heatsink absorbs the cold from the water block which is very inefficient :\

 

[stamandster]

Cool! Could you use glycol instead and seal it up?

 

[mattd2]

...I worry more about using the peltier at high power and low efficiency then at low power and low efficiency. Currently it almost maintains temperature on it's own. I've been running it around 3.2V to maintain ~65F. I was thinking of staging two different voltage regulators, one at 3v and another at 5v and write a program to switch between the voltage regulators using a relay board for the Pi.

The most efficient voltage can be calculated by using this: http://www.overclock.net/t/249758/ultrasonic2s-tec-calculator. If you check the chart

TEC's take some research to understand everything. I jumped in without doing much technical research and wasted a few bucks.
...

Yeah I did a little bit of research into TEC a few years back - Deffinately a lot to learn and nearly an art to size them
Looking at the graphs in that link it does look like they are more efficient at a lower voltage - I guess you don't move as much heat but you produce way less doing the amount you do.

 

[Trillium]

Very interesting and well done. I though about doing something similar for a glycol chiller.

 

[PDX_T]

Looks like you've gone to a lot of trouble, so why not buy a thermowell and monitor internal beer temperature directly instead of jacket water temperature?

 

[nayr14]

Looks like you've gone to a lot of trouble, so why not buy a thermowell and monitor internal beer temperature directly instead of jacket water temperature?

I want to monitor it digitally but since this was my first brew I didn't want to risk anything. I need to find a way to run the fairly thick wire into my carboy. It won't fit through the airlock.

I just found this:
https://www.homebrewtalk.com/f11/keeping-hood-thermowell-carboy-299764/index2.html

Does anyone have a better suggestion for a stopper?

 

[Koffie]

Interesting

I'm also working on a peltier cooled fermentor, but I want to pump the cooling water through a cooling spiral which is hanging directly in the wort.

 

[PDX_T]

I want to monitor it digitally but since this was my first brew I didn't want to risk anything. I need to find a way to run the fairly thick wire into my carboy. It won't fit through the airlock. I just found this: https://www.homebrewtalk.com/f11/keeping-hood-thermowell-carboy-299764/index2.html Does anyone have a better suggestion for a stopper?

My LHBS sells a two hole stopper that has a stainless steel tube fused into one hole, and you put an airlock in the other. the tube is capped at the end. you put your probe in the tube and stopper your carboy with a normal airlock. Not sure if that's what you're talking about. It's simple, elegant, and easy.

 

[Koffie]

I made something like that for my same project: http://www.goudenswaen.nl/gallery/yeast-2000/

 

[nayr14]

I made something like that for my same project: http://www.goudenswaen.nl/gallery/yeast-2000/

Looks good!

I started another batch and while it was rapidly fermenting (a day or two after pitching) I sanitized and dipped a thermometer into the carboy to measure the temperature. I was only reading ~1F (0.3C) more than the external water temperature.

I now feel safe not placing a thermometer in the carboy... the next step is a digital potentiometer controlled by SPI on the Raspberry Pi to replace that analog pot that I have to use a screw to adjust. That way I can write a simple program that looks at the last few temperature readings, checks the last few actions (step up/down voltage) and checks a designated temperature to adjust the voltage of the the peltier.

 

[Koffie]

The pics i posted shows a 1-wire DS18B20 sensor connected to the RPi. I plan to connect 2 more 2 sensors, 1 for ambient and one for the waterreservoir. My waterreservoir is about 4 to 5 liters big, and then it is pumped through a very large double spiral (something like this one: https://tapclean.nl/index.php?action=article&aid=13800&group_id=10000245&lang=nl#.Un3dfPnmP6c).

Why would you step down the voltage? If your amount of water is big enouh, you should keep a stady temperature.

I'm very curious about your plans and programs for the RPi peltier system!

At the moment I'm writing a program for the RPi to control and monitor the peltier fermentation. Everything is breaked down in small steps.
You start with scheduling a job. A job is unique for the wort you brewed. Once the job is finished it is never used again. However, a job points to a certain program. A pprogram can be used very often, it is even possible to control 2 peltier system from one RPi running 2 different jobs pointing to the same program.
A program is made of rules. A rule tells the system what the setting temperatur is, and how long this should be kept. When the rule is finished, the program flows into the next rule, until the program is finished.
Every sensor/device is logged (wort/ambient/water temps, peltier cool or heat, pump on/off) so you can plot a very detailed graph for the whole fermantation.

Biggest problem for me at the moment : I'm waiting from the begin of Oktober for delivery of my relays to connect to the RPi and peltiers/pumps

 

[nayr14]

I want to control the temperature of the water if the ambient temperature changes. I could turn the peltier on/off with a relay, but currently I use a 'buck' voltage regulator to turn up/down the voltage to control the power of the peltier to meet a certain temperature. It would be nice to have it run constantly, assuming ambient temperature will always be higher than the target wort temperature, and adjust the voltage to regulate the temperature.

Your program sounds interesting, I might try something similar if I have a chance.

 

[Koffie]

But that would assume you allways need to cool the system? What if you want to heat the system?
If I meassure the temperature of my wort direct by inserting a DS18B20 into the vessel, it is not allways higher than ambient temp.
A Colon of 30 liters wort is very steady in keeping temperature. Dropping ambient temperature by 3 degrees, will drop down wort temp by only a few numbers after the period in a couple of hours.