Home Back to Blog Photography Sailing CV Contact
It’s no secret that temperature control is one of the key steps to making great beer, and it’s something my process had been lacking with some time.
I’d evaluated a number of approaches to this problem, with some important criteria. Whatever I chose had to be:
I’d evaluated a number of potential systems to tackle this problem, and ruled them out for various reasons.
I came across peltier cooling as part of researching this topic on Reddit’s Home Brewing subreddit.
Turns out an Intel Engineer called Rob Swanson had built a peltier cooled conical system based off his Stout Tanks Conical, and put together a PDF of the build[1]. This set the idea in motion, but I first spent many more months researching & flip-flopping between ideas.
I was very dubious of Peltiers, and had seen a lot of negativity around their use. In the end, it was Dan Curtis (another invaluable source in this build, his document[2] was a huge help) who convinced me to try it.
First up was deciding on a conical for the build. I decided to go with the Stout Tanks model Rob Swanson used in his original build. These conicals are great value, and Stout Tanks sell an off-the-shelf peltier kit for these. I bought mine used from Home Brew Talk classifieds, but afterwards regretted doing so. The conical was a little rusty when it arrived, and had no Stout Tanks branding at all. Stout claimed they didn’t even manufacture this particular model. Oh well - my one and only bad homebrewing classifieds fieds experience!
If you’re considering this build, I’d still recommend the Stout Tanks conicals - but the cooling method will likely work on any conical tank.
First was a way of building a temperature controller capable of heating & cooling the conical. Ranco do a pre-wired two-stage controller, but it’s $150.
Instead, I went the well established homebrew route of building an STC-1000 temp controller. I bought a flashed STC1000 which gives neat brewing temperature profiles from The Black Box, and wired this myself using readily available online tutorials.
Black Box also sell a pre-wired version - it’s rarely ever in stock, but if it ever is, this is the route I’d take. I’m no electrician, so to wire it up took me a few hours.
My finished temperature controller
Now, the good stuff - how to build the peltier cooled conical.
This parts list assumed building 2X cold plate heatsinks each with 2x peltiers, which should be enough to see 40f drop below ambient.
Kudos to Dan Curtis’s document[2] here, almost all this list is from his document.
You can skip these parts if you buy the Stout Tanks pre-milled peltier kit, but you’ll still need everything above.
This is by far the most difficult part - to mill these parts takes an experienced machinist, which is very difficult to come by. Reminder that Stout tanks sell a pre-milled heat sink & fan combo to fit their tanks - it’ll likely work out more expensive, but if you have one of their tanks, buy this and skip to the assembly step.
I couldn’t, since I wasn’t sure my tank was made by Stout.
Finding somebody to CNC mill this took me almost a month of dead ends, but I eventually found an incredibly helpful Reditor to mill the parts for free - thanks William!
Had I to go to a commercial metal shop, this project would be dead in the water. It would have almost certainly been cost prohibitive, and I can’t caution this enough - find somebody willing to help with the CNC milling before beginning this project. The fit between the conical and the metal heat sink needs to be really precise.
With that caveat out of the way - the aluminium needs to be milled to fit the curvature of the conical. As part of this missing process, we also measured the screw hole pattern of my CPU fans, drilled & tapped holes for these as well.
We also included a notch in the edge of the aluminium heat sink to loop the strap into. All this is done for you using the Stout Tanks part.
Measuring the aluminium against the tank for fit before milling - this is the part that needs to be curved
CNC Milling 1/2 of the the heatsink parts
Now that the parts have been milled, before fitting to the tank, it’s time to assemble the full heat sink.
The PSU is almost ready to go - you’ll just need to create a jumper on your PSU to cause it to start without a motherboard - see here for instructions, skip the switch.
Then, snip the black molex connector (it’ll melt under the heat) from the CPU cable of the PSU. This provides 4x black-yellow 12v rails for the 4x peltiers.
The CPU Molex connector
At this point, it’s a good idea to wire one-by-one each of the 4 peltiers, to test and verify they cool OK. Only run for a very short period of time, seconds. You should be able to identify the hot and cold side - but be careful, these get hot!
Parts needed for this step - CPU coolers, peltiers and fans. Not pictured: thermal paste
Depending on your aluminium blocks, and how new your CPU coolers are, now might be a good point to lapp both the cold plates and the copper bottom contact surface of the CPU cooler. Start with 500 grit, then 1000, then 2000 grit sandpaper. Clean off any residue, and we’re ready to assemble.
You’ll need to cut the insulation to fit - I used an Exacto Knife.
That should be it - your peltier cooled conical is ready for cooling!
Finished, but before tidying up the wiring
Now that the cooling setup is assembled and wired, it’s worth tidying up the wiring with cable ties. I’d like to add small junction boxes at some point to my conical, but haven’t gotten around to it.
Now that I’ve installed the cold plates, my heating pad no longer fits - I need to find a smaller one, but this isn’t a priority for me - my ambient is 70f year-round, so I never need to heat.
I filled the conical with water, and tested by running for ~24hours. I don’t have the ability to graph the temperature drop, but I’m able to get a ~33f drop from ambient. I should be getting closer to 40f, and this I’ll try and figure out in coming months.
This means, without having the fans run an unreasonable amount (i.e. not constantly), I can chill 70f water down to ~40f.
I could probably ferment for the first stages of a lager, but could not lager longer-term, something I hope to fix.
Hopefully you’ll see better results still!
Pressure transferring wort under CO2 - no oxygenation, little sediment.
An over-active airlock
Here’s a list of troubleshooting steps I performed when at first I was seeing sub-optimal performance. Again, big thanks to Dan Curtis who exchanged many emails with me helping debug.
While I’m very pleased with how the conical is working out, I think there’s room for some improvement. Here’s inspiration for anybody taking on this project to learn from my mistakes!
Current setup:
+----------+ +----------+
| | | |
| +------+ | | +------+ |
| | | | | | | |
| |Pelt | | | |Pelt | |
| | | | | | | |
| +------+ | | +------+ |
| | | |
| | | |
| | | |
| +------+ | | +------+ |
| | | | | | | |
| |Pelt | | | |Pelt | |
| | | | | | | |
| +------+ | | +------+ |
+-+------+-+ +-+------+-+
Proposed improvement:
+----------+ +----------+ +----------+ +----------+
| | | | | | | |
| +------+ | | +------+ | | +------+ | | +------+ |
| | | | | | | | | | | | | | | |
| |Pelt | | | |Pelt | | | |Pelt | | | |Pelt | |
| | | | | | | | | | | | | | | |
| +------+ | | +------+ | | +------+ | | +------+ |
| | | | | | | |
+----------+ +----------+ +----------+ +----------+
Overall, I’m very pleased to have gone the peltier cooling route. It suits my needs - particularly the size constraints.
If I lived in a house, or had easy access to a basement, I might not have gone this route. I’d much rather use a large old fridge or chest freezer to control temperature - it’d be much cheaper, and no doubt easier to build!
None the less, I’ve no doubt this is something I’ll continue to use wherever I live. It’s so incredibly convenient to be able to set a temperature profile on the black box, walk away, and forget about the beer until fermentation ends.
Again - my eternal gratitude to both Dan[2] for his excellent document, from which I sourced all my parts - and William, who took a good chunk of his Saturday to mill my cold plate.
I’d strongly recommend this as an approach to temperature control - it’s worked out great for me, and hopefully more people become interested in this as an approach.
[1] Rob Swanson’s guide, the document where this all started. http://conical-fermenter.com/TemperatureControlledBuildSetup.pdf [2] Dan Curtis’s guide, where I learnt how to do this. https://dl.dropboxusercontent.com/u/1268608/Peltier%20Fermenter%20Cooling%20Project.pages
**Other useful resources: **