RDWC Glycol Chiller Experiment

What is a glycol chiller?

This term generally refers to a chiller that consists of a reservoir of cooled liquid. Typically, the liquid consists of a mixture of propylene glycol and water.
The cooled reservoir solution is then pumped through a heat exchanger (cooling loop) to chill a target solution or the surrounding atmosphere (air conditioning).
In many applications, the chilled reservoir is cooled to a temperature that is much less than that of the desired target temperature. A control system is utilized to adjust the flow into the heat exchanger to manage the target temperature. By chilling the reservoir to a low temperature, the system provides a large heat capacity and allows it to quickly react to large changes in the target temperature.
An additional feature of a glycol chiller, sometimes found on chillers for brewing systems, is that they are able to control multiple target temperatures independently. For each target, a separate pump and control system is utilized to control the target temperature. For instance, two loops could be utilized with one loop controlling the temperature of a RDWC reservoir and the other loop feeding an air heat exchanger (air conditioning).

What is propylene glycol?

Propylene glycol is a synthetic, water-soluble, generally non-toxic, and easily metabolized chemical that has a variety of uses. One use is that it acts as an anti-freeze and is used, such as in this application, to reduce the freezing point of water. It acts as a safety feature to prevent freezing of the reservoir and potential damage to the chiller itself. It also has anti-bacterial properties and acts to buffer metals against corrosion. Do NOT confuse this with ethylene glycol which is quite toxic if ingested by humans, pets, etc.

The experimental set-up
In order to build a glycol chiller, there are a several pieces of equipment that are needed

  1. a chiller with sufficient capacity to handle the maximum expected heat load (in this case a TECO aquarium chiller)
  2. an insulated reservoir (camping cooler)
  3. a heat exchanger (such as a beer wort coil, use stainless steel for anything metal contacting the nutrient solution)
  4. a temperature PID controller with thermocouple
  5. submersible aquarium pump(s)

Other miscellaneous items include

  1. hose, tubing, or pipe for interconnecting the components
  2. interconnect fittings as needed
  3. bulkhead fittings for the reservoir (see Extra Long Bulkheads)
  4. gylcol (optional) + water

This is the chiller I choose to use:

The glycol reservoir is simply a modified camping cooler (the better the cooler, the better the efficiency):

Connections to the chiller are made and insulated (somewhat) with some neoprene:

And, the tubing is routed to a couple of bulkheads installed in the reservoir (camp cooler). The holes for the bulkheads were drilled using a step drill bit (a large one):

Within the reservoir we have:

In this case, you’ll notice that there are two small aquarium pumps. The TECO chiller does not have an internal pump. One of the pumps purpose is to recirculate the glycol reservoir solution through the chiller. This pump currently operate in a continuous manner. The chiller itself monitors the temperature of the solution flowing through the chiller heat exchanger and will cycle on/off automatically based on the set-point. I have the set-point set to 48 degrees Fahrenheit. This will be the temperature of the glycol reservoir solution.

The second pump is used to pump the chilled glycol solution to an external heat exchanger (located in the RDWC reservoir). This pump is controlled by a PID controller which will cycle the pump on/off (PWM). PID controllers are “closed loop” controllers which “learn” how the system is reacting to changing conditions. The PID controller adjusts the on/off times and durations in order to precisely control the temperature. The PID controller which I’m using is an Inkbird (controller works but there are better options):

The pump to the external heat exchanger is connected to a “wort” chiller coil located in the RDWC reservoir:

The semi-transluscent tubing is an air line that is described in another topic, here: Airpumpless Sluckets

And, the interior of the RDWC reservoir:

The glycol solution never comes into contact with the nutrient solution and, likewise, the nutrient solution never leaves the RDWC reservoir. The heat exchanger (wort chilling coil) does the work of removing heat from the nutrient solution and transferring it to the glycol solution. Notice that the heat exchanger coil which is in contact with the nutrient solution is stainless steel. Avoid other common metals such as copper as these can slowly dissolve into the nutrient solution eventually lead to plant toxicity.

The results

In this first graph you’ll see the environmental temperatures. The red trace is the temperature in the grow area. The green trace is the temperature outside of the grow area. For this test, the lights, in a closed room, were ramped to full power and then ramped down. You can see how it affected the room temperature:

In the second graph, you can see the RDWC solution temperature at the point where it returns to the RDWC reservoir (system set-point is 68 degree Fahrenheit):

You’ll also notice how the PID controller reacted to the change in heat load. When the heat load is removed (lights off), the temperature overshoots then undershoots while the PID controller tries to learn the new situation with the solution temperature.

Here is another example where the lights are being cycle on/off and random intervals:

And, the solution temperature:

In both cases, the temperature set-point is 68 F and is controlled to within +/- 0.5 degree Fahrenheit. Additional tuning of the PID loop or using a better PID controller can improve the performance further.

Do you need this level of control? Probably not, but if you are running experiments this appears to be a good way to control the temperature variable.


Nice! I run 2x glycol chillers for the 40-gallon conical fermenters we have in our homebrewery, but they are going straight to a fermenter water jacket. This is neat. If I end up loving dwc a similar build will be in the works :thumbsup:


It was a fun project. Not complicated to build but perhaps a bit overkill for normal RDWC use. @Larry3215 and I were trying some dissolved oxygen experiments and, as it turns out, we needed good temperature control in order to understand what the heck was going on in the experiments.

FYI, I need me a brewing set-up! Some day…


If you ever decide to go down the beer rabbit hole, hit me up. We’ve been doing this for 15 years now, got lots of links sorted out… down to CIP and inline wort oxygenation…


Nice one, I love the teko chillers, def. want one to my lil DWC tent.
Thanks for sharing and the details + pics

1 Like

Most certainly will. Thanks.

I had a choice at the time, either spend some cash on building out a new RDWC set-up or brewing. Opt’ed for the grow room and supporting equipment. When I’m flush with funds again, a brewing set-up will likely be on the agenda. :+1:


Very interesting @Northern_Loki

I needed one if these in my rwdc days. Well written and very informative.


As usual, an excellent write up! This (almost) makes me wish I needed a chiller. You’ve given me ideas for doing a cheapo version using an old mini-fridge or window AC unit and MacGyvering the crap out of it :slight_smile:


The most expensive part to this, by far, is the chiller itself.

I’ve seen some folk that’ll take an air conditioner apart and place the cooling coils directly into a reservoir for a more economical solution. Kinda ugly and could present some problems over the long haul but if it works why not.

I await someone with some brazing skills, can cobble together a homebrew A/C system, charge and balance it, etc. That would make for the ultimate purpose intent solution…

1 Like

That was in fact our first attempt. We took a window shaker apart and split it, plumbed it, and recharged with r12a (too cheap to pay an hvac guy to come over and do it proper under vacuum). That refrigerant is supposed to handle some abuse, right? Well…

…didn’t last long. Worked while it lasted though. Compressor took a dirt nap hard.

So we tried it again. Try anything once, twice to be sure, right? Still burned put the compressor…

Lesson learned - must actually evacuate windo shakers and fill with proper refrigerant :point_up:


I would be concerned about the copper cooling coils corroding in the nute solution. After my copper test, I want no part in that.

Darn! I was thinking of doing the same thing. I have a kit for re-charging auto AC units. Its very easy to DIY, so I figured why not do it for a window AC unit. I can solder/braze pretty well, so modifying the AC unit would be easy, but if it kills the compressor, that wont work.

My other idea was to just box in the cooling coils in an insulated chamber, ice chest, etc then run your glycol cooling tubing through that chamber and into the rez with a pump.

With a mini-fridge you could do the same thing. Put your glycol rez in the mini fridge and plumb that over to the rez. Or, just put the entire rez in the mini fridge. A small chest freezer should also work. Just stick the rez in the freezer.

Control temps by using an Inkbird to turn the fridge or freezer ON/OFF as needed.

You can pick up used chest freezers or mini-fridges really cheap sometimes.


I think the idea of using a mini fridge and keeping the glycol rez in it is by far a better option over frankensteining a system.

R12A works great in most cars, that is the exact setup we used - a kit for recharging cars. However I think they run kn R410(?) From the factory, and the lubricity/compressibility of the organic R12A is just not the same. We have not tried the R134a because it is way expensive and pretty much impossible to find in containers conducive to DIY. Could have tried butane, but that would be sketchy AF :wink:


Actually, why not use a chest freezer as the nute rez itself? They are plastic lined inside and most have drains built in. You should be able to find one used pretty cheap on Craigs List, etc. The 5 cu ft ones go for like $200-$300 new, so a used one should be much cheaper than a chiller.

It would be a bit more work cutting holes for pipes, etc, but not that hard. Just be careful not to drill through any pipes. wires etc. I would seal any holes, etc with 3M 4200 Marine sealant.

A cheap InkBird should work well to control it. They have built in time delays for switching AC units to avoid damage to the compressor from too rapid a re-start, so your all set.


That is a really freaking cool idea! No pun intended. We use a bunch of these as keg fridges with no issues on Ebay controllers. The only thing we would need to figure out is the rez level – the evaporator core is kinda woven around the upper sides of the units, I can see the frost up there, so perhaps elevate the buckets and keep the freezer topped up?

This is brilliant
I’m going to go invade Craigslist now


The freezer shouldnt run long enough to form frost - I think. Once things are stable inside, I think the controller should only be running the compressor long enough to cool things down just a bit, but not form frost - maybe.


What I meant is that I see frost inside my freezers when they are operating as freezers - when I get excited, or intoxicated, I forget details like that… this is one of those nights. Fine, and dandy :slight_smile:


A glycol chiller generally has two stages. First stage cools a glycol reservoir. This is where the copper coils would go. The second stage circulates the glycol solution through a heat exchange sitting in the nutrient solution. This would be stainless steel, for instance. Copper components in the nutrient solution is a not suggested, as you’ve noted.

P.s. I’m not promoting that idea, just saying that’s what some others have attempted. Also, consider that the standard refrigerator/freezer compressors are probably not designed for the duty cycle / heat load you’d be dealing with for RDWC (which has a lot of mass to cool). So, there’s a good chance things won’t quite work out but that’s the point of experimentation, I suppose.


Great write-up @Northern_Loki!

This actually made me think a bit about my planned design. I have the SS wort coil, chiller, and insulated res, but I had planned on using a single pump through the chiller + coil and back. I noticed you’re running the chiller at a much lower temp (48 degrees) and using the PID to cycle.

Just would like a bit more on your thought process there. I was planning on a single pump w/ the chiller set to around 65 degrees or so. I assume you keep a lower glycol temp to affect changes in the nutes faster when the pump is triggered?


When I was a kid, we had a used chest freezer that initially sat outside, on our car port, in Phoenix, Arizona until my dad enclosed that end of the car port. It was in full sun in the afternoons for several hours every day. Summer temps were often over 100 F. Even after enclosing the end of the car port, it must have been well over 100F in that space for large parts of the day in summer. We kept that thing full to the brim too, and it lasted for years.

I think thats a good point about the heat load, but Im wondering if its more heat load to maintain a 70F-90F differential on a freezer full of meat, or to just barely cool a larger mass of warm water? I think it will depend on how many gallons of water in the RDWC system and where you put the freezer.

If its inside the grow space, then it will have a harder time, because it will be pumping that heat right back into the grow room - making for more heat to get rid of. Plus it will heat your grow space.

If you can put it in a different space, out side the grow room or tent, then it will only have to deal with a few degrees of temp difference. Typical room tamps are in the mid 70’s, and you only need to cool the rez water to just under 70F…

Im guessing that large mass of water with a small temp drop will not be much worse, if any, than a smaller mass of frozen ground beef with a large temp difference.

Insulating the pipes and buckets, etc will help too.

1 Like

Yes, I would call that the standard way most folk would utilize a chiller. Perfectly acceptable.

Yes, pretty much the thought process.

I was looking to reduce the duty cycle (on/off cycles) of the chiller overall while having a relatively precise control of the temperature even under step load conditions. A couple of reasons why this is important in my case, I occasional like to run experiments. Being able to control a major variable (temperature) is very helpful when interpreting data. Second reason, I thought it would be neat to see how well one could control the temperature in a rather complicated thermal system. Third reason, ahh well some OCD maybe?

By having a relatively large delta in the two temperatures, the system can react quickly to step changes in the environment. The rate of heat flow depends on the temperature difference between two substances (the thermal gradient). If the thermal gradient between to substances is small, the rate at which heat is transferred is likewise small. The higher the difference, the quicker heat moves.

The other possible advantage to keeping a gylcol reservoir much cooler is if you want to control the temperature on more than a single cooling loop (while using a single chiller). For instance, I have a couple of Hydro Innovations IceBox heat exchangers (https://www.amazon.com/Hydro-Innovations-Ice-Box-6/dp/B002JLAC3I). I could utilize these with the current glycol reservoir to cool/dehumidify the grow chamber while at the same time controlling the solution temperature. This is with two different temperature set-points that could also be enabled/disabled independently from each other. Or, maybe chilling two independent RDWC reservoirs to different temperatures, etc. A variety of possibilities. Additional flexibility is the result. All that would be needed is an additional pump dropped into the same reservoir and a controller for that pump.

Perhaps, a minor advantage, is that the fluid flowing through the chiller is clean and non-corrosive. Better protection of the chiller, easier to keep the chiller heat exchanger loop clean and sanitary.

The disadvantages include 1) less efficient since we are keeping a large mass of liquid at a high differential to the surroundings. Recall, heat will want to move across a thermal gradient. The larger the gradient, the faster the transfer of energy. This is where good insulation on the reservoir matters 2) an additional pump = an additional point of failure to consider 3) more expensive

In reality, whether a glycol reservoir would provide benefit to your particular system is a question of your goals. For instance, if you don’t mind having a hysteresis of several degrees, you can reduce the chiller duty cycle (good for the chiller) with the disadvantage of the nutrient solution cycling across several degrees. And, perhaps at a slower rate of temperature response to step changes (e.g. at lights-on).

If you are looking to expand your overall system in some future way, need a certain amount of precision, or foresee a need for a second chiller in the future; then consider a gycol reservoir. That’s all I could think of at the moment.