Sam Stelfox

Thoughts from a software engineer, systems hacker and Linux gubernāre.

Glycol Cooling

DIY Glycol Power Pack (Chiller) for Beer Line Cooling

This content originally appeared on the following link: I'm glad I mirrored it as the articles have since been returning 404s.

First Post

When planning for my nano-brewery, I broke the project down into three categories: brewing, fermenting and end product storage/serving. While waiting on some parts to get the brewing operation ready to go, I decided to move forward on the serving portion of the project. Rather than bottle the beer - which would be a daunting task given the volumetric capacity of my brewing equipment - I decided to go straight to kegging. Many homebrewers end up kegging once their operations get more formalized, so I figured I would just skip to the end.

Since I currently don't have any product ready to keg, I put a local craft brewery's beer on tap. I quickly realized that the 10 foot run from my DIY kegerator to my tap tower was allowing the beer to warm up too much giving rise to foam and nasty warm beer!

In commercial tap systems, a refrigerated solution of propylene glycol is circulated along side the beer lines to keep the product cool in transit. I decided to replicate this method. After three attempts, I finally have a system that is working perfectly and replicates every aspect of a commercial system except one: the cost!

One of the primary features of a glycol cooled tap system is the trunk line. Generally, a trunk line consists of some number of product lines, wrapped together with a glycol supply and return line, then wrapped in an insulated jacket. I have five 5/16" beer lines and two 1/2" vinyl tubes for the glycol. These are wrapped together using cellophane packaging tape, then inserted into a length of pipe insulation. The pipe insulation has 1/2" thick walls.

Based on the length of the trunk line run, the insulation's thermal conductivity, and the tap box's heat gain characteristics, I calculated a very rough total heat gain of 250 BTU/hr. This is the amount of heat I will need to pull out of the trunk line and tap box in order to keep the product at 35 degrees from keg to tap. I chose 35 degrees so that once poured into a room temperature pint glass, the end result will be 42 degree beer.

The first iteration of my glycol cooling setup featured a simple reservoir in the kegerator in the form of a 5 gallon plastic pail. A submersible pump would send the glycol up a 1/2" vinyl tube in the trunk line, circulate it through a coil in the tap tower, then return it to the reservoir through a second tube in the trunk line.

The glycol started out at 35 degrees (the set temperature of the kegerator) but quickly rose to near ambient temperature. I suspected the submersible pump was adding heat to the glycol, so I switched to a different pump: a Taco (pronounced 'take-oh') cartridge pump designed for hydronic heating applications. Although still a wet-rotor pump, which I knew could impart heat to the coolant, it was not submersed. However, I got a similar result. Not having another pump handy, I decided to abandon the reservoir idea and go with an air-to-glycol heat exchanger in the kegerator.

Using an old motorcycle oil cooler and three high velocity computer fans, I constructed a forced air heat exchanger. Again, I used the same Taco pump and managed to achieve slightly better results. I measured about 44 degree coolant temperatures at the tap tower, but the temperature differential was insufficient to affect a timely drop in temperature in the product lines or tap tower. Additionally, the kegerator would only shut off for a few minutes each hour, when previously it would only run for a few minutes each hour to keep the kegs cold!

At this point, I knew that the heat exchanger was working and suspected the pump. Knowing that the next iteration of this project would take me the route of a compressed refrigerant route, I decided to go ahead and buy a new pump to see if the Taco pump was introducing an unacceptable amount of heat.

I selected a Little Gaint 2-MD pump as it is rated for mildly corrosive liquids, has an air cooled motor, and physical separation between the pump head and motor. It was the most cost effective option before going to something more exotic.

After testing the air-to-glycol setup again with the new pump, it was clear the Taco pump was not my limiting factor as I got nearly the same results with the new pump. I suspect the kegerator (a converted chest freezer) was simply not designed to vent the added BTUs the heat exchanger was introducing. Afterall, the interior and exterior of a chest freezer have limited surface area when compared to an actual air-to-liquid heat exchanger. Rather than chase my tail further, I began to explore the compressed refrigerant option.

Others in the home brewing community, as well as those looking for DIY options for aquarium chillers, have converted window A/C units for use as liquid chillers. Even small window air conditioners are rated 5000 BTU/hr and greater. Given I needed to move approximately 250 BTU/hr, I found that to be overkill. Had I went that route, I probably would have ended up with something only running for a few minutes each hour, and as any mechanical engineer will tell you, compressed refrigerant systems are most efficient in steady state running and not cycling frequently.

I did however find one web site that gave me the idea of converting a dehumidifier instead. Much smaller in regards to BTU capacity, and without the potentially troublesome fins on the evaporator (cold) coils, I immediately decided to attempt this route!

I began by disassembling a 20 pint/day that I had laying around. It was too small to effectively dehumidify my basement, so it was no real loss if it didn't work. Normally the evap coils are infront of the condenser coils. Ambient air is drawn over the evap (cold) coils, thereby condensing water out of the air, then through the condenser (hot) coils to cool the compressed refrigerant, then exhausted to the room.

To directly utilize the cold produced by the evap coil, I carefully bent the coil down and into a 10 qt cooler that would serve as the glycol reservoir. The aluminum coils were very easy to bend. It freaked me out a little, so I was wearing a face shield and heavy welding gloves just in case the R134a got "liberated"! Regardless, some gentle persuading and the end product looks like this:

Second Post

The initial test was to hot wire the dehumidifier controls and turn it on with a tiny fountain pump in the cooler circulating the glycol. In 20 minutes it had dropped the temperature of the solution to 18 degrees. At 30 minutes, the solution began to slush at 9 degrees. There was also no lid on the cooler at the time, so I don't know how much heat gain I got from just ambient air. Regardless, a casual calculation based on temperature change put the unit at around 1200 BTU/hr, meaning I should be able to run at about a 20% duty cycle to chill my lines.

Having had success with the initial test, I cut the lid as necessary, installed plumbing and hooked it up to the pump. After only 10 minutes of running, my glycol line temp in the tap tower was 27 degrees. Keep in mind though the glycol was still chilled from the first test. In fact, I had transferred it back to the jugs so I could do my plumbing and such, and after an hour of sitting in the jugs, the surface of the jugs were still cold enough to form frost!

After 35 minutes of running, the glycol was 20 degrees and the beer line was

  1. At that point I started to manually cycle the dehumidifier on and off (so as to not freeze the beer), and ended up achieving 35 degree beer temp at the tap after about 50 minutes. Poured into a pint glass, a perfect 42 degrees!

I've nicknamed my creation R2-GlyCool, since I think it looks a little like the famous Star Wars droid R2-D2, but with a "beer belly" (the red cooler/glycol reservoir). I still need to insulate the glycol lines that supply the trunk line as they form an amazing amount of frost, and of course, waste energy.

Third Post

With regards to power consumption, the pump has a steady state draw of about 100 watts. The dehumidifier has a steady state draw of about 220 watts. Currently, the pump runs 100% of the time (which I don't expect to change), and the compressor with uninsulated lines runs 50% of the time. That averages out to 153 KW-hr/month or about $16/month at my current rates. Naturally that will drop once the lines are insulated.

The reservior temperature is regulated by a very simple PIC microcontroller, 1-wire temperature sensor and solid state relay. I plan to improve on this design so as to run multiple pumps to chilled fermentation vessels for hefeweisens and lagers.

I also have a redesign of the tap tower in mind that, by coupling the lines more closely and eliminating air space, will further improve efficiency and temperature stability.

I should also note that I had a brief flirtation with a thermoelectric Peltier cooler in the tap tower that totally didn't work. Without forced air (resulting in noise) over the heat sinks, it performed very poorly. I didn't include details on that in this post as it was such an under performer, I don't think it would ever be suitable for this application, nor was there anything more to learn from that experience than that.

Update 3/30/2011: The lines are now fully insulated and the compressor is cycling on and off at only a 16% duty cycle! Now if only I could find a more efficient way to circulate the coolant....