HERMS Design Specifics

Ok, after considering the Design Logic, I decided to build a HERMS based system with a stand alone heat exchanger. The design of this heat exchanger will be the key to the system. In practical terms, the heat exchanger should respond to heat input quickly, require less than 20 Amps of current on standard 120 V household voltage, and provide enough heat to heat a mash for a 10 gallon batch. These requirements immediately rule out using the HLT as the heat source. I know this goes against conventional wisdom, but let me explain. The mass of water in the HLT would require so much electrical power to heat and respond quickly, that it would just make it impractical. A good heat exchanger would be of a fairly small volume so as to get the most use out of heating elements that fall within the electricity limitations of 120 V household voltage. Now, many homebrewers manually heat up the HLT with propane and use the controller to control a set of solenoid valves to activate a "by-pass" loop around the coil to control wort temperature. I did not like this approach. I think it adds extra plumbing that isn't necessary and eliminates the temperature controller from directly controlling the temperature of the wort. A by-pass loop is a band-aid for a system with a poorly designed heat exchanger (yet another biased opinion from yours truly).

After some quick calculations, I determined that about 2000 Watts was the max power I wanted to deal with as far as heating elements. This makes the current draw about 16 - 17 Amps. Any more than this and I'd probably trip a breaker. So now, I needed to select the actual size of the container. I first thought that a 1 gallon paint can would be a good vessel. This seemed to be a practical size and something I could lay my hands on pretty easily. It was, until I tried to put standard sized heating elements in it and found that the tips of the elements stuck up above the rim of the can. That wouldn't do at all. The whole element needs to be submerged otherwise the element cannot dissipate it's' heat and it will burn up. I needed a slightly taller can. I was able to find a 5 quart container that had the same diameter as a paint can but was about 2 - 3 inches taller. Perfect. Once I put the coil in, I would end up with about 1 gallon of water. As a rough estimate of the power of the heat exchanger, you can calculate the Watts/gallon of water. In my case, I had about 2000 Watts/gallon. A similar power source for a 15 gallon HLT would make the power requirement 30,000 Watts pointing out the impracticality of using the HLT and trying to quickly change the water temperature with electricity (NOTE: 30,000 Watts equates to about a 100k BTU/hr burner, which is what most people using the HLT use to heat their water, quite a coincidence, eh?). Now, due to the small mass of water of my heat exchanger, the water will cool off more quickly as heat leaves via heated wort and heat loss to the environment, however, this merely means that the elements may have to fire up a little more often - but the water will react much faster to heat input because of the massive Watts/gallon ratio. Another important aspect to heat exchanger performance is water agitation. Definitely make plans for agitating the heater water to achieve optimum heat exchanger performance. This is basically something to constantly stir the water. I have details on my methods for this on the Heat Exchanger page. This makes a huge difference in your heat exchanger efficiency, don't overlook it.

Now, how much coil should go in. For this, throw away the calculator. The answer is, as much as you can stuff in there. Just remember, everything has to fit, the elements, the stirring mechanism, and the coil. Your choice of coil size will be a determining factor. In practical terms, you will choose either 1/2", or 3/8" copper tubing. Anything larger is overkill, anything smaller just won't allow enough flow. Don't bother with stainless steel, the thermal conductivity of copper is 30 times greater, plus, stainless is an absolute bitch to work with. 1/2" copper will provide more flow, but you won't get as much length in the same space. Plus, you will be limited in flow rate to about 1 gallon/minute due to the risk of grain bed compaction. I chose 3/8" and managed to get about 29 feet of tubing into the can. Temperature measurements show that the wort exiting the coil is about 3 - 5 F below the actual water temperature at full pump flow. This is pretty good, I wouldn't worry about getting much better than this. If it turns out that yours is higher than this, don't worry about it, your heating elements will just have to heat the water a little more, no big deal.

Now comes the issue of temperature probe placement. This is an issue of some fierce debate, and of course, I think I'm right in my opinion on this. In my opinion, you want to measure the temperature of the wort exiting the coil. "But, why wouldn't you put the probe somewhere within the mash, after all, isn't that the temperature you are concerned about?" Yes, it is, however, remember that we are dealing with a recirculating system. As long as you don't stir the mash (there are good reasons for not stirring, see HERMS Limitations), and have a good return manifold that gently deposits heated wort on top of the mash bed, ALL of the wort will be recirculated every X minutes (wort volume/flow rate). A probe half way down in the mash is seeing wort that was heated several minutes ago. If your mash temperature is 148 F, and your controller is set at 150 F, your probe calls for heat. It will continue to heat until it sees the proper temperature. By the time the probe sees 150 F wort, the heat exchanger may have heated the wort exiting the coil to 155 F, but the probe doesn't know it because it's seeing wort that was heated a few minutes ago. When the probe finally sees the 150 F wort and cuts off the heat, the wort temp around the probe will actually increase for the next few minutes because the probe is now seeing the overheated wort. By the time it sees wort at 149 F and calls for heat, the wort exiting the coil my be down to 145 F and the cycle repeats itself. The farther you put the probe away from the heat source in the direction of wort flow, the larger the time lag will be and so will the potential for temperature swings during mashing. The absolute worst case for this is locating the probe at the mash tun exit.

Now, it is true that when using a PID type controller, the controller can be programmed to "learn" your particular system and anticipate the entire system's reaction to heat input. This means that the PID knows that there is a time lag and adjusts itself to anticipate the scenario I described above. Problem solved, right? No. The PID sets its parameters to anticipate the system's reaction to heat input based on a fixed set of variables that include flow rate, batch size, and heat loss to the environment. If the flow rate from batch to batch is faster or slower and the probe is half way down into the grain bed, then the PID will see temperature changes faster or slower than it's supposed to and will anticipate the system's reaction to heat input incorrectly, leading to temperature swings. Likewise, a larger or smaller batch size will also affect the PID's ability to anticipate the system's reaction to heat due to the larger wort size. This would be particularly bothersome if you find yourself switching between 5 and 10 gallon batches, but even going from an English Mild to an IPA would be troublesome simply due the differences in the size of the grain bill and differing amounts of water added. Lastly, the ambient temperature, such as brewing in the winter vs. the summer, can affect the heat loss from the mash tun and also confuse the PID, although a well insulated mash tun can minimize this greatly. But, if the probe is located right at the coil exit, then all of these factors are moot.

So, in summary, HLT's make lousy heat exchanger vessels if you want your controller to directly control the wort temperature. A well designed heat exchanger is small, well insulated, and has a Watt/gallon ration in the 1500 -2500 range. The coil in the heat exchanger is made of copper and is long enough to provide around a 3 - 5 F differential between the heat exchanger water and the exiting wort temperature (this differential temperature is not that critical, if it turns out that its more than that, your elements will just have to heat up a little more, no big deal). Lastly, the temperature probe is located at the coil exit which minimizes the temperature swings in the mash and eliminates nearly all of the variables that can affect the PID settings from batch to batch such as flow rate, batch size, and ambient temperature.