Thermowell vs Taping Probe to Outside

[srice]

I recently switched my fermentation freezer over to a BrewPi from an STC1000 based controller. At the same time, I added a thermowell to my 30L Speidel fermenter. Previously, I had just been taping my probe to the outside of the fermenter and covering it with a piece of 1/2" packing foam. The BrewPi gives me great data logging and I noticed that it had an option for a 3rd sensor - a room temp sensor - that would be logged, but would not affect the temperature control. My setup for this experiment involves 3 sensors with the first 2 being used by the BrewPi to control the system. The first is the chamber sensor- it is simply hanging free in chamber and responds very quickly to the air temperature in the chamber. The second sensor is the beer sensor and it is inserted into a thermowell extending into the fermenting beer. The third sensor is the labeled as room temp, but I taped it to the side of the fermenter just for data logging. Here is how I connected it. A single piece of duct tape was used to adhere the sensor to fermenter. I then covered the sensor with a piece of foam held in place by more duct tape. The third pic shows the fermenter in the chamber with all 3 sensors available.

My ground water was a little warm yesterday so I was not able to get the wort all of the way down to pitching temps. I let the chamber finish the job for me. The blue line is the free hanging probe in the chamber, the green line is thermowell probe and the gray line is the probe taped to the outside of the fermenter. There is a little spike about 2-3 hours after I started the data logging - that is where I added oxygen and pitched my yeast. The BrewPi does an excellent job of holding the thermowell probe temperature at the setpoint. The probe taped to the outside of the fermenter has about a +/- 1 degree F swing around the setpoint. I opened the fermenter a couple of times and used my trusty Thermopen to check actual temperature and it is right at the setpoint of 60F.

The one test I did not run is to control the system with the probe taped to the outside of the fermenter and just monitor the thermowell temp. I may switch the probes around after the fermentation is complete. I suspect that they will track very closely to each other since the thermowell is in the middle of a much larger thermal mass.

From what I see, I don't think it much matters if the probe goes in a thermowell or is taped to the outside. I will stick with the thermowell since it is so much easier than taping a probe to the fermenter. There are no wires attached to the fermenter when I am moving it around. Once I get the fermenter in place, I just drop the probe down the thermowell and I'm ready to go.

The BrewPi I built is an excellent tool. I highly recommend putting one together.

 

[atoughram]

On my 17gal conical the thermowell made a much bigger difference. Maybe I'll try your experiment next time I have a batch working.

 

[ASantiago]

This is an awesome post! Thank you very much for taking the time.

I'm trying to get some understanding of the conclusions from your experiment. My main question is: Is it better to control the cooling/heating based on ambient temperature measurements or direct beer measurements (like through a thermowell)?

I ask because I've tried both and eventually settled on ambient measurements. I did this because I noticed that thermowell measurements resulted in significant lags in control, meaning cooling/heating would be on for a fairly long period of time while essentially the entire thermal mass of the liquid changed temperature. Further, once it did, there would still be a lot of "residual" cold or heat in the chamber (from the long period on) that would continue to further affect the temperature of the thermal mass. That would then result in an opposite reaction (e.g, heating on due to the liquid becoming colder than target), with it's associated swing to the other extreme.

Ultimately, the liquid would reach the target temperature, but not without a series of swings. This caused me concern because I didn't want to have fermentation stalling or proceeding too fast due to the swings. Since I do stepped fermentation (start with lower than target, keep at target, increase over target after main fermentation, cold crash), I would see these swings each time I made a change.

By changing to ambient measurement the temperature changes are more gradual, since the chamber quickly reaches the target temperature (resulting in turning off cooling/heating), but with the thermal mass then affecting that temperature and the cycle beginning again until equilibrium is reached. It's a process that definitely takes longer, but results in less swings. So I'm trading off the latter for the former.

I'm not ready to say one is better than the other, hence my question. Since you are doing these experiments and doing specific measurements, it seems to me you are in a better position to provide a more educated answer.

EDIT: I should say that looking at the graph you posted the temperature of the beer seems stable after the initial changes, without the swings I'm referring to. Do you do stepped fermentations?

 

[gromitdj]

Sounds like you need a Brewpi. Controlling the overshoot Is what it was built for. Because it is PID driven, it learns your system and turns the fridge (or heater) off prior to reaching your set point, so that it doesn't over shoot. It also has fermentation scripts that it will follow for your stepped fermentation procedure.

There are a few on this board that recommend controlling the ambient temperature for the reason you are experiencing. I've always taped the sensor to the side of my vessel. If/when I go the Brewpi route,I'll likely go with the thermowell.

 

[srice]

Yes, the thermowell sensor isn't seeing the overshoots due to the PID control. I can definitely imagine larger over/undershoots if using a bang/bang controller like the STC-1000.

The more that I think about it, the more that I realize I need to do the opposite experiment - use the externally taped sensor to control the chamber and just monitor the thermowell temps.

And yes. I do stepped (really ramped) temperature fermentations. I've already changed the profile of the current fermentation to ramp up 7F over 7 days and then hang out there.

 

[doug293cz]

The one test I did not run is to control the system with the probe taped to the outside of the fermenter and just monitor the thermowell temp. I may switch the probes around after the fermentation is complete. I suspect that they will track very closely to each other since the thermowell is in the middle of a much larger thermal mass.

From what I see, I don't think it much matters if the probe goes in a thermowell or is taped to the outside. I will stick with the thermowell since it is so much easier than taping a probe to the fermenter. There are no wires attached to the fermenter when I am moving it around. Once I get the fermenter in place, I just drop the probe down the thermowell and I'm ready to go.

srice,

I would love to see the results of this second experiment. Please do run it for the benefit of everyone here. And, thanks for running and reporting the first experiment.

Brew on

 

[atoughram]

The more that I think about it, the more that I realize I need to do the opposite experiment - use the externally taped sensor to control the chamber and just monitor the thermowell temps.

That could be an interesting expiriment...

Brewpi keeps my thermowell temps with .5f and usually less. I'll see if I can post some data later.

 

[srice]

OK, curiosity has gotten the better of me and I just swapped the probes

 

[MindenMan]

I use a STC-1000 controller in my ferm box, and tape the probe to the outside of my ferm. bucket. When I started using the STC unit, I was having overswings and underswings, too hot and too cold for what seemed like a couple of days. It may have only been 1 day, as the gift of ADHD affects the sense of time. Anyway, I took a guess at a 5* F differential as a starting point, and here is what I did: If I needed to cool the wort significantly, I would set the temperature to 5* F warmer than I really wanted, unplugged the heat,, and within 3 or 4 hours the temp would be where I had set it. I then set the temperature to where I really wanted it to be, and a couple of hours later the temp was stable, and pitched the yeast.

 

[Wynne-R]

Srice, I’m looking forward to your data. I think you’re on to something. ASantiago has a good point too.

On one end of the scale, you could control the air temperature, with frequent compressor cycles and tight temperature control. The problem with that is that the heat of fermentation is going to have a long lag time heating the air. Unless you correct for this manually (guess at it) you might end up with a substantial fermentation peak at the worst possible time.

The other extreme is using a thermowell. It’s well coupled to the beer temperature, but not so much to the air. By the time the beer hits the right temperature, the air is much too cold. So you add the differential of the controller with the overshoot of the too-cold air. Your beer temp will ramp up and down.

Is there a way to split the difference? Yes there is, though I try hard to convince anybody. It’s a classic engineering trade-off, though most engineers don’t seem to get it.

Instead of choosing between air and beer, how ‘bout a bit of both? When I did a lager in a mini-fridge I used two layers of duct tape on the carboy underneath and one on top. My thermowell temperature was within ±.1°C, if I remember correctly.

Now I do it in water and it’s less than .1°C with an STC-1000. That’s with the probe near the carboy, but not touching it.

 

[srice]

Is there a way to split the difference? Yes there is, though I try hard to convince anybody. It’s a classic engineering trade-off, though most engineers don’t seem to get it.

Instead of choosing between air and beer, how ‘bout a bit of both? When I did a lager in a mini-fridge I used two layers of duct tape on the carboy underneath and one on top. My thermowell temperature was within ±.1°C, if I remember correctly.

Now I do it in water and it’s less than .1°C with an STC-1000. That’s with the probe near the carboy, but not touching it.

That's the beauty of a BrewPi. It uses a free air probe in the chamber as well as a beer probe. The controller has enough smarts to take both into account so that the heating cooling decisions don't wait until the beer thermal mass doesn't have too much momentum to keep from overshooting the set point.

 

[srice]

OK, I grabbed some more data and then flipped the probes back to where they started - I do have 6 gallons of beer in there!

What I saw is with the BrewPi trying to control things with the probe taped to the fermenter was that the BrewPi was keeping the actual temperature of the beer (gray line) about 0.5F below the setpoint (red line). The beer wasn't seeing large over/undershoots, but it wasn't being controlled as tightly as if the probe was in the thermowell. 0.5F probably doesn't matter, but the thermowell is easier to set up/use so I will stick with that.

 

[day_trippr]

Did you happen to reload the default controls and settings between probe swaps? If BrewPi adjusted parameters while operating with the thermowell'd probe I'd expect there to be at least minor anomalous behavior - until BrewPi re-learned "how things work" (assuming it does)...

Cheers!

 

[srice]

Nope, just let the brewpi fend for itself. When I put the beer probe back in the thermowell, it quickly regained control and the thermowell is right on the set point.

My main goal was to see if the beer was over/under shooting when the temp was being controlled by the probe taped to the side.

 

[day_trippr]

Looking at the last graph I'd say BrewPi was shrinking the gap between beer temperature and set point before the plug got pulled, so it looks like it was "learning" about the new beer temp probe placement. I bet the tracking with a well-insulated strapped-on probe would be nearly indiscernible from a thermowell, given a fresh load of default settings/configs as a starting point...

Cheers!

 

[Reisende]

This is all great info, especially for those of us who are still planning/dreaming about a more automated system! I am in the thermowell camp, and I long for the day when I have automation in place so I can focus more on drinking my beer, but until then I have some personal experience/opinion to share for those using a simple digital bi-polar controller (STC-1000), like myself. I have built two STC-1000 based temp controllers for use with freezer/fridges for fermentation chambers, and I have worked through the initial frustrations to arrive and what, IMHO is a fairly accurate system for non-PID. I use the 7.9 gal plastic buckets from AHS exclusively for primary, with a chest freezer and a small fridge as enclosures/cooling source. I use incandescent light bulbs in metal paint cans for a heat source and have a fan running continuously to keep the air well mixed and at an even temperature. Here are some tips:

1. Thermowell - I wont go into a lengthy opinion-based rant here about thermowell vs. taped to the side, but here is some food for thought. The same reason that is given for taping your probe to the outside instead of using a thermowell (less swing) also works in reverse so to speak; your ambient air temperature will change drastically faster than your thermal mass will. Measuring the air is not measuring the wort, especially when active fermentation is an exothermic process! By applying a heating/cooling load to the ambient air to then effect a change in a thermal mass, we are already handicapping ourselves in terms of control. This is understood and accepted, as few of us have cooling coils submerged in our fermenters (I wish I did though). But we can still achieve a fair degree of accuracy, but only by measuring what we actually want to control.

I use a thermowell, and what works for me is placing the probe less than the full depth into the well (about 2.5-3 in seems to work best in my 7 cu chest freezer). This makes your thermal mass temperature more responsive. This suggestion is similar to what Wynne-R said in post #10, but instead of using a separate, smaller thermal mass or insulating your probe, you just measure closer to the "edge" between the air and the thermal mass. Trust me, it works. Deeper, less responsive; shallower, more responsive.

2. Set-point and dead-space - The best thing you can do to avoid massive temp swings is to have your wort as close to your fermentation temp as possible before placing it your chamber. I have started using a pre-chiller and ice water to do this, since I currently have warm well water. I would also set your dead space as small as possible. Some people use dead space to reduce swings, but this also reduces your precision. Adjust the depth of your probe instead, and find that sweet spot. Stepped/ramped fermentation profiles will obviously have to be done manually, just do it in small steps instead of one big step. Alternatively, add a switch on your controller to disable your heat source and let it rise naturally (I added this feature on my second controller).

3. Build a real controller - I love the STC-1000 for simplicity and low-cost, but it really isn't great out of the box. Obviously it needs an enclosure, but there are two more things that will really help. A. you need a mechanical relay on the cooling loop if you are switching an inductive load (fridge/freezer/motor). The solid-state relay on board will die prematurely if you don't. This model works great, and will last forever: K10P-11AT5-120. B. you should build a shielded temp probe cable if it will be near any electrical noise. I used Belden 9320 (expensive, but I have a source...) to build a cable, I just made a good solder connection to the probe that came with the STC-1000, leaving about 6" of OEM wire attached. The drain wire is connected to ground, at the controller end only (actual, "green wire" ground). This makes a HUGE accuracy difference, especially is your probe cable is anywhere near your heating/cooling source power wires.

I can control a standard 5.5 gal batch of wort, no matter how vigorous fermentation is, to within -.1/+.3 C (less than 1F total swing) with my system. This is good enough for me right now, until I build my dream brewery some day. Every system is different, and I spent many hours figuring mine out before I put any real wort into it. Here is a pic of my first controller, feel free to PM me with any questions. I can probably provide a BOM if someone wanted to copy it. In the meantime, RDWHAHB.

Prost!

 

[BeerdedBrewer]

thanks Srice for posting, and to all others for their contributions. It appears, for the time being, I will be a temp. probe in a cup of water kind of guy!

:beard: