PIDs & Heat Sensors

PID Controller used by Mr. T. for a homemade Sous Vide Cooker: Willhi – WH1436A Digital Temperature Controller.

Posted at SMF on 1/20/2017 by Sigmo of Central Wyoming

This is a great paper on the design of ovens and heat sensors as well as what happens during a cook. All credit for this is Sigmo’s.

I know this is long, so bear with me if you want to:

It probably isn’t a bad assumption to think that they’d use the same type of sensor element for both of these probes since it would simplify the design of the controller.  So they might both be thermistors of the same type.

And it’s not a terrible idea to have a moveable sensor for the controller so that you could force the unit to maintain the temperature at your desired test point to be what you want.

But you would want to be very cautious with any such modification and the implementation of it.

It’s important to realize that the position, shape, thermal time constant, etc., of the sensor that reports back to the controller so it can control the power to the heater are all part of the design of the system.

The position of their sensor, right above the heating element, and its size and speed of response have all been “tuned” along with the program in the controller to achieve the response that the designers want.

We can imagine that the sensor, right where they have it, “sees” the heat coming up off of the heating element fairly directly.  This causes the system to shut off the heater sooner than it would if the sensor was located somewhere farther from the heating element, and out of the direct path of warm air rising up from it.  So the speed of cycling on and off of the heating element is likely faster with the sensor right where it is than it would be if you placed the sensor farther away from the heating element.

If you place the sensor in some place that you feel is more representative of the actual temperatures that the meat is experiencing, that has the advantage of forcing the smoker to drive that temperature to the setpoint you’ve programmed.  And this seems, at first glance, to be a good thing.

But what you have to remember is that this could cause a dangerous situation, and further, it’s not necessarily what you really want in terms of making your cooking match what you see in various recipes.

As an extreme example:

Let’s say we swap the meat probe and the controller sensor on one of these smokers.

Now, we want the unit to control the temperature that our pork butt is actually experiencing.  So we put the meat probe fairly near to the pork butt, and fire up the smoker.

Right away, the heating element comes on full blast.  But it stays on for a very long time because the sensor is near to the pork butt, and the cold temperature of that butt, along with the moisture in it, creates an area around it that is fairly cool compared to areas in the smoker farther from the butt.

So the temperatures seen by the hottest places in the smoker (presumably down near the heating element, and other areas farther from the big piece of meat) end up being far higher than the setpoint we’ve punched into the controller.  The controller thinks it needs to keep the heater on, yet the temperatures in the hottest parts of the smoker may be dangerously high. High enough to start decomposing the foam insulation in the back or sides of the smoker nearest to the heating element.

Further, even if the temperatures don’t become dangerous, the heater may cycle on and off very slowly.  Perhaps staying on for half an hour or an hour at first before our probe “sees” the temperature we wanted.  And then, once it does “see” that temperature, so much heat might be stored in the system that the heating element will now remain off for a very long time before things cool down enough to cause it to cycle on again.

Cooking recipes take into account the phenomenon of the actual air temperature right near the big thing being cooked taking a long time to come up to match the “oven temperature”.

With a so-called convection oven (misnamed IMO because they use the OPPOSITE of convection, and instead rely on a fan to actively circulate the air). things cook faster than in a “normal” oven.  This is because the active fan-stirring of the air gives you a more uniform temperature throughout the oven and constantly forces this “proper temperature” air over the item being cooked.  So it doesn’t have this layer of cool air surrounding it.  Instead, it’s continuously exposed to air that is closer to the actual oven temperature setting.

So you have to consider the effects on cooking time in all of this.

Further, from a control-system tuning point of view, consider this:

In many laboratory drying ovens, incubators, etc., I have seen the temperature sensors placed in the direct path of the heated air coming off of the heating elements.  At first, I thought this was silly because everyone knows that you don’t put your home heating system thermostat right above a heat register, or it’ll shut off too early and never get the main living space up to the correct temperature.

But there is a method to their madness!  It’s part of how they “tune” the controller that maintains the oven’s temperature.  It effectively speeds the response of the system, and lessens the amount of temperature swing between the “ON” and “OFF” parts of the cycle.  It also acts as a safety feature because although parts of the oven are, indeed “too cool”, we know that no part of the oven is ever “too hot”.  And that’s the tradeoff they make.

As we cook something in an oven or smoker (or as we dry samples in a laboratory drying oven), we do NOT want any part of the food or samples to get above the oven senpoint temperature.  Food can burn or dry out, and samples might go over the upper temperature limit set forth in a particular analytical method, and thus give invalid results because you drove off things that would not be volatile if you stayed below the prescribed upper temperature limit.

So always keep in mind that unless you’re prepared to reprogram the controller and carefully troubleshoot all of the side-effects of tampering with the factory design, you probably should keep things pretty much as they are.

The easy method people have used on here is to simply set the controller to a temperature a bit above what you normally would if you want a higher cooking temperature.  And they base this “offset” on what they read with their moveable probe thermometers/loggers.  Still, they need to keep in mind that many of the recipes we see are based on an unmodified smoker that was set using the factory control system with no adjustments or modifications.  And measuring the temperature in one particular place may not really be telling us the whole story.

Still, I do sympathize with people thinking that their smoker’s actual temperature is incorrect.  And clearly, these inexpensive smokers don’t have the best possible control systems, and I’m sure some are way off.  Just be careful and realize that measuring the actual temperatures in an oven is a bit more involved than we might think.  There can be HUGE variations in temperature from one place to another, especially if there is no active stirring of the air.  I’ve seen laboratory ovens with 15 degrees C (27 degrees F) of difference between probes placed several INCHES apart!  I’ve had PhD chemists complain to me that my monitoring system was way off due to this, only to later prove to them that the system, and a number of separate probes were, in fact, all in agreement with a NIST traceable thermometer to within 1/100th of a degree C!  Indeed, it was simply variations in temperature within small distances in these ovens that was real.

So we need to be extremely diligent in our assessment of temperature accuracy.  What we see with our fancy external thermometer or data logger may well be true.  But only for the very place and time where it’s probe happened to be.  So I don’t like to be too quick to impugn the accuracy of the factory probes or controllers in any oven or smoker.

But again, I also know they aren’t perfect.  And for what these systems cost, I’m sure people get real lemons.  My MES’s sensor/controller is quite non-linear.  And while it’s very accurate at room temperature, it gets farther and farther off as we get up into cooking temperatures.  But so far, I have found that if I set the unit to the temperature the recipe calls for, even though I’ve proven that its temperature sensor is not linear, and is off, perhaps 20 degrees at 225, the food cooks to the correct internal temperatures in times that match up pretty well with what other people post in their recipes and other posts on here.

Also, we all find that we have to make trial-and-error adjustments to our recipes anyhow because our humidity, ambient temperature, elevation, etc., all play into this.  So none of it is exact.  Our smoker calibration is just one of the variables we have to account for with our trial and error.

What I DO want to be dead-on is my meat probe, though!  And fortunately, in my MES, the meat probe is within a degree or so over the whole range.  Maybe I lucked out.  I don’t know.  Maybe the meat probes are different.  And maybe, just maybe, the bizarre non-linear calibration of my MES’s controller sensor was done on purpose to make the actual response in the main part of the cooking chamber end up being correct.  Perhaps that’s giving MB too much credit.  But the fact is, I just don’t really know.

But because my MES’s meat probe IS so accurate, I can use it in different places in the chamber to see what the temperature is doing, and then make manual adjustments to the controller’s setpoint to get what I want during different parts of a smoke, or just to keep an eye on it.

I once placed the meat probe just below where a large pork butt was being smoked, suspended in some paper clips so it was about an inch below the butt.  It was very interesting to look at that temperature and see it go up very slowly during the first hours of the smoke.  I didn’t change anything based on those readings, but it was interesting to see how the temperature right there stayed fairly low, initially, probably in part because of the moisture on the surface of the butt.  Then, as it formed a pellicle or bark, the temperature right under the butt finally started to rise.  It’s interesting information.  Not because I wanted to adjust the temperature of the smoker, but because it helped show me what the temperature near the butt was actually doing during a typical smoke in a non-stirred oven environment.

It’s nice to have a known accurate, moveable probe that you can use to tell you what’s going on in different places within our smokers.  Just don’t be too quick to think you need to change things.  Millions of non-technical people make millions of great meals in their factory-stock smokers.  Ignorance is bliss, I guess.

Still, I’m a tinkerer, and for me, and clearly a lot of us on this forum, we LOVE to play with all of this, and we want to understand what’s going on, and we want to optimize things.  And really, that is a big part of the fun in all of this. And I do not dispute that some of these smokers really do have calibration problems.

In the spirit of experimentation, hobbyist enjoyment, and just plain Smoker Nerdism, I am planning a serious modification of my MES, using different sensors, and a stirring fan from a convection oven.  So I am right with you in this!

But I will leave the snap-disk safety cut-off in the circuit that feeds the heating element.

And I’ve already modified the air intake and heating element area with different baffling to get better pellet burning along with heat and smoke distribution in the smoker.  But I have made sure that some of the heat from the heating element still has a good “straight shot” up to the sensor.  I don’t want to circumvent the control system’s tuning too much until I do the full modifications I have planned.  At that point, a lot of things will change, particularly because of the stirring fan.

What I’ve got in mind is to use the stirring fan to keep the temperature and smoke fairly even throughout the chamber.  And I’ll also use a different sensor to inform the controller of the oven temperature as well as a couple of “free to move” sensors to give me information about what the temperature is doing in selected places within the smoker.  And finally, I’ve got a new meat probe as well because I want the controller (and me) to be able to keep track of the meat probe temp, and I have no interface that can read the factory style thermistors.

Hopefully, as I do this, it will yield some interesting and useful information about what the temperatures really do in different places inside of one of these smokers. Two of the thermocouple sensors I got have alligator clips so you can easily move them and hang them in different places.  They came from the manufacturer that way!  How cool is that?

I love experimentation.  Just be safe, and consider the consequences of various modifications.

All of this is great fun, IMO!

Edited by Sigmo – 1/20/17 at 6:29pm