Working with precision temp sensors

[geek with fire]

As I've mentioned on here before, one of my current passion projects (things that consume my every thought and extra time) is a control system for my smoker. I currently submitted my latest system to a robotics contest and won a gift certificate to get even deeper into this thing (boy is my wife happy about that!)

Anyhow, as I've been researching quite a bit of theory, i'm finding interesting things about the differences between thermistors (the temp probes that most oven and Q electronic readers use) and thermocouples. So, I've taken this gift certificate and bought a couple of digital cards that will let me take readings from precision K-type thermocouples and connect them to my Control System on my laptop. I also picked up 2 new thermocouples from ThermoWorks (the company that makes Thermopen). What I'm hoping to accomplish is to read both the temperature of the smoker and the meat at faster intervals, as to make faster and more accurate adjustments.

I'll be posting more as my testing progresses, but was interested if anyone on here had any knowledge of either type of sensor and wouldn't mind sharing dialog.

Thanks,

 

[ba_loko]

Dang, Josh...that sounds fascinating. I've never gotten too far into the control side of our business, so I can't really help you much or offer dialog. I will be paying close attention to what your testing reveals though.

 

[Deer Meat]

Sorry I can't be much help to you either. But I am really interested in what you have up your sleeve for this project.

Your smoker project was awesome.

 

[deejaydebi]

Sounds kewl Josh. Keep us posted.

 

[richtee]

Heya Geek!

The control I have for the Beast is a "K" type. I have not got to that part yet, or even fired up the controller, but it's basically an out of box app, sooo..well, I should say I'm not ANtICIPATiNG any trouble...heh. But anyway, sounds like it's just a question of sample times, and all current A/D's can do that in the millisecond range easily. The K types have fairly quick response, but any available A/D is factors faster than it. K's are a good temp range too... like -50 to like 900 with like 1.5% linearity, I believe.

 

[placebo]

I have worked on laser printers for years and they use a thermister and thermoswitch to control temps in the fuser(heat) roller. It's been a few years since I've rebuilt fusers so that technology may be somewhat different now. For the one's I used to rebuild they used the thermister to monitor the fuser temp, and the thermoswitch to regulate current to the heat lamp, by switching on when temps get too low and off when temps get too high, which would enable it to maintain a temp in the range of 220 degrees F. I don't recall the exact tolerences but probably somewhere in the neighborhood of ten degrees or so.

P.S. I have plenty of thermisters and thermoswitches so if you need some just let me know.

 

[starsfaninco]

I'm curious how that laptop is gonna stay clean sitting close to a smoker :) Being an IT guy and a gadget guy in general (I'm looking in to weather monitoring stuff myself) I'm curious how this progresses too.

 

[geek with fire]

Yea, I smoke on the front porch so that I can run the USB cables through the window, and therefore keep the computer in the house. But the "crap", as my wife calls it, on the front porch looks kinda' hill-billy; which I don't have a problem with, but she does. This kina' piggy-back's on the thread I started the other day about building a shed around a smoker....that's pending too.


Thanks for the offer Placebo. Actually the thermistors work well, but the problem is, their response time is really slow. A person can test this by plunking their probe in a glass of ice water. It takes about 20-30 seconds to get down to temp. I'm hoping the response times of the thermocouples are a bit better.


Yep, the linearity is big to me. The cheap thermistor probes that I use have such a strange curve, I have to calibrate them every 5 degrees and store the data in a cross-reference database instead of using the correct forumulas to calculate the temperature. Also, the new thermocouples I got appear to be more durable, and have 6' cables to boot!

 

[richtee]

As with any system, Dtemp/Dt is directly related to the system's thermal capacitance. EG..don't mount your air temp K in a block of steel...heh.

 

[geek with fire]

Yea, the air temp K is a gator clip. The off the shelf controllers have you clip this to the meat probe. Is that the best way to do it?

 

[smokebuzz]

Are looking for temps only? or are you trying to control the temps in your pit, with a gearmotor of sorts. might a RTD be a better selection if trying to control temps?

 

[richtee]

Can't rightly say..I have yet to go thru that part of it meself ;{)

 

[geek with fire]

Yea, I'm adjusting the damper with a servo:
http://www.smokingmeatforums.com/for...ead.php?t=7731

On the RTD question, I can't really answer which one is better. From what I read, RTDs and Thermocouples should be more durable than the thermistors. And the only place I could fine comparisons between thermocouples and RTDs were by people who were sellin' RTDs; so not sure what to believe. The reason why I chose thermocouples was because I can easily interface a K type thermocouple reader into my existing Visual Basic application.

 

[smokebuzz]

Are you able to get a analog output by useing a thermocouple? the controls i work with, we have to use a RTD and a PID loop to maintain a precise temp. I relize your useing a custom application and you can apply the input to the output to your servo to maintain a precise temp.

 

[geek with fire]

yea, kinda'. I take the reading from the temp probe and run it through a PID loop in the software. The result is a percentage from 0 to 100. This gets sent to a function that controls the servo; which has been calibrated to the known points of the damper's 0% and 100% position.

 

[smokebuzz]

OK, you have now struck my curiosity, i have been collecting used/abandend parts, have the RTD's, a Siemens S7-200 with a analog card, that works , will have the gearmotors with linear scales soon. I hope to be able to work with this set up this winter when we slow down at work.

 

[starsfaninco]

So, if I'm understanding correctly, the thermistors should be more durable than the thermocouples, while the thermocouples would be faster to get readings. Since the probe is going to be inside the chamber full time and not just poked into meat, I don't see the need for quick readings and would think that the minor temp fluctuations would have the servo cranking one way or the other constantly. This would wear out the servo and the thermistor in very short order. I would think averages would be more the game here. Even in regular ovens, the temps vary 5-15 degrees between cycles (had my trusty ET-73 in my oven just to see how the temps might vary).
I'm just saying that you might be trying to make this more exact than is feasable.

 

[richtee]

Thermocouples output a milli-volt level signal. I have a small "self adjusting" PID controller I will be using on the Beast, but I kinda doubt I'll even need any I or D to stay within 5 degrees or so, as I'm actually firing a propane burner inside the cabinet.

 

[smokebuzz]

OK, before i say this, remember most of my PID loop exp. comes with VFD's on motors. I believe it's P,I,and D not P or I or D, the purpose of useing PID is so you can vary your out put proportionally, to maintain the target temp/speed/load. The typical temp controller will give a relay output, on off on off , all or nothing. one analigy i heard is, PID is like cruise control and contact closure is more like your right leg with a cramp coming on.

 

[richtee]

Heh...interesting analogy. All are present in the feedback loop, but you can apply weighting factors to the various functions of the controller. This can effectivley limit differential and intergal response to minimum values.