Temperature Probes?

Hi all, very new to microcontrollers and first real project with my arduino is building a temperature control rig for my Big Green Egg.

So far Ive got the fan and such working without any problem, and my next step is the temperature probe(s). Looking to do something similiar to what these guys, http://docs.google.com/Doc?id=dg2z33xt_11fk47bvhj, did with the turkey monitor. But I think they went a little overboard on the temperature sensors. Was hoping to use a couple of replacement probes that Ive found on amazon, like this http://www.amazon.com/AD-DTTC-Replaceme ... pd_sbs_k_1.

Im not looking for super accurate, just something to give me +/- 10 or so degrees on the inside of the smoker (which will be between 250 and 350 degrees F).

Im looking for some input as to whether the above probe would work, and anything I would need to put between it and the inputs to get a good reading.

peck

Posts: 17
Joined: Sat Mar 28, 2009 7:03 pm

Re: Temperature Probes?

I bet those probes on Amazon are thermistor bead probes. They will read-out temperature as a function of resistance, not current like the thermocouple probes the Turkey-guys are using.
If they are resistors, you can use them real easy by attaching them to the analog inputs on the Arduino through some sort of resistor bridge array. I would (if you have the time) buy one of the Amazon probes and experiment with it. You can research resistor bridges on the net and calculate what values your bridge will require for the temperatures and probe resistance.
Also, (and I bet you are ahead of me on this) look and see what other temperature projects using micro controllers you can find. Don't limit your exploration to Arduino... look at other stuff with attention to the circuits they use to "normalize" their probes.

Phil
phil.drummond

Posts: 125
Joined: Sun Feb 08, 2009 4:57 pm

Re: Temperature Probes?

There are 4 basic types of temp sensors. Thermistor is a good guess but I have no idea if that is what is in there. Sometimes thermistors aren't rated for that temp range. The other 3 types are thermocouple, RTD, and semiconductor type with voltage, current or digital output. So it depends what is in there. If it is a thermistor then it will have a particular curve that you can figure out through experimentation. You might be better off using something that you know exactly what it is! This is a fairly good explanation:

http://www.microchip.com/stellent/group ... 027213.pdf

zener

Posts: 4557
Joined: Sat Feb 21, 2009 2:38 am

Re: Temperature Probes?

one thing i know works is get one of these
http://www.digikey.com/scripts/DkSearch ... 90-1911-ND
and one of these
i believe the output is analog voltage, and its pretty easy to calculate. using an arduino would be pretty trivial, there's probably an example project out there already you can use

Posts: 12151
Joined: Thu Apr 06, 2006 4:21 pm
Location: nyc

Re: Temperature Probes?

Thermistor it is.
Currently just checking out the different reading with my DMM. Appears to be 271k Ohms at ~ 66 F and 126k Ohms at ~93 F... or at least thats what I think it says, DMM is on the 2M setting when readings were taken. So a few more measurements of known temps and I should start to see a curve forming.

At this point Im wondering the best way to hook this up to my arduino to read. 5V -> thermistor -> analogue in? Would I need a pull down resistor there as well?

Thanks for the responses so far!
peck

Posts: 17
Joined: Sat Mar 28, 2009 7:03 pm

Re: Temperature Probes?

You need to decide:

What range do you want to read (min to max temp)
How many bits is your A/D
What resolution do you want

Then you can figure out what voltage range to your A/D you can live with. Then you can decide what kind of circuit will do the job. You can make a simple divider with the thermistor and another resistor. You just need to make sure it will give you enough resolution over the range you want.

See if you can match your thermistor up to one of these curves:

http://www.vishay.com/docs/33004/revtemp.pdf

Maybe 12 ? Just a guess. Hard to tell without a 25C number. I don't suppose it matters really. You can plot some numbers and do a regression in Excel.

zener

Posts: 4557
Joined: Sat Feb 21, 2009 2:38 am

Re: Temperature Probes?

Idealy I'd like to use the same type probe to do both the temperature inside the smoker (between 200 and 500 F) and for the food being cooked (60 - 200 F). I know this particular thermistor is capable of that whole range, and has matched up well with other measuring devices when testing it previously, but I'm not too concerned about the top end of that.

Im lost on the bits in A/D. Very new to the electronics side of things.

As far as resolution goes, a +/- 1 F would be great, but a few larger than that would be fine as well, especially at higher temps where the differences arent as important.

the 12th curve is a fair approximation. Took 5 measurements of known temperatures and got y=116.91x * 102.61.
peck

Posts: 17
Joined: Sat Mar 28, 2009 7:03 pm

Re: Temperature Probes?

I think the Arduino is 10 bits. So that means it divides the voltage into 1024 pieces in between 0 and your Vref. You can adjust your Vref to fit your analog range if necessary. If you post what your thermistor resistances are at your temp limits then that will help someone to assist you more.

zener

Posts: 4557
Joined: Sat Feb 21, 2009 2:38 am

Re: Temperature Probes?

Ok, thanks.

Below are the resistance and temps I have respectively:

0.025 208
0.119 98
0.126 93
0.14 89
0.243 68
0.271 66
0.555 41

After adding a few more readings the best fit linear equation is y = -234.89x + 144.34, but its not a very good fit at all on the curve. Off ~ 70 degrees on the top end. Perhaps I did the measurements incorrectly?
peck

Posts: 17
Joined: Sat Mar 28, 2009 7:03 pm

Re: Temperature Probes?

So I did a little experimenting (because I dont have a clue what Im doing, and that seems like the best way to see whats happening.

I setup my circuit like the one here, http://www.arduino.cc/playground/Compon ... Thermistor, except with 250k of resistors instead of the 10k there. Things seems to be going well. Getting reading of ~520 at room temp (75 in here). and around 650 or so with my hand on it (measured earlier to be 93).

I assume taking several readings at a time 100 ms apart and averaging or taking the median of that group would be the best way to get an accurate temperature. Is best approach to generate a lookup tables based on values, or plot some scatter points with this data and determine a function?
peck

Posts: 17
Joined: Sat Mar 28, 2009 7:03 pm

Re: Temperature Probes?

Your circuit is good. Averaging samples is good for filtering. You can plot some points then do a quadratic regression. This will give you a formula of the format ax^2+bx+c which should be good enough. I would guess that is the easiest way. You sound like you know what you are doing.

zener

Posts: 4557
Joined: Sat Feb 21, 2009 2:38 am

Re: Temperature Probes?

You said in your initial post that you are interested in monitoring temps between 250 and 350F. Taking readings in the 80-100F range, fitting a quadratic (or even linear) model, and then extrapolating is not a very good approach and prone to huge errors. You really want your data points to be in the range of interest. Existing oven, solder gun, etc., something you have might work to provide a more appropriate reference.

Good luck.

Oops, just noticed you had measurements from 40-208F, but you are biasing the model by having the majority of points in the lower temps...above discussion remains true.
jdl2001

Posts: 86
Joined: Sun Mar 08, 2009 12:53 pm

Re: Temperature Probes?

jdl2001 wrote:You said in your initial post that you are interested in monitoring temps between 250 and 350F. Taking readings in the 80-100F range, fitting a quadratic (or even linear) model, and then extrapolating is not a very good approach and prone to huge errors. You really want your data points to be in the range of interest. Existing oven, solder gun, etc., something you have might work to provide a more appropriate reference.

Good luck.

Oops, just noticed you had measurements from 40-208F, but you are biasing the model by having the majority of points in the lower temps...above discussion remains true.

Completely agreed, Ill be taking some higher readings, and working with it as I go. Just looking to get the right start beforehand. Getting accurate higher temps is going to be a pain, have to find something thats alright very accurate up there, and things like the thermometer on the smoker are pretty good, but probably are +/- 10.
peck

Posts: 17
Joined: Sat Mar 28, 2009 7:03 pm

Re: Temperature Probes?

If you can ever fit your measurements to a known thermistor curve then it will help you out in what you are doing.

zener

Posts: 4557
Joined: Sat Feb 21, 2009 2:38 am

Re: Temperature Probes?

A typical implementation for RTD & thermocouple linearization in industrial temperature controllers is via lookup table with linear interpolation. The table just implements the published curves for one or more of the standard types (J, K, E, etc).

Thermister curves generally less linear and not quite as standardized.