Resistor on thermistor question

buntay

Senior Member
#1
This may be a dumb question but I will ask anyway since I am not a master of electronic design. In the manual for hooking up a thermistor it says you should add a resistor between the adc pin and 0 volts, but why? arnt you just adding resistance?
 

hippy

Technical Support
Staff member
#2
It's not dumb if you don't know why !

The thermistor is a variable resistance so with it between pin and +V and with a resistor between pin and 0V that gives a potential divider. As the temperature changes, the resistance changes, the ratio of thermistor resistance to that of the the pull-down determines the voltage put to the ADC pin.

You've now got a temperature-to-voltage device which can be read by READADC then you can convert the voltage ( READADC returned value ) back to a temperature.
 

buntay

Senior Member
#4
Thanks,
Hippy, your definition makes alot of sense, and electric the web site helped alot. So lets see if i got it. If I have a temp. probe with a min resistance of 100.1 Ohm and a max resistance of 102.1 Ohm using the formula on the web site I need a 10 Ohm resistor? max and min are correct values....Ohm not k Ohm
 
#5
Thanks,
Hippy, your definition makes alot of sense, and electric the web site helped alot. So lets see if i got it. If I have a temp. probe with a min resistance of 100.1 Ohm and a max resistance of 102.1 Ohm using the formula on the web site I need a 10 Ohm resistor? max and min are correct values....Ohm not k Ohm
A typical thermistor has a range of
thousands of Ohms.

See, for example:
http://www.kpsec.freeuk.com/components/other.htm

You will need to look at the
Datasheet for your thermistor.

Otherwise follow Manual 3, and use a 10k
then start experimenting. :)


e
 
#6
Thermistors come in all sorts of values.

If you take a look at the listings in Farnell, for axample, your mind will boggle.

Choose one which which has a value of a few k at your favourite temperature. As hinted at by Ec.
And PLEASE take his advice about Data Sheets.
And get a calculator and pen and paper.
We must take the GUESSWORK out of these basic designs.

At this stage in your 'career' avoid putting low resistors in series.
Why?
Check out the current calcs.
The current may heat up the thermistor which will screw your thermometry.
Or, worse, excess current will pop your thermistor - an unhappy day.

So, i would suggest that you have a look at a couple of thermistors , tell us your app, tell us your proposed circuit and give us some links to Data Sheets for the favourite thermistors.
 

hippy

Technical Support
Staff member
#7
If I have a temp. probe with a min resistance of 100.1 Ohm and a max resistance of 102.1 Ohm using the formula on the web site I need a 10 Ohm resistor?
Technically you can use any resistor but that will affect the voltage range which the voltage output will cover; at 5V, 10R pull-down, 100.1R => 0.4541V, 102.1R => 0.446V. That's going to give an extremely narrow range of readings from READADC10 even under perfect conditions.

For voltage divider use it would be best to choose a thermistor which has a wide resistance swing across the temperature's you're measuring.

For lower / narrower resistances I'd guess one would use an op-amp with the thermistor to adjust the gain of an applied signal and measure the result, or some Wheatstone Bridge configuration.

An indication of your application and what you're intending to do may help give better answers.
 
#8
In a Voltage Divider setup the Larger resistance always has the most Voltage across it,
= available voltage swing to measure.

E.g. 10K Resistor(Neg) + 100K Thermistor(Pos) (@25DegC) (5V Supply) = 0.45V and 4.54V,
with Equal Resistances the Max you can get is 50/50.

Also a voltage Divider wastes power, so using low value components will waste a lot
more power. (Hint - Battery Apps :( )
The total resistance of both components should be taken into account at all temperatures
that may occur during use.

Some Thermistors are as low as 10 or 27 Ohms (@25DegC), not really suitable for this type of App.
 
#9
It sounds to me like this part is a 100 ohm RTD, not a thermistor. With a span of only 2 ohms, it will be hard to get a workable range of values for the ADC with just a simple voltage divider. Some signal conditioning is required.
 

premelec

Senior Member
#10
@buntay note there is another issue when measuring temperature with a temperature variable resistance device - self heating. Note that a power of V*V/R is being dissipated in the device and this - depending on your needs of accuracy and thermal mass - may affect your readings - ideally you want low voltage on the temperature sense unit or a high resistance in that unit.

Good luck with your project!
 

buntay

Senior Member
#12
Thanks for all the info and things to take into consideration. now to answer all the questions....

"whats the aplication?", again this is for my work, we have what we call "air chillers" to cool the cooked product. currently they are using 1970"s technology with what alkar(company name) calls the "juno". I have written a windows program to replace this using the outputs from the printer port to control solid state relays for fans and coolant. but I need to get my product temp and chamber temp into the computer, I have tried two ways already. 1. using a thermistor interfaced with the game port, did not work cause not all computers have a game port. 2. DS1820, didnt work, the outside temp transversed down the wires and influenced the DS1820 probe messuring internal product temp to eratic temps that couldnt be compensated for. So I decided the best route was to try and use the existing probes.

Explination of the probe: when I look up the part number all I can get is it is a RTD probe. It has 4 wires, 2 red, 1 white and a sheild ground.

saught setup: to interface these probes to the picaxe, do the conversions and spit me out a Tempreture serially.

If anybody knows how this type of probe works, and can guide me as to what I need (and why) to make this work I would be forever grateful. As I can incorperate picaxe into the industrial field it would be a big boost, cause this chip is the best thing since the "300 in 1 electronics lab"
 
#13
RTD probes are very, very common devices. Google will be a big help. Here is a starter:

http://www.omega.com/rtd.html

Many RTD's have extra lead wires used to compensate for lead resistance. That's what the extra red wire is. If accuracy is important, you will need to linearize the RTD's output. This is not a trivial task.
 

buntay

Senior Member
#14
+/- one whole degree is acceptable. My software can compensate and only whole degrees are accurate enough. So does that make post 4 true or should something be added? and what?
 

buntay

Senior Member
#15
After reasearch I beleive this is what I need. but what are the resistance values of the resistors? and am i actually powering this circut plus the pin loop on picaxe, if so is it supposed to be suplied with a voltage higher than picaxe. my yester-years of "300 in 1 electronics lab" did not cover this :eek: :confused:
 

Attachments

premelec

Senior Member
#16
Are you trying to replace _everything_ in the controller except the RTD? You'll need to establish very stable current source and if you have 4 wires I'd think you want a Kelvin connection to determine the resistance of the RTD. 100 ohms at 10 milliampers gives you 100 x .01 or 1 volt to work with - if the RTDs are pretty large and can take 40 ma then 4 volts which READADC10 will divide into 1023 intervals - with the caveat of self heating. You may need to use an amplifier with an offset to get expanded scale readings - with the small resistance changes in the range you are interested in more care and precision become requisite. Do you have access to the previous controller's electronics? Seems like you are learning - you'll be up to the 999 in one electronics set in no time :)

I typed RTD measurement circuit in google and came up with Microchip's AN687 - many more references there too...

http://ww1.microchip.com/downloads/en/AppNotes/00687b.pdf
 
Last edited:
#17
After reasearch I beleive this is what I need. but what are the resistance values of the resistors? and am i actually powering this circut plus the pin loop on picaxe, if so is it supposed to be suplied with a voltage higher than picaxe. my yester-years of "300 in 1 electronics lab" did not cover this :eek: :confused:
This is Wheatstone Bridge. Once again, Google is your friend. It will not provide much of a signal to the Picaxe. The excitation voltage is separate from Vdd and Vss.

Here's what big daddy says about RTD's and circuits for them:

http://ww1.microchip.com/downloads/en/AppNotes/00687c.pdf
 
Last edited:
#19
2. DS1820, didnt work, the outside temp transversed down the wires and influenced the DS1820 probe messuring internal product temp to eratic temps that couldnt be compensated for.

snipped from post #12
I've just re-read this sentence and I'm still confused. :-(

e
 
#20
Good point, eclectic. In such a situation the usual fix is to coil up some of the lead length inside the area of the temp being measured. The problem here seems to be one of mechanical design. What's this probe go into? What's its size, and what does it look like?

It seems there should be an easier way of solving this problem than to use a sophisticated wide temperature device and circuitry on a job that should be able to be done on the cheap and easy.
 

buntay

Senior Member
#21
Here is how the setup works, we are a pork prosessing plant. The two items that are put into these chillers are hams and bacon bellies. when they come out of the smokehouse they are roughly 160 degree's USDA says we have to get them to 40 degrees in a given amount of time. At this time a USDA temp recorder is used to check product temp....BLAH, BLAH, BLAH. So here is where the problem came. in order to get this done the chiller they are put into has to run about 10 degrees. when the time was up my ds1820 meat probe (LOL cant coil up wire inside the ham) said 40 and the USDA recorder says 45 or so. USDA accuricy was checked and it was found to be on my end.

I had to make my own probe for the ds1820 due to stainless regulations so I solderd wires onto the ds1820 and epoxied it into a 1/4" stainless tube. The only thing I can figure out is that the wire outside the ham (exposed to 10 degrees) conducted some coldness into the ds1820. the probe was 5" long and had the ds1820 all the way in.
the chamber probe was made the same way and is +/- .5 degree when compared to certified thermometors. So the chamber temp had to be influencing the meat probe through the connection wires.

As I said this was checked several times and 2 different ds1820's and it was never the same amount off. I.e if they stuck it into a smaller ham the more it would be off, larger ham the less. that is too much of a veriable to compensate for.

beleive me, I would love nothing more than to use the ds1820 but it seems I need a conductionless conductor ....LOl
 
#22
As I said this was checked several times and 2 different ds1820's and it was never the same amount off. I.e if they stuck it into a smaller ham the more it would be off, larger ham the less. that is too much of a veriable to compensate for.

beleive me, I would love nothing more than to use the ds1820 but it seems I need a conductionless conductor ....LOl
A small but hopefully valid point.

ds1820
or
DS18B20?
e
 
#25
2. DS1820, didnt work, the outside temp transversed down the wires and influenced the DS1820 probe messuring internal product temp to eratic temps that couldnt be compensated for.
How is switching to an RTD gonna stop the same thing from happening? The same outside temp will traverse down the wires to the RTD.
 
#27
Bets are off until the experiment is retried with a DS18B20 - IIRC there are also some DS18B20's around from several years ago that didn't work either
 
Top