NTC thermistor "parameter change" failure mode

Thread Starter

strantor

Joined Oct 3, 2010
5,642
I am playing with a climate controlled (heated/cooled) seat out of a Range Rover, trying to get the heating & cooling working. This seat has a TEC (Peltier) device with an integral NTC* thermistor for feedback.

(*) NTC according to drawings

The thermistor (I believe, not confirmed) is not accessible without destroying the TEC to remove it. I am trying to resolve what thermistor it is, so that I know what resistance corresponds to what temperature.

The other day in ambient temperature of approximately 80f (27C) I measured the resistance at 1046ohms. I then turned on the fan (part of the heating/cooling assembly) and watched the resistance gradually go down to 1020ohms, where it settled and did not change for several minutes. The TEC had no power going to it. There should have been no heat generated apart from that of the fan itself, which would expect to have been more than canceled out by the air that it was moving over the TEC.

Then, with the fan still running, I applied 12V to the TEC and instantly (more accurately: in the time it took me to connect the wires and grab my DMM probes) the resistance shot straight to 1372ohms and did not change during the time i monitored it (about a minute, maybe a bit less).

Then, with the fan still running, I swapped polarity to the TEC and again "instantly" the value shot to 399ohms and remained steady there for nearly a minute.

I did not feel any change in the seat temperature during either test.

What I was expecting to see, was a gradual change in resistance in the first test, followed by another gradual change. The TEC (and the thermistor itself) must have some thermal time constant, so I am perplexed by the "instant" resistance changes and further by the steady prolonged readings.

I suspected maybe my thermistor is bad, so I did a bit of research and I find that one of the less common failure modes is "parameter change." I could not find any resource which was more specific than that, so I'm here asking for personal observations. Can "parameter change" include a NTC thermistor spontaneously gender changing into a PTC thermistor? Could the strange readings I saw be included in the broad description "parameter change?"

Or maybe I should be asking, are my reading even strange? Are my expectations flawed? Does a peltier TEC in fact make "instant" temperature changes? This Peltier is rated 5.5A, at (I believe) 12V. so not a small device.
 

Alec_t

Joined Sep 17, 2013
12,209
It's not unknown for datasheets/specifications to be wrong. Some manufacturers allegedly include deliberate 'mistakes' on drawings to discourage counterfeiting.
Perhaps your thermistor is indeed PTC.
 

Thread Starter

strantor

Joined Oct 3, 2010
5,642
It's not unknown for datasheets/specifications to be wrong. Some manufacturers allegedly include deliberate 'mistakes' on drawings to discourage counterfeiting.
Perhaps your thermistor is indeed PTC.
I did not consider that it was done deliberately but I did consider that (for some bizarre reason) it was a PTC. I googled for PTC thermistors with a 25C resistance near what I observed, and all of the sensors I found had about the same temp:resistance relationship, and according to the charts for those,
Initial: 1046ohms = 91F/32 (plausible)
After fan running: 1020ohms = 79f/26C (plausible)
Polarity A: 1372ohms = >212f/100C (charts end at 1342ohms/100C) (not plausible)
Polarity B: 399ohms = < -40F/-40C (charts end at 697ohms/-40C) (not possible)

I forgot to mention, immediately after the tests I disconnected the TEC and blower, and measured resistance again, a steady 1016 Ohms. So even if the TEC were capable of instantly swinging temperature in a given direction, it should still take some time to settle back to ambient once power is removed, right? It makes no sense to me.
 

Thread Starter

strantor

Joined Oct 3, 2010
5,642
I did some more testing; some more detailed testing on the other seat and determined that it is in fact a NTC thermistor. For some reason, turning the fan on makes it heat up a bit. That makes no sense to me*. The thermistor is on the climate-controlled side and when the TEC is powered up blowing hot air through the seat, the resistance is low; when blowing cooled air, the resistance is higher. This time around I used a bench power supply so that I could take measurements at several intervals, leave my DMM connected to the thermistor, and observe changes in real time. I found that the thermistor responds nearly instantly to changes in input power because it is bonded to the TEC between the TEC and heat sink.

* I did some more reading about TECs and I have a hairbrained theory. Seebeck effect. I don't really understand it but I think maybe when I blow air over the TEC I'm turning it into a TEG. It is developing a voltage, and that voltage is doing the same thing as an applied voltage would, and pumping a bit of heat (and generating some heat). Does that sound plausible?
 

Alec_t

Joined Sep 17, 2013
12,209
The other day in ambient temperature of approximately 80f (27C) I measured the resistance at 1046ohms. I then turned on the fan (part of the heating/cooling assembly) and watched the resistance gradually go down to 1020ohms, where it settled and did not change for several minutes. The TEC had no power going to it. There should have been no heat generated apart from that of the fan itself, which would expect to have been more than canceled out by the air that it was moving over the TEC.
I'm now thinking there would not have been any cancellation. The air blown over the TEC would have been above 27C, due to the fan-generated heat; hence the resistance drop.
 

Thread Starter

strantor

Joined Oct 3, 2010
5,642
I'm now thinking there would not have been any cancellation. The air blown over the TEC would have been above 27C, due to the fan-generated heat; hence the resistance drop.
As a human being whose moist skin is designed for evaporative cooling, im going to have a hard time convincing myself to agree with that. But Occam's Razor. But this is hardly more than a CPU muffin fan. But Occam's Razor. Ugh...

How dare you attempt to dash my beautifully convoluted explanation?! :p

At work I deal with large Roots blowers used essentially as large vacuum cleaners. Their air outlet does get very hot, despite not compressing air (as compared to the heat generated by a compressor head). So I suppose you're correct, although I can't fathom how this tiny fan makes such a big difference. On second thought I don't know how big a difference it actually makes, as I don't yet have any context for what the resistance readings represent. I have one of the TECs in the freezer now, since last night. This evening I will put it in the oven. I will soon have some values to represent known temperatures.
 

debe

Joined Sep 21, 2010
1,251
An un powered Peltier device can generate a voltage if it has a temperature difference betwean each side. I have noticed in a Peltier cooler if it has been running & switched off then the cooling fan will run for some time, untill the temperature across the device equilizes. The fan was connected to the device.
 
Top