Measuring thermistor: Getting radically different measurements?

[WeekendHobbiest]

Hello world, first post!

I have scavenged a PC heatsink and fan to test high power LEDs with and I want to build a LM317 based circuit to regulate the fan speed using the fan's thermistor as part of the '317's voltage divider.

I have three meters that are giving me totally different resistance measurements for the thermistor and I'm not sure what's what. My meters and measurements:

Mashtech MS8268: 29M
Fluke 27/FM: 4.3M
Fluke 8010A: 200K range = 79.2, and 20M range = 2.27 [I bought the 8010A about two or three weeks ago and I am still getting comfortable with it. Why am I getting two different readings?]

The meters give the same general readings when retesting. The meters agree with each other with resistors I pulled out of those assorted resistor kits, so I don't think there's a problem with the meters.

The thermistor seems to work when I give it a blast from my hot air wand (the resistance lowers). It is buried somewhere deep near the motor body and I can't get an eye on it (without breaking the fan) so I can't find a datasheet for it.

I am now lost and unsure how I should approach this. Help?

[Cubdriver]

I'd be inclined to confirm that you're actually making good contact and really measuring the thermistor.  Readings in the tens-of-megohms do not sound very thermistor-like to me, and your readings are truly all over the map if they range from 79.2k to 29M.  That's a huge difference - 29M is over 350 times greater than 79.2k.

-Pat

[WeekendHobbiest]

Hi Pat, thanks for the reply!

Yeah, that mega-Ohm resistance was a 'wtf' moment for me and I knew something was up. I'm pretty sure general purpose thermistors are probably 10K, tops.

I forgot to mention in my post that I removed the contacts of the thermistor out of the connector and I was measuring it using one of those high-insulated silicone alligator clips that slips over the meter probe's tip. I was switching meters by pulling the probe's connector from one meter into another.

I've retested again (rotating the contact in the jaw several times) and my results are the same. I don't see any oxidation on the thermistor contact either.

Ed: if the thermistor has kicked the bucket, what would cause the each meter to repeat the same inaccurate reading, but conflict with the other meters reading? Shouldn't the meters be ballpark on this?

[Cubdriver]

I'd say that might depend on how the thermistor has failed.  If, for instance, it's physically broken in such a way that a lead is not making good contact internally, the readings could be changing as you switch meters and move it slightly, changing the contact resistance.  Are the readings consistent to the meters as you change them (eg, meter one measures ~100k each time, meter two reads ~3M each time and meter three reads ~427.65 ohms each time), or do they vary considerably as they're switched out (meter one reads 100k one time, then 375k another, meter two reads 10M then 1M) etc., etc?

And are you sure it's a thermistor, and not something else like some kind of temperature sensing IC instead?

-Pat

[magic]

Does it measure the same in both directions? If not, it's not a thermistor.

I suspect it's some diode or IC which simply doesn't have a well-defined resistance and different meters and ranges show different results because they use different test currents.

[RandallMcRee]

Your measurements are consistent with the hypothesis that the thermistor is not connected properly.

In that case each meter measures its own high impedance emi input and gets different dreck.

So, yeah, a working 10k thermistor will not read in the megohms. Common sense, etc.