# Offsetting a thermistor resistance value

#### john2k

Joined Nov 14, 2019
219
I have a thermistor that is built-into a car seat part of the heating element that provides feedback to a control box that heats a 12v heating element. This is a upgraded seat from a newer model of same brand car and it appears the thermistor on this new seat at room temperature reads a resistance value of 8.55k ohms. Whereas the original seat thermistor in the exact same room temperature environment starts at 10.55k ohms.

So that means the newer seat thermistor has a potential offset of 2k ohms. Because of this, the heating element is turning off sooner than usual and not warming up as much. Is there a way to somehow inline in the wire to add some sort of resistor to increase the resistance to offset this by 2k? So basically whatever the thermistor is reading, the reading going to the control box should be plus 2k ohms on top of that.

Thanks.

#### BobTPH

Joined Jun 5, 2013
9,169
It would work if the goal was to heat the seat to that exact temperature. At another temperature, the resistance difference would be different. It might be close enough or it might not.

#### Jerry-Hat-Trick

Joined Aug 31, 2022
582
Maybe start with a lower series resistance, perhaps 100R and try it? Since the room temperature resistance is lower and it switches off too early it must be an NTC (negative temperature coefficient) thermistor but the curve is very non-linear. The switch off resistance for the original seat thermistor may be only a few hundred ohms so if your series resistor is too high the heater may not switch off at all.

Better still, could you use a 1K potentiometer in series with the thermistor and adjust it to switch off the heater when you are happy with the temperature - then replace the pot with a fixed resistor when you know the value

#### Irving

Joined Jan 30, 2016
3,992
Can't find a standard thermistor at 8.55k nor 10.55k, but near enough is 8.2k and 10k both with type J curve and beta of 3982... so using an online calculator we get:

So you don't want to put any additional resistance in series - as it stands the seat will run 5degC cooler at the top end

#### Alec_t

Joined Sep 17, 2013
14,379
Can't find a standard thermistor at 8.55k nor 10.55k, but near enough is 8.2k and 10k
I suspect the originals were 8.2k and 10k because thermistors are usually rated at 25C, whereas the OP measured them at 'room temperature' (which may well have been lower than 25C).

#### Irving

Joined Jan 30, 2016
3,992
I suspect the originals were 8.2k and 10k because thermistors are usually rated at 25C, whereas the OP measured them at 'room temperature' (which may well have been lower than 25C).
I guessed as much...

#### AnalogKid

Joined Aug 1, 2013
11,146
Yes, adding a 2 K resistor in series will correct the new thermistor to mimic the old part ***at one specific temperature only***.

The resistance-versus-temperature characteristic of an NTC thermistor is highly non-linear. Above and below the corrected temperature point, the new-part-plus-resistor will not agree with the old part, but within something like +/-20 degrees it should be better than the new part alone. If you know the temperature at which you want the best accuracy, you can change the fixed resistor value to match that.

ak

#### Jerry-Hat-Trick

Joined Aug 31, 2022
582
So you don't want to put any additional resistance in series - as it stands the seat will run 5degC cooler at the top end
Or put a 1K resistance in series to get close to the same switching temperature? See my conservative approach suggested in post #3

#### Irving

Joined Jan 30, 2016
3,992
Or put a 1K resistance in series to get close to the same switching temperature? See my conservative approach suggested in post #3
Yes, it certainly needs to be much less than 2k otherwise the seat will never get anywhere as near as warm. Not sure what the top internal temp for a heated seat should be? 45C, 50C?

#### Jerry-Hat-Trick

Joined Aug 31, 2022
582
Yes, it certainly needs to be much less than 2k otherwise the seat will never get anywhere as near as warm.
Isn't it the other way around? Suppose the target temperature is 35C. So the 10K thermistor would switch at 6.5K. If you add a 2K resistor in series with the 8.2K thermistor it would switch at a thermistor resistance of 4.5K which from the table means around 39.5C, i.e. hotter.

A 1.2K resistor would bring them in line at 35C. If 40C is the switching temperature then a series resistor of 1K would do the trick, but I favour starting with a lower value of series resistor to be sure the system doesn't overheat. Your table suggests my original suggestion of trying 200R won't make much difference.

#### john2k

Joined Nov 14, 2019
219
OK so i'm a bit confused now lol. Just to provide a bit of context: the 12v control box that has the power going to the heating element is regulated by two independant feedback. The first is a setting volume type switch so the user can select 1, 2 or 3 heat setting. and the other is the resistance value from the thermistor located inside the seat. If i was to short circuit the thermistor input pins on the control box then the heating element continues to increase in temperature. But right now the heating is being regulated too early and the element is barely getting hot enough to even notice it.

The difference I mentioned at room temperature (room temp being around 24/25) is when i measured the values of the old thermistor vs the new thermistor almost at resting position without the heating element being on and this was reading at around 8.55k on one and 10.55 on the other. And as I powered the heating element, the resistance value started to come down. So even if I was to lets say add a 2k resistor to make it offset by 2k, then when the original resistor comes to 0 the lowest the reading will ever go to would be 2k meaning that it will never reach 0 to tell the controller to completely shut off? But then again at a lower resistance the controller seems to be cutting power before it reaches 0 anyway.

#### Jerry-Hat-Trick

Joined Aug 31, 2022
582
Okay, now I’m confused. You say that if you short circuit the thermistor, thereby reducing its resistance to zero, the heater keeps going? But surely the heater turns off when the resistance drops to a specific value, which is significantly higher than zero.

The thermistor will actually not drop to zero. By putting maybe a 1K resistance in series with the thermistor will raise the temperature at which the heater trips off.

#### MrChips

Joined Oct 2, 2009
30,979
By room temperature, we expect that to mean 25°C.

Take three resistance measurements, one at 25°C, one below 25°C, for example 20°C, and one above, for example, 30°C.
Then we will have a good idea of how the resistance varies with temperature.

#### MrAl

Joined Jun 17, 2014
11,575
I have a thermistor that is built-into a car seat part of the heating element that provides feedback to a control box that heats a 12v heating element. This is a upgraded seat from a newer model of same brand car and it appears the thermistor on this new seat at room temperature reads a resistance value of 8.55k ohms. Whereas the original seat thermistor in the exact same room temperature environment starts at 10.55k ohms.

So that means the newer seat thermistor has a potential offset of 2k ohms. Because of this, the heating element is turning off sooner than usual and not warming up as much. Is there a way to somehow inline in the wire to add some sort of resistor to increase the resistance to offset this by 2k? So basically whatever the thermistor is reading, the reading going to the control box should be plus 2k ohms on top of that.

Thanks.
Hi,

Back in the day we had to do some funny tricks with thermistors because the military did not want to use anything too new they wanted to stick with older, trusted methods.

One nutty thing you can try if you need 2k more of 'thermistor' is to put five 10k thermistors in parallel and then in series with the 8k thermistor. That gives you 10k overall.

If you have 10k and you need 8k, then you can try four 10k thermistors in series, then place that in parallel with the one 10k thermistor. That gives you an 8k thermistor.

Of course they all have to be NTC if the original is. If you can find them all with the same "B" value then you should get reasonable results.

Thermistors are not especially expensive.

#### BobTPH

Joined Jun 5, 2013
9,169
If the thermistor is accessible, just change it for the right one.

#### panic mode

Joined Oct 10, 2011
2,812
thermistors are embedded into the heater blanked that is just under the fabric/leather finish of the seat. normally that is close to back side of the seat cushion where the wire harness connects to the blanket. of you fold the seat fully forward, you may be able to reach the thermistor. but... i would not try to extract it, you stand more chance of damaging seat. i would simply find another one and just slide it into the seat under leather/fabric few inches from the back.

btw seats i worked with are using thermistor that is 8k at 20degC. resistance is nonlinear (exponential), at 0degC about 27k and at 100degC bout 540 Ohm. but... anything above 50degC is a problem - you do not want to melt things. adding fixed resistor in series will correct things for one temperature but as long as you set the offset for that target temperature (and not room temperature) you should be fine.

#### MrAl

Joined Jun 17, 2014
11,575
Thermistor short equation:
B is usually given in the data sheet.
Ro is resistance at temperature To, R is resistance at temperature T.

#### Attachments

• 3.8 KB Views: 1
Last edited:

#### panic mode

Joined Oct 10, 2011
2,812
The only caveat is that those things are part of the datasheet and automotive companies are not going to share the info unless one is working on a related project (seat tester for example).
Some have specified To different value etc
Fortunately, curve is only simple exponential so with couple of measurements and a bit of math one can derive sensor characteristic

#### Irving

Joined Jan 30, 2016
3,992
Here's a simulation of two J-type thermistor based on 10k @ 25degC and 8.2k @ 25degC. When used in the top half of a voltage divider with a fixed bottom half it produces a nearly linear voltage swing from 2 to 7v for 0 to 45degC. The cubic polynomials were generated from a trendline by plotting the graph of values in post #4 in Excel, gives a fit R = 0.9998 which is good enough.

Now change the bottom resistor on the 8k2 thermistor to a 10k. This causes the voltage to rise above that of a 10k thermistor resulting in cooler output. A better solution to putting a resistor in series with the thermistor is to use a silicon diode to offset the voltage. This tracks much closer to the 10k result.