NTC temperature probes in parallel. How?

[desA]

I have a heat-pump application where there are two NTC temperature probes (Carel NTC0150055), but only one spare input into the controller.

I'd like to find a way in which the two probes could somehow be joined (bridged, or other), to try & achieve an 'average temperature'. The output of this will then feed to the controller input.

In the 'old days', this was relatively straightforward with thermocouples, but, I've read that a direct bridge for thermistors is not possible, as one leg will steal all the current from the other parallel leg.

Is there an 'averaging' device on the market which can accept two NTC thermistors & provide a useful output signal acceptable by the controller (Carel)?

[Yuri B.]

Hello DesA. Nice question. Two resistances in parallel R1*R2/R1+R2 will be of course unacceptable for the controller. I would, adding some constant resistances, experemantally bring the value between the two probes to an acceptable level. However, this being a problematic solution, there MUST be a flawles one. Let's wait for more posts.

[desA]

Thanks, Yuri.

[chemi-cool]

The resistance of NTC is 10KΩ, If you connect two in parallel, resistance will not change but I am not sure about average temp reading.

[Argus]

.




It's the same as any bridge circuit.

To obtain the average of any set of resistances, you need to have four sensors, it cannot be done with two.

They must be connected in two pairs; R1 and R2 in series, then R3 and R4 also in series. Next you connect each pair in parallel.

Each sensor can be strategically spaced throughout a large room, but as the cable is usually 5 vdc, avoid voltage drop in long cables and avoid also RF pick-up. It's best to use screened cable, earthed at one end only, and keep the wires away from lighting chokes and power cables.




.

[Gary]

Please correct me if I am wrong.

If we wire two thermistors in parallel at point A and two thermistors in parallel at point B, then wire the two sets of thermistors in series, the result should be an average of points A and B.

[Gary]

Great minds think alike.

[Yuri B.]

Four of 10kOhm in a serie with a bridge (not at the input of course).

[Yuri B.]

Indeed, 10kOhm in the result.

[Yuri B.]

If we name them (in the serie) A,B,C,D, then A and C should be palced at one controlled place and B and D at another. Am I right or wrong?

[Yuri B.]

Sorry, of course, two pairs of parallelly connected probes joined in a serie, but still, the pairs should not be placed together? (but interchangeably)

[Gary]

One parallel pair in the first location, the other parallel pair in the second location.

[Gary]

Hmmmm... now that I think about it, it really doesn't matter which two sensors are at each location. The result will be the same, i.e. an average of the two locations... although the physical wiring of the circuits may be easier if the parallel sensors are at the same location.

[Yuri B.]

I am sorry, but I believe, probes intercanged in the pairs would work compensationally for temp differences at each controlled place. (so far as we are looking for the average temp)

[Gary]

You may want to test this theory and let us know if there is any difference.

[Yuri B.]

Anyway, thanks Argus and all others sharing with other people their knowledge.

[brunstar]

what you may need to do is get a multi input controller which will read both and give an average output.

[desA]

Thanks for your input everyone - this is really useful information. The bridge aspect makes very good sense.

what you may need to do is get a multi input controller which will read both and give an average output.

The controller in use has two inputs - one is already in use for another task, I'm only left with one at this point. TEchnology on a budget.

Next thought to consider:
What if one of the sensors goes down?

Does the whole array go down? What happens to the input signal to the controller? How to manage this in the controller?

[Yuri B.]

The controlling proces will go astray. The resistance of the entire chain will increase what will (falsely) perceived by the controller as temp increase. Other probes will be impaired, controller too.

[Yuri B.]

That is, as temp decrease of course.

[desA]

^ Thanks Yuri.

For a heat-pump, with the water loop driving up the measured temperature, on failure of one NTC probe, what would the connected array signal do?

[Yuri B.]

Oh, those English negative prefixes! Will stay unimpared of course.

[Yuri B.]

As temp will be falsely perceived as decreased (as if it became colder), heat pump will keep on driving the temp up.

[desA]

^ This would be problematic, for a heat-pump, I would expect.

Even with a reset timer, it would then keep trying to re-start & drive in heat.

Not good.