NTC thermistor and noise

[brewdexta]

Do you know what resistance the NTC is? If it's something like 10k then a few ohms of connecting wire will make little difference.

It is 10K according the spec sheet

Accuracy in this use isn't paramount, but a constant temp is. Obviously accuracy is preferred but if the missus feels cold when in her workshop, she'll either get a jumper or turn it up.

However, in the brewery I'm using PT100 probes, that's a different story as I need to be within about 0.5 degrees if I want to get all nerdy

 

[brewdexta]

That pretty much all we would use for rtd use, 0.75 or 1 mm usually but it wouldn’t harm being bigger. If you have some handy it may be worth trying on the floor before running the cable - there may be something more fundamental that is wrong with using a 2. wire thermistor at a distance with that type of instrument

That’s the unknown for me, as I’ve only ever seen thermistors used as rough temperature indicators and always pretty local to the instrument - that’s not to say it won’t work but I just haven’t seen it done before. Be interesting to see what you find.

I'll get a bit of 3 core 1.5mm2 and give it a whirl

 

[brewdexta]

I've gone for 2 core, 1.0mm CY cable as its fully screened compared to SY which "has a degree of screening"

Its a bit cheaper too, and being a bit less chunky I can use it in the brewery where I need to extend a couple of PT100 probes and it will fit through existing cable glands.

I shall report back when it arrives, probably next weekend.

Thanks for the help guys

 

[m_c]

Hmm well, I don’t know what the resistance is at the temperature it’s measuring but even a small resistance can give quite a big error - hence why for any accuracy 3 or 4 wire rtd are usually used to eliminate/compensate for the cabling resistance. That and the fact in 20 odd years of doing temperature control for a living and never having seen thin comms cable used to wire up a resistance temperature sensor tells me that it *might* be a factor, though as I say I’ve no direct experience of using a thermistor for this kind of application, but the theory should be the same as for an rtd, just with a different resistance/temp curve.

How you do things in a professional install, and an install where you just want things to work with what you have laying around can be quite different.

As @daedalusminos has said, a few ohms resistance in wiring will make little difference to a thermistor in the several kOhm range.

 

[daleyd]

How you do things in a professional install, and an install where you just want things to work with what you have laying around can be quite different.

As @daedalusminos has said, a few ohms resistance in wiring will make little difference to a thermistor in the several kOhm range.

That’s the thing - we didn't know it was in the k ohm range until the last post. If it is then maybe it maybe wont make much difference but there’s obviously some reason it’s not working with the addition of a length of cable - take the professional part out (it’s not something I’ve ever been accused of before ) but I’ve genuinely never seen cat 5 cable used, even on some cobbled together jobs I’ve been on - I assume thats for a reason.

 

[brewdexta]

That’s the thing - we didn't know it was in the k ohm range until the last post. If it is then maybe it maybe wont make much difference but there’s obviously some reason it’s not working with the addition of a length of cable - take the professional part out (it’s not something I’ve ever been accused of before ) but I’ve genuinely never seen cat 5 cable used, even on some cobbled together jobs I’ve been on - I assume thats for a reason.

The cat5 is my fault, my day job is networks so I use it like farmers use bailer twine

 

[brewdexta]

I got around to installing the 2 core, 1.0mm CY screened cable today. I tried it with and without earthing the screen at the relay end. No good, it was worse when earthed. I didn't get around to earthing both ends. However, I then tried a 100nF capacitor across ground and input, a little better. So I upped it to 4.7uF and all's well. temp variation has gone from around 4c to 0.2c which is good enough. I'll have a play with earthing both ends in case its acting like an antenna tomorrow. I'll also try the capacitor on the link to the other building where I can't swap out the cable as it goes underground.

 

[DSR]

I think the problem is more than likely due to the high resistance of the thermistor - 10K and the length of the wire. The wire will be picking up all sorts of electrical noise, the current through the thermistor will be low and a lot of the noise will be superimposed on the signal.

4-20ma signals are used because they are very resistant to noise, you ideally need to pass several ma through your thermistor so that it masks the stray noise. You would need 40V to pass 4ma through the thermistor !!!

Screened cable would improve the situation, a single twisted pair would be better than unshielded straight cable. I'd also consider using a differential input rather than single ended input to the amplifier, differential inputs tend to reject common noise i.e. the stray signals picked up in the cables.

As an aside this is my 63rd post - that's 2^6 - 1 or 111111 in binary (It's a geek thing)

 

[brewdexta]

I'm happy enough with the effect of the cap, it's good enough to stop the relay kicking in and out. It's now almost on par with a sensor directly connected to the relay. But I will have a bit of a play with earthing both ends without the cap to see how much difference it makes. If the cap works on the cat5 to the other building, that's a bonus

 

[DSR]

The cap is a good idea - it will provide a low impedance path to ground for the nose and spikes - don't use an electrolytic cap though - use a non polarised cap e.g. polyester. 4.7uF are readily available but are more expensive than electrolytics, however electrolytics have different characteristics, e.g. greater leakage current, inductance etc. that make them less efficient in this case. The nature of the signal from the thermistor (slowly changing) also helps here. If it was a rapidly changing AC signal, the capacitor would effectively short it to earth so make the situation worse.

 

[brewdexta]

I used used a ceramic cap, as I have lots ranging from 1pF to 10uF. The 4.7uF just happened to be the first large value cap I found, the labels are really small for tired old eyes, I had to get the magnifier out.

 

[Chris Snow]

Other things to try if needed.

Decade filters. IE multiple capacitors in parallel like 100n 1u etc. Always ceramic.

LC filters

Common mode chokes


That's the usual things I try when fixing problems when doing immunity testing.

 

[DSR]

I used used a ceramic cap, as I have lots ranging from 1pF to 10uF. The 4.7uF just happened to be the first large value cap I found, the labels are really small for tired old eyes, I had to get the magnifier out.

Yes. I started wearing specs a few years ago and it still irks me that I need aids to read stuff that I just used to be able to glance at. I need a glass to read colour coded 1/4W resistors.
I'm slowly working through my small components collection and sorting them - I recognise the 'just happened to be the first large value cap I found' syndrome