Zener diode as temp sensor

[bigrockcandymountain]

Ok im looking for a little help here.  I sell solar pumps and they have a pcb control box.  They are set to shut down below -15 degrees and i need them to work down to -40

The first version was easy to fool.  They used a ntc thermistor 10k and i just soldered a 51k resistor in parallel.  Then the controller thought it was always above -15 even though it was colder.  It also has a high temp cutoff of 60 degrees.  The 51k resistor trick would lower that a touch, making it even more protected. 

Is there any way to solder something in parallel or in place of a zener that would fix me up without losing the high temp protection? I'm in way over my head.

I know the best solution would be to change the code to allow lower temp and leave the hardware alone.  I'm in even deeper over my head on that.

[Mary B]

There has to be some external circuitry around the diode to use it as a temp sensor. Stand alone it will not do the trick. A diode as a temp sensor is reverse biased and used as a current source that is measured across a resistor to get a voltage to drive further circuits.

Have a schematic/model number? Simple trick to fool this is put a resistor next to the sensor, apply a voltage to the resistor so it is warm.(uses a 1 or 2 watt resistor) and thus warms the temp sensor. Yes this uses some of the energy budget...

[bigrockcandymountain]

Sorry, no schematic or model number.  I'm not sure why i thought it was a zener diode except that it is labelled z2 on the board.  It was labelled z2 when it was a ntc thermistor too, so that shows that i only know enough to be dangerous. 

It tests like a diode.  It is a small surface mount package with a wk on it. 

Anyway, could you explain a little more how it would work reverse biased? I can kind of picture how it would work foreward biased using the voltage drop as the signal.  In that case, it might be possible to mimic with a resistor. 

I tried a 51k resistor parallel to it like i did with the earlier version with no luck.

[bigrockcandymountain]

Here is a shot of the board with an arrow pointing at the component where the thermistor used to be. 

[Bruce S]

What's the specs on the R2?
It looks like Z2 is already in parallel with R2, could be just the pic, but the solder lines make it look like it's already in parallel.

[Mary B]

details here... it requires an op amp to amplify the changes to make it useful and I don't see one near that circuit... the pic is to blown out/low rez to be able to make out component markings...

https://www.arrow.com/en/research-and-events/articles/using-a-simple-diode-as-a-ballpark-temperature-sensor

[Mary B]

(Attachment Link)

Here is a shot of the board with an arrow pointing at the component where the thermistor used to be.

Looking at that it looks like it might be biasing Q9... R1 might change that bias...

[Scruff]

Resistors sound awful fancy and delicate. I'd be more inclined lay a reptile tank heating matt somewhere clever..like on a heat sink or defiant diode. You could automate it to not cook with a cheap ebay adjustable thermostatic relay board...I dunno if their is such an ebay board but sounds entirely plausible.

Ain't broke logic on the pump controller board.

[joestue]

more likely its a diode, not a zener.

-2mv/degree C change.

[bigrockcandymountain]

Yes i think r2 is parallel with z2.  It is 100k resistor i think.  R1 is 150k resistor i think and in series with the z2 diode.  They have full system (24v) across them.  The funny thing is that there aren't any small traces that lead anywhere on either side from these components.  They are hooked to a transistor q9 but that is all. 

It is looking like i might have to just heat the board. 

[OperaHouse]

I remember a pump board that had to operate in the cold, below what the microprocessor was rated for.  They put a resistor over the micro and bonded it with something to conduct heat. It was just powered all the time. Not all your boards may work when it gets that cold. Simple very small thermo switches that turn on at freezing may be needed to heat entire space. That sensor wasn't put there judt for fun.

[bigrockcandymountain]

Ok so the consensus is that i should warm it up rather than try to fool the sensor.  I was being hard headed, but i guess i have been told enough times that i will try it.  I would like the heating to cut in about -10c.  Any recommendations for a thermostat that is super reliable?

Just heating the micro would be great, but i think the sensor is on the other side of the board from the micro, so I'll probably have to heat the whole board. 

[Mary B]

I use a bunch of these for temp control on things around the house https://www.amazon.com/Inkbird-Temperature-Controller-Fahrenheit-Thermostat/dp/B0152LYY0I/ref=sr_1_5?dchild=1&keywords=temp+controller&qid=1631901635&sr=8-5

[joestue]

Ok so the consensus is that i should warm it up rather than try to fool the sensor.  I was being hard headed, but i guess i have been told enough times that i will try it.  I would like the heating to cut in about -10c.  Any recommendations for a thermostat that is super reliable?

Just heating the micro would be great, but i think the sensor is on the other side of the board from the micro, so I'll probably have to heat the whole board.

some circuits use a constant current source to send current through the diode to get better temp accuracy.

as far as heating the board, use a ptc resistor. as it warms up it will draw less current. no brains required.

[bigrockcandymountain]

Mary B those look good but won't fit inside the enclosure they need to.

Joestue, i looked at a few ptc resistors, but couldn't find any with low temp curves.  I didn't look terribly hard. I would be very interested if you know of some that start to cut out at freezing. 

I ordered some bimetal disc thermostats with 0c rating and wire wound ceramic resistors 5w 330 ohm.  That should give me 2w or so at 24v. 

I'm thinking wrap the controller in insulation and 2w will keep it above -15c. I don't really like the bimetal disc because in my experience, they aren't good for many cycles, but I'll give it a try.  It's cheap and available, and should fit in the space no problem. 

[Mary B]

They are actually fairly small... and I would make an external controller in a separate box, that way if it fails it is a heck of a lot easier to repair.

[DanG]

how about pinning an LED, zero distance from SMD (attached) so that it shines on the component?

[Astro]

And we found our electronics guys. I am lost just reading this post, but feel much better that you all are here should I ever have a question.

[bigrockcandymountain]

Yes, there are lots here that can do some amazing things with electronics.  I am a consumer of the knowledge, not one of the gurus.  I would name names of the extremely talented ones, but there are a lot and i would miss someone for sure.

Thanks for the help on this problem everyone. There are lots of good ideas, so if what i try doesn't work, I'll come back and read and try something different.