# Help! Which temperature sensor should I use with my PLC?

A

#### Alan K

Hello all!

Small issue on my road to control instrumentation knowledge.

I need to know what temperature sensor/assembly to use with my FL1D-H12RCC IDEC nano-PLC (Siemens rebadge, I believe).

Input is 0-10V analog PNP. I need to measure the temperature of the air right near the PLC - maybe two feet away inside a small enclosure (terrarium). There is a large 200W fluorescent bulb with attached ballast about two feet from the point of measurement, I don't know if that qualifies as a source of interference.

The overall application is maintaining an air temperature ceiling of 24 degrees celsius, via PLC/contactor activated vent fan. The lamp heats things up a bit, as does the heat pad. I keep reptiles.

I'm kind of new to this stuff. Should I direct wire, use a thermocoupling with some sort of amplifier? Should I use a temperature transmitter? Is there some device I am unaware of that is ideal for this? I'd like to spend as little money as possible.

Thanks for any help, and if I'm missing anything in my description, feel free to ask, I'll reply promptly.

-Alan

C

#### curt wuollet

Well you can go from $1 to$1000 easily. Most suggestions here will be towards the latter. I suggest you attach a LM35 temperature sensor. It works from 24 V and puts out a linear 10mV/degree C voltage, and it's very simple to use. You should be able to find one from 69 cents to \$10 depending on grade and greed. The diagrams are free on the National Semiconductor site and it will easily do what you want with good accuracy. You could also just use a thermistor, but then you have to solve a matrix to linearize it or use a lookup table. The LM35 is a lot easier.

Regards
cww

A

#### alan

cww, thanks! So far I've been looking at thermocouples with transmitters, but this seems rather cheaper. This thing needs no intermediate signal conditioner, then? Crikey!

C

#### curt wuollet

No, it will work in a simple 3 wire configuration and be certainly close enough for air temps, which are hard to be precise about anyway. It would benefit from being fed less than 24V to limit self heating. A 12-20V zener diode in series with the power lead will suffice as the output is very voltage independent and the current is on the order of 100 uA. The nice linear 10mV/degree C makes scaling pretty simple and it's absolute and requires no calibration. Gluing it to a small aluminum plate for some thermal mass helps average the readings, and makes it easier to mount. There is still a place for analog in our digital world.

Regards
cww

A

#### alan

I've been reading up on it and I can't believe how simple this is. Everywhere I've been told to use a thermocouple with a conditioner, but when I looked up the LM35 and read some of the tutorials, it surprised me how simple and cheap this was. Checked my plc's settings and sure enough they are fully compatible and easy as pie to program input response. It seems as if the controls industry has a serious case of expensive-propietoritis!

Thanks again.

Alan

M

#### M Griffin

If the signal is 10mV/degree C I think he also may wish to have an amplifier with a gain of about 10 or 20, and possibly some offset to put it in the correct range. That sounds like a cook-book op-amp circuit. Some pots in the circuit would allow calibration with a screwdriver.

At 10mV/degree C, the instrument range would cover 1000 degrees C, which doesn't give a lot of resolution if the measuring range only covers 5 or 10 C. A 10 bit A/D would only give 1 count per degree C, which would mean the signal would probably continually fluctuate by at least +/- 1 degree. That would be a fairly large chunk of the likely control range.

C

#### curt wuollet

An op amp certainly wouldn't hurt, but it does raise the complexity quite a bit. The LM35 could use a buffer if the A/D is low impedance too. That would kill two birds with one stone. But, for the application stated, control to more than a couple degrees is likely impossible anyway. Still air is notoriously difficult to measure and control. An op amp for higher resolution will let you control the temperature of the sensor more closely, but what relation that will have to the actual air temperature 2" away is debatable. I think the low res cheap approach should control to well within the reptilian comfort zone. A small fan to continuously circulate the air would make much improvement, but I have it on good authority that the fan would be more irritating than the deadband or uncertainty. Reptiles seem to be good heat integrators:^)

Regards
cww