LM334 as temperature sensor to PLC

[panzerschrek]

Hey everyone, I've got a question for you all.

I'm a college student here in Canada, and we're doing a power distribution control system, which uses as one of it's temperature sensor a cheap LM334 which we need to connect to a not so high impedance analog PLC input. I've already decided I would use a simple opamp as a buffer, but I was wondering...  The sensor, at 25°c outputs 64mV, and is directly proportional to temperature in kelvin. My question is quite simple I guess, I would like to be able to have a larger swing in voltage, since the datasheet states that it outputs 227uV/K, and that we only need to monitor temperature up to 100°c, the max output the PLC would read would be 373.15K * 227*10^-6 = 0.085V or therabout. I'm not sure I would be able to get good results with such a small voltage swing on an analog input, so again, is there a way that I can obtain, for a given small voltage swing, a larger voltage swing at the input of my buffer, which would be read by the PLC?

We don't have access to a lot of parts, generic opamps, TL082, 741, etc, and some passive of course. The goal of the project isn't to design a complete board, but rather designing a whole power system.

Anyway, I've heard of instrumentation amps, but I don't think we have access to these, maybe I could research into differential amps but we never saw those type of circuits so I'm not sure where to look at.

So, if you need anymore info, just ask, and if you have any tips and tricks, I'll willingly take them, but keep in mind that again this isn't a complete board design, simple a small circuit to somehow traduct a small signal swing into a larger one.

[katzohki]

You can probably get away with a simple opamp if you're not counting on reading the temperature at a fast rate. Decide on your target temperature range (e.g. 0-50°C), figure out what voltage range that would give you and then set up your amplifier so it will give you a wider output (e.g. 0-5V) that you can use.

[panzerschrek]

As a positive feedback amplifier? The problem I see with this, most likely because I have very little knowledge in this subject, is that it would obviously "amplify" the signal, say to simplify, 0.01V to 0.1V, to, if a gain of 10 is applied, 0.1V to 1V. It would actually offset it, but the swing would still be relatively small.

What I may end up doing is using multiple opamps with precision trimming potentiometer to set a voltage at their inverting input, so I can compare the output from the sensor and have a discrete output fed into the PLC. So instead of making analog treatment in the PLC, I would do it outside, that may be a way of dealing with it.

[Paul Moir]

I did as above:  offset and amplify using two op amps (LM358) for a ~.2-3.5v swing from 0 to 100degC (you have to stay well off the upper rail with the LM358).  But fortunately I was using a LM335 which is 10mV/K.  That's a lot easier to deal with.
 
One op-amp is a simple voltage follower, rigged up with a trimmer for fine adjustment to set the 0 point (2.73v in the case of the LM335 at 0 degrees).  The other was setup as a non-inverting summer, with gain added via the feedback resistor.  The voltage follower and the LM335 were summed.  I implemented the virtual ground the summer needs with two resistors, one to Vcc and one to GND, rather than the one resistor to ground.  Didn't get too fancy there, but it was good enough.

[ivan747]

See if you can get a hold of some OP07 op amps. They are available from several manufacturers and they are the generic precision op amp. If you don't have access to those, convince the people in charge of stocking it, since it's the best choice when you need a decent, cheap, popular op amp with low offset voltages and all of that. They are $1.50 in single quantities. It's pin to pin compatible with the LM741, with the exception offset null circuitry, which is optional in the end. Let's say it's the 741 of precision op amps.

[panzerschrek]

Just an update on my situation, I've actually completed the project, and simply wanted to explain what I did.

I went very easy with it, since the main part of the project wasn't building a sensor circuit, but designing a whole system. I started by using the LM334 as it was intended, a current source. I used a 10ohm resistor to set it's current, and a 1k resistor used to read a value. A current of 1 milliamp passing through the 1k resistor would equal a 1v voltage drop. For the values I needed, I obtain 6.67v, 7.67v, and 8.47v. Since these values aren't too close to each others, I decided to simply use the opamp as a voltage follower since the PLC input isn't exactly high impedance.

What I learned here is to always look upon more than only one way to approach a circuit. I had two main choices, read the voltage across the set resistor (which was very small) or use the current source itself. The right choice was obvious with what I had at hand at the time.

Anyway, thank you guys for your ideas, I definitely appreciated them.

Have a good day/evening,
Simon

[ivan747]

Thanks for the follow up, glad to see that you got it working and found another way around the problem. Although it doesn't sound heroic, sometimes the best solution is to give up on your current solution and try something else.