scaling sensor outputs....

[Guest]

Hi,

I need to scale the output of a temperature sensor using op-amps. Can someone point me to a circuit that allows for independent adjustment of the offset and gain? I think i can do this with two non-inverting op-amps, but it's been too many years.....

Thanks,

Chuck

[PCM programmer]

Type it into Google. I typed in: opamp level shifter
and got this schematic as the first hit:
http://www.daycounter.com/Circuits/OpAmp-Level-Shifter/OpAmp-Level-Shifter.phtml

[Humberto]

Also you will try this one, it has gain and offset control.
It has a linearized response if your sensor is a thermistor.

http://www.edn.com/archives/1995/120795/25di5.htm


Humberto

[Guest]

Hi Guys,

Thanks for the input. It's not a thermistor, BTW, it's a solid state sensor. It's output goes negative for negative temperatures, so I need to offset the voltage, and then scale it to obtain my desired output. I want the "offset" and "gain" to be as independent as possible. I believe that a single stage circuit (like the ones suggested so far) have a small interaction between adjustments, which is not desirable in my application. I'd like to avoid that even if my circuit becomes more complex!

Thanks,

Chuck

[Ttelmah]

How 'perfect', is 'perfect'...
The offset, for an inverting amplifier is simple. All you do is run an op-amp, as a 'virtual earth' amplifier. You feed your 'offset' voltage into the +ve input, and the amplifier, amplifies the difference between this, and the input. This input becomes the 'earth' point for the op-amp. You adjust this voltage to set your zero point. Then you have a negative feedback connection from the output, to the -ve input, and make this into a potential divider, so that the ratio of this divider, sets the gain.
Now, if the op-amp was 'perfect' (infinite gain), an had infinite input impedance, there would be no 'interaction' here. However in the real world, there will be a _tiny_ interaction, as the op-amp itself, slightly loads the sources, and has less than infinite gain. However with a modern carefully selected op-amp, this really will be infintessimal. A few uV...
As a question, remember that 'adjustments', are expensive, in terms of your time (and even more so, in production equipment, in terms of 'setup' time). Do you really need adjustments?. If you know the tolerances of the sensor, you can determine the worst case limits of the output, and chose a zero point, and gain, which just puts both ends of this range, inside the limits of your op-amp (allowing for the tolerances of the resistors involved as well). You can then do a 'software' correction, putting the sensor into two known temperatures, and letting it calculate the slope, and zero-point of the sensor, and store these in the processor EEPROM, as 'calibration' values.

Best Wishes