sensor voltage conditioning

[tom86951]

I have a sensor that produces 0 - 5 vdc. I would like to create a circuit to change this output so that anything below 2 volts translates to 0vdc, and so that the the sensor's range from 2 to 5vdc translates to 0 to 5 vdc. Seems like I need to start with something that subtract 2 vdc from the sensor's output (down to but not less than 0vdc), and then multiples that result times 5/3. Any thoughts on how to do that with op amps and such...?

 

[Mikebits]

This sounds like a classic case for a comparator.
Read all about it.
**broken link removed**

 

[crutschow]

This sounds like a classic case for a comparator.
Read all about it.
**broken link removed**

A comparator will give you a logic output. I believe what's wanted is a linear output over the range of 2 to 5V while ignoring inputs below 2V.

You can use a non-inverting op amp with a gain of 5/3 and a negative 2V offset.

Is this homework or a real problem?

 

[Mikebits]

I guess your right, he wants to ignore anything from 0-2v, and then output a proportional voltage from 2-5 volts where 2 volts =0 and 5 =5. I did not read the post carefully enough.

 

[ericgibbs]

I have a sensor that produces 0 - 5 vdc. I would like to create a circuit to change this output so that anything below 2 volts translates to 0vdc, and so that the the sensor's range from 2 to 5vdc translates to 0 to 5 vdc. Seems like I need to start with something that subtract 2 vdc from the sensor's output (down to but not less than 0vdc), and then multiples that result times 5/3. Any thoughts on how to do that with op amps and such...?

hi,
I designed this offset circuit for a similar sensor, if you recalculate the resistors etc, this circuit will give you,
within a few mV's of zero and +5V, for the +2V thru +5V input signal.

Keep the load resistance that the 358 drives over say 10K, [which a PIC is]
then you should get as low as 5 to 10mV, for a "0V" out.

Can you recalc the values OK.?

A CA3140 or CA3240 OPA will be a suitable alternative, but NOT a 741..

 

[tom86951]

A comparator will give you a logic output. I believe what's wanted is a linear output over the range of 2 to 5V while ignoring inputs below 2V.

You can use a non-inverting op amp with a gain of 5/3 and a negative 2V offset.

Is this homework or a real problem?

It is a real world project. I have a pressure sensor that measures from absolute zero (vacuum) to 30psi and produces a linear 0-5vc. I want to use that sensor to drive a pressure gauge that takes a 0-5v signal and reads from atmospheric pressure to 30psi. To do that, I need to ignore the first few volts of the sensor (which is vacuum below the gauge's ability to display) and then amplify the signal above 2v by a 5/3 ratio.

 

[tom86951]

hi,
I designed this offset circuit for a similar sensor, if you recalculate the resistors etc, this circuit will give you,
within a few mV's of zero and +5V, for the +2V thru +5V input signal.

Keep the load resistance that the 358 drives over say 10K, [which a PIC is]
then you should get as low as 5 to 10mV, for a "0V" out.

Can you recalc the values OK.?

A CA3140 or CA3240 OPA will be a suitable alternative, but NOT a 741..

Eric, thanks once again for your help. It looks like you are on to something, but I need a little help understanding the circuit and how to recalc the values for my 5/3 gain and -2 offset. Can you dumb it down for me?

 

[OutToLunch]

I would suggest using an instrumentation amplifier. I will assume that the output impedance of your transducer is rather low, but you want very high impedance on the input to your gain circuit. See the wikipedia page on instrumentation amplifiers:

Instrumentation amplifier - Wikipedia, the free encyclopedia

Put a 2V reference (many ways to do this) on the V1 node and put the output of the transducer on the V2 node. Set the gain to 5/3. Again, there are a number of ways to do this - one would be to set R1=Rgain and then set R3=5k and R2=9k.

This assumes that you are using a set of op-amps that have a supply that exceeds 5V on Vdd and is also supplied by a negative rail on Vss. If this is the case, then I would suggest adding another stage to your output: Tie the output of the instrumentation amp to the non-inverting input of an op-amp set up as a voltage follower through a 10k resistor. On the non-inverting input of the voltage follower, you can tie the output of a comparator (biased from a positive rail and referenced to ground) that is comparing the input to that same 2V reference. The comparator will pull the input to the voltage follower to zero volts if Vin<2V, otherwise you get the output of the instrumentation amp. The op-amps can be in a quad package with the comparator being on its own.

If you are biasing the op-amps with ground on Vss, then you'll have to go with some type of rail-to-rail op-amp and figure on having some accuracy issues near the rails. With this set up, though, you only need the three op-amps set up as an instrumentation amp.