problem w/ op-amp front-end to adc

School project: I am using a National LMC660 op-amp as a unity gain
high-impedance front-end to a TLC2543 adc. I intend for the sensor
input to the op-amp/adc is to be user selectable for either 0-5VDC or
1-5VDC. In testing, everything works fine until I disconnect the
sensor input from the op-amp. With the input to the op-amp open, the
output floats around 2 - 2.5 volts. I want the output of the op amp
to stay at the minimum value (0 or 1VDC), depending on range
selected).

Thanks in advance for any help you can give me!

That's the problem with a high impedance input, it picks up any old rubbish - is it in a screened metal box?, and are your input cables screened as well? (this will reduce it, but not cure it).

How about posting your circuit, so we can see how it's wired as well.

Be aware that in industrial process control situations is is usual for an input to be shown as either less than 0% or over 100% when the sensor is not there, thus preventing the situation where a faulty or disconnected sensor give a reading within normal range.

JimB

Jim/Nigel - Thanks for your posts.

Nigel, I don't think the main issue is noise because of the magnitude of the error I'm seeing. I think the issue is that I have an infinite impedence (open circuit) ahead of a unity-gain voltage follower op-amp.

Jim, you make a valid point regarding <0% and >100%. For the project I'm working on, if the sensor is not there (i.e. broken wire) I just want the op-amp output/adc input to go to zero. I guess what I'm looking for is a method to detect an open input and drive the input to the adc to the minimum value.

I hope this makes sense. Thanks again for your help!

RobH

cntrl.freak said:

Jim/Nigel - Thanks for your posts.

Nigel, I don't think the main issue is noise because of the magnitude of the error I'm seeing. I think the issue is that I have an infinite impedence (open circuit) ahead of a unity-gain voltage follower op-amp.

Click to expand...

I would say it was, it's picking massive amounts of noise up - don't you even have an input resistor to limit it somewhat? - which is why I asked for a circuit.

OK, if you want the output to go to zero, just put a low value resistor on the input.

Of course the big problem is, how low is low.
That depends on the input circuit and the sensors driving the inputs, and maybe how much inaccuracy you can tolerate due to the resistor.

JimB

JimB said:

OK, if you want the output to go to zero, just put a low value resistor on the input.

Of course the big problem is, how low is low.
That depends on the input circuit and the sensors driving the inputs, and maybe how much inaccuracy you can tolerate due to the resistor.

JimB

Click to expand...

If the resistor is only for getting zero volts out when the sensor is disconnected, he can use a large value, like 10Meg, unless that will put an unacceptable load on the sensor (such as with a pH transducer).

Ron H said:

If the resistor is only for getting zero volts out when the sensor is disconnected, he can use a large value, like 10Meg, unless that will put an unacceptable load on the sensor (such as with a pH transducer).

Click to expand...

Exactly, a low value resistor could in reality be of quite a high value.

JimB

JimB said:

Exactly, a low value resistor could in reality be of quite a high value.

JimB

Click to expand...

Jim Einstein's Theory of Relativity.
I was pretty sure that's what you were implying. I just wanted to clarify it.