Behavior between different Op-amps

[lcodar]

I would appreciate some practical advice on a problem I'm having with a project I'm working on.

I have implemented a simple voltage adding circuit using an op-amp. (see http://homepages.which.net/~paul.hills/Circuits/Adder/Adder.html for the generic circuit)

Using one amp of a LM348 op-amp, the circuit performs as expected.

The application for this is to add the output from an array of detectors into one signal...and that works just fine at low frequencies. As I increase the frequency of the signal, the magnitude of the op-amp output drops.

I think it is due to the slew rate of the LM348 being about 0.6V/uS. Which at 40kHz does not leave a lot of "breathing room" to get a 5V p-p signal ramped up.

So I head out and look for a new "better" op-amp with improved slew and GBW parameters. I'm trying the TI OPA4228 (see http://focus.ti.com/docs/prod/folders/print/opa4228.html )

I drop it in the circuit (having been forwarding-looking enough to use a socket instead of soldering the Op-Amp right to the board)...and I get no output. (well.more accurately I get a DC output at some level) I try a couple of the op-amps all with the same behavior.

What in the world could cause such a dramatic difference in the behavior between the op-amps?

What can I do to trouble shoot the problem?

Or what can I do to understand if a given model op-amp will work in my circuit before I just write Digi-key and tell them to send me one of everything they've got .

Thanks for your help.

 

[berkeman]

Sounds like I should buy stock in Digikey!

You probably already checked the obvious stuff like pinout, allowed supply voltages, etc., right? If the opamp circuit is not working, you should follow the usual debug procedure for opamp circuits, which is to first check the power supplies and make sure the part is not getting hot (indicating a bad connection), and then check for the proper operation of the "virtual ground" effect. That is, make sure that the voltage on the - input due to feedback matches the voltage that is on the + input. If it doesn't something is wrong with the feedback, or the opamp is bad/blown.

If you still can't figure out what is wrong with the circuit using the 2nd opamp, just build a simple follower or inverting amp with the 2nd opamp separately, and make sure that it works. That will help you find any subtle pinout issues, etc. Let us know what you find.

 

[berkeman]

Couple more things. Could you clarify which package you are using for the LM358 and OPA4228? AFIK, the LM358 is a dual opamp in an 8-pin package, and the OPA4228 is a quad opamp in a 14 pin package.

Also, what are you doing with the unused opamp in the package? It should be tied off for noise reasons.

And what are the values of your input resistors and feedback resistors? At some point, the RC lowpasses formed with the input capacitance of the opamp (and related interconnect wiring) will roll off your frequency response.

 

[lcodar]

Good questions:

In this case I'm using the dual op-amps in the 8-pin package. I should have said the OPA2228 for the TI op-amp.

Input resistors are 5k, feedback resistor is 10k.

 

[lcodar]

OH...and the Digi-key stock might be a good idea. I'll let you know when I get home tonight!

 

[berkeman]

What power supply rails are you running? The schematic you show has a ground reference for the - input. The LM358 is specially designed to tolerate inputs near ground when running off of a single supply (like 0V and 5V), but the OPA2228 family of opamps is designed for use with split supplies, and the input Vcm only comes within 2V of the rails.

My guess now is that the input voltage is not within the allowed range for the OPA2228 part. Try it with split supplies, like +/-12V, and I'm betting she works.

 

[lcodar]

Ah...interesting. I am running 0 to +12 volts supply. I'll try the split voltage trick tonight.

I'm not too sharp on this, what is the input Vcm?

 

[berkeman]

Ah...interesting. I am running 0 to +12 volts supply. I'll try the split voltage trick tonight.

I'm not too sharp on this, what is the input Vcm?

It's defined in the datasheet. Something like the common-mode voltage of the inputs (the average of the two input voltages), but I usually also try to keep both inputs from going outside the allowed Vcm just to be safe. Violating the Vcm limits is a very common mistake with opamps. Of course, it's never happened to me...

 

[lcodar]

I found this article (http://pdfserv.maxim-ic.com/en/an/AN656.pdf), which I think is on target with what you suggested.

I might try the "conversion" that is represented by the differences in these circuits: http://focus.ti.com/lit/an/sloa093/sloa093.pdf

Thanks for your insight...I'll let you know how it goes