I have a question for all the experienced engineers of the forum.

There is a small debate I am having with my colleagues about testing an op-amp's PSRR and CMRR. I am sure many have seen or at least heard of the servo loop being used in such tests.

The test team recently cited MIL-STD-883 for the CMRR and PSRR tests (methodology and circuit design) stating that if the tests are not done dictated exactly as MIL-STD-883 then we are violating JEDEC standards.

I am not convinced that MIL-STD-883 is a JEDEC standard; although, it may be referenced and/or incorporated by other JEDEC standards. MIL-STD-883 is a standard from the DoD and focuses on Military and Aerospace applications.

Anyone have their opinions or can set me straight?:

 

Is the original specification you're using JEP121A?

MIL-STD-883 is called out in it so it is required.

You have to subscribe to get free access to the JEDEC document but here is the mil-std

http://scipp.ucsc.edu/groups/glast/electronics/mil-std-883.pdf

Why would you want to deviate from the specification? They are usually called out because an engineer knew it was needed or it's required by contract. If you feel the testing is excessive you should bring that to the attention of your team and possibly find and provide a more relaxed specification.

 

Is the original specification you're using JEP121A?

MIL-STD-883 is called out in it so it is required.

You have to subscribe to get free access to the JEDEC document but here is the mil-std

http://scipp.ucsc.edu/groups/glast/electronics/mil-std-883.pdf

Why would you want to deviate from the specification? They are usually called out because an engineer knew it was needed or it's required by contract. If you feel the testing is excessive you should bring that to the attention of your team and possibly find and provide a more relaxed specification.

We do not specify that we test to MIL-STD-883 or any other JEDEC standard. The argument was that since CMRR and PSRR (among other tests) are spelled out in these standards then we are violating industry practice. The assumption is that everyone else and their mother tests op-amps to this standard.

I maintain that the DC servo loop and the test methodolgy that goes along with it may be incorporated into these standards but it is not THE standard and since we and no one else explicitly states that they test to 883 or JEDEC then the point is moot. We can vary the test any way we see fit since we do not claim 883 or JEDEC compliance.

No datahseet that I have every seen lists 883 or any other JEDEC standard under Vos, Ib, CMRR, PSRR, etc.

 

Unless you specify that your parts meet MIL-STD-883 then you are not violating
*your* specification. If a customer specifies MIL-STD-883 and the company
agrees to meet it then you are bound by that agreement.

I have written many component test programs (including op-amps) and I *only* work off of the published specification or a customer specification.

If testing to MIL-STD-883 does not increase your test time or require a lot of
development time it may make sense to do it. That way when the salesman asks
if you can do it then you are ready. If it is going to slow down all production or require
major changes to existing hardware/software it doesn't make sense.

Also just testing to MIL-STD-883 is probably not enough to get a MIL-STD product.
I am sure there are a variety of other process requirements to meet MIL-STDs
(like a long burn-in).

(* jcl *)