Should calibration range of Oscilloscope cover the test ranges?

[Ajit Basrur]

I have this question on the calibration of oscilloscopes.

In performing an internal audit, we’ve come across a couple of instances where oscilloscope calibration ranges may not appropriately cover certain test ranges. As these are complex test instruments, calibration service providers perform a subset of tests outlined in the manufacturer’s calibration procedure, after which the device is certificated as “meets manufacturer’s specifications”. Our concern is that this subset of tests does not explicitly call out some of the ranges we are testing in.

As an example, DC gain accuracy is tested at ranges from 1mV/div up to 1V/div, but our tests are performed at 5V/div. There rationale may be that if oscilloscopes are within calibration at the lowest extremes, it implies that they are within manufacturer’s specifications at higher settings – this seems reasonable based on the internal workings of oscilloscopes, but we’re curious as to how others address this situation.

Thank you Covers

 

[BradM]

Re: Should calibration range of Oscilloscope cover the test ranges ?

This is the same issue I had with handbook 44 for scales. They allow large scales to be tested at only a fraction of their full range.

If there is a detailed (validated) test plan where it has been demonstrated that the high range doesn't need to be verified, I might accept abbreviated testing.

Otherwise, I think an instrument needs to be tested at points that includes the use range.

 

[Marc]

Re: Should calibration range of Oscilloscope cover the test ranges ?

Other than the scales aspect which I am not familiar with, I agree with Brad but hesitated to reply because I have no experience in oscilloscopes in testing since the 1980's working at the then Cincinnati Electronics. And knowing you're as knowledgeable as I am in calibration, I figured you know my understanding of "standard" acceptance of calibration in the range of measurement use.

If it were me, I would contact the manufacturer to get their justification. What you say does sound reasonable, but at the very least I'd have something in writing to back it up as Brad says.

I was thinking there was someone who visits here who has dealt with electronic test equipment but I can't remember who.

 

[Ajit Basrur]

Re: Should calibration range of Oscilloscope cover the test ranges ?

Yes Marc, we have written an identical email to the manufacturer who did this calibration.

 

[normzone]

Re: Should calibration range of Oscilloscope cover the test ranges ?

I've little to add here, but I did have an instance where after the scope was returned from our cal house certified as in spec, one EE told me he knew of an error in the scope and he did not mention it because he wanted to see if the cal house would find it.

Of course, the cal house did not find it. We sent it to the manufacturer for servicing.

 

[Mikishots]

Re: Should calibration range of Oscilloscope cover the test ranges ?

I've little to add here, but I did have an instance where after the scope was returned from our cal house certified as in spec, one EE told me he knew of an error in the scope and he did not mention it because he wanted to see if the cal house would find it.

Of course, the cal house did not find it. We sent it to the manufacturer for servicing.

Expensive experiment.

 

[dwperron]

Scope calibration is a blast!
It all depends on the instrument, and the manufacturer knows the details and incorporated them into their calibration / verification procedures. Trust them.

For instance, the scope in question is checked from 1 mV to 1 V/div, but it goes to 5 V/div. Chances are the amplifier stages are 1,2,5 steps, so you would check at 1,2,and 5 mV. For higher ranges the scope switches in X10 attenuator steps, so you would check each of those as well. The top attenuator step would cover 1,2, and 5V/div. Since you have already checked the 2 and 5 amplifier gain steps there isn't a requirement to check them again for the 2 and 5V/div ranges, but the calibration is still valid for those ranges due to overlap in what is being tested.

That's the same rationale used in multimeters, where all attenuator ranges are checked but amplifier linearity and frequency response are often only checked on one range.

 

[Hershal]

Yes, it should, short answer. Your accredited calibration provider has a requirement under ISO/IEC 17025, Clause 4.4, to understand your requirements and needs.

Your scope absolutely should be calibrated to the range of the test requirements. Now, modern scopes may be using the same concept as modern balances, which is 2-3 points and the algorithms extrapolate from there. That means in theory that the scope will read correctly at any point. But best to make sure the cal provider does cover the range.

Hope this helps.

 

[notemadres]

I would just send it to the manufacturer for servicing.