Handheld meter robustness testing

[Fungus]

I am mostly curious if the standard omits the transient tests for inputs like ohms

I believe it specifically includes them, that meters have to "fail safely" with all possible combinations of selector switch and input jack.

as a way to explain how a high end meter like the keysight could pass third party testing, but still fail at your less demanding tests.

CAT rating is a safety rating. If the user was in no danger then it's a pass!*

Re: The requirement to show hazardous voltages:

Is there a hard requirement in the standard to indicate unsafe voltages or is is the requirement something more like "never show lower voltages than are actually present"? A blank screen would meet that requirement - if it's not showing you any voltages at all then it's not telling you any lies.

(*) I'm not an expert, I haven't read the standard, this is just my understanding from reading Internet forums. I love being corrected.

[NoMoreMagicSmoke]

I am mostly curious if the standard omits the transient tests for inputs like ohms

I believe it specifically includes them, that meters have to "fail safely" with all possible combinations of selector switch and input jack.

But that's your opinion. My question is regarding the actual wording in the specification. How do you know that the meter must be tested on all possible combinations if you have not read the spec?

as a way to explain how a high end meter like the keysight could pass third party testing, but still fail at your less demanding tests.

CAT rating is a safety rating. If the user was in no danger then it's a pass!*

Again, how do you know that the user would not be in danger? Joe's tests have tiny amounts of energy behind them compared to the actual CAT ratings. All his tests demonstrated is that the meter had a breakdown internally and was irreparably damaged at just above half the CAT rating. How do you know that the meter would not fail catastrophically if there was sufficient energy behind the pulse? If you can tell this from Joe's test you are a better expert than I am.

All we know from Joe's testing is that an ESD event with a 5KV or greater potential can irreparably damage the meter. I can tell you one thing. I would NOT be comfortable using a meter that can breakdown internally at half its rating if I were working in an arc flash environment. In that environment it's certainly possible that a minor breakdown inside the meter could turn in to a safety hazard real fast! Again, are you really able to extrapolate safety from Joe's tests? I always thought safety testing must be done with a true combo generator.

Is there a hard requirement in the standard to indicate unsafe voltages or is is the requirement something more like "never show lower voltages than are actually present"? A blank screen would meet that requirement - if it's not showing you any voltages at all then it's not telling you any lies.

(*) I'm not an expert, I haven't read the standard, this is just my understanding from reading Internet forums. I love being corrected.

I am not one to speculate so I will leave this for people who have actually read the standard to answer.

[Fungus]

CAT rating is a safety rating. If the user was in no danger then it's a pass!*

Again, how do you know that the user would not be in danger? How do you know that the meter would not fail catastrophically if there was sufficient energy behind the pulse?

If you go and look up the certifications for this meter you get this:



ie. Keysight are stating that it meets the relevant standards.

https://regulations.about.keysight.com/DoC/DoC_U1280ANU1280AU1281Aa_10-Aug-2021-05_56_58.pdf

Edit: I'd be happier if they published some certificate numbers as issued by some independent testers, but they don't. All they say is "The products were tested in a typical configuration with Keysight Technologies test systems."

Maybe Keysight are too cheap to do independent testing and that's why the only marking on the back of the meter is the "CE" mark.

I would NOT be comfortable using a meter that can breakdown internally at half its rating if I were working in an arc flash environment.

So thank joe for making this video and don't buy a Keysight.

Again, are you really able to extrapolate safety from Joe's tests?

Who says I'm extrapolating anything from joe's tests?

[joeqsmith]

...
All we know from Joe's testing is that an ESD event with a 5KV or greater potential can irreparably damage the meter.
...

This is not correct. 

I posted data for the grill starter along with the gun I designed to simulate an ESD event based on the IEC standard.  The transients these both produce are a few ns wide.    Part4 shows applying transients from both.   

I refer to the transient derived from the generator shown in Part 5 as a "surge" but I make it very clear that it is not the same as what is defined by the 61010 low voltage directive.   I've also explained the reason for this.    What damaged the U1282A was this surge waveform, not the ESD. 

[joeqsmith]

That's why I am curious about how the standards are worded, and I thought you had a copy (I do not have access to a copy) which is why I asked you how it is worded.

I do and believe I had already covered this.   The standards are available for purchase if you are interested in obtaining them.

[joeqsmith]

I would NOT be comfortable using a meter that can breakdown internally at half its rating if I were working in an arc flash environment.

So thank joe for making this video and don't buy a Keysight.

It certainly made a noise but hard to say.  I ran the U1231A without the case and switch to get some idea what may have happened.

Going back and watching the videos, it appears to be a very similar if not an identical failure mode.  I had ran it with that half cycle generator as well. Not a lot of energy but still gives us some insight.   A few viewers suggested that the damage to the switch contacts what the cause of the problems during the switch life cycle testing.   This time, the plan is to wait to do any further destructive testing until after the switch has been cycle tested. 

[joeqsmith]

Mismatches of inputs and ranges are examples of REASONABLY FORESEEABLE MISUSE, even if the documentation or markings prohibit such mismatch.

The attached image was taken shortly after design an construction of the transient generator (2016).   Shown with 5kV peak, 100us FWHH, no load.  This is the transient that damaged both Keysight meters.   

[AVGresponding]

Mismatches of inputs and ranges are examples of REASONABLY FORESEEABLE MISUSE, even if the documentation or markings prohibit such mismatch.

The attached image was taken shortly after design an construction of the transient generator (2016).   Shown with 5kV peak, 100us FWHH, no load.  This is the transient that damaged both Keysight meters.

Just out of curiosity, have you looked at the waveform of the grill starter output, across a nominal load of say 10MΩ, to simulate what it would look like to the front end of a typical handheld DMM?

[joeqsmith]

Just out of curiosity, have you looked at the waveform of the grill starter output, across a nominal load of say 10MΩ, to simulate what it would look like to the front end of a typical handheld DMM?

Well, I am not sure what "typical" would be but I don't think I have a meter that could be simulated with a simple 10M resistor.   Case and point, I modified the UNI-T UT181A's layout to attempt to improve its ability to handle an ESD event.   That PCB is part of the circuit and has inductance, capacitance....   That ESD pulse reaches its peak current in less than 1ns! 

https://www.esdguns.com/content/11-waveform-verification-iec-61000-4-2-waveshape

That said, I used the grill starter to demonstrate why the UT61E was damaged and how it could possibly be improve.  That video shows the waveforms while I injected the transient into the meter.   

The other video was Part2 for the UT181A where I show how I actually measure the ESD transients.   Pretty low end setup using my home made targets but that's a side effect when watching a hobbyist at play vs an actual lab.

***
Consider that the surge waveform I apply has a rise time of just over a 1us, compared with the ESD at < 1ns, we can see these two transients are not at all the same.

[joeqsmith]

My home made target, a section of PCB with embedded resistor, costs maybe a buck.  Add attenuators, connectors, cable call it maybe $100.    Real target, $3500 USD.   

https://www.esdguns.com/esd-calibration-test-targets/118-cst2-4-ghz-2-ohm-esd-target-current-sensing-transducer.html

[AVGresponding]

10MΩ should get you in the ball-park though? Maybe with some MOVs or GDTs in parallel...   

My understanding is these piezo ignitors don't pack much of a punch, energy wise, but can deliver maybe 10kV or so, happy to be corrected if wrong.

[joeqsmith]

10MΩ should get you in the ball-park though? Maybe with some MOVs or GDTs in parallel...   

My understanding is these piezo ignitors don't pack much of a punch, energy wise, but can deliver maybe 10kV or so, happy to be corrected if wrong.

Right.  So a few points.  When I show the UT61E, I show that while the MOVs would be fairly fast and have a fair amount of capacitance, their inductance was too high to make any difference at the frequencies we are considering.   I think I talked about shunting it right at the source (small cap right at the inputs) but that IMO, would be a very bad choice.   I have seen handhelds with intentional spark gaps designed into the artwork, right at the front end.  Bat shit crazy stuff.   In the end, I talked about like any problem, there may be many ways to solve it.  Still, each meter is going to be unique.   


But now I think we get to the crooks of what you are driving at.  You suspect the grill start is putting out 1MegaVolt which unfairly wipes out all of those UNI-T meters.   

I don't think I ever did a video on it but not all grill starters are alike.  Even if I have two what appear on the outside to be identical grill starters, the waveforms they produce can be drastically different.  They can also change a lot as the cheap springs they use take a set.   If we use 1000V per mm, 10mm air gap for 10kV, I have no doubt that I have one grill starter that will exceed that. 

I wonder too, if I could find a much stiffer spring and maybe shim it, how far I could push one.   I wouldn't be surprised to see one jump 20mm or more depending where you live.   

So, yes, I have a standard reference grill starter that I calibrate with my home made target.  It has indeed damaged many UNI-T meters along with a few others.       I imagine I could tone it down further if the goal is to get more meters to survive it but we wouldn't be able to compare results....

[joeqsmith]

Just in case you're thinking about it, at my first job we had one of those tommy guns (keyence).  I knew a guy who wanted to look at the output of it and decided it was smart to zap the 10M scope probe.    We had one of those Tektronix huge mainframe DSOs with the external keyboards and a 100MHz plugin.   High end scope for it's day.   I didn't get to see it happen but in a small company, word travels fast.   Don't be that guy! 

****
Something like this old relic:
https://www.ebay.com/itm/133826840305?epid=96961633&hash=item1f28b386f1:g:cWUAAOSwNpFg-g2c&amdata=enc%3AAQAHAAAAwDpmdjpLrVZGGVtQ85eqNKRffOFpTXoVuHAYnQ5BHZtiSsDTYPxm88Pvs1fI0DXpt09My8XuFi2g%2FluXb%2FwEL%2FUxDsCFMlae8kr5aFC6MsDlWqR0ssDC%2FQlh8JXw2tiennHHqNwQFJMEtMIN0PrwS21oz%2Bsl5GpTpRGigAHqWIaKGDCnm9MAl0WrBgdJuovXW0AayyCVDsW2GoVAdE5aE3FlEKIAKdsdX33v5k2dTPl%2FhHOsK83LbMXb6kz0Fcwk6A%3D%3D%7Ctkp%3ABk9SR9jdhoreYA

[Fungus]

But now I think we get to the crooks of what you are driving at. 

"crux" 

[joeqsmith]

 
****

............ the crooks of what you are driving at.  You suspect the grill start is putting out ...

 

Also, stater.  Guessing many others... Dig deeper.

[Fungus]

Just keeping things on the straightened arrow...

[joeqsmith]

Nice.  I've been known to use such poor spelling/grammar that I couldn't even make sense of it the following day.  Worse, I have no drug problems to blame it on.       

[AVGresponding]

A megavolt? Not hardly. That would allow for a spark across a gap of a metre, never mind a meter! As I said, maybe 10kV.

What I was curious about is whether the the output of these things is too fast for the protection circuitry to react to, as it is obviously going to be way lower energy than the IEC standard tests for CAT ratings, and less than the 20J your transient generator puts out. My point therefore is, given these facts, how does it damage (for example Uni-T) meters? My only guess is it's too fast for the clamp(s).

[joeqsmith]

A megavolt? Not hardly. That would allow for a spark across a gap of a metre, never mind a meter!

You suspect the grill start is putting out 1MegaVolt which unfairly wipes out all of those UNI-T meters.   

I didn't expect you to read that as a literal statement but ok.

As I said, maybe 10kV.

As I wrote: 

If we use 1000V per mm, 10mm air gap for 10kV, I have no doubt that I have one grill starter that will exceed that. 

  Maybe the post was too long for you.

What I was curious about is whether the the output of these things is too fast for the protection circuitry to react to, as it is obviously going to be way lower energy than the IEC standard tests for CAT ratings, and less than the 20J your transient generator puts out. My point therefore is, given these facts, how does it damage (for example Uni-T) meters? My only guess is it's too fast for the clamp(s).[/color][/size][/b]

You seem confused.  You start out asking about ESD then you switch to CAT ratings and 20J.   ESD has nothing to do with CAT ratings.  It is called out in IEC 61326, not 61010.   

The small 20J transient generator I show has a 1.2us risetime as can be seen in the scope shot above.   This is something I took from the IEC standards which I based my generator on.   

Yes, I could tone down every test I run.  Maybe chemical test with tap water, drop them from 1mm, ESD test them with my finger while wearing a wrist strap....   

No doubt I could just unbox the meters, blab for a half hour telling people how great the products are and then give them 5 stars.  Finish up with Don't forget to like and subscribe.  Consider supporting me on patreon so I can continue to bring you this great content. 

That's not something I have any interest in.  I do like seeing how the different brands hold up against one another.   The fact some brands are more sensitive and prone to failures than other under certain conditions is pretty much the whole point of this thread.   

[joeqsmith]

I did write Dave about the U1282A that he swam with, drove over, tossed off the bridge, threw against the wall to see if he would be willing to part with the PCB to possibly salvage this meter.  Maybe we will get lucky.

Until then, it's time for the next phase.

[AVGresponding]

I am not confused, I never mentioned ESD, that was Fungus and latterly yourself. I am asking a question. What does the transient generated by a grill starter look like across the inputs of a DMM?
You partially answered me by pointing out that piezo spark generators are very inconsistent in their output; fine, thank you for that information.

I then asked why the grill starter damages meters; is the transient too fast for the clamp, or is it something else? I mentioned CAT ratings and your transient generator merely to compare the energy levels involved.

I am asking for your considered opinion, or if you know the answer, that is all. I am not asking you to perform any additional tests.

[joeqsmith]

I am not confused, I never mentioned ESD, that was Fungus and latterly yourself. I am asking a question. What does the transient generated by a grill starter look like across the inputs of a DMM?
You partially answered me by pointing out that piezo spark generators are very inconsistent in their output; fine, thank you for that information.

I then asked why the grill starter damages meters; is the transient too fast for the clamp, or is it something else? I mentioned CAT ratings and your transient generator merely to compare the energy levels involved.

I am asking for your considered opinion, or if you know the answer, that is all. I am not asking you to perform any additional tests.

To be clear from the context of the tests I show,  Grill starter = ESD != 20J.   In simple terms, the grill starter damages meters (mostly UNI-T) as they are not very robust.  Why each particular meter was damaged is far beyond the scope of this thread.   However,  I did do a more detailed analysis of the UT61E and talked about the UT181A and reasons why they were damaged.  These are not so much my opinion as I backed up my comments with data. 

For your question

What does the transient generated by a grill starter look like across the inputs of a DMM?

, I assume you want to look at the voltage rather than the current.   While I did show this from the UT61E's controller IC's perspective, it is not what you asked.  Again, I would expect this to be different with each meter.    I don't have a good way to look at it.   It suspect it far exceeds any probe I have.   

If you are interested in trying to measure it yourself with a few different meters and grill starters.  I am interested in seeing your results.     Just be aware of my friend who damaged that DSO attempting to do something similar.   

[joeqsmith]

Life cycle testing has begun.  How many cycles do you think the detent spring will last?  Cast your vote...

Plan is to let it run the full 50k as then sweep up what's left.

[Fungus]

Life cycle testing has begun.  How many cycles do you think the detent spring will last?  Cast your vote...

I say "no more clickys" at 1600.

[joeqsmith]

Remember our friend who said they put 10's of thousands of cycles on the first Keysight I looked at every year.   He even made a video to prove it for us.  You don't believe??   

It only has about 5000 cycles on it so far and lets just say things don't look Sound good.