Handheld meter robustness testing

[joeqsmith]

I did not try this test in the off position, only in the Amps range that was being tested at that time.   

Looks like I could borrow a supply what would get us to 150mA.  The standard calls for 500VA but it may be enough to damage the meter.    I think the first step is to trace this section out and see what these top Keysight engineers have done to cause such a high load.

You high voltage supply must be very weak, I had expected the meter to start smoking (There is nothing in the meter that can handle more than a few Watt for much time).

I think this is exactly the problem!!  I pulled the meter back apart and they are using a 2.5M to limit the current.  Should be well under a mA.   I believe I had turned down the current limit at one point.   I'll recheck that function before moving forward.     

[joeqsmith]

Looks like that was the problem.  Makes much more sense now.   I readjusted the supply.  Shown at 2kV with 1Meg load.  Their 2.5M resistors should handle the 1.6W.   

[AVGresponding]

Looking forward to part 5!

[Fungus]

Such a difference in how we tested it.

Doesn't it deserve to dropped off the roof like the others?

[tautech]

Such a difference in how we tested it.

Doesn't it deserve to dropped off the roof like the others?

Why ?
He's already dead Jim Fungus according to Joe. 

[joeqsmith]

Such a difference in how we tested it.

Doesn't it deserve to dropped off the roof like the others?

I would like to run the life test on the switch before doing anything else.  The question now is do all the prongs on the detent spring crack like before?   

[Fungus]

I would like to run the life test on the switch before doing anything else.  The question now is do all the prongs on the detent spring crack like before?

Of course... 

[joeqsmith]

I would like to run the life test on the switch before doing anything else.  The question now is do all the prongs on the detent spring crack like before?

Of course... 

I suspect the same but that's why we test them. 

[Fungus]

I suspect the same but that's why we test them. 

On the bright side: It might be a short test... 

[joeqsmith]

On the bright side: It might be a short test... 

Well...  I am thinking even if the spring is damaged as before, we leave the meter together and just let it run out the full 50,000 cycles to get an idea what else goes wrong.   Like the first Keysight meter I looked at, this one was damaged bad enough that I doubt I will attempt to repair it.   If it were Brymen, I am sure I could get parts but based on my previous attempts to open a conversation with Keysight, I doubt I would get anywhere.  If Dave would provide me with the broken up PCB from his, maybe the ICs  could be salvaged to save it.   I had asked before about getting it to pull the front end chip.  No luck.  So I am fully expecting it to hit the recycle bin.   

[joeqsmith]

Not much to say beyond she's dead Joe.

[NoMoreMagicSmoke]

Not much to say beyond she's dead Joe.

Well that's disappointing. I was hoping it would do better than that.

One thing that (too late now...) would have been interesting to know is if it was damaged in that hit when the meter shut off, or if the final hit after it turned off is what did the actual damage. What I wonder is if the fact that the DMM chipset would have (likely) been turned off for that last hit, thus taking the range resistors out of circuit, did the meter get damaged worse that it would have if the chipset was still powered? Again it should have survived, but that last hit was (in my opinion) a little unfair due to the fact that no-one would be trying to measure with a meter when it is so obviously in a non-functional state with the display off.

[Muttley Snickers]

 

I think it may have been mentioned already, but perhaps you could ask Dave what he did with the U1282A he took down the canyon and if he would be willing to post you the board.

It would be nice to see the rotary selector test done with a fully operational meter to monitor it for any weird stuff during the cycle testing. With the types of testing I use these meters for I am probably more likely to wear out the selector switch well before being anywhere near a 5kV transient.

[Trader]

Failing in the 5KV Transient Voltage test, means that this DMM is CAT IV for up 150V, but CAT III for 151-300V ?

https://www.digikey.com/en/blog/what-are-multimeter-cat-safety-ratings

[Fungus]

Failing in the 5KV Transient Voltage test, means that this DMM is CAT IV for up 150V, but CAT III for 151-300V ?

The CAT documents only say that meters must protect the user. They don't say that the meter has to survive.

OTOH a meter that survives seems more desirable to me.

[joeqsmith]

The CAT documents only say that meters must protect the user. They don't say that the meter has to survive.

OTOH a meter that survives seems more desirable to me.

History once again repeats.  Again, the thread is not about meeting a safety standard.  Of course it does keep coming up.  The question I have for you  is if you finally took the time to speak with someone who may actually have the background needed to answer this?  Or it is just the same old posting your feelings about a papers I doubt you have ever seen let alone read?   Welcome to the internet.   

I think it may have been mentioned already, but perhaps you could ask Dave what he did with the U1282A he took down the canyon and if he would be willing to post you the board.

It would be nice to see the rotary selector test done with a fully operational meter to monitor it for any weird stuff during the cycle testing. With the types of testing I use these meters for I am probably more likely to wear out the selector switch well before being anywhere near a 5kV transient.

I asked him about it years ago as a parts meter to repair the prototype 121GW.  No luck.  Too bad really as it could have very well saved this meter from the recycle bins. 

I've damaged a lot of meters that were not able to be salvaged.  Most of the higher cost meters had enough protection that at least saved the parts that your not going to find at the local distributor.  Even the UNI-T UT181A could be repaired.  Keysight doesn't appear to attempt to protect these unique parts like many of their competitors.   
 
For the life cycle test, the meters are unpowered, their batteries pulled.   I collect data on the switch contact resistance and look at how the switch wears.  It would be nice to be able to power it up after the face but it's certainly not a requirement.   

Well that's disappointing. I was hoping it would do better than that.

I'm also very disappointed but not at all surprised by the outcome.   It's an old design and I would assume the same group designed their entire DMM product line.  When you copy paste, this is what you get.   On the other hand, I am pleased that the testing is so reproducible.   

One thing that (too late now...) would have been interesting to know is if it was damaged in that hit when the meter shut off, or if the final hit after it turned off is what did the actual damage. What I wonder is if the fact that the DMM chipset would have (likely) been turned off for that last hit, thus taking the range resistors out of circuit, did the meter get damaged worse that it would have if the chipset was still powered? Again it should have survived, but that last hit was (in my opinion) a little unfair due to the fact that no-one would be trying to measure with a meter when it is so obviously in a non-functional state with the display off.

  I would imagine there are many people who feel several of the tests I conduct are a little unfair!!  After a few years, I'm sure I've heard it all.   I do wonder at times if making these results public has had any impact on the companies who develop these meters or the people who buy their products. 

I am looking for what level the meters are damaged.   Just that simple.  Because the transient generator is semi-automatic, I will typically walk away from it.   If you were to watch where I have damaged other meters, I suspect you will note several cases where the meters are subjected to the full 5 cycles no mater the outcome of each individual transient.   I dare say that in some cases I will even go so far as to finish up the remaining transients for a given level.    If it's damaged, its damaged.   If I roast it to a crisp or pop a SOT23, what is recorded is still the same, the meter failed at level X.   

[NoMoreMagicSmoke]

One thing that (too late now...) would have been interesting to know is if it was damaged in that hit when the meter shut off, or if the final hit after it turned off is what did the actual damage. What I wonder is if the fact that the DMM chipset would have (likely) been turned off for that last hit, thus taking the range resistors out of circuit, did the meter get damaged worse that it would have if the chipset was still powered? Again it should have survived, but that last hit was (in my opinion) a little unfair due to the fact that no-one would be trying to measure with a meter when it is so obviously in a non-functional state with the display off.

  I would imagine there are many people who feel several of the tests I conduct are a little unfair!!  After a few years, I'm sure I've heard it all.   I do wonder at times if making these results public has had any impact on the companies who develop these meters or the people who buy their products. 

I am looking for what level the meters are damaged.   Just that simple.  Because the transient generator is semi-automatic, I will typically walk away from it.   If you were to watch where I have damaged other meters, I suspect you will note several cases where the meters are subjected to the full 5 cycles no mater the outcome of each individual transient.   I dare say that in some cases I will even go so far as to finish up the remaining transients for a given level.    If it's damaged, its damaged.   If I roast it to a crisp or pop a SOT23, what is recorded is still the same, the meter failed at level X.

My curiosity not withstanding, running fully automated is the best way to ensure a consistent test across all meters!

[NoMoreMagicSmoke]

The CAT documents only say that meters must protect the user. They don't say that the meter has to survive.

OTOH a meter that survives seems more desirable to me.

History once again repeats.  Again, the thread is not about meeting a safety standard.  Of course it does keep coming up.  The question I have for you  is if you finally took the time to speak with someone who may actually have the background needed to answer this?  Or it is just the same old posting your feelings about a papers I doubt you have ever seen let alone read?   Welcome to the internet. 

I have no read the specs, so I can't comment on the need for the meters to survive, but I do remember Joe showing snippets of the standard and one thing that stands out in my memory is the statement that "the meter shall remain capable of indicating hazardous voltage". This meter certainly was not capable of indicating hazardous voltage after it failed.

Joe one thing that seems to hold true is the meters usually fail on the inputs other than the primary voltage input, usually ohms or (as in this case) mv inputs. How is the specification worded regards those inputs? I know it must take the full input voltage, but does it explicitly state that these secondary (ohms mv, etc) inputs must also be able to survive the same transient testing?

Again I know your testing isn't to test the CAT rating. I am just always amazed when NRTL tested meters fail your robustness tests at voltages lower than the CAT transient tests.

[Fungus]

I have no read the specs

Nor many people have. They print them on unobtainium.

I do remember Joe showing snippets of the standard and one thing that stands out in my memory is the statement that "the meter shall remain capable of indicating hazardous voltage".

Oh, yeah, that part...

This meter certainly was not capable of indicating hazardous voltage after it failed.

How do you know?

It's certainly possible that there's an LCD segment that turns on via an independent circuit, not via the main CPU. This isn't something that I recall joe ever checking after a meter has failed.

It seems very unlikely, but it really ought to be explicitly tested. Hook it up to +50V with the batteries in and see if anything happens anywhere on screen... 

[joeqsmith]

...  Because the transient generator is semi-automatic, I will typically walk away from it.  ...

My curiosity not withstanding, running fully automated is the best way to ensure a consistent test across all meters!

Automating the knob rotation for each meter, IMO would be a lot of work with no value added.  Designing a generator that supports both polarities vs swapping the leads, again, I see as little value.   Of course, if we are going to switch out the generator and then automatically test the meters between each level, again a bit of work for what I see as no value added.   
Just an FYI, we do not do this in the real world.


I am certainly not apposed to seeing someone else step up and run similar tests.  I wouldn't mind seeing someone procure an actual combo generator that also supports modes like burst.   Add an ESD gun, calibrator, maybe setup some sort of EMI chamber....   Maybe the holy grail and setup an arc flash lab?   It's all far outside the scope of what I am looking for but it
would be fun to watch.

...
Joe one thing that seems to hold true is the meters usually fail on the inputs other than the primary voltage input, usually ohms or (as in this case) mv inputs. How is the specification worded regards those inputs? I know it must take the full input voltage, but does it explicitly state that these secondary (ohms mv, etc) inputs must also be able to survive the same transient testing?

I'm pretty sure I went over this early on but you would need to read them to know for sure.  I can tell you in Joe's world, they are going to be tested this way before I will ever consider them robust.   

Again I know your testing isn't to test the CAT rating. I am just always amazed when NRTL tested meters fail your robustness tests at voltages lower than the CAT transient tests.

I am amazed when meters are damaged by that stupid little grill starter or if they fail at levels lower than what damaged that $50 Amprobe AM510.   The worse, IMO, are meters like this Keysight that carry a high price tag an are rendered non-repairable. 

[Chance92]

I wouldn't mind seeing someone procure an actual combo generator that also supports modes like burst.

May I ask what a combo generator is? I've googled it and diesel generators showed up. How are diesel generators going to fit into these tests?

[joeqsmith]

I wouldn't mind seeing someone procure an actual combo generator that also supports modes like burst.

May I ask what a combo generator is? I've googled it and diesel generators showed up. How are diesel generators going to fit into these tests?

I was referring to burst and surge transients from the safety standards that people are talking about.  Have a look:
https://www.ametek-cts.com/products/productgroups/transient-generators-surge-and-burst/surge-generator

That said, you may need a genset if you plan to setup an arc flash test lab.
https://www.ecmag.com/section/safety/lets-blow-it-arc-flash-testing

****
It doesn't hurt to repeat it once again.

I want to be very clear that I really have no interest in the safety standards or if a meter is safe or not.  I do not work in a environment where I deal with high energy circuits.

What I am interested in is if a meter will survive low energy transients.  I am more interested in IEC 61326 than 61010.   This dates to my first DMM (Fluke) that I damaged a few times from low energy transients and it cost a fair amount to repair.   You couldn't give me a Fluke meter up till the time Fungus convinced me to buy that 101.  I had full intentions of watching that meter burn to the ground but it proved to be a very worthy opponent. 

I run the tests out of my own interests and look for things that have damaged my own meters.   This is why you see them being exposed to methanol and gasoline.   It's why I cycle the switches.   And it's why I run these low energy transients.

I'm not suggesting any of this information is useful to viewers.  Maybe it provides some level of entertainment?  Maybe some education?  Increased awareness?  Who knows.   It's not very popular and I am certainly not in it for the money!   

***
Site doesn't seem to like dollar signs...

[Chance92]

I wouldn't mind seeing someone procure an actual combo generator that also supports modes like burst.

May I ask what a combo generator is? I've googled it and diesel generators showed up. How are diesel generators going to fit into these tests?

I was referring to burst and surge transients from the safety standards that people are talking about.  Have a look:
https://www.ametek-cts.com/products/productgroups/transient-generators-surge-and-burst/surge-generator

Is it similar to a pulse generator like this one:

[joeqsmith]

Is it similar to a pulse generator like this one:

You would need to ask them what standards they adhere to.   The ones I am referring to support both the 1.2us/50us voltage as well as the 8us/20us current waveforms (hence the name, combo generator or combination). 

Doing a search on their site, I do not believe they have support for the IEC standards we are talking about.

[NoMoreMagicSmoke]

...
Joe one thing that seems to hold true is the meters usually fail on the inputs other than the primary voltage input, usually ohms or (as in this case) mv inputs. How is the specification worded regards those inputs? I know it must take the full input voltage, but does it explicitly state that these secondary (ohms mv, etc) inputs must also be able to survive the same transient testing?

I'm pretty sure I went over this early on but you would need to read them to know for sure.  I can tell you in Joe's world, they are going to be tested this way before I will ever consider them robust.   

Again I know your testing isn't to test the CAT rating. I am just always amazed when NRTL tested meters fail your robustness tests at voltages lower than the CAT transient tests.

I am amazed when meters are damaged by that stupid little grill starter or if they fail at levels lower than what damaged that $50 Amprobe AM510.   The worse, IMO, are meters like this Keysight that carry a high price tag an are rendered non-repairable.

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 agree completely with your assertion that a robust meter should survive the tests you throw at them. I am mostly curious if the standard omits the transient tests for inputs like ohms, etc 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.