Handheld meter robustness testing

[joeqsmith]

Thanks.  It's getting there but that 4ns edge is still too slow.

[tautech]

Thanks.  It's getting there but that 4ns edge is still too slow.

Oopps, right, yeah I did read that the other day.

An IEC pulse has a rise time of less than 1 ns and dissipates most of its energy in the first 30 ns.

[joeqsmith]

Trying a few new ideas today.  It's come a long way since I started working on it.  Shown zoomed out and in.   Around 1.5ns now. 

I wonder if I inject this across the two leads, would ANY meter survive.   

[joeqsmith]

Here are five overlaid shots.  10A peak with a sub 500ps rise.   Also shown zoomed out.   The current is way too high at the 60ns and I have detuned it a little further.  My plan will be to keep the waveform lower than the IEC standard calls for.   

The next step will be to make some sort of box for it.

[joeqsmith]

Looking at the distribution of 50 shots, about 70ps of error on the rise time.  Note how the scope misses finding the edge.  I have it throw out this data. 
Amplitude moves around about a half amp or about 5% error.

Scope showing about 10A peak, 7A@30ns and 3A@60. 

[joeqsmith]

With the same setup, just the RG400 moved to the higher BW scope.  We can see the peak currents are closer to 11A.   The rise time is about 100ps less as well. 

IMO its pretty good for a home made gun. 

[joeqsmith]

It's almost finished.

[rsjsouza]

Joe, this progress is incredibly interesting. Are you getting these values under load or open circuit?

I have cooked some ideas to create spark gaps years ago but had to postpone my plans due to newborn babies along the way. Give it a few years and "I'll be back".

[joeqsmith]

Joe, this progress is incredibly interesting. Are you getting these values under load or open circuit?

I have cooked some ideas to create spark gaps years ago but had to postpone my plans due to newborn babies along the way. Give it a few years and "I'll be back".

Hello.  The IEC standards call out the calibration method for the guns.   The don't have much on the open circuit voltage but we know they want 4KV.   In that last picture, you can see my home made 2 ohm target and attenuators that I am using to get the current waveforms. 

[joeqsmith]

It's pretty much ready now.  Sorry John, this one's not getting a nice wooden case.   

Besides the pulsed output, the gun also has a DC output.  The display in the upper right is a ballpark of the voltage in KV.  Shown at 1.4KV and hooked to the Brymen using my high voltage attenuator.  Also shown are the target and grill starter.

Now I need a new free HF meter to try it out on....

[joeqsmith]

The completed gun..

[Daruosha]

Keep up the amazing work man! I enjoyed a lot and your videos encouraged me to do some more research on ESD protection standards and tests. Thanks again.

[joeqsmith]

Keep up the amazing work man! I enjoyed a lot and your videos encouraged me to do some more research on ESD protection standards and tests. Thanks again.

Thanks and glad you are enjoying them.

[Daruosha]

It seems that I am having a hard time to understand and differentiate the safety vs robustness.

You managed to kill two fluke 87v's by puting a 1,500V pulse on their resistance measurement mode, however CAT IV 600V class, claims protection against 8,000V input surges an yes, at least the flukes failed safely.

Is it correct to conclude the CAT rating has nothing to do with the robustness and it is only a parameter of how a multimeter will fail in terms of safety?

[tautech]

It seems that I am having a hard time to understand and differentiate the safety vs robustness.

You managed to kill two fluke 87v's by puting a 1,500V pulse on their resistance measurement mode, however CAT IV 600V class, claims protection against 8,000V input surges an yes, at least the flukes failed safely.

Is it correct to conclude the CAT rating has nothing to do with the robustness and it is only a parameter of how a multimeter will fail in terms of safety?

Correct.
Joe's robustness testing demonstrates how a meter can survive (or not) ESD and common or not so common misuse.

[joeqsmith]

And if I may add, I based my testing loosely around the IEC surge test.  This is part of the safety standards and the CAT rating you mention.  I did not do this to attempt to confuse or mislead people into thinking I was looking at safety.   There just are no industry standards that really covered what I was attempting to show (how robust the meters are compared with one another).  Of course, calling the thread, hear kitty kitty not that kind of CAT did not help I'm sure but I assumed people would figure it out. 

The big difference between what I am doing and what the standard calls for is the amount of energy available.  So while I now basically follow the open circuit voltage waveforms, I do not use the short circuit current waveforms.   I do use the 2 ohm source and limit the rise times.     Of course, if a meter did not break down, the current would not mater.   It would absorb very little of the energy available. But some meters do breakdown and rather than the meter possibly violently coming apart, they may spark a little is all.     

Way back when I first started testing, I was not using the standard waveform shape.  Another member had access to a standard generator and offered to test a Fluke 101.  Knowing this, I changed over to the waveform, but doubled the decay time from the IEC standard and I went up another KV.  Basically, I wanted to know that the Fluke 101 was actually going to survive.  The member repeated the tests and we got to see a Fluke 101 survive the real transients.   

More often than not, when a meter is damaged, I can't repair them.  I added a new generator that basically simulates the energy in a half cycle 440 CAT III bus.  I will put the dead meter into the AC volts mode and apply the high voltage wave that caused the meter to break down.  If the meter does break down, the half cycle simulator turns on and feeds all that energy into the meter.  I don't keep any metrics on that test and it's really only there to give some sort of perspective how the meters handle higher energy levels. 

[Daruosha]

Thanks for sharing the idea behind the tests. Well, that's shane there is no ratting for multimeter robustness and no one the industry (i'd like to exclude you) has properly tested and rated the handheld multimeters for robustness. In fact it's quite surprising!

Anyhow your tests are pretty unique and I believe if you document all your tests methodology and their results, perhaps it could lead to a new trend and attract knowen manufacturers. I hope.

It sounds to much to ask, can we have all your test results in a single sheet including torture pulse duration and voltages? I found an Excel sheet, in the first post,  but it looks just a bit outdated and not well documented. Also how the various multimeter failed, they just failed with no explosion or something nasty happened too?

[joeqsmith]

Thanks for sharing the idea behind the tests. Well, that's shane there is no ratting for multimeter robustness and no one the industry (i'd like to exclude you) has properly tested and rated the handheld multimeters for robustness. In fact it's quite surprising!

Anyhow your tests are pretty unique and I believe if you document all your tests methodology and their results, perhaps it could lead to a new trend and attract knowen manufacturers. I hope.

It sounds to much to ask, can we have all your test results in a single sheet including torture pulse duration and voltages? I found an Excel sheet, in the first post,  but it looks just a bit outdated and not well documented. Also how the various multimeter failed, they just failed with no explosion or something nasty happened too?

If you watch my videos, the last meter I tested was the CEM DT-9939.  This is the last meter entered into the spreadsheet, so it appears up to date.   Looking at the header,  the peak amplitude (voltage), source impedance and FWHH (I would guess what you are calling pulse duration)  are shown for each level.  The videos for each test are online so if someone wanted to dig into what was done, you can go so far as to watch it.  There you would find that I subject each mode of the meter to five transients each both positive and negative (again right from the IEC standards).   I already covered what my goal was so I am not sure why you would ever think a meter would explode during my testing.  There is a link to a FAQ on my YT About page.  That may provide some additional information.

I spoke with a few companies about their testing and wanted to know if they would make their data public and in all cases was told no and it is one reason I started running these tests and making them public.  If you look at the reports that are available to the public, what information do you find?   Not too long ago I linked one from TUV and that had a lot more detail than most!

If someone wanted to really run this test on a bigger scale, personally I would just follow the IEC surge test using all of the standard equipment.  I would get that generator on the AC mains and I would build a nice blast room to go with it!     I would step the voltage in small increments and functional test the meter between each step as I do and document the results.   Simple enough to do, but a huge investment.  For a company, what would be the advantage to making this data public?  Some companies like Fluke may be able to make a marketing spin with it but then they would also have to show their 87V as well.  It's one thing for a hobbyist to say the 87V is not very robust, it would be a whole different game for Fluke to come out with data showing where it fits.   I just don't see the companies ever doing this.   Even if they did, I would be concerned the testing would be corrupt.   Look at that TUV Intertek report where they ESD test the meter without the test leads.  The average person may not care but if bought a meter and knew this was how they got it to pass, I would not be happy.    Really it seems it should be a not for profit, independent group running the tests as a watchdog public service.  This would require some major cash flow. 

I would like to see it on a large scale.   Public education would need to be part of it. I really don't see how you would turn it into a business.  The money needs to come from somewhere.    Look at the comments in this thread and consider this is a tech sort of audience.   I suspect the majority of people who would purchase a meter could care less about this sort of thing, so they are not going to fund it.  You need to make enough to feed and house yourself along with your staff.   Seems like a rough road. 

Just blabbing...   If you would like to make a go of it your self I would be interested in hearing about it.

[joeqsmith]

i wanted to try to understand the amount of energy released into a DUT under test, i ended up finding this https://www.onsemi.com/pub/Collateral/TND410-D.PDF

so in the  IEC 61000-4-2 level 4 contact test, 8kV (it says 30A max, i thought there is no limit since it is suppose to be a surge?), i calculate that the amount of energy the DUT must absorbed if 100% discharged in 30ns = 28.8J, but that doesnt look right as there is a discharge curve and the actual amount of energy absorb look more like 50% (15J?) in 30ns and maybe 75% (22J?) in 60ns? is my guess somewhat  in the right direction?

i assume if DUT survived the 30ns (or 60ns) 8kV discharge, the DUT wins (irregardless of how much energy was not absorbed but left over in the discharge capacitor) am i right?

and if based on human body model, the max current it states is 5.33A for 8kV, so much lower (energy going thru DUT = 5.1J ?).

with this, i assume that if i include in circuit design some kind of TVS with 50J of absorption capability, it is sure way to survive 99% of mishaps? yes?

Think mJ.  My big generator puts out about 20 when cranked up.

[kcbrown]

and if based on human body model, the max current it states is 5.33A for 8kV, so much lower (energy going thru DUT = 5.1J ?).

with this, i assume that if i include in circuit design some kind of TVS with 50J of absorption capability, it is sure way to survive 99% of mishaps? yes?

Think mJ.  My big generator puts out about 20 when cranked up.

Drat.  I was hoping it would be MJ.

"The meter didn't survive the test, but at least the scope in the room on the other side of the house did!"

 

[joeqsmith]

My scope's LCD was wigging out while I was working on the new gun.   It's also common for the USB devices to have to be power cycled in my office. 

For high energy transients, I like watching this guy's channel. 

[EEVblog]

I only have two of the latest hand soldered pre-production prototypes, I ain't sacrificing one to the impulse gods 
it hasn't been Intertek tested yet.

[joeqsmith]

That's fine.  When I posted the comment about my offer, I was thinking we would run nondestructive tests on it. Are they getting both the EMC and safety certs?

[joeqsmith]

What's in the box.

[bitseeker]

Joe, thanks for the tour of the box. Pretty handy checker.