Handheld meter robustness testing

[joeqsmith]

What about some lighter solvents like IPA? I don't often use racing fuel on my bench but the IPA comes out quite often.

Then there's the acetone ... I try not to use that one the bench though.

I expect acetone would make a mess of just about any meter but these are supposed to be tough and "industrial". Maybe try it on some of the less important lettering. 

I doubt I will add any others to the test but you are certainly free to run your own tests with what ever cocktail you can come up with.   Please post the results if you do.   

I had damaged the plastics on my meters that I use in the garage.   That Fluke 97 scope meter that was used in a garage has a lot of damage from I expect gasoline or brake cleaner.   This is really why I started looking at them.    A few people have asked about brake cleaner as well which can be pretty hard on plastics as well.   I think it was the methanol that stripped the lettering off the 121GW where that VP fuel had no effect.   

[Fungus]

The other Keysight slightly underperformed in the zapping tests.

I'll go with 6-7kV, slightly under the 8kV of the CAT IV 600V rating.

There's no way it's going to go up to 10kV+ like the Flukes and Brymens. 

Edit: Maybe I could do some research and look at some teardown photos before opining, but that wouldn't be me. 

[rsjsouza]

Non-destructive testing of the Keysight U1282A.   

Great testing, Joe! I only wish you had put in the chamber the meter with the same Hycon chipset to see if we could further isolate the root cause of the issue down to the processor/chipset itself (my best bet, since the meter became "frozen") or another component (XTAL, voltage regulator, etc.). From where we see, there is a chance that Brymen's chipset might be better fabricated than Hycon, but we can only speculate.

Another aspect: when you rotated the switch after the low temp tests, was the meter already back at a specified temperature of at least -20ºC? The reason is that the material might have become brittle at such low temperature and rotating it might degrade it (microfissures?) which would add another variable to the rotary switch tests.

[rsjsouza]

AFAIK the vertical reading tends to be preferred by this audience as well (in contrast to sideways).

I have an Ideal branded clamp meter (Pretty sure it's actually a Uni-T  ). And they have this really neat dual display. I was expecting it to be gimmicky, but I have used the display on the end of the meter often.

That is pretty neat indeed and I can see its practical aspects.

Although I wonder if it weakens the housing.

I tried to find the OEM but to no avail. I tried CEM, Mastech, Uni-T and Appa.

It looks like they did the housing pretty well. There is additional plastic ribbing to reinforce the display and input jacks.

I can't say for sure who the OEM is, but the interior build looks suspiciously similar to my Uni-T 210e. It appears to be (the scratched off the label) the same DM1106 controller chip. If they weren't made by the same OEM, they appear to have been based off the same reference design.

Thanks for sharing the pictures; indeed the bottom display seems sturdy.

Looking at the PCB silk screen, it seems to me an original design from Ideal: the model number 61-757 is shown in both PCBs. 

[NoMoreMagicSmoke]

Great testing, Joe! I only wish you had put in the chamber the meter with the same Hycon chipset to see if we could further isolate the root cause of the issue down to the processor/chipset itself (my best bet, since the meter became "frozen") or another component (XTAL, voltage regulator, etc.). From where we see, there is a chance that Brymen's chipset might be better fabricated than Hycon, but we can only speculate.

May I suggest comparing the Hycon chip datasheet/pinout with the "custom" chipset in the BM789. 

[NoMoreMagicSmoke]

AFAIK the vertical reading tends to be preferred by this audience as well (in contrast to sideways).

I have an Ideal branded clamp meter (Pretty sure it's actually a Uni-T  ). And they have this really neat dual display. I was expecting it to be gimmicky, but I have used the display on the end of the meter often.

That is pretty neat indeed and I can see its practical aspects.

Although I wonder if it weakens the housing.

I tried to find the OEM but to no avail. I tried CEM, Mastech, Uni-T and Appa.

It looks like they did the housing pretty well. There is additional plastic ribbing to reinforce the display and input jacks.

I can't say for sure who the OEM is, but the interior build looks suspiciously similar to my Uni-T 210e. It appears to be (the scratched off the label) the same DM1106 controller chip. If they weren't made by the same OEM, they appear to have been based off the same reference design.

Thanks for sharing the pictures; indeed the bottom display seems sturdy.

Looking at the PCB silk screen, it seems to me an original design from Ideal: the model number 61-757 is shown in both PCBs. 

It's definitely an original design. Ideal has a patent on that display layout so only their meters have it. My meters is one of the first ones sold and I had a minor issue when I first got it. I ended up on the phone with one of their engineers who confirmed that they contract with "A major Asian meter manufacturer". The wouldn't say which one, so any exact OEM guess is speculation.

[joeqsmith]

Non-destructive testing of the Keysight U1282A.   

Great testing, Joe! I only wish you had put in the chamber the meter with the same Hycon chipset to see if we could further isolate the root cause of the issue down to the processor/chipset itself (my best bet, since the meter became "frozen") or another component (XTAL, voltage regulator, etc.). From where we see, there is a chance that Brymen's chipset might be better fabricated than Hycon, but we can only speculate.

Thanks. 

While I do provide several opportunities for viewer's to offer their suggestions BEFORE conducting these tests,  "hindsight is 20-20", "Coulda Shoulda  Woulda",  "it is what it is" all come to mind.  Now if you are planning to repeat the test,  one viewer had asked about injecting a signal at a fixed frequency to see if the meter's clock  is showing down.  This seems like it would have some merit and something for you to consider.   I am sure a few of us would enjoy watching it but for me, I've seen what I wanted to see.     

Another aspect: when you rotated the switch after the low temp tests, was the meter already back at a specified temperature of at least -20ºC?

From the video, the meter was stored at -40C for well over an hour and then thermal shocked when I rotated the switch to ohms.   From the video we can see the internal temp was reporting  -24 when I tried it again,  then 0. 

The reason is that the material might have become brittle at such low temperature and rotating it might degrade it (microfissures?) which would add another variable to the rotary switch tests.

   
If Keysight's choice of materials have a problem with these few rotations, that's really poor.  From my early days we designed to -40 operating temps and we used plastics in many of these designs.   Many of these parts were not stationary.   Now, don't get me wrong.  Seeing that glass filled spring fail prematurely as it did doesn't give me any confidence in their mechanical designers abilities.   Still, I also have no plans to start treating this meter with kid gloves.   If the plastics fail because of the temperatures, vapors from the gasoline seeping into the plastics, from the UV off the sun,  or the shock of being dropped so be it.     
         
Obviously one shouldn't expect my home hobby testing is going to be conducted like a professional PV test where we may run 50 samples under a very controlled environment.  Of course, the costs in setting up such a lab and the time invested in conducting such tests would far exceed what I would be willing to donate.   I suspect this is why we don't see many reviews doing much more than collecting free meters to unbox and give their 5 star professional rating.   Doing even the small bit I show takes a lot of time, effort and resources.       

[joeqsmith]

Great testing, Joe! I only wish you had put in the chamber the meter with the same Hycon chipset to see if we could further isolate the root cause of the issue down to the processor/chipset itself (my best bet, since the meter became "frozen") or another component (XTAL, voltage regulator, etc.). From where we see, there is a chance that Brymen's chipset might be better fabricated than Hycon, but we can only speculate.

May I suggest comparing the Hycon chip datasheet/pinout with the "custom" chipset in the BM789. 

I'll leave the reverse engineering to others.   It's certainly something you could do.  Anytime I have dove into a design, it was to get some idea why a meter failed my tests or how to improve it. 

[NoMoreMagicSmoke]

Great testing, Joe! I only wish you had put in the chamber the meter with the same Hycon chipset to see if we could further isolate the root cause of the issue down to the processor/chipset itself (my best bet, since the meter became "frozen") or another component (XTAL, voltage regulator, etc.). From where we see, there is a chance that Brymen's chipset might be better fabricated than Hycon, but we can only speculate.

May I suggest comparing the Hycon chip datasheet/pinout with the "custom" chipset in the BM789. 

I'll leave the reverse engineering to others.   It's certainly something you could do.  Anytime I have dove into a design, it was to get some idea why a meter failed my tests or how to improve it.

I wasn't expecting you to reverse engineer the meter. My comment was to rsjsouza in reference to his question that the Brymen chipset might be fabricated better than the Hycon.

[joeqsmith]

Ran a few tests that viewer had asked about including looking at some smaller capacitors and measuring the burden voltage.    I also started running the potentially destructive tests. 

[Trader]

Pretty cool tests, seems to be a very robust DMM.

Thanks for explaining about the testing box.

[joeqsmith]

Pretty cool tests, seems to be a very robust DMM.

Thanks for explaining about the testing box.

If you were confused about how "...useless is to use a "crappy box" with a lot of uncalibrated "references"..",  chances are good that there are others who are also not understanding what is being shown.  While the threads now several years old and the basics are documented and have been explained many times,  I certainly understand not everyone is going to take the time to do any research before chiming in with their expert opinion.   

From my viewpoint, it's too early to say if the U1282A is electrically robust or not.  It's well known that I used the runner up from the first $50 shootout to set the bar for what I consider robust.  This is why the transient generator is limited to 5.8kV.   If your product can't survive to the level that the $50 Amprobe AM510 can, IMO, that's not very impressive.    Then again, if all you had were UNI-T products to compare with it, the Keysight is surely a notch above them as the grill starter normally does them in.   Of course, that's not my position and I have a lot of data to back it up.  I would have liked to ran the Keysight open fused at 2kV like the standard calls for and just see what happens.  It's not been a problem but this is the first time I have seen a meter load my supply like this. 

From a chemical robustness, hard to say.  It's rare we have seen these two chemicals cause any problems.  I was very surprised that a company like Keysight would not have worked with a company that knew what they were doing.  Obviously they are out there. 

I'm not sure about the switch until we run it.   The one to beat is still that Fluke 17B+.   Both Brymen and Fluke have shown us they know how to design a switch that will last.  Even that last UNI-T meter I looked at held up much better than I was expecting.    The first Keysight made it a few thousand cycles was all.  I can tell you that the spring feels very stiff.  We will know soon enough. 

[Trader]

I don't think the testing box is "useless", yes you are using it just to know if the meter still working; but I'm sure that could be much better if you at least provide a table with the "correct" value of each reference, maybe the meter could be a little off after the tests.

I know, nothing there is accurate and can vary a little when the temperature or power supply changes, but at least is better than not knowing the correct value.

[joeqsmith]

I don't think the testing box is "useless", yes you are using it just to know if the meter still working; but I'm sure that could be much better if you at least provide a table with the "correct" value of each reference, maybe the meter could be a little off after the tests.

I know, nothing there is accurate and can vary a little when the temperature or power supply changes, but at least is better than not knowing the correct value.

"Useless" and "crappy" were your choice of words, not mine.   If you like calling viewer's pussies and such just be aware we are reading... 

I have not yet seen a meter that when damaged exhibited a subtle shift in readings.   Still, these are relative measurements, not absolute.   It sounds like you may not understand that there is a difference.   

Correct values?   Let's see, there's a pot that I set manually to get some idea if the temperature circuit is still functional.   There's even a mark on the box for roughly 500C.  For the resistors, I provide their tolerances.  At one time, I published the values of the caps using my RLC meter but again, it's not in cal.    It would add no value to do anything beyond this.   Again, I assume the meters are within their manufactures specs when I buy them.   Had my goal been to verify their calibration,  I would rent a calibrator.   That said, this does seem to be Defpom's area of interest.    Mine is how robust these meters are.  While I would expect reviewer's like Defpom would rarely damage a meter during a review and should have a very large collection,  there's a reason I never show a Keysight handheld meter as a comparison.  Mine went to the recycle bin.  It just wasn't robust enough to make the cut.     

[NoMoreMagicSmoke]

Again, I assume the meters are within their manufactures specs when I buy them.   Had my goal been to verify their calibration,  I would rent a calibrator.   That said, this does seem to be Defpom's area of interest.    Mine is how robust these meters are.   

And that's a good thing. Having different reviewers that specialize is great for the folks watching. It allows a reviewer to concentrate on and get good at their area of interest and we can then watch a few different reviews to gain indepth knowledge from all.

It goes back to the saying "Jack of all trades master of none". When watching a review my preference is to watch different folks that are good at what they care about so that I know I am getting good Indepth information from each.

Honestly the one thing I wish was covered in the accuracy review videos is long term drift. I would be great to see the drift for each meter at 1, 2, and 5 years on... Of course this is wholly impractical, so I realize it will never happen! 

[NoMoreMagicSmoke]

I would have liked to ran the Keysight open fused at 2kV like the standard calls for and just see what happens.  It's not been a problem but this is the first time I have seen a meter load my supply like this. 

Wait... It looks like your Applied KilloVolts power supply is capable of 40ma. That would imply that the 1282A is dissipating about 80W somewhere. 

For comparison the BM78x uses a of 5MOhm resistor per amp jack for the lead sense which if my math is correct it would be expected that the current draw with 2KV applied would be around 0.4ma and Brymen used 2W resistors to handle the power (expected 0.8W) dissipated.  That's a significant difference between the different meters.

Did you run this test on the BM78x meters? I don't remember seeing it in any of the videos.

[joeqsmith]

“A jack of all trades is a master of none, but often times better than a master of one.”

Attached showing the last time I checked my UT61E, representing almost 6 years of data.  Of course, this meter is highly modified for educational purposes only.   

Wait... It looks like your Applied KilloVolts power supply is capable of 40ma. That would imply that the 1282A is dissipating about 80W somewhere. 

You're catching on!!  If I had a larger supply, I may have ended that meter right then and there.   It's certified and would have been subjected to this.... or is there more to the story..

For comparison the BM78x uses a of 5MOhm resistor per amp jack for the lead sense which if my math is correct it would be expected that the current draw with 2KV applied would be around 0.4ma and Brymen used 2W resistors to handle the power (expected 0.8W) dissipated.  That's a significant difference between the different meters.

Did you run this test on the BM78x meters? I don't remember seeing it in any of the videos.

You know I did!   That poor meter was under the microscope, just like every other meter I look at. 

Maybe Defpom has a larger supply and would be willing to sacrifice their free meter?   I paid for mine and would be more than willing to run it.  Sort of the whole point of the channel.   

[joeqsmith]

To save you time searching.

Chemical: 21:44
Open Fuse: 22:55
Finally, the meter is damaged at 12kV (more than double what my first Keysight was damaged at):  57:47
After 50,000 full rotations on the switch: 1:15:00

[Fungus]

Wait... It looks like your Applied KilloVolts power supply is capable of 40ma. That would imply that the 1282A is dissipating about 80W somewhere. 

Yep. This needs investigating... where is all that power going?

Does it do it with the switch in the "off" position?

Edit: I guess it survived though, and an 8kV transient is only going to push something like 200mA through it for an instant, it will probably survive that, too. 

OTOH it doesn't inspire confidence. It doesn't say "we made it as good as possible!", it asks "where else did they cut corners?"

[NoMoreMagicSmoke]

To save you time searching.

Chemical: 21:44
Open Fuse: 22:55
Finally, the meter is damaged at 12kV (more than double what my first Keysight was damaged at):  57:47
After 50,000 full rotations on the switch: 1:15:00

Thanks for the link. I probably don't remember that test because it.was too uneventful. 

I honestly think that poor meter got the most scrutiny of any of your videos I watched. It certainly says something when a meter takes that much poking and proding and comes out that good. It certainly seems the Flukes and Brymens are in a special league with little company!

[NoMoreMagicSmoke]

Attached showing the last time I checked my UT61E, representing almost 6 years of data.  Of course, this meter is highly modified for educational purposes only.   

Thanks for that. It's interesting data.

It's.too bad the Uni-T aren't very robust. They seem like great measuring meters and certainly fall under the bang for buck category. If they made them just a bit more robustly they could be great meters! I really wonder how muchlre it would cost them to make them just a bit better. It seems they are willing to spend money on parts in their better meters.

[joeqsmith]

It's.too bad the Uni-T aren't very robust. They seem like great measuring meters and certainly fall under the bang for buck category. If they made them just a bit more robustly they could be great meters! I really wonder how muchlre it would cost them to make them just a bit better. It seems they are willing to spend money on parts in their better meters.

I say that about every single meter I look at. It's very possible that they just lack the engineering skills. 

Thanks for the link. I probably don't remember that test because it.was too uneventful. 

I honestly think that poor meter got the most scrutiny of any of your videos I watched. It certainly says something when a meter takes that much poking and proding and comes out that good. It certainly seems the Flukes and Brymens are in a special league with little company!

It's possible that because both the BM786&9 were still in the development stages (not yet available for purchase) that I spent more time going over them both.   It's rare we get to look at preproduction products and have an opportunity to provide a company feedback that could be used to possibly release a better product.   Dave had shown us the meter long before I became involved. I guessing he was testing them and uncovering a few problems as well.   

Dave provided me with the same opportunity with the 121GW but the end results were much different.  You may remember I had pulled those videos after Dave's comments about them confusing viewers.   After waiting two years after the 121GW's release, I bought two new ones and was surprised of how many of my original findings still existed.  This is the difference Brymen brings.  Brymen has proven they are interested in producing a good product. 

[Fungus]

It's.too bad the Uni-T aren't very robust. They seem like great measuring meters

There's meters for half the price that are "good measuring meters".

This is the difference Brymen brings.  Brymen has proven they are interested in producing a good product. 

QFT.

[joeqsmith]

Wait... It looks like your Applied KilloVolts power supply is capable of 40ma. That would imply that the 1282A is dissipating about 80W somewhere. 

Yep. This needs investigating... where is all that power going?

Does it do it with the switch in the "off" position?

Edit: I guess it survived though, and an 8kV transient is only going to push something like 200mA through it for an instant, it will probably survive that, too. 

OTOH it doesn't inspire confidence. It doesn't say "we made it as good as possible!", it asks "where else did they cut corners?"

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.   

[HKJ]

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).