Handheld meter robustness testing

[joeqsmith]

Keep in mind, using your 500nF, 40kHz, 1k resistor that the power dissipation for the PTC is 36 Watts.

Well, they did unsolder themselves in your case!  But we don't know the power dissipation because we don't know what the phase angle between I and V across the PTC is.  The resistor can't escape the heat so easily. This is one of the things I'd like to test on part samples, rather than cooking a nice meter.

Yes, the resistor is key at low voltage and much higher frequencies. 

I agree that the resistor is key.   Using your 0.5nF, but 2 PTCs in series, and 1MHz @ 25Vrms.    I get  Xc = 318 ohms and a current of 39.3mA.  Each PTC would dissipate 490mW.   Not much.   

At 120MHz @ 25Vrms, where I damaged the meter, Xc = 2.65 ohms.   Current is now 4.7A and the power of each PTC is 59Watts.   

Staying with the 120MHz @ 25Vrms, but swapping one PTC for a 1k resistor.   The current drops to 25mA.   The resistor dissipates 622mW and the PTC 1.65mW.   

While I have a few RF amplifiers as well as some low frequency ones, I don't have anything that will create 600Vrms at 10s-100s of kHz.   Your idea of using a VFD drive seems like it may be the solution if you can find one that would run up that high. 

[joeqsmith]

She has a whole 200 cycles on her.  50,000 cycles  from off to NCV back to off as one full cycle.  None of that half cycle rubbish.  No cleaning the switch every few cycles.  No sticky notes, but instead an actual meter to read the contact resistance.   Listen to her click! 

[joeqsmith]

What a difference a day makes.   It has a long way to go...

[Fungus]

Sounding like a Keysight.

[joeqsmith]

Sounding like a Keysight.

But we were told by the one reviewer, they put 10s of thousands of cycles on the same Keysight meter every year.   Their rebuttal video on the Keysight where they couldn't even bother to show the video detent spring was the last time I watched any of their content.    Looks like they reviewed this meter as well.  Based on the Keysight, I bet the UNI-T is 5 stars and the best meter out there.   

It's a bit over half way now.  Do you think the contacts are wearing as good at the Brymens?  Think it will out perform the Fluke 17B+, after all this is a + model too!


Someone had asked about the life of the test fixture.  It's a direct drive stepper stuck in a PanaVise, with an aluminum fork  mounted to the end.  The only wear would be the stepper's two ball bearings.   The software adjusts the motor's torque while learning the meter.  The meter's present basically no load to the stepper.   I looked at the manuals and did not find any MTBF data.  4 year warranty.     

[Per Hansson]

What a difference a day makes.   It has a long way to go...

Haha, perfect for a quiet lab, it's so silent now! 

[joeqsmith]

Into day 3 of the life cycle.  I wonder if there is anything left of the PCB or contacts. 

The UNI-T uses grease on the contacts where the 17B+ was dry.  Surely the grease will make the UNI-T outperform the Fluke.   

[rsjsouza]

Into day 3 of the life cycle.  I wonder if there is anything left of the PCB or contacts. 

The UNI-T uses grease on the contacts where the 17B+ was dry.  Surely the grease will make the UNI-T outperform the Fluke.

I forgot: you constantly evaluate the staus of the contacts during the ordeal, right?

[joeqsmith]

Yes, it's being recorded.  Like before, I will compare this new data against the others I have looked it.

[floobydust]

Uni-T did a ripoff of the Fluke rotary switch patent US8946571 and US8093516, as far as the detents, teeth and spring...  copied. So it should work just as good    minus the nickel-plated beryllium copper...
It looks like there are still a few via's on the rings contact path though.

[joeqsmith]

Uni-T did a ripoff of the Fluke rotary switch patent US8946571 and US8093516, as far as the detents, teeth and spring...  copied. So it should work just as good    minus the nickel-plated beryllium copper...
It looks like there are still a few via's on the rings contact path though.

I have no doubt that some of the UNI-T fanboys would actually believe the UNI-T UT61E+ is every bit as good as the Fluke 87V in every way.  They are just lower cost is all.  And, if you happen to own an 87V, you're considered a snob.       It does say something about the market UNI-T targets.   

50,000 full cycles on the clock when I woke up.

***
Inspecting the UT61E+ under the frankenscope. 

[joeqsmith]

UT61E+ life cycle test.

[AndrewBCN]

Into day 3 of the life cycle.  I wonder if there is anything left of the PCB or contacts. 
...

So in summary, the contacts and the PCB traces survived the 50,000 cycles in relatively good shape, the (nylon?) spring lobes were slightly worn out and the two small endstop plastic posts broke off, which you "repaired" in less than 15 minutes.

Wouldn't it be correct to point out that 50,000 rotary switch endstop-to-endstop cycles correspond to > 20 years of daily normal use for any of us here?

[Fungus]

Wouldn't it be correct to point out that 50,000 rotary switch endstop-to-endstop cycles correspond to > 20 years of daily normal use for any of us here?

I think it would be far more correct to point out that:
a) The little posts snapped off much sooner.
b) The meter stopped working right after they snapped off (see contact resistance graphs)
c) The dial stopped making a clicking sound after they snapped off which means it wasn't turning the white part of the selector properly.

[rsjsouza]

Well, the meter survived quite well the torture - thanks for doing this.

A few unknowns, especially due to the dramatic self-healing, but that happened with the other meters as well.

After 50k the status of its contacts and PCB are quite interesting. No way to compare with the 17B+, but still...

[joeqsmith]

So in summary, the contacts and the PCB traces survived the 50,000 cycles in relatively good shape, the (nylon?) spring lobes were slightly worn out and the two small endstop plastic posts broke off, which you "repaired" in less than 15 minutes.

With the wiper contacts not fully rotating, they and the PCB were not subjected to the same level of testing other meters endured.   I have no idea how long it would take to repair the meter but if I were trying to use the meter and these pins snapped, it would have gone to the trash as the meter can't be trusted.  Would I just happen to have a soldering iron, some scrap plastics and other tool need to repair the meter with me at the time, I doubt it.  The delay in the test could be much longer.   Then again, who is going to trust using a UNI-T product for something that is time and data critical...   

Wouldn't it be correct to point out that 50,000 rotary switch endstop-to-endstop cycles correspond to > 20 years of daily normal use for any of us here?

We talked about this early on.  Similar to the transient testing, it evolved.  If you watch the very first video where I proposed running this test, I was looking at 2000 cycles and was not monitoring the resistance.     Someone posted an ad for a UNI-T meter that was rated for 30,000 cycles on the switch. 

https://www.eevblog.com/forum/testgear/finally-received-my-uni-t-ut17b-pro-im-pleasantly-surprised/msg3180548/#msg3180548

b) The meter stopped working right after they snapped off (see contact resistance graphs)

Technically yes, but misleading.  No doubt that the pins breaking is a failure and makes the meter unusable.   I have no idea how many cycled were on the meter when the pins broke or how it correlates with the contact resistance.  What we can see is that the contact's resistance  was opening up at 5700 cycles.     No doubt that when the pins broke it limited the movement of the wiper contact.   

[AndrewBCN]

...
With the wiper contacts not fully rotating, they and the PCB were not subjected to the same level of testing other meters endured.
...
No doubt that when the pins broke it limited the movement of the wiper contact.

So, the 50,000 cycles "torture test" is to be considered inconclusive with regards to wear of the PCB and contacts?
And would you say the two small plastic pins/post broke of because your test bed setup was hitting the end stops thousands of times?
Does that correspond to a real use case?

It seems you concluded in the video that the UT61E+ is "more robust" than the discontinued UT61E, but that's not saying much, is it?

[joeqsmith]

Well, the meter survived quite well the torture - thanks for doing this.

For a UNI-T product, overall, I couldn't agree more but not when compared with the better brands I have looked at. 

Then again, having the meter modified may have saved it from the total destruction we are use to seeing with this brand.  It's not good enough for me to want to invest more time into looking at the brand again.   After 13 products, I think we have learned our lesson. 

A few unknowns, especially due to the dramatic self-healing, but that happened with the other meters as well.

Well, let's be clear.  If we are talking about the contacts opening up (>1kohm) some meters have self healed.  Cycle3 shows the five worse meters I have looked at.  All of these went open before 7000 cycles.  It's good to see Keysight in the mix with UNI-T and the free HF meter.      

But consider the other five meters in Cycle4.  None of these meters ever opened up during the entire duration of the test.  Note that even Dave's 121GW with it's cobbled switch and countless contacts hung in there.    While Dave's test limited the measurement to 10 ohms, I like being able to see the higher values and trusting the HP bench meter.   

After 50k the status of its contacts and PCB are quite interesting. No way to compare with the 17B+, but still...

True but again, the wiper contacts were not locked to the knob.   How would the meter have held up had the pins not broke?  What if they hadn't spec'ed the frequency counter to 200MHz and the meter still had the original components?  Would a small ESE event have damaged it? 

[joeqsmith]

...
With the wiper contacts not fully rotating, they and the PCB were not subjected to the same level of testing other meters endured.
...
No doubt that when the pins broke it limited the movement of the wiper contact.

So, the 50,000 cycles "torture test" is to be considered inconclusive with regards to wear of the PCB and contacts?
And would you say the two small plastic pins/post broke of because your test bed setup was hitting the end stops thousands of times?

During the video, I demonstrate how the torque is adjusted before starting the test, slowly reducing it until the motor stalls (slips).   The torque is then increase a small amount and this is used for the remainder of the tests.  The video then shows how the software then searches for the dead stops.  It then limits the range of rotation within a few degrees.   

Does that correspond to a real use case?

How would we know?  There is a big difference between the guy taking the meter out of a drawer twice a year to look at some clock batteries and someone working as a professional in an industry were the meter is in constant use. 

I would not be surprised if most users would turn the knobs to their dead stops with far more force than I subject them to.  But then I would not be surprised if most UNI-T products are not used in a professional environment and see minimal use compared with what I am demonstrating.   I loaned out my CEM meter to someone who forced it beyond the dead stops.  It's one of the weak points of that particular meter.  I guess that was real world.

It seems you concluded in the video that the UT61E+ is "more robust" than the discontinued UT61E, but that's not saying much, is it?

Yes, discontinued but still available for purchase brand new.   And yes, 61E+ more robust after mods than the stock UT61E.   From the factory, I have no idea.   
***
SP

[Neutrion]

Joe, do you know which are the plastics paired in this meter? (The white spring, and the dark one it rubs on.)
And in the Keysight?

If we have the angle of the half rotation(from off to full stop to the right), and the angle of wobble of the wiper pad without the two pins, than we have to double this second value (if we suppose that the wipers were not moving at all while in a "floating position) and deduce the value from the full angle of the rotation.
It can be that the switch had actually 30% or even less rubbing movement than it would normally had without the pins broken.

Edit:
If I remmber right not even the Brymen is using plain PA6 for the case. Is that because of the possibility of picking up water?

[joeqsmith]

I am not a materials engineer and have no idea what they are using.  The Keysight appears to be a glass filled material. 

For the rubbing movement, lol... yes the detent spring will move with the knob once it hits the knobs dead stops (not referring to the dead stop in the case).   The ends of the rotation will have less rubbing movement than the center.  I would have no idea how much less as I am not sure when the pins broke.  With the contacts moving to a different location, it will effect the resistance measurement.

[AndrewBCN]

Joe, do you know which are the plastics paired in this meter? (The white spring, and the dark one it rubs on.)
And in the Keysight?
...

Like Joe, I am not a material/plastics engineeer, but I believe in the UT61E+ the white spring is nylon, whereas the dark plastic it rubs on is polycarbonate.

https://en.wikipedia.org/wiki/Nylon

https://en.wikipedia.org/wiki/Polycarbonate#Applications

[Fungus]

A chain is only as good as the weakest link. It turned out that the wipers and PCB weren't the weakest link here, but so what? The meter broke in a way that would have most people throwing it away.

It's bad design. The physical integrity of the dial depends on two small posts for no good reason - the posts could easily be made much bigger for zero additional cost.

[joeqsmith]

I agree with fungus on this one.   It's better to just buy products that meet your needs.  This meter isn't something I would use but it could be a very good fit for someone else.  I would assume the pin breaking problem was known as someone had left a comment about it.

If this meter could survive the transients from the factor, they stopped with the false ads (freq counter), improve the mechanical design and lowered the price.....   

Do you know how many meters I say something similar about?  Every single one of them!!      

[AndrewBCN]

...
It's bad design. The physical integrity of the dial depends on two small posts for no good reason - the posts could easily be made much bigger for zero additional cost.

Yet again, that's your "armchair expert" opinion from watching Joe's video. It holds as much value as the two plastic posts that broke off... 

You should try devising your own test procedures, and building your own testing apparatus, and accumulating a few years of experience, because that's where the value of Joe's videos lies, as I am sure you realize?