Handheld meter robustness testing

[dcac]

Dr. Frank, Dcac; joeqsmith, Sleppy et al:
   
I bought a 121GW that I received about a week ago.   Today I noticed unusually poor capacitor readings (FW 2.02); now finding this capacitance discussion, I'm having buyer's remorse. 

A 100 pf cap in my meter for several hours drifted only 1-2 pf, so drift isn't the issue.   But at 67 pf, it's just way off. 

Reading between the lines, perhaps, I get the impression that it will not be possible to get reasonable sub-100 pf readings.  Since most of my work is 'portable', I do not want to drag a $10 meter along as a fix.

I'm not thrilled by the prospect of re-engineering the meter I bought to do work, nor am I qualified to do so.

What's your best suggestion?

Fabrice
_______________

I can't recalibrate ... for some reason using a 10nF like in instructions result in 0.000nF readings in this range for any cap , so the only option is to reload the saved calibration . Maybe it's a bug , I tried many times even with different values , I don't think I did something wrong 

Also pay attention if you try to re-calibrate the capacitance ranges.
There is no entry for zero value calibration in the small table inside the chapter ZERO OFFSET CALIBRATION.
In the big colored table, though, there is a scrambled entry, 'R1 : ', which might indicate that the first calibration point with open leads sets the offset, and the 2nd one with the nominal reference capacitor sets the gain.
That might explain CDaniels problem, that 0.000 is displayed for a 10nF capacitor, after calibration.

Frank

I'm not sure if the capacitance calibration procedure ever been properly documented, I can't find it anyway, but here is what worked for me:

First make sure...

Keep in mind the test leads play a major part if you measuring capacitors and you want pF accuracy or resolution. If you watch Joe’s videos when he check meters against his reference box he’s using relatively short leads and is not touching them when reading the values in the pF range.

And here he’s using a ‘double’ BNC adapter to connect the capacitor, just for the reason of taking the test leads out of the equation.



This is really the only way to do it if you want to determine if you meter has a problem with small cap values. Also keep in mind even the BNC adapter adds 8-10pF to your reading.

[joeqsmith]

Dr. Frank, Dcac; joeqsmith, Sleppy et al:
   
I bought a 121GW that I received about a week ago.   Today I noticed unusually poor capacitor readings (FW 2.02); now finding this capacitance discussion, I'm having buyer's remorse. 

A 100 pf cap in my meter for several hours drifted only 1-2 pf, so drift isn't the issue.   But at 67 pf, it's just way off. 

Reading between the lines, perhaps, I get the impression that it will not be possible to get reasonable sub-100 pf readings.  Since most of my work is 'portable', I do not want to drag a $10 meter along as a fix.

I'm not thrilled by the prospect of re-engineering the meter I bought to do work, nor am I qualified to do so.

What's your best suggestion?

Fabrice

Keep in mind the test leads play a major part if you measuring capacitors and you want pF accuracy or resolution. If you watch Joe’s videos when he check meters against his reference box he’s using relatively short leads and is not touching them when reading the values in the pF range.

And here he’s using a ‘double’ BNC adapter to connect the capacitor, just for the reason of taking the test leads out of the equation.

This is really the only way to do it if you want to determine if you meter has a problem with small cap values. Also keep in mind even the BNC adapter adds 8-10pF to your reading.

Also, just note that I had shown adding that 1nF as a way get the meter above 0 so I could null it out.  This part would remain in place, the meter was nulled, then you would add the part you want to measure. 

It's too bad Dave hasn't released the older alignment procedure for the capacitance as it can be improved as I demonstrated.   

[joeqsmith]

I have continued to work on the damaged 121GW (Production 2) and now have it working for the most part.

This meter was cycled from 5 deg C to 45 and back down to 5.  Sweep1 is showing the entire data set.  I was using a 1mV source.   White is the chamber's temperature, Red is the 121GW's ambient temperature and Green is the reported voltage. 

Sweep2 is looking at the voltage.  You can see as the temperature is changing how the meter responds. 

[joeqsmith]

That temperature drift from -10 to 60C is very concerning, and what the hell is causing it.
...
I just don't get it.

I allow it to sit at the two extremes for a half hour to stabilize.  There's not a lot of room in that box and there is a decent sized fan.  At -10 we had 1.003mV, 20ish 0.997mV and 60C 0.928mV.   

It’s difficult to tell exactly as this part of your video is so time compressed, and is perhaps the overlayed chamber camera not really in sync, but:
...
121gw showing 0.871mV at 20.9c, control program (chamber temp) is shoving about 43c:
....
so I get 0.899mV when chamber temp is shoving about 20c and then 0.997mV at 20c “room temperature”, why the big difference?

Sweep3 is showing the temperature rising from 5C to 45C.  (Zoomed into the 20C area)  With all of the data synchronized, we can see that Production 2 is pretty much dead on with Production 1 during the rising ramp. 

Sweep4 is showing the temperature falling from 45C back to 5C.   (Zoomed into the 20C area)   If the meter were allowed enough time to settle at this temperature, the voltage would read the same. 

[joeqsmith]

Production 2's reported ambient temperature seems to track the chamber very close.   After cleaning up Production 1, I took the meter back down to -10C.  Oddly enough, it now behaves more like Production 2.   

Screenshot of camera, 121GW interface software and chamber control software.

[fabrizio_fabrice]

Dr. Frank, Dcac; joeqsmith, Sleppy et al:
   
I bought a 121GW that I received about a week ago.   Today I noticed unusually poor capacitor readings (FW 2.02); now finding this capacitance discussion, I'm having buyer's remorse. 

A 100 pf cap in my meter for several hours drifted only 1-2 pf, so drift isn't the issue.   But at 67 pf, it's just way off. 

Reading between the lines, perhaps, I get the impression that it will not be possible to get reasonable sub-100 pf readings.  Since most of my work is 'portable', I do not want to drag a $10 meter along as a fix.

Keep in mind the test leads play a major part if you measuring capacitors and you want pF accuracy or resolution. If you watch Joe’s videos when he check meters against his reference box he’s using relatively short leads and is not touching them when reading the values in the pF range.

And here he’s using a ‘double’ BNC adapter to connect the capacitor, just for the reason of taking the test leads out of the equation.

This is really the only way to do it if you want to determine if you meter has a problem with small cap values. Also keep in mind even the BNC adapter adds 8-10pF to your reading.

Also, just note that I had shown adding that 1nF as a way get the meter above 0 so I could null it out.  This part would remain in place, the meter was nulled, then you would add the part you want to measure. 

It's too bad Dave hasn't released the older alignment procedure for the capacitance as it can be improved as I demonstrated.   

Yes, I understand the lead/stray capacitance issue, so I used no leads.  I formed j-loops on the capacitor leads and plugged them directly into the meter.  About 1" long each.  That demonstrates a meter issue; as you say, probably calibration.  I'd better tolerate a few pf of drift or ten seconds settling time than a constant 30-40% error on a 100 pf reading.

If I understand the suggestion, other than for calibration, perhaps, there's no need for a 1.0000 nf reference capacitor: any stable cap that gets the meter to read somewhat above 0 even with test leads, would allow stable nulling if it remains connected in parallel with the test subject.  If that's correct, then a dual banana plug and jack fixture holding the cap would work and be much easier to carry than another meter.  Alternatively, a trimmer cap on the fixture to bring the reading to zero?  Even so, I'm fixing equipment rather than fixing broken gear.

It's not obvious to me that is there consensus on the cal process.  Am I wrong??  I do have a large assortment of "fully-aged" [old] caps, and could select some of various values, based on either/both an old Elenco LCR meter or a Fluke 83III, to effect re-calibration, even if not to the accuracy specs.

But this has given me an entirely new take on meter specs.  What I presumed was simple documentation error (which I'd try to correct, along with fixing bad page links), now presents the impression of perhaps being a clever way to reduce the accuracy required for a chicken dinner. 

I'm retired, and reflect on the high-end gear I saw discarded through corporate bankruptcy, take-overs, and downsizing to zero.  So much equipment and so fully depreciated that they were not even trying to monetize it.  I condemn my own lack of foresight.  My only defense is that I didn't work the electronics, I worked the optics.      (Is that irony?)

But I really appreciate your collective thoughts and experience, since your dedication to the subject shows it more central to your interests than mine.  So, now what?

-Fabrice

[CDaniel]

Joe is refering to something else , just a trick to make it work , the proper recalibration is done  as "dcac" described .
If you have LCR bridges it is easy to sort an exact 1nF and 10nF .
Just save the old calibration and read first ten times the calibration procedures from the manual 

[joeqsmith]

....

If I understand the suggestion, other than for calibration, perhaps, there's no need for a 1.0000 nf reference capacitor: any stable cap that gets the meter to read somewhat above 0 even with test leads, would allow stable nulling if it remains connected in parallel with the test subject. 

That is correct.  Try it.

If that's correct, then a dual banana plug and jack fixture holding the cap would work and be much easier to carry than another meter.  Alternatively, a trimmer cap on the fixture to bring the reading to zero?  Even so, I'm fixing equipment rather than fixing broken gear.

In the latest manual, which I am sure you have read, it states for the 10nF range it has a resolution of 10pF and an accuracy of +/-2.5% +5counts.  Of course latest firmware has a resolution of 1pF, not 10pF as the manual suggests.  If it was +5 counts or 50pF in addition to the 2.5%, seems like it still would be the same, just with a higher resolution.    That's a guess.  The manual needs to be updated. 

It's not obvious to me that is there consensus on the cal process.  Am I wrong??

There is the manual, which does not appear correct.  There is dcac's post:

Enter calibration mode, select lowest cap range and connect a 1nF cap, meter will display a value probably somewhat less than 1nF, press Setup to perform offset cal, after count down meter will still display an offset value. Now change the 1nF to a 10nF cap, press Mem to perform gain cal, after count down meter will store both offset and gain calibration and should now display 10nF.

I believe this matches the old documentation that Dave had supplied me.   Again, I am not sure why they changed the procedure other than the additional cost to perform that extra step.   

[joeqsmith]

For fun, I ran a fairly long temperature test using Production 1 after I had cleaned it up.  This is the undamaged meter that was temperature cycled in Part 6.  Again, from my post:

https://www.eevblog.com/forum/testgear/hear-kitty-kitty-kitty-nope-not-that-kind-of-cat/msg2837862/#msg2837862

I had started out at -10C.  I then stepped the temperature in 5 deg increments up to 55C, holding at each temperature for 2 hours.  I wanted to make sure the meter was somewhat stable.    I then took it back to 0C for a few hours and then took it back to 50C and finally back to 25.    About a day and a half to run.   

[joeqsmith]

In the last transittion from 0 to 55C, where the chamber's slew rate was not limited,  I again zoom into the 20deg C area.   Again, we see that the meter behaves very close to Production 2 as well as what was shown in Part 6.   

[joeqsmith]

A bit more interesting is that when we look at the 121's measured voltage, it's not linear.  Once we get above 30C or so the error starts to increase.

[joeqsmith]

Looking at the 121's measured voltage during the transitions between the two extremes, we can see the most error.   I wouldn't be too surprised if I actually thermally shocked it, say for example, leave the meter in the car during winter and bring it into the warm house and try to use it, we would see something worse.   The change would be less but the ramp would basically be the thermal mass of the meter. 

When I was trying to shock(ish) some of the other meters, that UT181A has proven to be one of the more stable meters. I doubt it has anything to do with the changes I made to improve it's front end.    It could just be an odd ball meter. 

[CDaniel]

And the cause for this thermal anomaly ?

By the way , shitty switch indeed , I just had to open it , clean it and resettling those contacts , resistor/cap position developed an intermitent fault , no power or reseting if I touched it . And of course the meter wasn't used much at all .

[dcac]

I have continued to work on the damaged 121GW (Production 2) and now have it working for the most part.

This meter was cycled from 5 deg C to 45 and back down to 5.  Sweep1 is showing the entire data set.  I was using a 1mV source.   White is the chamber's temperature, Red is the 121GW's ambient temperature and Green is the reported voltage. 

Sweep2 is looking at the voltage.  You can see as the temperature is changing how the meter responds.

I was amazed how fast the 121 seem to react to temperature changes. So here I have taken the green mV trace, flipped it vertically and overlayed on the two temperature traces where red being 121 internal temp and white the chamber temp.

I’ve placed the time/sample axis in sync but the vertical axis for the green mV trace is now only relevant for the (inverted) trend.

As it seems the 121‘s internal temperature is lagging behind as you’d expect for the temp change to make its way through the meter casing. But the mV changes almost directly on a temperature change i.e. at the orange vertical markers I’ve placed.

Edit: correction! those two screenshots I've merged perhaps wasn't from the same sweep and therefor not in sync to begin with.

[joeqsmith]

And the cause for this thermal anomaly ?

By the way , shitty switch indeed , I just had to open it , clean it and resettling those contacts , resistor/cap position developed an intermitent fault , no power or reseting if I touched it . And of course the meter wasn't used much at all .

I doubt I will dig into it but it could very well be one of the clamps. 

Sorry to hear about the switch failing.  What was your switch,
1) single dimple on all contacts
2) double dimple on outside contacts, single on inner
3) double dimple on all contacts

Platted?  Shim orientation?  Shimless with thin PCB?     Shimless with thick PCB? 

So many to choose from..

[joeqsmith]

I was amazed how fast the 121 seem to react to temperature changes. So here I have taken the green mV trace, flipped it vertically and overlayed on the two temperature traces where red being 121 internal temp and white the chamber temp.
....
Edit: correction! those two screenshots I've merged perhaps wasn't from the same sweep and therefor not in sync to begin with.

Left scale is voltage, right temperature.   White is chamber temp, Red 121's ambient, Green reported voltage.  Samples are roughly at 2Hz and lag is about 170 seconds.  Of course, the time to settle is actually well over an hour in this case..

[joeqsmith]

If you want some idea how long it took to settle after that step, the total time is about 142 minutes.  The chamber took about 33 minutes to ramp.   So about 109 minutes or say about an hour and a half in this particular case.   

[CDaniel]

And the cause for this thermal anomaly ?

By the way , shitty switch indeed , I just had to open it , clean it and resettling those contacts , resistor/cap position developed an intermitent fault , no power or reseting if I touched it . And of course the meter wasn't used much at all .

I doubt I will dig into it but it could very well be one of the clamps. 

Sorry to hear about the switch failing.  What was your switch,
1) single dimple on all contacts
2) double dimple on outside contacts, single on inner
3) double dimple on all contacts

Platted?  Shim orientation?  Shimless with thin PCB?     Shimless with thick PCB? 

So many to choose from..

Don't be sorry for me  , at this point I don't care too much and I can sort this kind of problems by myself ... This is just for science , I doubt this meter will be ever 100% functional , maybe if somebody has too much time to rewrite the firmware . Even the cap in circuit measurement is not working right , very easily fooled by the components on the board when other meters have no trouble , but no point in complaining anymore .
Some time ago I put some picture  for you , it is single dimple on all contacts with pre-installed shim . Now I put in reverse like in David's video , a little more tension on the contacts  I see some wear in the plastic where is touching the pcb inside the hole and fingers .

[joeqsmith]

Don't be sorry for me  , at this point I don't care too much and I can sort this kind of problems by myself ... This is just for science , I doubt this meter will be ever 100% functional , maybe if somebody has too much time to rewrite the firmware . Even the cap in circuit measurement is not working right , very easily fooled by the components on the board when other meters have no trouble , but no point in complaining anymore .
Some time ago I put some picture  for you , it is single dimple on all contacts with pre-installed shim . Now I put in reverse like in David's video , a little more tension on the contacts  I see some wear in the plastic where is touching the pcb inside the hole and fingers .

Don't be surprised to find your pictures in some of my videos.  I saved the one's people uploaded and use them for references.  This was how I knew there had been so many changes.   

I saw where someone had Dave if they would verify the proper direction of the shim but there was no response.   With yours in the same direction as mine, it seems like Dave just placed it in upsidedown in his video.   This make sense just looking at it.   Then again, maybe some meters need that little extra preload.   

On the meter I ran, the switch was obviously in pretty poor condition after life cycling it, so I used some 600 paper and rubbed down the contacts, while they were still attached to the switch.  Basically to get them flat.  I then polished them with some 1000.   Then used some Kapton tape to make a cover for the shim and the housing.  So now it's Kapton riding on Kapton.  This brought  the preload up enough to get the switch usable.   I may measure how deep some of the grooves are in the pads.   I was surprised that some of those hadn't cut through to the fibreglass.   

[joeqsmith]

I was going to attempt to look at the switch with the profilometer but the new software crashed on me several times and I eventually gave up.   

[maxwell3e10]

Maybe it has already been discussed before in this thread, but I am wondering why there are no cheap Fluke clones with good electrical robustness. Most of the techniques are in plain sight. Just copy the board with big clearances, large fuses and MOVs. It may cost a few more dollars for parts, but the margin for "professional" multimeters is much larger.

[Fungus]

Maybe it has already been discussed before in this thread, but I am wondering why there are no cheap Fluke clones with good electrical robustness. Most of the techniques are in plain sight. Just copy the board with big clearances, large fuses and MOVs. It may cost a few more dollars for parts, but the margin for "professional" multimeters is much larger.

Because:
a) There's no need, sales of unsafe meters are doing just fine.
b) Certification costs money and you're not going to sell many meters without certification marks.
c) Once a meter is certified you're not allowed to change a single component supplier or any part of your production line without going through re-certification.

I suspect (c) is a real problem for Chinese meter makers.

[maxwell3e10]

They seem to have no problem putting certification stamps on existing meters, whether outright fake or just shady. With plenty of review sites and videos, it wouldn't be hard to get a reputation for real robustness with a few large protection components.

[rsjsouza]

Maybe it has already been discussed before in this thread, but I am wondering why there are no cheap Fluke clones with good electrical robustness. Most of the techniques are in plain sight. Just copy the board with big clearances, large fuses and MOVs. It may cost a few more dollars for parts, but the margin for "professional" multimeters is much larger.

Because:
a) There's no need, sales of unsafe meters are doing just fine.
b) Certification costs money and you're not going to sell many meters without certification marks.
c) Once a meter is certified you're not allowed to change a single component supplier or any part of your production line without going through re-certification.

I suspect (c) is a real problem for Chinese meter makers.

Fungus, "c" is only valid if you apply for their "listed" program (UL, TÜV, Intertek, etc.). The listing requires annual inspection to the manufacturing plant audits and process reviews. Most Chinese low cost meters do not go through that, but that does not solely explain lack of safety: several japanese brands (Samwa, Hioki) have excellent quality and reputation and also do not have markings. It boils down to the manufacturer's reputation and seriousness.

A reasonable meter with decent safety is the UT139C.

[floobydust]

[...] several japanese brands (Samwa, Hioki) have excellent quality and reputation and also do not have markings. It boils down to the manufacturer's reputation and seriousness.

No- Sanwa, Hioki etc. will "design to" 61010 but never tested, no approvals, so it's another gamble. One spacings mistake and you learn why testing is done in the first place.
Using gas-tubes for protection is a known no-no because of the follow-through current that exists with real mains-transients. OVC aren't just a high-voltage blip. Once a gas-tube is lit due to a transient, you have mains to deal with afterwards and in the 1/2 cycle arc flash happens, the part explodes.