Handheld meter robustness testing

[joeqsmith]

I have finished with the updates to the BM789 and ran through a basic functional test.  It seems no worse for the wear.   I plan to start working on the review this weekend.  With it being very similar to the BM786, I plan to forego the drop, chemical and selector switch life cycle test.   If there is anything specific anyone would like to see, feel free to ask. 

[joeqsmith]

The meat packing box is box out.   

[joeqsmith]

It's the middle of summer and the office is fairly warm.  Even with all the insulation, the small Peltiers are having a difficult time reaching -5C. 

[rsjsouza]

It's the middle of summer and the office is fairly warm.  Even with all the insulation, the small Peltiers are having a difficult time reaching -5C.

If you were making money out of these tests, a chest freezer would be your christmas gift! 

[joeqsmith]

I had seen a small compressor type cooler that looked like it would be perfect for this kind of work.   

Consider I held the meter at -5 / 60C for a half hour.  It took about 6 hours to cool down to -5.  Ramp up was about an hour.   Basically it's a full day.   Now the best part, that's about 1 minute of video!       It's actually a worse time ratio than the switch life test.   

[joeqsmith]

I may show off my vintage Fluke 8506A that I saved from the dumpster to compare the AC performance.   This meter has a thermal RMS converter.   It needed some repairs and I had using my old HP34401A to do a poor mans alignment on it, except for the AC sections.    The AC specs on this meter far exceed anything I own and I suspect I would have done more harm than good. 

[joeqsmith]

We ran into a bit of a snag with the BM789.  Brymen suspects the problem is alignment related caused from changing out the IC to update the firmware.  They are providing a brand new BM789 which should address it.   

Because it'sonly the alignment,  I plan to continue with some of the other tests (battery life, transients...) where this isn't critical.  Hopefully I can at least keep the effort moving ahead.     

This is what happens when you do more than unbox a product, provide a half hour blab fest and give it five stars.  So be patient. 

[joeqsmith]

A brand new BM789 arrived.   

Good job Brymen.   

[joeqsmith]

The Brymen BM789.

[joeqsmith]

This video is for you vintage analog collectors. 

[Cymaphore]

Thank you for the great video! Your videos are really a different class of review.

About the resistor autoranging test: When I tried to make a comparison of the B1023 autoranging performance, I used a couple of photo resistors (demonstrated here). Two of them, and maybe more, can be easily trimmed to near equal values to compare meters in an insulated fasion without transistor sideeffects.

[joeqsmith]

I doubt using better resistors and a good relay would make any difference for the settling time.   Originally, I would just short them but it seems Dave had posted some data with a 10 ohm and I follow suit to try and get a decent comparison.   

There was a person who had posted about trying to use one of my videos to come up with an actual time.  After that I setup a high speed camera with a counter as a reference.  at 1000fps, you would watch the LCDs operate.   It was fairly accurate but not very useful information, to me anyway. 

[Cymaphore]

I doubt using better resistors and a good relay would make any difference for the settling time.   Originally, I would just short them but it seems Dave had posted some data with a 10 ohm and I follow suit to try and get a decent comparison.   

There was a person who had posted about trying to use one of my videos to come up with an actual time.  After that I setup a high speed camera with a counter as a reference.  at 1000fps, you would watch the LCDs operate.   It was fairly accurate but not very useful information, to me anyway.

I don't think it would impact the settling time, but I wanted something that is as close to a bare resistor as possible. That's why I tried out that LDR-solution.

Operating multiple LDRs by a single light source is a nice thing. Depending on the LDR in question it can be controlled quickly and on a very wide range. The ones that I used for toying have ~6 kOhm (Full brightness) to ~600k Ohm (Complete darkness), with a little tweaking that touches 4 resistance ranges on the BM789 and it can be controlled quickly and easily by just changing a single light source.

Other LDRs have different ranges.

But you're correct, for your test it probably would not make any difference. I just wanted to mention it, because it was really interesting for me to compare the B1023 to the BM789 in that way and it is maybe a good idea in case you ever need such kind of a controllable resistor.

[HKJ]

I uses a opto-fet for switching a resistor in when testing DMM's, this way I avoid any noise or leak current and because it is a dual-fet I can also use it for capacitance ranges.
Instead of comparing different meters I find how long time I need to switch the resistor on before the meter will measure it correctly, this is easier to do and more precise, but it would not look as interesting on video.

My circuit: https://lygte-info.dk/info/DMMTesterACSwitch%20UK.html

[joeqsmith]

The only use I had for an LDR with the meters was to detect the ambient light for controlling the backlight of a check UNI-T meter.   

I was wanting to run the meter's entire range consistently.

I uses a opto-fet for switching a resistor in when testing DMM's, this way I avoid any noise or leak current and because it is a dual-fet I can also use it for capacitance ranges.
Instead of comparing different meters I find how long time I need to switch the resistor on before the meter will measure it correctly, this is easier to do and more precise, but it would not look as interesting on video.

My circuit: https://lygte-info.dk/info/DMMTesterACSwitch%20UK.html

You want to avoid leak current but stick a 10M probe across it?  I wanted to see the meters go full open range on down.   The scope may show the current source changing but it won't show the meter settling.  Dave's 121GW was interesting in that one of the things they were constantly doing was playing with the software filters.  It was easy to detect with the camera. 

I had not thought about trying to look at the capacitance this way.  Some of the meters are again VERY slow to settle at the lower ranges. 

Fully agree about the scratching leads not being a very useful way to compare them which was why I started using something a bit more controlled.   

[HKJ]

You want to avoid leak current but stick a 10M probe across it? 

No, the probe was only to because I wanted to show the curve, in real test I do not stick anything across.

[joeqsmith]

Instead of comparing different meters I find how long time I need to switch the resistor on before the meter will measure it correctly, this is easier to do and more precise, but it would not look as interesting on video.

I'm not understanding your above comment.   If you do not time them with the scope, I doubt leakage would be a problem with these handhelds.   Without the scope, I assume you monitor the settling with the display.   How is what you are doing more precise.  I guess I need to take the time to read your entire blog. 

With your jig using a 10uF, I guess I don't see the point of the two FETs.   I was looking my fixture using the first four meters and capacitance was about 1nF.  Even with these large vintage FETs (I think I have some BUZ parts in there), 10uF would be no problem to switch in.   

The meters do not always use the same polarity and I need to flip the leads which is where the dual FET would be an improvement.       

[HKJ]

Instead of comparing different meters I find how long time I need to switch the resistor on before the meter will measure it correctly, this is easier to do and more precise, but it would not look as interesting on video.

I'm not understanding your above comment.   If you do not time them with the scope, I doubt leakage would be a problem with these handhelds.   Without the scope, I assume you monitor the settling with the display.   How is what you are doing more precise.  I guess I need to take the time to read your entire blog. 

There is no reason to look at the pulse, I know the pulse with from my generator and when the pulse width is long enough for the meter to get a correct reading I have my time. With some meters this time varies a bit and I increase the time until I get a stable reading for each pulse.

With your jig using a 10uF, I guess I don't see the point of the two FETs.   I was looking my fixture using the first four meters and capacitance was about 1nF.  Even with these large vintage FETs (I think I have some BUZ parts in there), 10uF would be no problem to switch in.   

Many meters uses some sort of AC to measure capacitors, with two FETs I switch both polarities on/off.

[joeqsmith]

Instead of comparing different meters I find how long time I need to switch the resistor on before the meter will measure it correctly, this is easier to do and more precise, but it would not look as interesting on video.

I'm not understanding your above comment.   If you do not time them with the scope, I doubt leakage would be a problem with these handhelds.   Without the scope, I assume you monitor the settling with the display.   How is what you are doing more precise.  I guess I need to take the time to read your entire blog. 

There is no reason to look at the pulse, I know the pulse with from my generator and when the pulse width is long enough for the meter to get a correct reading I have my time. With some meters this time varies a bit and I increase the time until I get a stable reading for each pulse.

Again, I guess I am missing the your point of it being more precise.  More precise than scratching the leads?  More precise than using a high speed camera with a clock for a reference?   

Does it not look as interesting as what on video?  Scratching the leads?

Looks like are basically doing the same thing. 

With your jig using a 10uF, I guess I don't see the point of the two FETs.   I was looking my fixture using the first four meters and capacitance was about 1nF.  Even with these large vintage FETs (I think I have some BUZ parts in there), 10uF would be no problem to switch in.   

Many meters uses some sort of AC to measure capacitors, with two FETs I switch both polarities on/off.

I assume you ran into problems with some meters when turning off the single FET and the meter still reading the 10uF or some other high value.   I had looked at four of the meters using a 10uF in series with the single FET and it seemed fine.  Any idea which meters had the problem with the single FET?  Maybe I have one.   

[HKJ]

Instead of comparing different meters I find how long time I need to switch the resistor on before the meter will measure it correctly, this is easier to do and more precise, but it would not look as interesting on video.

I'm not understanding your above comment.   If you do not time them with the scope, I doubt leakage would be a problem with these handhelds.   Without the scope, I assume you monitor the settling with the display.   How is what you are doing more precise.  I guess I need to take the time to read your entire blog. 

There is no reason to look at the pulse, I know the pulse with from my generator and when the pulse width is long enough for the meter to get a correct reading I have my time. With some meters this time varies a bit and I increase the time until I get a stable reading for each pulse.

Again, I guess I am missing the your point of it being more precise.  More precise than scratching the leads?  More precise than using a high speed camera with a clock for a reference?   

Does it not look as interesting as what on video?  Scratching the leads?

Looks like are basically doing the same thing. 

With my method I do multiple tests, this means I do not risk getting the fastest time the meter can do.

Looking at a couple of meter and seeing which is fastest may be more interesting for some, I like the numbers because I can compare all the meters I have tested (>100). If you measure frames you can, of course, do the same, except for continuity.

With your jig using a 10uF, I guess I don't see the point of the two FETs.   I was looking my fixture using the first four meters and capacitance was about 1nF.  Even with these large vintage FETs (I think I have some BUZ parts in there), 10uF would be no problem to switch in.   

Many meters uses some sort of AC to measure capacitors, with two FETs I switch both polarities on/off.

I assume you ran into problems with some meters when turning off the single FET and the meter still reading the 10uF or some other high value.   I had looked at four of the meters using a 10uF in series with the single FET and it seemed fine.  Any idea which meters had the problem with the single FET?  Maybe I have one.   

I did not bother checking capacity before I had the AC switch, only disconnecting one way may lead to errors (I have no idea if it do, I have never bother comparing).

[joeqsmith]

With my method I do multiple tests, this means I do not risk getting the fastest time the meter can do.

Looking at a couple of meter and seeing which is fastest may be more interesting for some, I like the numbers because I can compare all the meters I have tested (>100). If you measure frames you can, of course, do the same, except for continuity.

I used a frequency counter, tied to the arb that would output a 1kHz clock once triggered.   The counter is cleared on start and counts up while with a 1mS resolution.    The load is applied or removed at the trigger time.    The camera has just shy of a 1000fps sample rate.  With this setup, I can get some decent numbers if it were needed.   I can run up to three meters this way in parallel and it makes no difference as far as the absolute data that could be collected.  I'm looking at what was captured with the camera is all.   The fact we can see the meters side by side real time is just an added plus.   

The downside to this method is my camera can only capture in bursts of two seconds.  I currently trigger the camera using the same trigger.  Some meters take longer than this to settle.  In these cases, I have used the cameras manual post trigger to capture the data.   A pain.   I had written the guy who designed that high speed camera when he announced the prototypes and offered to buy one.  That camera would have been a big help with the testing.  He gave them away as part of a promo.  Dave did a review of the one that was sent him. 

I understand you want numbers and are your not using the scope to get them.  I understand your jig, tied to the arb but I am not understanding how you make this precise measurement.  Could you please explain how you actually get the settling time?

I did not bother checking capacity before I had the AC switch, only disconnecting one way may lead to errors (I have no idea if it do, I have never bother comparing).

So you haven't checked if there was actually a problem with the single FET.  More just trying to prevent a possible problem.  Makes sense.  Adding a capacitor may be of interest but I would like to go down to 200pF ish   and up around 1000uF   as most of the meters I have looked where slowest at the two extremes.
 
 

[HKJ]

Could you please explain how you actually get the settling time?

When the pulse is long enough for the meter to show the correct value (It is easy enough to do a long pulse first to see what the correct value is).

So you haven't checked if there was actually a problem with the single FET.  More just trying to prevent a possible problem.  Makes sense.  Adding a capacitor may be of interest but I would like to go down to 200pF ish   and up around 1000uF   as most of the meters I have looked where slowest at the two extremes.

I do not bother with the small values for speed tests, but test up to 70mF if the meter supports it. I have a few large capacitor I use to test with and decades/PCB's with lower values.

[joeqsmith]

I think I get it.  So you keep changing the pulse time until you find the spot where the meter is stable, triggering each time.   To get any sort of precision it must take a very long time to run.   If you had a way to monitor the meter (HS camera) one long pulse and you would have it.

[HKJ]

I think I get it.  So you keep changing the pulse time until you find the spot where the meter is stable, triggering each time.   To get any sort of precision it must take a very long time to run.   If you had a way to monitor the meter (HS camera) one long pulse and you would have it.

It takes a few minutes, precision is not really an issue (2.5s or 2.53s do not matter).

[joeqsmith]

I think I get it.  So you keep changing the pulse time until you find the spot where the meter is stable, triggering each time.   To get any sort of precision it must take a very long time to run.   If you had a way to monitor the meter (HS camera) one long pulse and you would have it.

It takes a few minutes, precision is not really an issue (2.5s or 2.53s do not matter).

30fps or 33ms would do it.  Basically any cell phone should handle it.  One cycle would get you there.  Maybe a counter in the background for the time reference. 

***
Does seem like with your eyes being used for the  feedback, 30ms error would be conservative.   Maybe not.