Poll

How many cycles will the KeySight U1281A's detent spring last?

0-2000
7 (17.9%)
2k-4k
5 (12.8%)
4k-8k
14 (35.9%)
8k-16k
7 (17.9%)
>16k (most rubust meter ever made)
6 (15.4%)

Total Members Voted: 36

Author Topic: Handheld meter robustness testing  (Read 1150068 times)

0 Members and 3 Guests are viewing this topic.

Offline Neutrion

  • Frequent Contributor
  • **
  • Posts: 305
  • Country: hu
Re: Handheld meter robustness testing
« Reply #4525 on: August 10, 2022, 03:32:49 pm »

I'm not aware of any meters I have that are perfect.  If you feel there were, let me know which and I can have another look.   It is very possible that someone has a specific case they were running and I had meters that would handle it but let's be clear that doesn't suggest they wouldn't have other problems.   As shown in the photo, you could see that Keysight is no exception.   
I did not state that any meter was perfect, it was only about this specific test. Maybe the extech was the only one with ac+dc mV scale which passed it.

Quote
You were having the meters in series and changing the test current. Some meters were responding really Slowly.
You may have noticed from watching reviews for this Keysight that it is very slow.    I think the Martin (??) guy had made a video showing how it compared.  I think Dave may have made one as well.   You may want to try searching YT and see what's out there.  It's an old meter and many people have looked at it. 

Quote
Now that you made your wife to be the standard switch-feeling tester, it is a pity that you didn't show her the dt830 meter.
I didn't want to waste too much of her time but thought it was interesting that the Fluke 87V and UNI-T shown were both subjected to the 50,000 cycle life test and still have a better feel than the Keysight.   I never would have guessed she would like the BM869s.  Many people complain about how tight the switch is and the small knob.
[/quote]

Possibly the waterproof washer is the problematic part. Still, nex time show her the dt830 as well, just to have an objective oppinion :)
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4526 on: August 10, 2022, 03:33:24 pm »
The SW generally works well for me (although I am on Windows 8 and not the newfangled 10), but the biggest annoyance is that is requires MS Office installed for any reporting/export functions. That is a PITA as I haven't installed MS Office in years in my PCs (I use some really excellent clones such as the Softmaker Freeoffice).
That may be why the report generator locks up.   I have Office installed.... 97!   :-DD   I refuse to change it as IMO, it started going down hill after this.  Anytime I have to use the latest Office tools today, part of my soul dies.   

Do you find that the system will glitch as I saw for longer captures?   

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4527 on: August 10, 2022, 03:37:47 pm »

I'm not aware of any meters I have that are perfect.  If you feel there were, let me know which and I can have another look.   It is very possible that someone has a specific case they were running and I had meters that would handle it but let's be clear that doesn't suggest they wouldn't have other problems.   As shown in the photo, you could see that Keysight is no exception.   
I did not state that any meter was perfect, it was only about this specific test. Maybe the extech was the only one with ac+dc mV scale which passed it.

You would have to be more specific as I've covered a lot of ground. 

Quote
Possibly the waterproof washer is the problematic part. Still, nex time show her the dt830 as well, just to have an objective oppinion :)
I don't think that's going to happen any time soon but you are certainly free to ask your wife to run what ever test you like. 

Offline Neutrion

  • Frequent Contributor
  • **
  • Posts: 305
  • Country: hu
Re: Handheld meter robustness testing
« Reply #4528 on: August 10, 2022, 05:15:35 pm »
To be more specific:
Pulled out my favorite meter of all time, the lime green TIP 194 II.  Now keep in mind this meter is currently damaged and has high leakage on one of the controllers port pins that effects the resistance measurements. 

In DCmV the meter over ranges at +/-500mV. 
In ACmV the meter over ranges at 1.417Vp-p or again 500mV RMS (60Hz sinewave)

With 490mVRMS biased with 400mV DC, both the AC and DC  millivolts readings are correct. 
Raising the DC bias to 499.9mV with 490mV RMS applied, the meter provides the correct readings for both.
Next I apply 600mV of bias to our 490mV RMS signal and of course the DCmV reading is now out of range.  But the ACmV reading is now effected again with no indication that there is a problem.   With 1VDC bias the ACmV is all the way down to 374mV.

The TPI 194II does support the AC+DC and is a tri-display.  But sadly this is not supported in mV ranges.  It's too bad I could not get a controller to replace the one on the meter.  I would have then attempted to improve it like I show with that low end UT61E. 

**********************************************************************************************
Looking at the Extech EX540
In DCmV the meter over ranges at +/-412mV. 
In ACmV the meter over ranges at 1.168Vp-p or again 412mV RMS (60Hz sinewave)

With 409mVRMS biased with 400mV DC, both the AC and DC millivolts readings are correct. 
Raising the DC bias to 500mV, the meter over ranges on the DCmV but there is a slight effect on the ACmV.
With 1VDC bias the ACmV is down to 399mV which is better than the TPI 194 II but it's an error and there again is no indication that there is a problem. 

This meter also supports AC+DC calculations and appears to show an over range for any combination that goes above 412mV. 

**********************************************************************************************
And I'm sure everyone wants to know the UNI-Ts top of the line UT181A does.  Again this meter was damaged like the TPI with one hit from the grill starter.  I was able to repair it and made some pretty drastic mods to the PCB.  This meter is not original but I do not believe I changed anything that would effect this test. 

In DCmV the meter over ranges at +/-612mV. 
In ACmV the meter over ranges at 1.731Vp-p or again 4612mV RMS (60Hz sinewave)

With 610mVRMS biased with 600mV DC, both the AC and DC millivolts readings are correct.
Raising the DC bias to 700mV, the meter over ranges on the DCmV but the ACmV is correct. 
With 1VDC bias the ACmV is down to 587mV and again there is no indication that there is a problem. 

This meter will calculate AC+DC and has a tri-display.  In this mode with the same 1VDC bias, the AC+DC and mVDC are both over ranged but the mVAC shows the 587mVAC which is of course not correct.  But if the OL were not enough to sway the user into thinking there may be a problem with their measurement, they go one step further and enable the yellow triangle with the lightning bolt.   Still, really it's not full proof and the ACmV should show an error as well or you just know someone is going to trust that number...

**********************************************************************************************

Scott's going to ask about the Fluke 97.  That's a special snowflake as it auto ranges both the AC and DC values.  You exceed the mV range, it will switch.


I don't disagree with the OP that there is a problem with some meters showing inaccurate results under certain conditions without any warning to the user.  Hopefully running these few meters shows that this is not something unique to one brand.
 

Offline Chance92

  • Regular Contributor
  • *
  • Posts: 52
  • Country: gb
Re: Handheld meter robustness testing
« Reply #4529 on: August 10, 2022, 05:34:07 pm »
Sorry for the late reply. It's been a long day. A inverter drive I was using decided to give up and released some magic smoke, which is quite strange since we were only running it at no more than 10% of its rated power.

I'm surprised you would consider using a 2-wire approach below 10.

I agree using the 4-wire method is more appropriate. This is a side project and the support we can get is quite limited. We need something that can log the measurement automatically. It is either this or an old general purpose data logger. We do have some Keithley bench meters but we don't have any GPIB compatible PC to log the data. Hopefully the situation will change soon, if my colleague can get some funding.

When using the this meter, I will do some software tricks to improve the measurement.

That sounds interesting. I wonder if you could kindly share some details about these tricks?

I would expect what you are asking to be a paper study more than anything.

Not purely on paper. I would say this is a preliminary experiment to test the viability of liquid hydrogen cooling. However it is still far away from full scale experiment or even a technological demonstrator.

Shown after 7 hours with a 40R0000 0.005% +/-1ppm S102K attached.   Meter just setting on  my desk in open air (no temp control).   After 3 hours, something happens that we see a spike down to 1.239 ohms.  There is nothing going on during this time.  Office is closed off.   So I am going to suggest the software/firmware/hardware has a problem.      Zooming into this area, and ignoring the glitch we can see the meter changes 2 counts.   I have a low tempco 0.5ohm I'll attach and let it run for the day.

Thanks for sharing these. The glitch looks similar to what I experienced before on a Keithley 2700. Other than that the drift doesn't look so bad. Definitely better than my Gossen calibrator. Did you connect the resistor directly to the banana socket?
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4530 on: August 10, 2022, 05:55:39 pm »
To be more specific:
Pulled out my favorite meter of all time, the lime green TIP 194 II.  Now keep in mind this meter is currently damaged and has high leakage on one of the controllers port pins that effects the resistance measurements. 

In DCmV the meter over ranges at +/-500mV. 
In ACmV the meter over ranges at 1.417Vp-p or again 500mV RMS (60Hz sinewave)

With 490mVRMS biased with 400mV DC, both the AC and DC  millivolts readings are correct. 
Raising the DC bias to 499.9mV with 490mV RMS applied, the meter provides the correct readings for both.
Next I apply 600mV of bias to our 490mV RMS signal and of course the DCmV reading is now out of range.  But the ACmV reading is now effected again with no indication that there is a problem.   With 1VDC bias the ACmV is all the way down to 374mV.

The TPI 194II does support the AC+DC and is a tri-display.  But sadly this is not supported in mV ranges.  It's too bad I could not get a controller to replace the one on the meter.  I would have then attempted to improve it like I show with that low end UT61E. 

**********************************************************************************************
Looking at the Extech EX540
In DCmV the meter over ranges at +/-412mV. 
In ACmV the meter over ranges at 1.168Vp-p or again 412mV RMS (60Hz sinewave)

With 409mVRMS biased with 400mV DC, both the AC and DC millivolts readings are correct. 
Raising the DC bias to 500mV, the meter over ranges on the DCmV but there is a slight effect on the ACmV.
With 1VDC bias the ACmV is down to 399mV which is better than the TPI 194 II but it's an error and there again is no indication that there is a problem. 

This meter also supports AC+DC calculations and appears to show an over range for any combination that goes above 412mV. 

**********************************************************************************************
And I'm sure everyone wants to know the UNI-Ts top of the line UT181A does.  Again this meter was damaged like the TPI with one hit from the grill starter.  I was able to repair it and made some pretty drastic mods to the PCB.  This meter is not original but I do not believe I changed anything that would effect this test. 

In DCmV the meter over ranges at +/-612mV. 
In ACmV the meter over ranges at 1.731Vp-p or again 4612mV RMS (60Hz sinewave)

With 610mVRMS biased with 600mV DC, both the AC and DC millivolts readings are correct.
Raising the DC bias to 700mV, the meter over ranges on the DCmV but the ACmV is correct. 
With 1VDC bias the ACmV is down to 587mV and again there is no indication that there is a problem. 

This meter will calculate AC+DC and has a tri-display.  In this mode with the same 1VDC bias, the AC+DC and mVDC are both over ranged but the mVAC shows the 587mVAC which is of course not correct.  But if the OL were not enough to sway the user into thinking there may be a problem with their measurement, they go one step further and enable the yellow triangle with the lightning bolt.   Still, really it's not full proof and the ACmV should show an error as well or you just know someone is going to trust that number...

**********************************************************************************************

Scott's going to ask about the Fluke 97.  That's a special snowflake as it auto ranges both the AC and DC values.  You exceed the mV range, it will switch.


I don't disagree with the OP that there is a problem with some meters showing inaccurate results under certain conditions without any warning to the user.  Hopefully running these few meters shows that this is not something unique to one brand.

As I read it again, each meter had problems.  Maybe you were not understanding what I had posted.   

One side note, the TPI 194II has since been repaired.  Well, I replaced the damaged IC.  It was never aligned after this but at least now throws up some OKish sort of numbers.     


Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4531 on: August 10, 2022, 06:12:26 pm »
Sorry for the late reply. It's been a long day. A inverter drive I was using decided to give up and released some magic smoke, which is quite strange since we were only running it at no more than 10% of its rated power.

I'm surprised you would consider using a 2-wire approach below 10.

I agree using the 4-wire method is more appropriate. This is a side project and the support we can get is quite limited. We need something that can log the measurement automatically. It is either this or an old general purpose data logger. We do have some Keithley bench meters but we don't have any GPIB compatible PC to log the data. Hopefully the situation will change soon, if my colleague can get some funding.

When using the this meter, I will do some software tricks to improve the measurement.

That sounds interesting. I wonder if you could kindly share some details about these tricks?

I would expect what you are asking to be a paper study more than anything.

Not purely on paper. I would say this is a preliminary experiment to test the viability of liquid hydrogen cooling. However it is still far away from full scale experiment or even a technological demonstrator.

Shown after 7 hours with a 40R0000 0.005% +/-1ppm S102K attached.   Meter just setting on  my desk in open air (no temp control).   After 3 hours, something happens that we see a spike down to 1.239 ohms.  There is nothing going on during this time.  Office is closed off.   So I am going to suggest the software/firmware/hardware has a problem.      Zooming into this area, and ignoring the glitch we can see the meter changes 2 counts.   I have a low tempco 0.5ohm I'll attach and let it run for the day.

Thanks for sharing these. The glitch looks similar to what I experienced before on a Keithley 2700. Other than that the drift doesn't look so bad. Definitely better than my Gossen calibrator. Did you connect the resistor directly to the banana socket?

The resistor is mounted inside some insulation with banana connectors and plugged directly into the meter.  I am doing the same with the 0.5ohm.  The tempco for that part is not near as good and I am expecting more drift.   I may have some more stable parts if you feel there is something useful to be gained from running them.   

When looking at low resistance (ohm),  I will typically wear gloves and clean the parts with Proclean or methanol.  I will normally do a 2-point cal first.  Even though the equipment is under calibration,  I keep a set of standards on-hand that are not in the calibration system,  just for a sanity check.   I mark where I am going to measure from and always repeat on that location for every measurement.   My software will make multiple measurements waiting for the standard deviation to fall below a certain criteria while the parts become stable.   Some times I will handle the parts with tweezers to avoid transferring my body heat into the parts and speeding things up.   Once stable, I will take several more readings.   I normally automate this so it's not as bad as it seems.  I am a fan of the beach towel as well to remove air currents.   

I would imagine all basic stuff for people attempting to tighten up their measurements. 

I'll let that resistor run a few more hours then post the data for it.   Maybe from there, you can decide if there is something else you would like to see. 

Offline Neutrion

  • Frequent Contributor
  • **
  • Posts: 305
  • Country: hu
Re: Handheld meter robustness testing
« Reply #4532 on: August 10, 2022, 07:35:11 pm »
To be more specific:
Pulled out my favorite meter of all time, the lime green TIP 194 II.  Now keep in mind this meter is currently damaged and has high leakage on one of the controllers port pins that effects the resistance measurements. 

In DCmV the meter over ranges at +/-500mV. 
In ACmV the meter over ranges at 1.417Vp-p or again 500mV RMS (60Hz sinewave)

With 490mVRMS biased with 400mV DC, both the AC and DC  millivolts readings are correct. 
Raising the DC bias to 499.9mV with 490mV RMS applied, the meter provides the correct readings for both.
Next I apply 600mV of bias to our 490mV RMS signal and of course the DCmV reading is now out of range.  But the ACmV reading is now effected again with no indication that there is a problem.   With 1VDC bias the ACmV is all the way down to 374mV.

The TPI 194II does support the AC+DC and is a tri-display.  But sadly this is not supported in mV ranges.  It's too bad I could not get a controller to replace the one on the meter.  I would have then attempted to improve it like I show with that low end UT61E. 

**********************************************************************************************
Looking at the Extech EX540
In DCmV the meter over ranges at +/-412mV. 
In ACmV the meter over ranges at 1.168Vp-p or again 412mV RMS (60Hz sinewave)

With 409mVRMS biased with 400mV DC, both the AC and DC millivolts readings are correct. 
Raising the DC bias to 500mV, the meter over ranges on the DCmV but there is a slight effect on the ACmV.
With 1VDC bias the ACmV is down to 399mV which is better than the TPI 194 II but it's an error and there again is no indication that there is a problem. 

This meter also supports AC+DC calculations and appears to show an over range for any combination that goes above 412mV. 

**********************************************************************************************
And I'm sure everyone wants to know the UNI-Ts top of the line UT181A does.  Again this meter was damaged like the TPI with one hit from the grill starter.  I was able to repair it and made some pretty drastic mods to the PCB.  This meter is not original but I do not believe I changed anything that would effect this test. 

In DCmV the meter over ranges at +/-612mV. 
In ACmV the meter over ranges at 1.731Vp-p or again 4612mV RMS (60Hz sinewave)

With 610mVRMS biased with 600mV DC, both the AC and DC millivolts readings are correct.
Raising the DC bias to 700mV, the meter over ranges on the DCmV but the ACmV is correct. 
With 1VDC bias the ACmV is down to 587mV and again there is no indication that there is a problem. 

This meter will calculate AC+DC and has a tri-display.  In this mode with the same 1VDC bias, the AC+DC and mVDC are both over ranged but the mVAC shows the 587mVAC which is of course not correct.  But if the OL were not enough to sway the user into thinking there may be a problem with their measurement, they go one step further and enable the yellow triangle with the lightning bolt.   Still, really it's not full proof and the ACmV should show an error as well or you just know someone is going to trust that number...

**********************************************************************************************

Scott's going to ask about the Fluke 97.  That's a special snowflake as it auto ranges both the AC and DC values.  You exceed the mV range, it will switch.


I don't disagree with the OP that there is a problem with some meters showing inaccurate results under certain conditions without any warning to the user.  Hopefully running these few meters shows that this is not something unique to one brand.

As I read it again, each meter had problems.  Maybe you were not understanding what I had posted.   

One side note, the TPI 194II has since been repaired.  Well, I replaced the damaged IC.  It was never aligned after this but at least now throws up some OKish sort of numbers.     

So  again: The problem is when a meter  does not give any indication on the ac+dc mV scale (dc coupled)that either AC or DC overranged, and displaying false ac values.
And the extech EX540 shows that it overranged. It was not PERFECT, but it gave some indications.
The problem was NOT whether the meter can read AC perfectly with unlimited DC bias which is usually specified (a least in case of Brymen) that it can not. The problem is the missing of any kind of indication that the DC bias is overranged to take the AC values with a grain of salt.

Joe Smith also didn't like  this with a Fluke:

What about Fungus's favorite meter, the Fluke 101?  Fungus recommended that meter to me for the very first $50 shoot em up test I ran and it cam out a clear winner.  The one I have has never been apart. Not even for a peak!  It has survived every thing I have thrown at it to date and works great. It's an averaging meter but again not a big deal for what we are doing.  Surely this is our perfect meter!

It has a mVAC range and appears to over range at 660.0mV AC RMS 60Hz sinewave.  Injecting 1.7VACp-p meter reads roughly 600mV as expected.  Adding 300mV DC of bias, the meter reads 13mV high.  With 600mV bias, the meter reads 10mV low.  At 1VDC bias, its now reading 480mV.  At 1.9VDC bias, its all the way down to 85mV and no indication there is anything wrong!  Big fail for the Fluke 101!

 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4533 on: August 10, 2022, 09:12:43 pm »
So  again: The problem is when a meter  does not give any indication on the ac+dc mV scale (dc coupled)that either AC or DC overranged, and displaying false ac values.
And the extech EX540 shows that it overranged. It was not PERFECT, but it gave some indications.
The problem was NOT whether the meter can read AC perfectly with unlimited DC bias which is usually specified (a least in case of Brymen) that it can not. The problem is the missing of any kind of indication that the DC bias is overranged to take the AC values with a grain of salt.

Joe Smith also didn't like  this with a Fluke:

What about Fungus's favorite meter, the Fluke 101?  Fungus recommended that meter to me for the very first $50 shoot em up test I ran and it cam out a clear winner.  The one I have has never been apart. Not even for a peak!  It has survived every thing I have thrown at it to date and works great. It's an averaging meter but again not a big deal for what we are doing.  Surely this is our perfect meter!

It has a mVAC range and appears to over range at 660.0mV AC RMS 60Hz sinewave.  Injecting 1.7VACp-p meter reads roughly 600mV as expected.  Adding 300mV DC of bias, the meter reads 13mV high.  With 600mV bias, the meter reads 10mV low.  At 1VDC bias, its now reading 480mV.  At 1.9VDC bias, its all the way down to 85mV and no indication there is anything wrong!  Big fail for the Fluke 101!
:palm: :palm: :palm:   While I would have thought with all of my other comments about none of this being any sort of concern for me everyone would have recognized the heavy sarcasm.   :-DD

I would imagine I could find a case where I drive that Extech to 0 with voltage applied, just like I have shown with others.  Again, more entertaining than anything but of no practical value, at least to me. 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4534 on: August 10, 2022, 09:19:32 pm »
I'll let that resistor run a few more hours then post the data for it.   Maybe from there, you can decide if there is something else you would like to see.

Showing the 0.5ohm part over the 8 hours.  Looks like 8 counts.   Again, this part isn't near the same specs as the first 40 ohm I ran.   Nice thing is there was no glitches in the data this time. 

Let me know if you would still see any value in looking at it with a DC reference.  I have a very old Fluke 731B but outside of that I really don't have much for the way of standards.   
 
The following users thanked this post: Trader

Online rsjsouza

  • Super Contributor
  • ***
  • Posts: 5980
  • Country: us
  • Eternally curious
    • Vbe - vídeo blog eletrônico
Re: Handheld meter robustness testing
« Reply #4535 on: August 10, 2022, 09:27:17 pm »
Do you find that the system will glitch as I saw for longer captures?
Glitch? I haven't seen any issues. Long captures were not a problem as shown below (albeit with my U1273A)

Vbe - vídeo blog eletrônico http://videos.vbeletronico.com

Oh, the "whys" of the datasheets... The information is there not to be an axiomatic truth, but instead each speck of data must be slowly inhaled while carefully performing a deep search inside oneself to find the true metaphysical sense...
 
The following users thanked this post: Trader

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4536 on: August 10, 2022, 10:15:31 pm »
Do you find that the system will glitch as I saw for longer captures?
Glitch? I haven't seen any issues. Long captures were not a problem as shown below (albeit with my U1273A)

As in that first shot.  Keep in mind you are only three minutes of data.  I was showing 7 hours.   That glitch was after 3 hours with a 10 second sample rate. 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4537 on: August 11, 2022, 01:05:56 pm »
I let the meter run for about 14 hours with a 1mV source attached.  Again, the meter was just sitting on my desk without any temperature control.   This time the sample rate was set to 1Hz.    Once again, we see a downward glitch.   Also, like the last time this is during the night with everything shut down.

If we ignore the glitch and zoom into the baseline, it looks like it drifts about 8 counts.  Keeping in mind these measurements all include the standards and I am not suggesting this is all the meters doing.   

Let me know if this answers your question.  I am not happy seeing these two glitches in the data.  The fact I have seen it twice now and once with nothing more than a resistor attached, it seems there is something wrong with the software/firmware/meter as this isn't something I would expect to see.   
 
The following users thanked this post: Chance92

Online rsjsouza

  • Super Contributor
  • ***
  • Posts: 5980
  • Country: us
  • Eternally curious
    • Vbe - vídeo blog eletrônico
Re: Handheld meter robustness testing
« Reply #4538 on: August 11, 2022, 10:48:42 pm »
Do you find that the system will glitch as I saw for longer captures?
Glitch? I haven't seen any issues. Long captures were not a problem as shown below (albeit with my U1273A)

As in that first shot.  Keep in mind you are only three minutes of data.  I was showing 7 hours.   That glitch was after 3 hours with a 10 second sample rate.
That was an older capture, but the total was more than twenty minutes of data without a glitch (not three). I can retry with the U1282A, but I won't promise a runlength of hours.
Vbe - vídeo blog eletrônico http://videos.vbeletronico.com

Oh, the "whys" of the datasheets... The information is there not to be an axiomatic truth, but instead each speck of data must be slowly inhaled while carefully performing a deep search inside oneself to find the true metaphysical sense...
 

Offline Chance92

  • Regular Contributor
  • *
  • Posts: 52
  • Country: gb
Re: Handheld meter robustness testing
« Reply #4539 on: August 11, 2022, 11:03:40 pm »
I let the meter run for about 14 hours with a 1mV source attached.  Again, the meter was just sitting on my desk without any temperature control.   This time the sample rate was set to 1Hz.    Once again, we see a downward glitch.   Also, like the last time this is during the night with everything shut down.

If we ignore the glitch and zoom into the baseline, it looks like it drifts about 8 counts.  Keeping in mind these measurements all include the standards and I am not suggesting this is all the meters doing.   

Let me know if this answers your question.  I am not happy seeing these two glitches in the data.  The fact I have seen it twice now and once with nothing more than a resistor attached, it seems there is something wrong with the software/firmware/meter as this isn't something I would expect to see.

Thanks for sharing this. I guess the voltage across the resistor during the resistance test was much higher than 1 mV, so the drift wasn't picked up. Do you think if you could do another resistance test with the 0.5 Ohm resistor to see if the drift would appear?
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4540 on: August 11, 2022, 11:25:00 pm »
Thanks for sharing this. I guess the voltage across the resistor during the resistance test was much higher than 1 mV, so the drift wasn't picked up. Do you think if you could do another resistance test with the 0.5 Ohm resistor to see if the drift would appear?

Guessing that's a typo as I have shown a 40 and 0.5 ohm already.   

That was an older capture, but the total was more than twenty minutes of data without a glitch (not three). I can retry with the U1282A, but I won't promise a runlength of hours.

If you have saved other collections that were longer, looks like you can import them.     
***
Also, I am using software version 3.1.51130.01 
« Last Edit: August 11, 2022, 11:31:26 pm by joeqsmith »
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4541 on: August 12, 2022, 02:14:10 am »
Spent a few minutes, literally, to put a simple program together that just reads the meter and plots the data at 1Hz.  Shown with the 40ohm inserted.  Note that they send up one more digit.  I'll let this run overnight and see if a glitch shows up.   

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4542 on: August 12, 2022, 01:05:46 pm »
After letting the Keysight U1282A run with the 40 ohm resistor,  shown is roughly 10 hours of data collected with my simple program.  Once again we can see the problem.  Zooming into both glitches, they are a single data point.   While it could be the driver, I think this rules out their software.   

Odd such a high end meter would have such a problem.  Surely Keysight tests their products.  It would be nice to know if it was unique to this particular meter. 

Offline HKJ

  • Super Contributor
  • ***
  • Posts: 2899
  • Country: dk
    • Tests
Re: Handheld meter robustness testing
« Reply #4543 on: August 12, 2022, 02:06:57 pm »
I tried on mine, it is a fairly old 1282 with the original firmware:



Dropout after 20 minutes, but to 5 not to 0.
 
The following users thanked this post: Fungus

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4544 on: August 12, 2022, 02:32:37 pm »
Thanks for repeating this test.   Going back and looking at the first time I ran it, they both dropped to around 1.5.   So not exactly 0.  So far, every time I have seen it glitch, it is always towards a lower value. 

I'm surprised they wouldn't have caught something like this but then again, that counter spec doesn't give me any comfort.   

With them recently sending that one reviewer a meter, they may be able to provide them feedback. 

Chance92 and others would have to decide if its going to be a problem for them.   I can imagine if you are running a long term experiment and the signal is not stable as you expect, then your spending time trying to track down if it's something on your end or the equipment, that will get old fast.   I guess, just something else to be aware of if you buy one of these meters. 
 
The following users thanked this post: Chance92

Online Fungus

  • Super Contributor
  • ***
  • Posts: 16562
  • Country: 00
Re: Handheld meter robustness testing
« Reply #4545 on: August 12, 2022, 02:49:47 pm »
Have you tried this test with any of your Brymens?

It would be interesting to see the signal at that point to see if it's data corruption or if the meter is sending incorrect data.
« Last Edit: August 12, 2022, 02:51:50 pm by Fungus »
 

Offline Chance92

  • Regular Contributor
  • *
  • Posts: 52
  • Country: gb
Re: Handheld meter robustness testing
« Reply #4546 on: August 12, 2022, 02:50:31 pm »
Thanks for repeating this test.   Going back and looking at the first time I ran it, they both dropped to around 1.5.   So not exactly 0.  So far, every time I have seen it glitch, it is always towards a lower value. 

I'm surprised they wouldn't have caught something like this but then again, that counter spec doesn't give me any comfort.   

With them recently sending that one reviewer a meter, they may be able to provide them feedback. 

Chance92 and others would have to decide if its going to be a problem for them.   I can imagine if you are running a long term experiment and the signal is not stable as you expect, then your spending time trying to track down if it's something on your end or the equipment, that will get old fast.   I guess, just something else to be aware of if you buy one of these meters.

Thank you very much for taking the time to do these tests. A glitch like this is indeed surprising. I just looked up the firmware update. The latest version is 1.05, which came out in 2017. Release notes indicate they probably haven't address this issue so far.
https://www.keysight.com/us/en/assets/9018-18550/release-notes/9018-18550.txt?success=true
Quote
Firmware revision updates:-


Version 1.02
-    Initial release

Version 1.03 fixes several issues present in the previous release.
-   Change microcontroller sampling period from 25ms to 12.5ms.
-    Bug fix for CAP+NULL+AUTO wrong range.
-   Faster LED response for continuity.

Version 1.04 fixes several issues present in the previous release.
-   Change Ampere calibration limits from 70% to 80% to solve       production calibration fallout.
-   Improve Autozero limitation by recalculating Autozero and not       loading Autozero default value when inputs are OL.
-   Reset OP1 & AD1 before reading Autozero.
-   Zero Timer change from 4.2 seconds to 6.2 seconds.

Version 1.05 includes the following fix.
-   Disables DC Filter during measurement at ACmV mode.


Have you seen anything like this on other meters you have tested?
 

Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4547 on: August 12, 2022, 03:42:20 pm »
Have you tried this test with any of your Brymens?

It would be interesting to see the signal at that point to see if it's data corruption or if the meter is sending incorrect data.

I have not.

Have you seen anything like this on other meters you have tested?

Normally I am using my old HP34401A with GPIB and GPIB to Ethernet.  That system is very stable.   I saved a very old Fluke thermal RMS meter from the recycle and repaired it.  That meter was using RS232 and is again very stable.

The handhelds I have looked at are the UNI-T UT181A, the CEM DT-9939, Dave's 121GW, Fluke 189 and the Gossen Ultra (now Prime, go Gossen).   

I did have a few problems with the 121GW but from what I remember they were self induced.   It seems like there was a problem with recording to the SD card.  I was mostly using BLE.  Some of that may have been corrected by now.   

The CEM also has a problem (seen in the attached video).  The other meters were stable.



Video showing some of these meters.



Offline joeqsmithTopic starter

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Handheld meter robustness testing
« Reply #4548 on: August 12, 2022, 07:26:35 pm »
Here's the vintage Fluke I trash picked after my repairs.  Shown with a 60 ohm using 4-wire for 20 hours.  Again, this is using RS-242 to USB.  The whole setup is very stable and I would expect that Keysight to be the same.   

While this meter is very old, big, heavy... the specs are still very impressive.   You may want to consider trying to find an old working bench meter for your experiment.       

Online 2N3055

  • Super Contributor
  • ***
  • Posts: 6454
  • Country: hr
Re: Handheld meter robustness testing
« Reply #4549 on: August 12, 2022, 08:42:17 pm »
Have you tried this test with any of your Brymens?

It would be interesting to see the signal at that point to see if it's data corruption or if the meter is sending incorrect data.

I did long data logging of voltage with both BM525S in standalone mode and BM869S with a IR cable..
Never had any problems..
 
The following users thanked this post: Fungus


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf