EEVblog 121GW Multimeter Issues

[Stonewatch]

With 400VAC 50Hz mains voltage, the output of the VFD can not be significantly above 400VAC. With the corner frequency of the VFD set to 50Hz at this frequency the output voltage should be close to the input.
The VFD option is lacking on the 121GW and the 1kHz lowpass filter is valueless for these kinds of signals. Both meters in the pictures attached, are 1 meter away from the VFD and connected in parallel.

[CDaniel]

You should look at that waveform with a scope , to see what "ugliness" is present there 
121GW is good for measuring AC true-rms up to 100KHz and beyond , no doubts about that .

[Kleinstein]

The VFD output is very likely a mixture of different frequencies: the 50 Hz part and as it looks like some 4 KHz contribution. So the frequency readings can be both right (the frequency is present) and wrong (it is not a simple repetitive signal). For the RMS part one has to interpret it together with the bandwidth.  Different BW settings / meters can thus give different results as they measure different frequency bands.

Besides the signal, there can also be EMI issues near an VFD.

[Stonewatch]

IMO the 121GW provides a unique combination of features at a moderate price point, but after having seen Joe Smiths videos, and my personal experience (VFD), it is not designed to be used in a workshop or an industrial environment.

[floobydust]

VFD measurements are dangerous and never connect a scope to one's output without a proper rated HV probe. It won't tell much anyhow.
400VAC gives about 700VDC bus voltage, which is modulated by the IGBT's for a 700Vpp square-wave phase-voltage. It's very hard for any test equipment to behave with that much AC.

To see if it's an interference problem, put the DMM on ACV manual range say 1000V, short the two input leads together and it should read around 0V. With the drive running, place both probes on ONE motor terminal (DMM should still read 0V). I have seen DMM's freak out, reboot or crash, or give silly readings during this test.

The DMM frequency reading is likely correct - the drive's PWM carrier frequency is programmable in F159 from 800-10,000Hz, it's the squeal you hear. This is not the motor speed or PWM fundamental i.e. giving 0.5-590Hz

Otherwise, maybe the 1kHz filter has some issues?

[Stonewatch]

To see if it's an interference problem, put the DMM on ACV manual range say 1000V, short the two input leads together and it should read around 0V. With the drive running, place both probes on ONE motor terminal (DMM should still read 0V). I have seen DMM's freak out, reboot or crash, or give silly readings during this test.

Thanks for the tip floobydust. I have just performed this test...121GW reads 0 even holding the instrument next to the VFD.

[floobydust]

But with both DMM test probes touching one motor terminal when the drive is running?

[Stonewatch]

Yes Floobydust:  both tips on one terminal with the motor running, no EMI issue.

Maybe some future SW revision could add a function to facilitate correct VFD measuring as BM869 , BM235 and Fluke 28II can do it. The Fluke 175 however, dislays nearly the same incorrect values (10% too high) as the 121 does.

[e0ne199]

any news about firmware update?

[EEVblog]

any news about firmware update?

I have a new version but I haven't tried it yet due to the holidays.

[firewalker]

any news about firmware update?

I have a new version but I haven't tried it yet due to the holidays.

Release it as stable and tested! 

Alexander.

[1anX]

any news about firmware update?

I have a new version but I haven't tried it yet due to the holidays.

Dave, does this mean you dont take your 121GW meter away with you on holidays ?
Shame on you 
Enjoy the holiday break 

[npelov]

Hey everyone,

I'm back from the surgery - alive and kicking. I can see you've repeated the strange SD card logging problem - thanks joeqsmith. I noticed that the sample repeat frequency is about 60 samples. I would have guessed 64, but looking at this image it's more like exactly 60 (I didn't look at the csv data - just eyeballing it in two screnshots). So that's nice round number - maybe the developers can relate this number to the code.

Happy holidays!

[joeqsmith]

Hey everyone,

I'm back from the surgery - alive and kicking. I can see you've repeated the strange SD card logging problem - thanks joeqsmith. I noticed that the sample repeat frequency is about 60 samples. I would have guessed 64, but looking at this image it's more like exactly 60 (I didn't look at the csv data - just eyeballing it in two screnshots). So that's nice round number - maybe the developers can relate this number to the code.

Happy holidays!

Good to have you back.  I did mention your findings in my last video.  If you need to get some sleep to help with your recovery, I suggest you watch it.   

[EEVblog]

any news about firmware update?

I have a new version but I haven't tried it yet due to the holidays.

Dave, does this mean you dont take your 121GW meter away with you on holidays ?
Shame on you 
Enjoy the holiday break 

I'm going away in a few days, so not sure when I'm going to have time to test this myself.
So might be best if I just attach it here and get feedback.
So consider this an internal not released try at your own risk if you dare firmware.
It fixes the resistance drift problem, i.e. 1M reading which slowly drifts down to the correct value over time, but only on the 2nd probing, it works fine after power-on and first measurement.

Also, time/date stamping has been added to the logging file.

[Dr. Frank]

Thank you, Dave.
The 1 MOhm range now works as intended, i.e. gives instantaneously a stable display on 2nd measurement.

The 600V_dc range, down ranging problem is not yet solved, i.e. it still down ranges at <= 49.9 V only.

Frank 

[e0ne199]

Thank you, Dave.
The 1 MOhm range now works as intended, i.e. gives instantaneously a stable display on 2nd measurement.

The 600V_dc range, down ranging problem is not yet solved, i.e. it still down ranges at <= 49.9 V only.

Frank

how about logging? did you test that too?

[Scottjd]

how about logging? did you test that too?

If your referring to the offing loop error in this thread, he didn’t mention that this firmware fixes it. This version just adds date and time stamp to the logs.
Joe, npelov, Dave and I spent a lot of time replicating this log error. I know I had a few days running logs for it.
As for me, I’m not looking to run that test again until Dave specifically states that issue was addressed in the firmware.

[joeqsmith]

how about logging? did you test that too?

If your referring to the offing loop error in this thread, he didn’t mention that this firmware fixes it. This version just adds date and time stamp to the logs.
Joe, npelov, Dave and I spent a lot of time replicating this log error. I know I had a few days running logs for it.
As for me, I’m not looking to run that test again until Dave specifically states that issue was addressed in the firmware.

I'm in the same boat.  There's also no mention of them correcting the auto range problems that prevent the meter from indicating a presence of high voltages.   I can't see spending any time retesting it and plan to pull the Duracells and box them up.   Maybe in another couple of years we can revisit what ever hardware they come up with.  I doubt we have seen the last of changes to the front end or that switch design.   Maybe they will get some proper lettering on the meter by then as well.   ANENG could help them out with that one. 

[CDaniel]

So far I see no difference , probably they resolved just that weird issue with Mohm range ...
For resistance is just as slow and showing that first bad value in 50Mohm range ...

[LiftedTrace]

Edit: After browsing the other stickies, I see I may have posted this in the wrong section as this is not a meter issue that I know of. Probably should have been posted in the discussion section. Sorry about that.


Hi all, I just got this meter, not really needing it, but wanting to support eevblog and the fact its hack-able and encouraged.I havent used many settings on it yet, but I was trying to get the temperature of my hot end on my 3d printer. Seeing a large discrepancy, I wanted to find a way to see if the meter was off. I am almost positive the majority of the error is due to not a good thermal couple. Either way, I dont know how to really calibrate these and the only thing I could think of was boiling water.
I noticed it said my boiling point was exactly 97C. Im not up in the mountains, so shouldn't this be closer to 100C. Is there a way to offset it? Is boiling water not a true way to check an accurate temp?

Also when looking at a K type theralcouple chart, my ambient temp is 19C, shouldn't I be measuring .758mV? I measure -.085mV when trying to check what the thermalcoupler is outputting.
Thanks.

[thinkfat]

Check polarity of the plug.

[LiftedTrace]

I tried the polarity. It only flips the sign, buts still a far cry away from.758mV.

[qft1967]

A thermocouple doesn't measure the absolute temperature, but only the temperature difference. (Keyword: Cold junction compensation) So in your case you would measure the temperature difference to its internal temperature of 19°C. If everything is in equilibrium then you should see 0mV!

[AVGresponding]

A thermocouple doesn't measure the absolute temperature, but only the temperature difference. (Keyword: Cold junction compensation) So in your case you would measure the temperature difference to its internal temperature of 19°C. If everything is in equilibrium that you should see 0mV!

This is not how thermocouples, or cold junction compensation works.

I suggest you read this as a primer.