Brymen 869s DC offset issue when reading ACmV

[J-R]

7.5VDC+100mVAC (RMS) 500Hz sine is definitely a tricky signal.  Out of my stash, the only DMMs that could actually provide those three fundamental values were:

Fluke 87V
Fluke 88V
Keysight 1233A
Sanwa PM300 (no mV range however)
Of course Fluke 287


Some notable "failures":

121GW
Brymen BM789
Of course Brymen BM869s

Like the BM869s, the BM789 displayed a garbage reading in mVAC mode.  No matter the amplitude of the AC component, it always displayed approximately 105mV.


The Keysight 1233A surprised me with the fact that while displaying the frequency, it uses the bar graph to display the AC voltage.  It's also very fast and pretty accurate (99.4mV).  It did lose the frequency measurement below about 30mV AC (RMS) (using the above stated signal).

The Fluke 87V & 88V were also quite accurate (100.17mV & 99.86mV).  They lost the frequency measurement below about 20mV AC (RMS).

No surprise I suppose, the Fluke 287 was the most accurate (100.02mV) and made it to 5mV AC (RMS) before dropping out.

[bdunham7]

7.5VDC+100mVAC (RMS) 500Hz sine is definitely a tricky signal.  Out of my stash, the only DMMs that could actually provide those three fundamental values were:

Just for fun I quickly tried every meter I could easily grab, with or without dBm/dB readings.

Meters with mVAC and dBm: Fluke 8846A, Fluke 289, Fluke 189, HP 403B w/ OPT 01 -- all read correctly and accurately.

Meters with mVAC but not dBm:  Fluke 8842A, Fluke 27, Fluke 8800A, Fluke 116 -- all read correctly except the F116 which displayed 'OL' in the mV range, but was exactly correct (0.100V) in the 6V range.  This is what the Brymen should be doing.

Meters without mVAC: Mastec/Centec (Harbor Freight) P37772, Fairchild 7000A, Simpson 270 (OUTPUT) --obviously all AC-coupled voltmeters should work fine here, and they do.  The Simpson 270 requires careful zeroing, using the special 2.5VAC scale and some squinting, but it is accurate to less than a needle-width.

The bottom line is that I can't find any meters in my stash that give me a wrong reading.  Something to at least be aware of and IMO, something that should be fixed if possible. 

[_Wim_]

Just for fun I quickly tried every meter I could easily grab, with or without dBm/dB readings.

Meters with mVAC and dBm: Fluke 8846A, Fluke 289, Fluke 189, HP 403B w/ OPT 01 -- all read correctly and accurately.

Meters with mVAC but not dBm:  Fluke 8842A, Fluke 27, Fluke 8800A, Fluke 116 -- all read correctly except the F116 which displayed 'OL' in the mV range, but was exactly correct (0.100V) in the 6V range.  This is what the Brymen should be doing.

Meters without mVAC: Mastec/Centec (Harbor Freight) P37772, Fairchild 7000A, Simpson 270 (OUTPUT) --obviously all AC-coupled voltmeters should work fine here, and they do.  The Simpson 270 requires careful zeroing, using the special 2.5VAC scale and some squinting, but it is accurate to less than a needle-width.

The bottom line is that I can't find any meters in my stash that give me a wrong reading.  Something to at least be aware of and IMO, something that should be fixed if possible.

Just did a quick test after reading your post, and indeed my other 2 multimeters (keithley 196 and gossen 22M) read the signal also correctly. The brymen 869s does read it correctly in Vac mode (0.0998Vac was indicated), but no frequency was detected in that mode.

So indeed it seems the Brymens are more the exception than the rule, I did not expect that. 

[Neutrion]

7.5VDC+100mVAC (RMS) 500Hz sine is definitely a tricky signal.  Out of my stash, the only DMMs that could actually provide those three fundamental values were:

Fluke 87V
Fluke 88V
Keysight 1233A
Sanwa PM300 (no mV range however)
Of course Fluke 287


Some notable "failures":

121GW
Brymen BM789
Of course Brymen BM869s

Like the BM869s, the BM789 displayed a garbage reading in mVAC mode.  No matter the amplitude of the AC component, it always displayed approximately 105mV.


The Keysight 1233A surprised me with the fact that while displaying the frequency, it uses the bar graph to display the AC voltage.  It's also very fast and pretty accurate (99.4mV).  It did lose the frequency measurement below about 30mV AC (RMS) (using the above stated signal).

The Fluke 87V & 88V were also quite accurate (100.17mV & 99.86mV).  They lost the frequency measurement below about 20mV AC (RMS).

No surprise I suppose, the Fluke 287 was the most accurate (100.02mV) and made it to 5mV AC (RMS) before dropping out.

The 87V has no mV scale so if we are talking about the V AC scale the Brymens read also correct on the v AC scale.

[Fungus]

Like the BM869s, the BM789 displayed a garbage reading in mVAC mode.  No matter the amplitude of the AC component, it always displayed approximately 105mV.

What did it display in normal AC+DC mode?

[Caliaxy]

The 87V has no mV scale so if we are talking about the V AC scale the Brymens read also correct on the v AC scale.

That's actually the interesting thing we noticed while having this discussion in the BM789 thread.

Fluke 87v has a DC mV position on the dial (with a single 600mV range). There is no AC mV position on the rotary dial per se, but there is a 600mV AC range on the AC V position of the rotary dial. I suspect they did this precisely to avoid the issue under discussion here without adding an extra position on the dial (the alternative would have been having a common AC/DC mV position on the rotary dial and switch between AC and DC with a soft touch button).

[bdunham7]

The 87V has no mV scale so if we are talking about the V AC scale the Brymens read also correct on the v AC scale.

It actually does, in fact IIRC in Hi-res mode it has a 200.00mVAC range, so 10uV resolution. 

[Caliaxy]

A quick fix for the OP (other than a capacitor) would be to use a lab power supply to subtract the offset. Set the output of the power supply to the offset voltage (7.5V), connect (V-) to the GND of the DUT and use BM689S to measure the AC voltage between the measuring point and (V+) on the power supply (as opposed to GND or V-).

[Fungus]

The 87V has no mV scale so if we are talking about the V AC scale the Brymens read also correct on the v AC scale.

It actually does, in fact IIRC in Hi-res mode it has a 200.00mVAC range, so 10uV resolution.

So the OP can simply use AC range (or AC+DC range) with no problems?

[Neutrion]

The 87V has no mV scale so if we are talking about the V AC scale the Brymens read also correct on the v AC scale.

It actually does, in fact IIRC in Hi-res mode it has a 200.00mVAC range, so 10uV resolution.

Maybe this one goes to the  87V but it would be still interesting to compare the 87Vs 200,00mVAC range(0,7%+1d) with the 789s 6,0000VAC range   0,5%+30d in an actual test.

Did by the way anyone tested this scale of the 87V with a DC bias, and for overrange indication?

[bdunham7]

Did by the way anyone tested this scale of the 87V with a DC bias, and for overrange indication?

IIRC yes, but it is AC coupled so there are no surprises there and not a memorable event.

You mentioned in another thread that this issue was 'common' and that even many Flukes exhibit it.  I can't comment on any that I don't have to test, but I'm not seeing that at all.  What might be observed is that some of the Flukes, for example my F116, will appear at first to have an error when you apply a DC bias to an AC signal in the mVAC range.  This actually takes effect even at lower signal levels and is not due to overloading.  What these meters with DC-coupled mVAC ranges are doing is correctly displaying the TRMS AC+DC value, not just the AC component.  That isn't clear just by looking at the meter, so demerits points for that.  However, keeping that in mind, I still haven't managed to get any of them to display a 'wrong' reading.  It is correct or OL.

[J-R]

To clarify, I only tested some of my handheld DMMs and didn't provide the full list of failures.  I have a few more from Amprobe, UEi, etc. that I didn't mention.

There are plenty of handheld DMMs that can get away with showing some/most readings.  There are also plenty of tricks to work around the "issue".

But I consider it a fail if it can't correctly display all three of the signal properties, AC, DC & frequency.  This also goes for bench DMMs that don't support frequency.

I gave the Sanwa a pass but the others get a hard fail for not being able to show the frequency, even despite the fact they can display 100mVAC correctly in VAC mode.

[bdunham7]

But I consider it a fail if it can't correctly display all three of the signal properties, AC, DC & frequency.  This also goes for bench DMMs that don't support frequency.

I gave the Sanwa a pass but the others get a hard fail for not being able to show the frequency, even despite the fact they can display 100mVAC correctly in VAC mode.

I don't see the point of 'failing' a device for lacking a specific feature that it doesn't purport to have.

[J-R]

But I consider it a fail if it can't correctly display all three of the signal properties, AC, DC & frequency.  This also goes for bench DMMs that don't support frequency.

I gave the Sanwa a pass but the others get a hard fail for not being able to show the frequency, even despite the fact they can display 100mVAC correctly in VAC mode.

I don't see the point of 'failing' a device for lacking a specific feature that it doesn't purport to have.

I was primarily referring to the handheld DMMs that DO claim to have that feature.  The bench DMMs that can't do frequency I am considering a fail for the OPs original request.

[bdunham7]

I was primarily referring to the handheld DMMs that DO claim to have that feature.  The bench DMMs that can't do frequency I am considering a fail for the OPs original request.

OK.  If you take the OP's stated needs to be AC-coupled mVAC, readout in dBm, 20,000 counts or more, frequency readout at ~100mVAC or lower and then you read in handheld and probably CAT-rated, that knocks out a lot of products.  If he doesn't need handheld and CAT III/IV ratings, then most modern bench meters will likely work. 

The only inexpensive (relatively) meter I can think of that may meet the OP's specific requirements in this case is the manual-ranging BK Precision 391A.  All the AC ranges are AC coupled, it has a 200mVAC range, the frequency range stated sensitivity is 50mV_rms and it is 20,000 counts. 

[Brumby]

A quick fix for the OP (other than a capacitor) would be to use a lab power supply to subtract the offset. Set the output of the power supply to the offset voltage (7.5V), connect (V-) to the GND of the DUT and use BM689S to measure the AC voltage between the measuring point and (V+) on the power supply (as opposed to GND or V-).

This idea also crossed my mind - but it does introduce some other concerns.

Not having any practical expertise in this area, here are my observations:

The PSU would, obviously, be floating, reasonably stable and have (ideally) zero ripple.  It wouldn't need to be precise - just close enough to the 7.5V offset that it brings the input to the DMM's ADC within range.

How practical it would be is another question.  Having extra leads floating around increase both mechanical and EMI risks.  There is then the time taken to set up such a rig - with the risk of errors doing so - plus the expense of purchasing another piece of gear.  Also, is there somewhere you could place such an item in the production floor environment?  Not to mention the fact that, if you are going to have some calibration history requirements, this PSU would need to be certified (as I understand it).


This sounds to me like a classic case of "Buy once, cry once".  As others have said I, too, would recommend the "bite the bullet" direction.  The pain will only exist at the initial point of expenditure - but if you go down the "cheaper" path, it will flare up repeatedly - costing you time, money and reputation - until you get to the point of wanting to rip your hair out.  In desperation, you eventually cave and fork out the funds to do it right then a month later you marvel at how smoothly things are running and kick yourself for not having done this in the first place.

This applies to ANY situation: Get the right tool for the job.  I daresay we ALL have had experience here - whether working on electronics, your car or any other task.

[joeqsmith]

And for 61E drifting problem, watch the following video from Joe.

As to the 61E comment. My issue with Uni-T is not temperature drift. It's time drift. Take a Ut-61E and wait 6 months and it will be way out of calibration, unacceptably so. That doesn't help me trust Uni-T.  Even the cheap Anegs and Harbor Freight meters hold calibration. Not so much Uni-T.

I continue to monitor the long term drift after making those modifications.   It appears it has been almost a year since the I last checked.

Pulled out the UT61E today to check it it's long term drift.  After six years of collecting data (not six months), I'm not seeing anything concerning.   Obviously I am not taking this data in a controlled environment but after compensating the reference, I doubt the bit the office changes plays into it too much.   

[joeqsmith]

I still haven't had anyone confirm this is really a 869 design error. It seems strange that it is since other meters handle this fine.

There was a long discussion previously on this exact subject, I don't remember which thread. ...

There's been several what I consider fringe topics like this.  My favorite so far was measuring a high value resistor right at the switch point.  Lots of drama on that one.  I attempted to replicate it with our 120VAC mains.  It took several twists of the leads around the lamp cord but I was able to replicate it.   Mixed AC/DC waveforms are a common problem for many meters I have looked at. 

I think this is the last time it came up:
https://www.eevblog.com/forum/testgear/brymen-bm789/msg3696892/#msg3696892

For the UNI-T UT181A fan boys.  What kind of meter reads zero volts when it's obvious there is a signal applied?  More than you may think!
https://www.eevblog.com/forum/testgear/brymen-bm789/msg3689206/#msg3689206 

Is the claim that the Fluke 189 is without a fringe case?   If so, seems like a challenge I could take on...     

[bdunham7]

There's been several what I consider fringe topics like this...

Is the claim that the Fluke 189 is without a fringe case?   If so, seems like a challenge I could take on...   

Well, it is apparently a real use-case and not an unrealistic expectation.  The problem is fairly simple though--lack of AC coupling along with an external block not being an adequate fix.  We sort of expect AC ranges to work with DC bias, probably because it is easy enough to accomplish--and has been widely implemented.  However, if you take a DC reading with a large AC bias, most meters won't handle that very well beyond a certain point and we just accept that as the way things are.

I have not been able to find any unexpected fringe case for the F189. 

[Fungus]

Is the claim that the Fluke 189 is without a fringe case?   If so, seems like a challenge I could take on...   

Or any other Fluke.

[joeqsmith]

There's been several what I consider fringe topics like this...

Is the claim that the Fluke 189 is without a fringe case?   If so, seems like a challenge I could take on...   

Well, it is apparently a real use-case and not an unrealistic expectation.  The problem is fairly simple though--lack of AC coupling along with an external block not being an adequate fix.  We sort of expect AC ranges to work with DC bias, probably because it is easy enough to accomplish--and has been widely implemented.  However, if you take a DC reading with a large AC bias, most meters won't handle that very well beyond a certain point and we just accept that as the way things are.

I have not been able to find any unexpected fringe case for the F189.

All fringe cases are real use-cases.  Everyone who finds one needs it fixed.  Every meter with a fringe case is garbage.   Other threads covered it. 

[Neutrion]

Did by the way anyone tested this scale of the 87V with a DC bias, and for overrange indication?

IIRC yes, but it is AC coupled so there are no surprises there and not a memorable event.

You mentioned in another thread that this issue was 'common' and that even many Flukes exhibit it.  I can't comment on any that I don't have to test, but I'm not seeing that at all.  What might be observed is that some of the Flukes, for example my F116, will appear at first to have an error when you apply a DC bias to an AC signal in the mVAC range.  This actually takes effect even at lower signal levels and is not due to overloading.  What these meters with DC-coupled mVAC ranges are doing is correctly displaying the TRMS AC+DC value, not just the AC component.  That isn't clear just by looking at the meter, so demerits points for that.  However, keeping that in mind, I still haven't managed to get any of them to display a 'wrong' reading.  It is correct or OL.

In the BM789 topic, what I think you are pointing to,(and the other older topic which I linked in) it was mostly about AC+DCmV capable meters  which above a certain point of AC or DC bias, or overloading would read entirely wrong values. Mostly in case of overload.
That is what we were discussing 10 pages long last time. Unfortunately not even te experts from here were able to come up with an easy to implement Idea, how it could be solved. I suppose from the discussion there, that possibly some extra parallel sampling would be needed only to be able to flag the the overload situation. Or maybe some really smart analogue circuit?
To not to talk about reliable ACmv reading with unlimited DC bias on the AC+DCmv scale which seems also like a challenging issue.


And yes I remembered wrong, the 87V had it's wrong reading problem in Joe's test(bm789 video) on the DC scale with AC "bias", not the other way around.

[joeqsmith]

Is the claim that the Fluke 189 is without a fringe case?   If so, seems like a challenge I could take on...   

Or any other Fluke.

You mean like this:
https://www.eevblog.com/forum/testgear/brymen-bm789/msg3688948/#msg3688948

That thread had a lot of info for those wanting to read it but obviously it's much easier, maybe more fun to repeat it.   


Interesting if you read:
https://www.eevblog.com/forum/testgear/brymen-bm789/msg3696619/#msg3696619
and
https://www.eevblog.com/forum/testgear/brymen-bm789/msg3696679/#msg3696679

Go back a few pages:
https://www.eevblog.com/forum/testgear/brymen-bm789/msg3689695/#msg3689695
Specifically:

But then again, I like that low frequency response that I can only get with DC coupling.....  Ah the choices we have. 

Fun with fringe cases.

[joeqsmith]

And yes I remembered wrong, the 87V had it's wrong reading problem in Joe's test(bm789 video) on the DC scale with AC "bias", not the other way around.

Of course I went after the gold standard of the industry.   

[Neutrion]

Maybe it's the colour.   However these days maybe even shining gold would be possible! On the other hand I have to admit that I like the yellow colour of the Fluke too, pity that Brymen chose the red. But it has some positive meaning in buddhist tradition if I remember right.