Brymen BM789

[joeqsmith]

Thanks for posting.

Not sure off the top of my head how not having the low voltage clamp would ever work in the mVAC mode with a robust handheld meter.   Maybe measure the current through the meter with the mVAC mode selected and look for the knee where the clamp starts to conduct.    Most of handhelds I have are DC coupled and could easily be measured with my bench supply. 

Still, I think we now have our first documented case and I wouldn't be surprised if there were not others.  Recently I had made a video where I used a cheap VNA to make some PDN measurements.   These are normally made with the unit turned off as well as powered up.   The VNA may not like getting exposed to the PDNs voltages, so wideband blocks are added.   You need to be very careful when bringing up the voltages and well as when turning them off.   No big deal for a $50 VNA.  Wouldn't want to tell my boss I damage a quarter million dollar VNA this way.   

[joeqsmith]

10nF with a 2kohm source at a kV.... Interesting...

Doesn't seem like it could damage much ... but your grill starter has killed a few meters that have some input protection on them (high speed clamps after a protective resistor).

Your talking about products that I would expect to be designed to handle some basic transients.  Now I'll certainly be the first to say that not all handheld meters have the same level of robustness.  And yes, there are some down right crap ones that the stupid grill starter with it's what, 5A peak waveform has damaged.     

Here we are talking about the meter becoming the source of the transient and your electronics may not be able to handle the transient it could produce.  I wouldn't be at all surprised if its happened a fair amount and people may not have have understood the cause.    Hard to say.   

Could the free meter from HF be damaged with a Fluke 189, charged to 1.5kV??   
***
Talk about your fringe case! 

[Fungus]

Could the free meter from HF be damaged with a Fluke 189, charged to 1.5kV??   
***
Talk about your fringe case! 

It would be interesting to see if any manufacturers have thought about this in their designs. Is this a place where the mighty Fluke can differentiate themselves by having an internal bleeder resister where other manufacturers don't? Will it be the other way around...?

If you've got the volts and probes at hand then it's a simple job to charge up a meter then discharge it and look at the waveform.

[bdunham7]

I think you'd have to start looking at some schematics to see if this is even a reasonable possibility.  For example, with the (original) F87, if you leave the meter in the AC range you might get, at the very most, a 0.01uF capacitor discharging through about a 10M resistor in parallel with a ~1pF capacitor.  If you switch to any other range, the capacitor is discharged instantly.  I think bench meters with a lower frequency limit, larger blocking caps and a 1M AC input impedance would be more likely to deliver such a jolt, if that is even possible.

[joeqsmith]

Could the free meter from HF be damaged with a Fluke 189, charged to 1.5kV??   
***
Talk about your fringe case! 

It would be interesting to see if any manufacturers have thought about this in their designs. Is this a place where the mighty Fluke can differentiate themselves by having an internal bleeder resister where other manufacturers don't? Will it be the other way around...?

If you've got the volts and probes at hand then it's a simple job to charge up a meter then discharge it and look at the waveform.

Bleed and high impedance don't mix.  Maybe change your habits to discharge the meter.  All these years and still new tricks to learn.      

Using a sting of LEDs to look at the waveform as well as conducting the finger test...

[AndrewBCN]

Thank you for yet another great video, Joe.

My takeaway from this thread, apart from the information that some DMMs use a blocking capacitor in their ACmV input circuit and others don't (in other words, some are AC-coupled and others are DC-coupled in their ACmV range): remember to always switch on my brains before I switch on the DMM to make a measurement. 

[bdunham7]

Bleed and high impedance don't mix.  Maybe change your habits to discharge the meter.  All these years and still new tricks to learn.      

Using a sting of LEDs to look at the waveform as well as conducting the finger test..

Interestingly I was easily able to replicate this and found that even if you switch to DC, the capacitor retains its charge.  If, for example, you go to measure 500V but the meter is in mVAC, if you switch to DC and measure your 500VDC, if you then remove the test leads and switch back to mVAC, you get the jolt.

[floobydust]

If the cap is on the high-side of the divider chain, it charged up to the DC component and the next action you take can cause damage. Moving the probe to a different circuit node, the cap discharges into it and any semi is usually killed due to the voltage.
Moving the probe to GND or a low resistance point to GND discharges the cap into the test equipment front-end. There are many RMS-converters like AD636 etc. in the graveyard as this kills their input, even though the blocking cap is post-divider string. There is no protection at that node.

With a scope, switching from AC to DC coupling shorts the blocking cap and the scope's switch makes a loud ka bang and arc. Tektronix includes a small series resistor.
The cap seems to be 27nF-220nF depending on the low freq. response of the instrument.
My practice is to touch the probe to a 1MEG resistor to GND and wait for the cap to discharge before proceeding.

[Fungus]

Cool!

I bet even Dave didn't know this. 

Using a sting of LEDs to look at the waveform

Is that the correct collective noun for LEDs?

[joeqsmith]

Bleed and high impedance don't mix.  Maybe change your habits to discharge the meter.  All these years and still new tricks to learn.      

Using a sting of LEDs to look at the waveform as well as conducting the finger test..

Interestingly I was easily able to replicate this and found that even if you switch to DC, the capacitor retains its charge.  If, for example, you go to measure 500V but the meter is in mVAC, if you switch to DC and measure your 500VDC, if you then remove the test leads and switch back to mVAC, you get the jolt.

I'm not surprised that the cap would be isolated when the function is changed.    It's too bad the manufactures no longer supply schematics.   

[kawal]

Joe
Good video - Didn't realize you can  get a  meter to damage your equipment or give you a shock.  Very informative. 

Kawal

[Fungus]

Good video - Didn't realize you can  get a  meter to damage your equipment or give you a shock.

Don't worry, I don't think anybody else did either.

(and we're all paranoid now...?)

[joeqsmith]

Let's summarize (again):
1. The OP found a way to "fool" his shiny new Brymen BM789 when set to the DC-coupled high-resolution ACmV range. This basically involves saturating the input circuitry such that the DMM cannot measure the (small) AC component because of a (large) DC bias. Under these conditions the DMM neither detects an overload condition nor provides a correct measurement of the (small) AC component.
2. joeqsmith has demonstrated that a number of different DMMs from different brands can be "fooled" using a similar method.

I think the lesson here, if there is one, is that using a DMM  (any model from any brand) to get correct measurements always involves a minimum of thinking - as 2N3055 reminded us. As well as spending a few seconds to read and understand the User Manual.

Edit: Sorry 2N3055, we seem to have posted our summaries almost simultaneously. I defer to yours which is much better than mine.

Edit2: I can think of a number of reasons why Brymen did not include a series capacitor in the ACmV range front end, and similarly for other DMMs from other brands. But that would be a matter for (reasoned) discussion in another thread, if you ever want to. I for one would be very interested in your opinion.

Thank you for the kind words.
I would not mind if you started that discussion, but in my opinion it would necessitate at least partial R.E of front end and switching circuit. Otherwise we would all have many ideas, but no confirmation. I guess Joe did quite some work on that already..
It might even be that meter could be bodged with a capacitor as an afterthought. But I would be very reluctant to do that to my meter and also to publish it on Internet. I have same opinion on that as Joe has about his "hardenings" of various meters. Maybe that is a Pandora box better be left unopened.
Best,

I have made my rules about making comments on my YT channel and have filters help the signal to noise ratio.  The rules are of course in the very first post:
https://www.eevblog.com/forum/testgear/hear-kitty-kitty-kitty-nope-not-that-kind-of-cat/

Once in a while I still get a few gems from the kiddies.   Attached is a recent post that relates to your comment.  I have removed their name as they run a small channel and see no reason to advertise for them.

They are correct that people continue to ask about my specific mods, even though the rules are clear.  They are also correct that I will normally provide a totally dick response.    I'm amazed that after so many years anyone would expect a different outcome.   

I like the part about having millions of viewers (love that drama), the reality is we are normally talking about a few thousand.   That's already a far larger group than I would expect for such a niche subject.   
 
While the FAQ covers this subject and even includes a few of the earlier like minded comments, my comment about John D. Rockefeller giving stock advice is a case were an ignorant person like our friend here makes a similar request only to file a lawsuit is something people should read about.   From my perspective, there is also a question of being responsible.  The meters being a safety device should not be tampered with.  YouTube may also frown on a channel promoting unsafe behavior. 

The easiest thing for me to do would be to not make any content, but there's a few of you who seem to enjoy it.  So for now, I plan to continue my unfunded, sparse and low quality content.   

[2N3055]

Let's summarize (again):
1. The OP found a way to "fool" his shiny new Brymen BM789 when set to the DC-coupled high-resolution ACmV range. This basically involves saturating the input circuitry such that the DMM cannot measure the (small) AC component because of a (large) DC bias. Under these conditions the DMM neither detects an overload condition nor provides a correct measurement of the (small) AC component.
2. joeqsmith has demonstrated that a number of different DMMs from different brands can be "fooled" using a similar method.

I think the lesson here, if there is one, is that using a DMM  (any model from any brand) to get correct measurements always involves a minimum of thinking - as 2N3055 reminded us. As well as spending a few seconds to read and understand the User Manual.

Edit: Sorry 2N3055, we seem to have posted our summaries almost simultaneously. I defer to yours which is much better than mine.

Edit2: I can think of a number of reasons why Brymen did not include a series capacitor in the ACmV range front end, and similarly for other DMMs from other brands. But that would be a matter for (reasoned) discussion in another thread, if you ever want to. I for one would be very interested in your opinion.

Thank you for the kind words.
I would not mind if you started that discussion, but in my opinion it would necessitate at least partial R.E of front end and switching circuit. Otherwise we would all have many ideas, but no confirmation. I guess Joe did quite some work on that already..
It might even be that meter could be bodged with a capacitor as an afterthought. But I would be very reluctant to do that to my meter and also to publish it on Internet. I have same opinion on that as Joe has about his "hardenings" of various meters. Maybe that is a Pandora box better be left unopened.
Best,

I have made my rules about making comments on my YT channel and have filters help the signal to noise ratio.  The rules are of course in the very first post:
https://www.eevblog.com/forum/testgear/hear-kitty-kitty-kitty-nope-not-that-kind-of-cat/

Once in a while I still get a few gems from the kiddies.   Attached is a recent post that relates to your comment.  I have removed their name as they run a small channel and see no reason to advertise for them.

They are correct that people continue to ask about my specific mods, even though the rules are clear.  They are also correct that I will normally provide a totally dick response.    I'm amazed that after so many years anyone would expect a different outcome.   

I like the part about having millions of viewers (love that drama), the reality is we are normally talking about a few thousand.   That's already a far larger group than I would expect for such a niche subject.   
 
While the FAQ covers this subject and even includes a few of the earlier like minded comments, my comment about John D. Rockefeller giving stock advice is a case were an ignorant person like our friend here makes a similar request only to file a lawsuit is something people should read about.   From my perspective, there is also a question of being responsible.  The meters being a safety device should not be tampered with.  YouTube may also frown on a channel promoting unsafe behavior. 

The easiest thing for me to do would be to not make any content, but there's a few of you who seem to enjoy it.  So for now, I plan to continue my unfunded, sparse and low quality content.

Oh, THAT guy...  hehehe well he doesn't like you, wonder why ...

Problem is, and nobody wants to hear this, is that you gave enough information to point out in a right direction what you think problem is.. For people that understand what is going on. They have enough info to experiment themselves.
Those that don't know enough to pick up the hints are probably not qualified enough to poke around inside meter.
Even if you publish the schematic and clear photos what needs to be done, will they follow the instructions to the letter?
Will they use same MOVs, PTCs, surge rated resistors, same Corona Dope, or Teflon inserts?
Add to that hordes of people that measure on the secondary of MOT, and then are baffled why the meter is smoking..
I wouldn't want to bear that responsibility. It's called social responsibility.
If people want to walk outside the line, it is their problem and responsibility. I don't want to be enabler.

We had same type of discussion some time ago, when people started promoting that people shouldn't buy cheap scopes from Rigol and Siglent because they figured out how 1000 EDU series from Keysight can be "hacked" and "you get real scope"..
There was heated discussion about it when it was pointed out that "hacking" Keysight scope involved physical removal of components and soldering on the motherboard that not only kills your warranty, but may lead to damaging brand new scope, that still cost more than some very nice scopes from competition. They were making it look simple, but it is not for beginners and people that don't have proper experience and equipment.
We still have few topics here on forum, where people are trying to repair Keysight scopes damaged in attempt to hack it.
Draw your own conclusions..

[AndrewBCN]

  It seems I got quoted? 

Actually 2N3055 this is what you wrote:

...
- Meters SHOULD AC couple in AC only measurements so you could measure 20mV AC riding on top of 100V
- For some reason Brymen doesn't do that on AC mV range. That is NOT good way to do it.
...

I believe Brymen does not AC couple the ACmV range because they don't want to have to deal with the kind of accident that Joe has clearly demonstrated can happen if you add a blocking capacitor in your DMM input circuitry, as Fluke did with the 189. It's simpler, safer and cheaper for a DMM manufacturer to just add a note in the User Manual that the meter will not provide a correct reading when the offset + signal > 1000mV, and make that range DC coupled, as Brymen does with most or all of their DMMs. And of course any user that knows a little bit what she/he is doing is going to use an external blocking capacitor when needed and be careful about discharging said capacitor after completing the measurement.

This is what I offered to discuss with you and others in a separate thread, but it seems you somehow misinterpreted my post as me wanting "to open Pandora's box". 

[2N3055]

It seems I got quoted? 

Actually 2N3055 this is what you wrote:

...
- Meters SHOULD AC couple in AC only measurements so you could measure 20mV AC riding on top of 100V
- For some reason Brymen doesn't do that on AC mV range. That is NOT good way to do it.
...

I believe Brymen does not AC couple the ACmV range because they don't want to have to deal with the kind of accident that Joe has clearly demonstrated can happen if you add a blocking capacitor in your DMM input circuitry, as Fluke did with the 189. It's simpler, safer and cheaper for a DMM manufacturer to just add a note in the User Manual that the meter will not provide a correct reading when the offset + signal > 1000mV, and make that range DC coupled, as Brymen does with most or all of their DMMs. And of course any user that knows a little bit what she/he is doing is going to use an external blocking capacitor when needed and be careful about discharging said capacitor after completing the measurement.

This is what I offered to discuss with you and others in a separate thread, but it seems you somehow misinterpreted my post as me wanting "to open Pandora's box". 

Yeah it's kind of misunderstanding or better say my prediction where that discussion would end up, not by you and me but maybe somebody else.

We cannot know whether Brymen did not include it because it was easy and lazy thing to do or there was intention and reason for it. Designer of the meter would know that, but we can only speculate. And in order to speculate on a level of "educated guess" instead of "some punters are crapshooting ideas" a reverse engineering could be done on meter inputs, and then you could see if decision to DC couple it had some simplification benefits (like less contacts on switch used, or simplification of layout, or whatever). It might have been because switch routing was simpler and provided more isolation distance or whatever.. Now I'm just throwing random ideas and that is useless..

Once reverse engineering of schematics was done, and alternative version that include capacitor was posited, there is a thin line between that and people that start cutting the board and adding capacitor into what is now a improvisation that outside looks  like a CAT IV meter and inside is a death trap..
I don't want to be complicit in something like that, and I personally don't care why. It is what it is, I use external cap if I need it, or lately just use MTX3293 if I need AC coupled mV. That one has the capacitor. But, mostly it is not an issue. I already spent too much time on it, and that only because I realized it was topic that was unknown to many, to my surprise. So it was good deed to spread the word, we also spoke a bit about good measurement practice (also good topic) and that's it.

If I ever decide to design my own meter design (hypothetically speaking, no intention to do so..) I would then think about it in more detail. I have no interest for it now.

[bdunham7]

I believe Brymen does not AC couple the ACmV range because they don't want to have to deal with the kind of accident that Joe has clearly demonstrated can happen if you add a blocking capacitor in your DMM input circuitry, as Fluke did with the 189. It's simpler, safer and cheaper for a DMM manufacturer to just add a note in the User Manual that the meter will not provide a correct reading when the offset + signal > 1000mV, and make that range DC coupled, as Brymen does with most or all of their DMMs. And of course any user that knows a little bit what she/he is doing is going to use an external blocking capacitor when needed and be careful about discharging said capacitor after completing the measurement.

I usually try to use logic and facts in these discussions, rather than characterization and hyperbole.  However, in your case I'm going to make an exception.

Your entire statement is an embarrassing fanboy apologist fantasy that exceeds, or a least meets, the highest levels of marketing spin that I've ever encountered.  The simplest explanation is that Brymen added mVAC to a range that would otherwise only be mVDC and mVDC+AC because they could do it at virtually zero additional expense using only software.

[bdunham7]

I would hate to think I left a few of you with the idea the having a meter show zero volts while in it's ACmV mode is unique to Brymen's new BM789,  here is another one for the UNI-T fan boys.   

Of course, switch it over to tri-display mode and we get a feel for what's really going on. 

I don't have any of the meters like the 789 to test, otherwise I would check myself--but has anyone confirmed that the DC overload issue is strictly confined to the dedicated mVAC range?  So the meters will all show 1VAC with 900VDC bias on the 6.0000V (or whatever is appropriate) range?

[2N3055]

I would hate to think I left a few of you with the idea the having a meter show zero volts while in it's ACmV mode is unique to Brymen's new BM789,  here is another one for the UNI-T fan boys.   

Of course, switch it over to tri-display mode and we get a feel for what's really going on. 

I don't have any of the meters like the 789 to test, otherwise I would check myself--but has anyone confirmed that the DC overload issue is strictly confined to the dedicated mVAC range?  So the meters will all show 1VAC with 900VDC bias on the 6.0000V (or whatever is appropriate) range?

I keep repeating. Yes it will. AC volts is AC coupled.

[floobydust]

For the Brymen and other's using the same chip, you wouldn't have a DC blocking cap. It can't go after the divider chain because well, it's inside the DMM IC. Outside the IC means it's on the hot side side of the divider chain, so a large expensive HV film cap plus another rotary switch contact is required, which costs more.

The real issue is the DMM IC saturating if DC is present and it's not reliably indicated by the firmware.  The op-amp in and out are available to the ADC, if the firmware looked.

[bdunham7]

For the Brymen and other's using the same chip, you wouldn't have a DC blocking cap. It can't go after the divider chain because well, it's inside the DMM IC. Outside the IC means it's on the hot side side of the divider chain, so a large expensive HV film cap plus another rotary switch contact is required, which costs more.

I keep repeating. Yes it will. AC volts is AC coupled.

I'm sorry if this seems repetitive or if I've missed something, but I'm not clear on what 'AC coupled' may mean in this case, I haven't seen anyone actually apply the test that I mentioned and in teardowns, I have not seen the HV blocking cap that I would expect.  So perhaps I'm not clear on some aspect of how the meter works.  I'm willing to believe 2N3055 is correct and I'm wrong, but I would like to know where I'm wrong.

First, I'll assume that like most of these, the AC input is actually not totally AC coupled and will in fact have a measured 10-11M input resistance when in the AC mode and measured with another meter.  If that isn't true, then that's where I'm wrong.

Second, I'll assume that indeed my observation that there is no 1kV HV blocking cap is correct and there is not in fact one lurking in there somewhere.  Again, if there is one, that's where I'm wrong.

Third, if the first two are true, then I'll assume that the AC coupling is added somewhere after the voltage divider and if that is a capacitor, it is a smaller, lower voltage variety which is only exposed to the voltage after the divider...but those dividers don't really work in such a way as to make it easy to shove a capacitor in there, do they?   So what does that look like?   And on the 6.0000V range, what would such a blocking capacitor be exposed to if the input voltage was 5VAC + 600VDC?

Just to be thorough, I tested my F116 with 5VAC + 100 to 600VDC.  Each step up resulted in a brief 'OL', then back to a 5VAC reading.  I took it apart and there appears to be a rectangular capacitor C1 near the switch that could plausibly be a blocking capacitor, although it is a bit smaller than others I've seen. 

[Caliaxy]

I don't have any of the meters like the 789 to test, otherwise I would check myself--but has anyone confirmed that the DC overload issue is strictly confined to the dedicated mVAC range?  So the meters will all show 1VAC with 900VDC bias on the 6.0000V (or whatever is appropriate) range?

I keep repeating. Yes it will. AC volts is AC coupled.

Right – the issue solely occurs on the "mV" position, in AC mode. The rotary dial "mV" position has 6 modes you can circle through by pressing the yellow soft “Select” button:

DC mV
AC mv
AC+DC mV (a combined RMS value)
frequency
duty cycle
dBm*

Interestingly, the “AC mV” position is the only one in which the meter can’t figure it’s overloaded (in the presence of a DC bias) and displays a false value. “DC mV” and “AC+DC mV” show, correctly “OL”.

As circling through these 6 modes doesn’t require moving the rotary dial, there is hope that the issue could be, in theory, fixed in software (by “fixing” I mean having the meter display “OL”, as it correctly does in "AC+DC mV" mode**). It won’t help the current owners, because the firmware is not user upgreadable, but still (it will stop people complaining about this issue and bring this thread to an end; not sure about the latter ).

*dBm also doesn’t show “OL” when it should, but that’s just a calculated value based on whatever the meter thinks it measures in AC.
**haven't checked extensively (and won't...), but the AC+DC mV mode seems to take longer to display OL when AC+DC overloaded, and it sometimes displays a first wrong value which is quickly updated to OL, suggesting that it does some extra measurements/calculations. Whatever it does, maybe it could be done in AC mV mode as well?

[floobydust]

Just look at the schematics, and for the big 1,000V film cap at the input to the divider string. Looking at a few:
34401a AC-coupled C301 0.22uF 400V max DC bias.
34410a AC-coupled C400 0.22uF 400V max DC bias.
3478a  AC-coupled C301 0.15uF 630V
Agilent U1273a AC-coupled C30 0.1uF 630V
Fluke 101 AC-coupled 0.01uF 1,000V C1
Fluke 189 AC-coupled 0.01uF 1,000V
Fluke 29/79II, III Fluke 87 old; AC-coupled 0.01uF 1,000V C3
Fluke 87V; AC-coupled 0.022uF 1,000V C1
Fluke 867; AC-coupled 0.1uF 1,000V C31
Fluke 27 AC-coupled 0.022uF 1,000V C3
Fluke 10/11/12 AC-coupled 0.01uF 1,000V C14
Fluke 17B DC-coupled
AN8008 DC-coupled
UT-61e DC-coupled

For the hardware to have a DC-blocking cap, if the CMOS switches and mV AC op-amp and true-RMS converter are all in the DMM IC, then you simply can't AC-couple post-divider without something getting swamped. Like the ANENG AN8008, Brymen 789, 121GW, Fluke 17B, UT61e. These all can get overloaded from the DC.
The issue is if these detect and display it.

Note the true-RMS DSP also loses the DC component, as well as an RMS converter IC i.e. AD637. So this is a second (downstream) location in the chain where the DC component is removed. BUT everything ahead of this can get saturated- unless there's a blocking cap at the input of the divider chain which the more expensive multimeters incorporate.

[joeqsmith]

I would hate to think I left a few of you with the idea the having a meter show zero volts while in it's ACmV mode is unique to Brymen's new BM789,  here is another one for the UNI-T fan boys.   

Of course, switch it over to tri-display mode and we get a feel for what's really going on. 

I don't have any of the meters like the 789 to test, otherwise I would check myself--but has anyone confirmed that the DC overload issue is strictly confined to the dedicated mVAC range?  So the meters will all show 1VAC with 900VDC bias on the 6.0000V (or whatever is appropriate) range?

So we want to overdrive the ACV input with high DC offset?  So we put about 2KVAC into the meter and adjust the offset by a KVDC?   Seems like fun.  I'll sit back and watch this one. 

If you're interested in more details on the 189, I took a quick stab at tracing out this area.   I also made an attempt to look at the current into a short. 

[bdunham7]

Just look at the schematics,

I have or have seen the schematics for the first 8 you listed, but I don't have and don't recall the details of any DC-coupled models.  If you have a 10M input resistor and can live with a 30-40Hz cutoff, the capacitor can be pretty small.

For the hardware to have a DC-blocking cap, if the CMOS switches and mV AC op-amp and true-RMS converter are all in the DMM IC, then you simply can't AC-couple post-divider without something getting swamped. Like the ANENG AN8008, Brymen 789, 121GW, Fluke 17B, UT61e. These all can get overloaded from the DC.
The issue is if these detect and display it.

They don't in mVAC, but I'm wondering if/how they cope with the issue in the higher ACV ranges.  And if they indeed use a pre-divider HV cap, then not incorporating that in front of the mVAC range has to be for other reasons.