Hi,
Edit: See feedback from Brymen at the bottom of this post.
I recently purchased a BM867 based on the EEVBlog video overview, comments on the forum and some other online reviews. It's a fantastic meter and I get a lot of joy using it.
However, I may have discovered a bug on it and would like to know if others could replicate it or if it is a problem specific to my unit. It has to do with Crest detect Mode in the AMP range.
The test is simple:
- short the common and 10A input jacks with a test lead
- switch to the A scale
- now press CREST button
- does yours also measure at least 0.65A as a max? (without having any current flow through it)
Can owners of the BM869 perhaps try the same test?
The above problem does not allow me to take current in-rush measurements!
Update from Brymen:Thanks for using our products and feeding back comment to us.
This is not its bug. Actually it is its nature out of its design. BM869 Crest mode design is targeted to capture the peak value short to 0.8ms in duration. To hit this target, very high speed sample rate is in need. In such case, its noise rejection can not be good as normal measurement mode naturally. That is the reason why we add an additional tolerance for CREST mode to manual specs as followings.
Crest mode (Instantaneous Peak Hold)
Resolution: 5000 counts
Accuracy: Specified accuracy +- 100 digits for changes > 0.8ms in duration
--deckert
I wonder if the Brymen BM869 also has this issue. I'm thinking about buying one as a second meter.
I suspect that the BM867 and BM869 may use the same main board. I recall seeing a tear down picture of the BM867 and it looked like there was a temperature stencil on the pcb for the mode selector knob where the BM869 has this mode setting. Temperature measurement is not a feature on the BM867 as I understand.
So maybe BM869 owners can also check for this issue.
The test is simple:
- short the common and 10A input jacks with a test lead
- switch to the A scale
- now press CREST button
- does yours also measure at least 0.65A as a max? (without having any current flow through it)
Can owners of the BM869 perhaps try the same test?
Test BM869:
- Short the common and 10A input jacks with a test lead.
- Switch to the A scale.
- Now press CREST button. -> The same, I get 00.83 Amp. (DC), and without test lead, 003.6 mA (AC).
Why short the inputs? You would shorts the V input to common to ensure zero input voltage, but it you want to absolutely ensure zero current, leave the inputs open. What do you see with the test repeated with open inputs?
(That said, I am struggling to imagine why shorting the input would cause this...)
Why short the inputs? You would shorts the V input to common to ensure zero input voltage, but it you want to absolutely ensure zero current, leave the inputs open. What do you see with the test repeated with open inputs?
(That said, I am struggling to imagine why shorting the input would cause this...)
With open probes: I get 00.79 Amp. (DC), and without test lead, 003.6 mA (AC).
Same thing on my BM857... With shorted/open inputs.
With open probes: I get 00.79 Amp. (DC), and without test lead, 003.6 mA (AC).
Yup, I get the exact same values with open probes on A and mA scale. Seems that the crest sampler might be starting up too soon, without first waiting for the A/D to settle (or something).
Now that I've confirmed this is not only my unit, I've contacted Brymen with a full explanation and step-by-step instructions to replicate it. But I wonder if this can be fixed at all - does the meter support firmware upgrades?
--deckert
My BM869 shows 0.65 A DC with COM and A shorted with a test lead, and 3.4 mA DC open circuit. It also shows 3.4 mA AC open circuit.
same thing on my 869. .94A shorted and ~9mA open for crests.
This bug is confirmed on my BM869. I will write an email to Brymen about this.
P.S.
If there is no fix for this, then it will certainly take Brymen down a notch in reputation. This does not happen on my BM257 BTW, at least not to this extent. It does pick up some noise though.
P.P.S. Interestingly enough this also happens on my UEi DM397. Perhaps it is a particularly difficult thing to design correctly.
Confirmed on BM857.
Open: 7.4mA
Shorted: 0.74A
Yup, I get the exact same values with open probes on A and mA scale. Seems that the crest sampler might be starting up too soon, without first waiting for the A/D to settle (or something).
Wouldn't it be more likely they use precision rectification and a cap?
Although it would be better if they used the ADC, that could be fixed in firmware.
I will contact with Brymen tomorrow, but I don't see how we can upgrade the firmware. Maybe through the (infrared) serial port?
Wouldn't it be more likely they use precision rectification and a cap?
Agreed - I don't know how DMMs work nowadays, but the ICL7106 that I studied years ago uses a dual-slope converter.
Edit: maybe the crest-mode uses an A/D converter, since it samples at 1000 times per second (and reduces resolution at the same time).
Although it would be better if they used the ADC, that could be fixed in firmware.
Well, I think this can easily be solved in firmware - just reset to zero after a short delay. The problem is that I don't know if one can upgrade the firmware without sending it back to the factory.
--deckert
This bug is confirmed on my BM869. I will write an email to Brymen about this.
Aye - the more people that write to them about this, the better the chances are that they will start to react to it.
P.P.S. Interestingly enough this also happens on my UEi DM397. Perhaps it is a particularly difficult thing to design correctly.
It really shouldn't be that difficult. As a work-around, I used a 0.2 ohm resistor and measured the voltage spike over the resistor using Crest mode on the voltage scale. That works like a charm.
--deckert
Oh yes there are work arounds, but we should not need to do this. Also if your peak is above the baseline erroneous reading then you still get the correct reading. I wonder how many other meters with a peak detect or "crest mode" have the same problem.
Oh yes there are work arounds, but we should not need to do this. Also if your peak is above the baseline erroneous reading then you still get the correct reading. I wonder how many other meters with a peak detect or "crest mode" have the same problem.
What I'm wondering is if they will do the same as OWON, i.e. fuck you, not my problem.
Just did the test on my BM869, didn't get the same result.
- Shorted COM and A
- Turned on to A scale, reading is about 0.0001A DC with 0.000A AC
- Pressed crest: OL (reading is DC only) on both MAX and MIN.
Hold on folks, I am getting very conflicting behaviour, It is looking like a PEBKAC situation to me. Testing a BM869 I initially could NOT confirm any failure, my unit was reading around 0.4 mA on the amp and mili-Amp range and then all of a sudden I got it to read .8 then .9 Amp on MAX and Min only but it is very inconsistent, as if it was sampling noise (on a large loop antenna). I could not duplicate on my BM857 at all but haven't played with it as much. When on AC amps the display takes a while to stabilize and wind down to negligible numbers as if a high impedance input stage is discharging.
edit:
Ok I had more time to play around and I can replicate now consistently. Only on the Amp range with a short present A<-->com. Open socket works fine and the mu-Amp range also works as expected.
BM857a reads .33 to .45 amps plus minus on crest max,min
BM869 reads .62 then .81 max, .94 min on crest
AC +DC crest is a ridiculous 6.021 amps max and 4.130 min
averaging seems to work fine on all ranges. I guess it confirms a multimeter doesn't make a good O'scope substitute. The crest feature is a bit of a dubious and gimmicky thing.
2nd Edit: more experimentation
I can get the BM869 to mess up by only plugging a lead into the 10 Amp socket and leaving COM open. InpErr not always detected if I then shift to mu-amp range. It appears that the lead detect circuit on the 10Amp socket interacting with crest measurement.
I can get the BM869 to mess up by only plugging a lead into the 10 Amp socket and leaving COM open. InpErr not always detected if I then shift to mu-amp range. It appears that the lead detect circuit on the 10Amp socket interacting with crest measurement.
I get the problem always, no matter what configuration I use. I found this issue while using it to measure actual DC in-rush current to my circuit. To test this, simply put it in series with any DC circuit, switch the circuit off and press CREST mode. It
should go to zero, then measure in-rush as you switch on the circuit.
Mine consistently measures 0.65A the moment I press the CREST button, sometimes going to 0.85A on a second "beep". This happens open-circuit or closed. My jack-detection works 100% of the time though.
By the way, seems you can check the firmware (or model number) by pressing and holding the Delta-button, then turning it on. Mine reads "869-P9". I wonder what that number represents.
--deckert
Mine also says 869-P9
Pressing and holding the Delta-button, while turning it on, mine reads "869-P9" too.
Although some multimeters can measuring inrush current, few have that capability advertised, as it can be a hit and miss affair. The best instrument for the job would be a clampmeter, with a special triggering circuit implemented.
The other way to measure that short duration current would be with a storage oscilloscope, using a current probe or a current shunt.
As far as I can tell, Brymen doesn't list inrush current measurement as a feature for their multimeters.
Brymen does list the "Crest" function as being intended to be used in the current modes. This is a failure of the intended use, or at the very least a limitation.
Anyone have any news from Brymen?