Brymen 869s DC offset issue when reading ACmV

[AnalogEngineer]

Howdy,

I recently got a Brymen 869s in the hopes we could find a lower cost alternative for the Fluke 289 / 189 we use on the shop floor.  We need at least 20,000 counts and dBV or dBm reading so the choices are slim since few lower cost non Fluke meters include dB readings.

There appears to be a show stopper bug and I wonder if anyone can confirm or knows a work around solution. I suspect the bug is the same causing this issue, but since that is an old thread I decided to start a new one.
https://www.eevblog.com/forum/testgear/brymen-multimeters-fault/

What it appears is that when using the AC mV reading the meter is not AC coupled. So if there is a DC offset that is more than the maximum DCmV scale the AC mV reading will not work and you get nonsense readings.  This also seems to affect frequency and AC / DC dual readings as well.  In my case I am reading a 100mV signal riding on top of a 7.5VDC offset. My Fluke 289 handles this perfectly for all measurements as does my ancient Keithly and Tektronix models. The Brymen 869 cannot and I have to go to the AC Volt reading (with less accuracy due to small voltage). If I add a >0.22uF capacitor in series then it works but takes a 5-10 seconds to settle down to an accurate reading (due I assume to the meters input impedance). Smaller capacitors induce error in the value and larger increase settling time.

Any ideas? This seems like such a huge fail on the meters part basically making it unusable to me in our application.
Anyone want to recommend a sub $300 meter with >=19,999 counts and dBV or dBm reading capability?

[Fungus]

[mqsaharan]

Anyone want to recommend a sub $300 meter with >=19,999 counts and dBV or dBm reading capability?

I am not sure about the price in your country but may I suggest UT181A. It is another blatant attempt to copy Fluke 287 externally and functionally like some others. Even though its manual does not state measurements in dBv or dBm, its VAC menu does include both. Please check out the following link for Joe Smith's video at 17:50 minutes mark



It has AC coupling cap in it. And I can only hope it will also have AC coupling on mVAC range because usually the meters that have an independent mVAC range are AC coupled in that range unlike the ones that share one dial position for both DC and AC mV function.
I hope some other member who owns one will confirm AC coupling as well as dBV and dBm capability on mVAC range for UT181.

[Fungus]

I am not sure about the price in your country but may I suggest UT181A. It is another blatant attempt to copy Fluke 287 externally and functionally like some others. Even though its manual does not state measurements in dBv or dBm, its VAC menu does include both. Please check out the following link for Joe Smith's video at 17:50 minutes mark

It's a good meter but the the fancy screen means the battery doesn't last very long and you can't use it while it's charging becasue it charges through the input jacks. This can make it very inconvenient to use.

(plus it's a LIPO battery so it won't even hold a charge a few years from now)

joe says the battery lasts about 40 hours in a new meter:

https://www.eevblog.com/forum/testgear/uni-t-ut181a-pictures/msg3327088/#msg3327088

[AnalogEngineer]

I am not sure about the price in your country but may I suggest UT181A.

Thank you. I had not considered that before. I wonder how stable it is.
I have two Uni-T ut61e meters now and they both drift like crazy and don't stay in calibration more than a few months. This is a known complaint with Uni-T. Even my cheap Anengs don't do this.  I wonder if this meter suffers the same issue. I also wonder if the UT61e+ has fixed this. Otherwise it seems to be a great meter. The 61E actually meets all my needs except the dB reading so worth considering.

Someone else mentioned to me the Vici VICHY VC8145  which seems possible.  Alas all this trying other brands is making me a bigger Fluke fanboy. They are all so close, but I keep seeing small details that does make the fluke much nicer. I just can't afford them for every tech at the shop. Sigh.

[mqsaharan]

If you are a hobbyist and like to tinker/fight with your equipment in addition to the project you are working on, try getting 8060A. Don't take me wrong for "tinker/fight" remarks. I got one and even though it took me some time to make it work properly it is all set now and working great. If you can get one with 4xxxxxx or above serial number in original condition, hopefully its electrolytic capacitors have not leaked and you only need to replace them. It is 20k count, TRMS and has dBm and dBV measurement capability. And unlike most DMMs out there it shows dBm and dBV for both AC and DC. dBm and dBV requires a little bit of effort on operators part (check the attached manual pages) but it is a very good quality DMM with low price tag. Only downside for some is its manual ranging. Please check out the following two links by modemhead regarding its capacitor replacement:
http://mrmodemhead.com/blog/fluke-8060a-repair/
http://mrmodemhead.com/blog/ibm-8060aaa-fluke-8060a-refurbish/

For a bench top, VC8145 is not a bad choice, again from a hobbyist perspective. Please take a look at its review by HKJ for more information: https://lygte-info.dk/review/DMMVici VC8145 UK.html

The biggest downside of UT181A is its battery as Fungus has pointed out. To me the other one is its screen. But if you are buying it for bench use you can make it work. As far as the drifting problem goes, I doubt it drifts like 61E. There are quite a few members on this forum who own it and so far no one has reported such a problem with it.

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

[_Wim_]

If I add a >0.22uF capacitor in series then it works but takes a 5-10 seconds to settle down to an accurate reading (due I assume to the meters input impedance). Smaller capacitors induce error in the value and larger increase settling time.

The BM869s has an input impedance of 10MOhm in AC-mode, so using a 0.222µF cap creates a high pass filter with a -3db point of 0.072Hz. This makes sense that this takes around 10 seconds to settle. I do not know what frequency the AC signal is you are trying to measure, but I expect the cap can be much smaller without any significant loss in accuracy.

With this online tool (http://sim.okawa-denshi.jp/en/CRtool.php) you can easily calculate the filter, and the step response will give you an idea about the settling time.

[2N3055]

If I add a >0.22uF capacitor in series then it works but takes a 5-10 seconds to settle down to an accurate reading (due I assume to the meters input impedance). Smaller capacitors induce error in the value and larger increase settling time.

The BM869s has an input impedance of 10MOhm in AC-mode, so using a 0.222µF cap creates a high pass filter with a -3db point of 0.072Hz. This makes sense that this takes around 10 seconds to settle. I do not know what frequency the AC signal is you are trying to measure, but I expect the cap can be much smaller without any significant loss in accuracy.

With this online tool (http://sim.okawa-denshi.jp/en/CRtool.php) you can easily calculate the filter, and the step response will give you an idea about the settling time.

10nF is just fine, unless you think meter can measure RMS of signals less than 5 Hz...
Even 4,7nF would be just OK for 10-15Hz and up.

[David Hess]

My Tektronix bench meters use a 0.1 microfarad AC coupling capacitor before the 10 megohm decade divider.  My Tektronix DMM916 handheld meter solves the problem for the millivolt range by only supporting DC.

Some meters include a warning in their documentation that high DC offsets will cause ranging when making AC measurements because they rely on their decade divider to bring the DC offset of the AC measurement into range.

[Caliaxy]

Anyone want to recommend a sub $300 meter with >=19,999 counts and dBV or dBm reading capability?

A Fluke 187/189 should be cheaper than $300 (in US, at least), if you are OK with buying used. Agilent U1272A should do it as well.

[bdunham7]

10nF is just fine, unless you think meter can measure RMS of signals less than 5 Hz...
Even 4,7nF would be just OK for 10-15Hz and up.

Unfortunately that neglects the input capacitance of the meter.  Any blocking capacitor must be much larger than the typical 100pF or so of input capacitance or you'll have an error.  10nF would cause the reading to be about 1% low and so on. 

[_Wim_]

Unfortunately that neglects the input capacitance of the meter.  Any blocking capacitor must be much larger than the typical 100pF or so of input capacitance or you'll have an error.  10nF would cause the reading to be about 1% low and so on.

10nF is indeed may be a little too low, but this would also speed up the current response time x22 which is probably also not needed. A response time around 1 sec seems acceptable (to me), so a 22nF  cap would achieve this. Taking into account the 80pF input capacitance on the 500mV range, this would create a 0.3% error which does not degrade too much the specified accuracy.   

[_Wim_]

For reference the Fluke 289 specs. This probably also just has a series cap somewhere between 22nF and 100nF (do not now the response time of the Fluke). I should thus be possible to create identical performance with the Brymen 869s


Edit: added the AC+DC specs for the 289, as this is really what the Brymen does.

[AnalogEngineer]

Just a couple points of clarity and thanx to people.
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.
The 0.22 cap was derived by testing. Any smaller and the error was too great for a 500Hz test signal and our calibration procedure.  Any larger was unneeded. No idea why but that is the reality.  I get it shouldn't need to be that big but that's the measurement I got.

I agree a 189 or 187 may be a good choice, but they seem to go for a lot and they are all getting pretty old.
I need something I can reliably get several of for the folks on the manufacturing floor so would rather not putz with old stuff.

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.

[bdunham7]

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 are many other meters that cannot be used to measure ripple on the mVAC range, including some Fluke meters like the 113-117 models, so that can be a bug or a feature or just something that isn't included.  The real issue, IMO, are the cases where you overload the mVAC range with DC but you just get a wrong reading rather than an OL indication of some sort.  There was no consensus IIRC, but IMO that is not an acceptable result from a professional tool.

I need something I can reliably get several of for the folks on the manufacturing floor so would rather not putz with old stuff.

Nobody ever got fired for buying Fluke, right?  The 289 can be fairly criticized--it is expensive, big, cumbersome to use according to some and it eats batteries like popcorn.  But for some uses it seems to be only current production tool I know of that works, other than a bench meter which would have another set of issues.  There may be alternatives that will perform to your expectations, but unless you get lucky, they won't be all that cheap anyway.

[mqsaharan]

I agree a 189 or 187 may be a good choice, but they seem to go for a lot and they are all getting pretty old.
I need something I can reliably get several of for the folks on the manufacturing floor so would rather not putz with old stuff.

If you are using it on a production floor, like bdunham7 said, I would also suggest you to use a quality instrument like 287. But if budget is the issue then you might have to compromise and use a cheaper quality instrument. But you'll have to keep it in check which sometimes is extra headache.
With AC coupled dB measurement requirement, your choices are limited in handheld DMM catagory. Mostly expensive models cover your requirement. If you glance at dB column of the DMM spreadsheet, you'll find many multimeters supporting dB function. But it is difficult to confirm which ones have AC coupled inputs.
https://www.eevblog.com/forum/testgear/multimeter-spreadsheet/

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.

The video I linked in my last post is a no nonsense video regarding drift. In that video you'll learn about the drift of UT61E, BM869s and UT181A along with a few others. If I remember correctly, a few years back when 61E's drifting problem was kind of a hot topic, 61E's drift boils down to temperature cycling. Then again, I don't remember that discussion. I didn't take notes as I don't own one.

[Fungus]

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. z with old stuff.

There was a lso a discussion of how that AC coupling capacitor can be charged up to hundreds of volts then used to zap other electronics. 

Nobody ever got fired for buying Fluke, right?  The 289 can be fairly criticized--it is expensive, big, cumbersome to use according to some and it eats batteries like popcorn.  But for some uses it seems to be only current production tool I know o
f that works, other than a bench meter which would have another set of issues.  There may be alternatives that will perform to your expectations, but unless you get lucky, they won't be all that cheap anyway.

Yep. The Hioki DT4282 can do it, too, but it's not much cheaper than the Fluke. You may just have to bite the bullet.

[_Wim_]

The 0.22 cap was derived by testing. Any smaller and the error was too great for a 500Hz test signal and our calibration procedure.  Any larger was unneeded. No idea why but that is the reality.  I get it shouldn't need to be that big but that's the measurement I got.

What accuracy where you aiming for in the calibration procedure? Can you describe the test you performed? From your first post I would presume a 500Hz 100mVac signal riding on top of 7.5Vdc and then checking the deviation?

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.

I do not think this is a design error they omitted the series cap, but this is related to the fact the BM869s has dual display capability on mVac mode where it can measure Vac and Vdc simultaneously (which many meters can't do). It would have been nice of course to have a dedicated AC-mode (with series cap), as this would allow them to calibrate out the capacitive divider formed by the series and parallel input capacitance. 

Related to settling time, I would expect other meters to also to have a relatively low settling time with your test signal, as otherwise they would have too much error when measuring low frequency AC signals. As stated in a previous post, the Tektronix bench DMM also uses a 100nF series cap, which would only half your settling time, and maybe the 289 uses something similar?

[_Wim_]

The real issue, IMO, are the cases where you overload the mVAC range with DC but you just get a wrong reading rather than an OL indication of some sort.  There was no consensus IIRC, but IMO that is not an acceptable result from a professional tool.

I agree that this could be handled better. I wonder if this could be related to the fact that the OP does not achieve the expected accuracy? Many he could try manual ranging and see if the accuracy improves on a high range for his test signal.

[Fungus]

I wonder if this could be related to the fact that the OP does not achieve the expected accuracy? Many he could try manual ranging and see if the accuracy improves on a high range for his test signal.

A BM869S can easily read a 100mV AC signal with 7.5V DC offset in the normal AC+DC range.

I think the problem is that he wants it in dBm

[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.   

[Martin72]

The 289 can be fairly criticized--it is expensive, big, cumbersome to use according to some and it eats batteries like popcorn.

 

So damn true....

[2N3055]

I wonder if this could be related to the fact that the OP does not achieve the expected accuracy? Many he could try manual ranging and see if the accuracy improves on a high range for his test signal.

A BM869S can easily read a 100mV AC signal with 7.5V DC offset in the normal AC+DC range.

I think the problem is that he wants it in dBm

And it can easily do the dBm too, and at the levels OP asked for.
It won't be accurate down to 0.1 dBm, mind you... It will be inside 0.3-0.4 dBm of reading.

But if you ask me, it this is critical testing, for pretty much the price of F289 + a bit more you can get DMM6500 and do it right...
Precision testing is not a handheld job. Also, you might need some kind of record or printout of "Passed" label....

[bdunham7]

And it can easily do the dBm too, and at the levels OP asked for.
It won't be accurate down to 0.1 dBm, mind you... It will be inside 0.3-0.4 dBm of reading.

But if you ask me, it this is critical testing, for pretty much the price of F289 + a bit more you can get DMM6500 and do it right...
Precision testing is not a handheld job. Also, you might need some kind of record or printout of "Passed" label....

F289 will do that measurement (100mVAC @500Hz, -17.78dBm) 10X more accurately (+/- 0.02dBm) if that's what he needs.  The errors get down to the point where the typical 1M input impedance of a bench meter may affect the results more than the lesser precision of the F289, if the source is actually 600 ohms.

[joeqsmith]

.... But if you ask me, it this is critical testing, ....

...Precision testing is not a handheld job. ...

I would be concerned if I worked for a company who was having problems buying a few new handheld meters let alone trying to save the cost of their production testing.