EEVBlog Rebranded Brymen BM235 & Backlight noise research

[idpromnut]

FYI, I did measure the capacitor when I had it off the PCB and mine came out to be a little over 14uF.

Ugh, either your measurement is off (it should be closer to 10uF), the tolerance on that part is horrendous, or it was from a fairly crap batch of capacitors

[Scottjd]

FYI, I did measure the capacitor when I had it off the PCB and mine came out to be a little over 14uF.

Ugh, either your measurement is off (it should be closer to 10uF), the tolerance on that part is horrendous, or it was from a fairly crap batch of capacitors

Ok Idpromnut, you have to stop proving me wrong, lol.
I just checked four ceramic caps and one tantalum. My fluke is constantly reading the ceramics way to high. Why is the Brymen making my fluke look like crap now? Oh, and the old extech also. I guess this is why people say you should own multiple meters.
I need to pull out the manual for my fluke and see what the tolerance is on capatamce readings, and if it's out of spec I guess it goes in for service and a free calibration. I did recently pull the batteries and when I replaced them I received the "clock and date not set" message and it's supposed to hold this information with a super cap. Maybe it sprung a leak?

If you take the average that the fluke is reading the caps high, then that would make the cap in the Brymen about 9.1uF. The other 10uF I found if I ended up replacing the cap also just showed close to 14uF on the fluke. So I would say the cap was a 10uF on the Brymen. By bad, or should I say my flukes bad?

[bson]

Got to thank Brymen for designing a meter with an inherent design fault. So much more fun than if they'd designed a perfect meter without any problems.

Well, clearly they know their audience! 

Can't wait for the next batch so I can have one, too!

[idpromnut]

FYI, I did measure the capacitor when I had it off the PCB and mine came out to be a little over 14uF.

Ugh, either your measurement is off (it should be closer to 10uF), the tolerance on that part is horrendous, or it was from a fairly crap batch of capacitors

Ok Idpromnut, you have to stop proving me wrong, lol.

I didn't prove you wrong; we identified a problem with your measurement set up, and now it's fixed! So the next capacitance measurement(s) you do will be spot-on.  Publishing, review and adjustment in action! 

[Robomeds]

So I have a Greenlee version of this basic meter family (DM-200A).  The Greenlee, in addition to being green, has an amber backlight.  As far as I can tell there is no noise.  Perhaps the lower voltage of the amber LEDs avoids the need for a charge pump.  Not that it maters much but I actually kind of like the amber better. 

[Scottjd]

So I have a Greenlee version of this basic meter family (DM-200A).  The Greenlee, in addition to being green, has an amber backlight.  As far as I can tell there is no noise.  Perhaps the lower voltage of the amber LEDs avoids the need for a charge pump.  Not that it maters much but I actually kind of like the amber better.

Without seeing the schematic of your meter is hard to validate. But some older meters that switched to LED backlighting was done by using a simple resistor to ensure the LED didn't receive to much current and was powered within its limits. This is an old way or running LEDs and usually restricted the time the LEDs could be turned on or powered this way before the LED had a thermal runaway and was damaged. This is also why some older meters used LEDs for backlighting was still restricted to the light being on for 10 to 30 seconds to ensure the LEDs didn't overheat and fail.
This was efficient for short backlight times and worked, but was not optimal from a power consumption point and still waisted energy using the resistor, and wasn't optimal for the lifetime health of the LED.

The BM235 has a backlight that can stay on for 10 minutes at a time because it's using a power efficient LED driver to properly run the LEDs at the optimal rated current and voltage for the LEDs. Using a low energy LED driver is not only more energy efficient but also optimal for the LEDs life time and keep them away from thermal runaway failures without the need to worry about the LEDs being powered by to much current or the need for additional thermal heat sinks.

Although LED drivers have become the best way to run LEDs some drivers must also act as a boost circuit to allow the driver to run the LED even when the power source (AAA Bateries, or 9V battery) becomes low. Part of this boost circuit uses other components such as inductors and capacitors to help maintain and regulate the specified current and voltage to the LEDs despite the batteries continually draining over time by using the meter. An example pointed out on the thread was a UNi-T meter, the manual stated the LEDs light may go out as the batteries drained, but the meter would continue to run properly. This UNi-T meter either uses an older LEDs driver that's not efficient enough for the task or uses the LEDs with a resistor.

Boost circuits are mostly not an efficient way of controlling power. So as LED drivers continue to get better, cheaper and more efficient were some don't need boost circuits and use what's called a charge pump buit into the LED driver. This is what's used in the BM235 allowing the LEDs to be run at the optimal current, voltage and temperature and also allowing the backlight to be on for 10 minutes while not effecting the battery lifetime of the unit. This charge pump required/used an external ceramic capacitor and this was the capacitor that was traced to be the source of the noise.

The explanation on how some ceramic capacitor vibrate/resonate  at and audible frequency that can be heard by the human ear is explained in Dave Jones EEVBlog latest video and some of his older videos. Most ceramic capacitors have become bigger with more layers allowing them to be used for higher voltage applications and because of the additional mass they may still resonate, but not at a frequency that can be heard by the human ear. It's not a common issue these days, or I haven't noticed it that often myself. In fact even the capacitor for the BM235 charge pump built into the LED driver isn't that loud, but some meters like the one I received had just enough room or maybe not the same amount of solder others had to cause the sound to amplify in some way making the noise a lot more noticeable in the case of my meter.

[Robomeds]

Scott,
Our forum friend MrModemhead has some great pictures of the inside of the same meter I've got
http://mrmodemhead.com/blog/gallery/dm200a/
It looks like the main boards have significant differences (not surprising given the differences in features).  The DM-200A backlight is limited to something like 2 minutes.  I've never timed so it might be only 30 seconds or perhaps more like 4-5 minutes.  I just know it's short enough that when I've used it I've often had to turn it back on. 

I do understand the current limiting resistor bit as well as the rest of the discussion.  I'm not at all questioning why the BM235 is making the noise, just noting that a similar meter doesn't but it also doesn't use white LEDs.  Not sure what the forward drop is across the white vs amber LED but I would assume the amber is less. 

[sangvikh]

So I have a Greenlee version of this basic meter family (DM-200A).  The Greenlee, in addition to being green, has an amber backlight.  As far as I can tell there is no noise.  Perhaps the lower voltage of the amber LEDs avoids the need for a charge pump.  Not that it maters much but I actually kind of like the amber better.

Without seeing the schematic of your meter is hard to validate. But some older meters that switched to LED backlighting was done by using a simple resistor to ensure the LED didn't receive to much current and was powered within its limits. This is an old way or running LEDs and usually restricted the time the LEDs could be turned on or powered this way before the LED had a thermal runaway and was damaged. This is also why some older meters used LEDs for backlighting was still restricted to the light being on for 10 to 30 seconds to ensure the LEDs didn't overheat and fail.
This was efficient for short backlight times and worked, but was not optimal from a power consumption point and still waisted energy using the resistor, and wasn't optimal for the lifetime health of the LED.

The BM235 has a backlight that can stay on for 10 minutes at a time because it's using a power efficient LED driver to properly run the LEDs at the optimal rated current and voltage for the LEDs. Using a low energy LED driver is not only more energy efficient but also optimal for the LEDs life time and keep them away from thermal runaway failures without the need to worry about the LEDs being powered by to much current or the need for additional thermal heat sinks.

Although LED drivers have become the best way to run LEDs some drivers must also act as a boost circuit to allow the driver to run the LED even when the power source (AAA Bateries, or 9V battery) becomes low. Part of this boost circuit uses other components such as inductors and capacitors to help maintain and regulate the specified current and voltage to the LEDs despite the batteries continually draining over time by using the meter. An example pointed out on the thread was a UNi-T meter, the manual stated the LEDs light may go out as the batteries drained, but the meter would continue to run properly. This UNi-T meter either uses an older LEDs driver that's not efficient enough for the task or uses the LEDs with a resistor.

Boost circuits are mostly not an efficient way of controlling power. So as LED drivers continue to get better, cheaper and more efficient were some don't need boost circuits and use what's called a charge pump buit into the LED driver. This is what's used in the BM235 allowing the LEDs to be run at the optimal current, voltage and temperature and also allowing the backlight to be on for 10 minutes while not effecting the battery lifetime of the unit. This charge pump required/used an external ceramic capacitor and this was the capacitor that was traced to be the source of the noise.

The explanation on how some ceramic capacitor vibrate/resonate  at and audible frequency that can be heard by the human ear is explained in Dave Jones EEVBlog latest video and some of his older videos. Most ceramic capacitors have become bigger with more layers allowing them to be used for higher voltage applications and because of the additional mass they may still resonate, but not at a frequency that can be heard by the human ear. It's not a common issue these days, or I haven't noticed it that often myself. In fact even the capacitor for the BM235 charge pump built into the LED driver isn't that loud, but some meters like the one I received had just enough room or maybe not the same amount of solder others had to cause the sound to amplify in some way making the noise a lot more noticeable in the case of my meter.

That sounds wierd! Are you sure about the thermal runaway issue?
I know resistors in series with LEDs have been used in a lot of circuits for decades. And the LED's have been running for decades too.

[Scottjd]

Scott,
Our forum friend MrModemhead has some great pictures of the inside of the same meter I've got
http://mrmodemhead.com/blog/gallery/dm200a/
It looks like the main boards have significant differences (not surprising given the differences in features).  The DM-200A backlight is limited to something like 2 minutes.  I've never timed so it might be only 30 seconds or perhaps more like 4-5 minutes.  I just know it's short enough that when I've used it I've often had to turn it back on. 

I do understand the current limiting resistor bit as well as the rest of the discussion.  I'm not at all questioning why the BM235 is making the noise, just noting that a similar meter doesn't but it also doesn't use white LEDs.  Not sure what the forward drop is across the white vs amber LED but I would assume the amber is less.

I'm not sure either if their is a different forward drop accosts the white or amber. I've never been asked to build an amber color LED flash light 😀

I looked at his pictures and it is different, after a closer look i saw that the main PCB on that green one is marked as a 251 model with the 25X riser board,for,the 251,252,255,257 meters. According to the manual only the 257 and 255  models have a backlight that turns off in 32 seconds. So your is probably a rebranded 255 or 257 Brymen meter.

[Scottjd]

So I have a Greenlee version of this basic meter family (DM-200A).  The Greenlee, in addition to being green, has an amber backlight.  As far as I can tell there is no noise.  Perhaps the lower voltage of the amber LEDs avoids the need for a charge pump.  Not that it maters much but I actually kind of like the amber better.

That sounds wierd! Are you sure about the thermal runaway issue?
I know resistors in series with LEDs have been used in a lot of circuits for decades. And the LED's have been running for decades too.

Well I didn't design any of these meters so I can't say for sure if the short times are for partial thermal or over all battery life deciding factors, or both.
I wasn't saying that the LEDs on the meter would have immediate failures from thermal runaway, but the fact that when their ran at a higher current and the power is not being regulated by a driver this causes the LED to heat up out of spec and shortens the life span.
Since most LEDs have a claim to fame with 50,000+ hours average life span so people have been running them slightly out of spec for years. Why pay for a driver and extra IC draining power when LEDs in series and a resistor still give the LEDs a long run time, even if it cut the lifespan in half of the specs runtime this usually is still a very long life time.

Even I have built flashlights that when they are on high output the LED throws more light and I know it's out of spec. Some people just want brighter, and as long as I manage the extra heat correctly to not effect the Li-Ion battery putting the user in danger at that point, then I still build it. I even bought a few (before I stated building them) that deliberately push the LED for the extra light. And the flashlights are using a driver.

It's the factor of more light on the flashlight for 24,000 hours runtime vs 50,000 runtime. It's been a common trade off in flashlights, so with backlights this is a feature that is probably used the least on a DMM (unless you work in the dark) and a shorter life span is still a long time. 50,000 hours is little over 6 years if the LED was constantly on. But now days the drivers have become more efficient, smaller, less heat, and cheaper it so I expect to see light being in longer times and the leads should run their full expected life time.

I think with the progression speed of technology advancements most wouldn't ever see the LED burn out from its normal use in its rated lifetime, and most likely something else on on the device will fail fist or people would just want to upgrade to a new versions of that device.

[Robomeds]

Scott,
Our forum friend MrModemhead has some great pictures of the inside of the same meter I've got
http://mrmodemhead.com/blog/gallery/dm200a/
It looks like the main boards have significant differences (not surprising given the differences in features).  The DM-200A backlight is limited to something like 2 minutes.  I've never timed so it might be only 30 seconds or perhaps more like 4-5 minutes.  I just know it's short enough that when I've used it I've often had to turn it back on. 

I do understand the current limiting resistor bit as well as the rest of the discussion.  I'm not at all questioning why the BM235 is making the noise, just noting that a similar meter doesn't but it also doesn't use white LEDs.  Not sure what the forward drop is across the white vs amber LED but I would assume the amber is less.

I'm not sure either if their is a different forward drop accosts the white or amber. I've never been asked to build an amber color LED flash light 😀

I looked at his pictures and it is different, after a closer look i saw that the main PCB on that green one is marked as a 251 model with the 25X riser board,for,the 251,252,255,257 meters. According to the manual only the 257 and 255  models have a backlight that turns off in 32 seconds. So your is probably a rebranded 255 or 257 Brymen meter.

I've looked into it before and figured it didn't have an exact Brymen counterpart.  When looking at the 251, 252, 255 and 257 none are really an exact match.  The 252 and 251 have no backlight yet they are the only ones with the 4 button (vs 6) layout of the 200A.  It's possible the 200A is a 251 with the backlight enabled.  Since these things are microprocessor driven I suspect all Brymen has to do is make a general PCB, populate the parts needed for a feature set, then program the correct code in the micro controller. 

The table here shows amber LEDs with a 2.0-2.1 forward drop while blue is 2.5-3.7.  I'm assuming blue for white since what I understand is most white is blue + phosphorus coating that responds to the blue wavelengths.  Anyway, given the system runs on ~3V an amber LED should be find with a current limiting resistor while a blue/white may need a charge pump. 

http://www.radio-electronics.com/info/data/semicond/leds-light-emitting-diodes/characteristics.php

[retiredcaps]

The DM-200A backlight is limited to something like 2 minutes.  I've never timed so it might be only 30 seconds or perhaps more like 4-5 minutes.  I just know it's short enough that when I've used it I've often had to turn it back on. 

I measured the DM-200A at 33 seconds, but manual states 30 seconds as per screenshot below.

[sangvikh]

Well I didn't design any of these meters so I can't say for sure if the short times are for partial thermal or over all battery life deciding factors, or both.
I wasn't saying that the LEDs on the meter would have immediate failures from thermal runaway, but the fact that when their ran at a higher current and the power is not being regulated by a driver this causes the LED to heat up out of spec and shortens the life span.
Since most LEDs have a claim to fame with 50,000+ hours average life span so people have been running them slightly out of spec for years. Why pay for a driver and extra IC draining power when LEDs in series and a resistor still give the LEDs a long run time, even if it cut the lifespan in half of the specs runtime this usually is still a very long life time.

Even I have built flashlights that when they are on high output the LED throws more light and I know it's out of spec. Some people just want brighter, and as long as I manage the extra heat correctly to not effect the Li-Ion battery putting the user in danger at that point, then I still build it. I even bought a few (before I stated building them) that deliberately push the LED for the extra light. And the flashlights are using a driver.

It's the factor of more light on the flashlight for 24,000 hours runtime vs 50,000 runtime. It's been a common trade off in flashlights, so with backlights this is a feature that is probably used the least on a DMM (unless you work in the dark) and a shorter life span is still a long time. 50,000 hours is little over 6 years if the LED was constantly on. But now days the drivers have become more efficient, smaller, less heat, and cheaper it so I expect to see light being in longer times and the leads should run their full expected life time.

I think with the progression speed of technology advancements most wouldn't ever see the LED burn out from its normal use in its rated lifetime, and most likely something else on on the device will fail fist or people would just want to upgrade to a new versions of that device.

That may be the case with high power leds running at their design limits.
There should be no problem selecting a resistor that limits the current to be within spec taking into account reduced Vforward with temperature etc.

Never really gave this problem a lot of thought. You learn something every day, especially on EEVblog forums

[hammy]

I had success with this replacement cap for C44:

Murata, GRM31CR71E106KA12L
10µF, 10%, 25V, 1206, X7R

Cheers
hammy

[fanOfeeDIY]

I had success with this replacement cap for C44:

Murata, GRM31CR71E106KA12L
10µF, 10%, 25V, 1206, X7R

Cheers
hammy

Nice!

It is also has information here too.
http://www.murata.com/en-global/products/capacitor/mlcc/solution/naki?intcid5=com_xxx_xxx_cmn_hd_xxx

[Scottjd]

I had success with this replacement cap for C44:

Murata, GRM31CR71E106KA12L
10µF, 10%, 25V, 1206, X7R

Cheers
hammy

Nice!

It is also has information here too.
http://www.murata.com/en-global/products/capacitor/mlcc/solution/naki?intcid5=com_xxx_xxx_cmn_hd_xxx

Interesting, putting the caps on little metal high rises  they confirmed what had was theorizing. Thanks for that link to that site and a different explanation.

To be fair making a thicker capacitor and increasing the thickness of the ceramic layers would have  strengthened the capacitor also causing the same effect and results. I would think also adding more solder/stronger solder, or bigger solder pads for more solder would also prevent the contraction and expansion as I mentioned in my video. So I guess both explanations are correct between Dave and I, we just went different ways of looking at things when approaching the same issue for troubleshooting.

But the name of the game in electronics seems to be how can we make things smaller, I wonder how small they can make those metal high rise pads?

If we didn't have people look at things different ways then new ways and inventions would never be made, and things would never advance. It takes all kinds of people to make things better or different. That's why we brain storm and sometimes multiple ways are created to fix the same issue, it doesn't make either way wrong, they can both be correct.

I miss the engineering brain storming, it's not about who's wrong or correct. It's about different views and new ideas. Maybe that's why I like this site and reading long threads, hearing different views from different people. One of the many things that makes EEVBlog and exiting platform to be part of.

SIDE Note and take a vote:
What do you think and what's your opinion. I wonder how Brymen decided to correct the buzz going forward.
1. Making bigger solder pads and keeping the same capacitor they have in stock?
2. Changing out the stock of capacitor and using a higher rated thicker capacitor?
3. Just adding more solder?
4. Using a different solder with more strength characteristics?

I'm sure they probably fixed it or have taken these comments into consideration for the next DMM design. After all everyone's comment did a little of the work for them. But I also know from experience with big companies that the lower cost solution usually wins.
What way do you think Brymen choose to fix the intermittent issue with some of these meters that are louder then others going forward?

[Flipflop]

Have Brymen fixed this issue on the latest batch of meters?

[thisguy]

Have Brymen fixed this issue on the latest batch of meters?

As of today there are multiple reports that the sound is gone, Dave says it's fixed in the latest shipment, although one person says he can still hear it quietly and high pitched if the meter is held directly at his ear. (Mine makes no sound even when held directly to my ear.)

[EEVblog]

Have Brymen fixed this issue on the latest batch of meters?

Yes. The capacitor was changed from 10uF to 0.1uF. Sound if now gone. Although it seems one person here has reported a very feint sound.
Obviously I can't open hundreds of meters and check them all.
Only the first batch of 300 was affected (those in the coloured boxes).

[deadlylover]

Yep those guys are probably referring to me. 

I have to hold the meter right up to my ear like a mobile phone to hear it, it's inaudible 6"/15cm away from my ear. Even up close it's veeery faint and it sounds like it's in the ~13kHz region.

It's a complete non issue IMO, it gets a thumbs up from me and keep in mind I'm in my mid 20s so my hearing is probably a bit better than most here.

[Flipflop]

I went to collect my meter today had a quick play with it and I can hear a very slight noise when I hold it close to my ear but, nothing that is going to bother me.

Overall very pleased with it.

Worked out cheaper (including import taxes) than buying a BM257 from TME.

[Jayeson]

I got mine this week and it is certainly audible to around 30cms or so. The level varies - not sure if it is the level from the meter changing or the orientation or just my ear tuning it out.

[Edit: Removed off-topic comments on the display - I did not see that there was another thread on this]

[imidis]

Mine only makes a very slight noise have to hold up to my ear to hear it, I agree some viewing angles aren't the best, however I have found that most of the ways I use it so far are at the better viewing angle. Quite happy with it all in all though, so far.