EEVblog #884 - EEVBlog BM235 Multimeter REPAIR

[Kean]

Just pulled mine apart again, fully this time, after mentions of the PCB mechanical design in this area due to the large holes, and I can see the PCB flex noticeably around L2 & L3 when inserting/removing F2 (the 11A HRC fuse).  The flex seems more pronounced around the nearby capacitors like C25, but the slots under L2 & L3 definitely create a weak area.

I think an extra moulded support and mounting screw for the PCB near the currrent shunt is required.  Currently there is only one support in this area, and it is closer to the other end of F2.  Another option, which wouldn't change the PCB, only the mould, would be to partly enlarge the mounting post near the COM terminal, so as to provide a support lip for the PCB.  That would only help with downwards pressure on fuse insertion, but on fuse removal the matching half of the moulded case should hold the PCB against the upwards pressure.  This assumes the fuses are installed after the PCB is in the case, which is probably not the case at the factory, so extra care may be required there.

In the meantime, I recommend being very careful when removing or inserting the HRC fuse.  As this is likely to be a common fault, it would be nice to know the factory values of L2 & L3 for replacement purposes, so as to keep up with specs/compliance.  I might remove them from mine for characterisation, or maybe that would be a great topic for an EEVblog video.

Because L2 has been substituted with an 0805 ferrite, I reckon it is more likely to survive fracture from PCB flexing than the 1206 L3.  It does seem that the substitution to 0805 for L2 was an afterthought, and it only just makes it onto one pad of the footprint (see attached photo).  So, on this meter at least, a hairline crack in the soldering of L2 seems much more likely than a broken component.

[TheSteve]

As I believe has already been stated I think the ferrite was just fine when reflowed but was subjected to flexing or received an impact blow during final assembly.

[jitter]

Actually, no. Dave's conclusion was a (probably) one off production fault with L3, based on the goo he found under and around that component.

But closer inspection (thanks Kean!) reveals some other issues in that area that may have helped the problem with the inductor to surface. I can imagine if that weak inductor had been in a more rigid area, we would never have known about it. It might have survived for a long time.

For now the more fundamental issues seem to be a weak board around the L2/L3 area, their placement smack in the middle of that area (across slots to make things even worse) and too small size for L2.

Another meter with the same symptoms has surfaced. Hopefully the inspection of that one can shed more light.

[Fungus]

Am I the only person expecting hoping for a picture of Rick Astley on the inside?

[coppice]

Why are there heatshrinks around PTC1 and PTC2 at 10:30? Sorry if it was explained in another video, I have missed it.

Thick heat shrink over PTCs and VDRs is a common practice. If you push these parts too hard they can fail explosively and rip right through a case. Restraining them at source greatl reduces the damage they can do.

[Fungus]

Why are there heatshrinks around PTC1 and PTC2 at 10:30? Sorry if it was explained in another video, I have missed it.

Thick heat shrink over PTCs and VDRs is a common practice. If you push these parts too hard they can fail explosively and rip right through a case. Restraining them at source greatl reduces the damage they can do.

They also make a mess.

Quite a few PTC were turned to dust in the meter robustness testing thread. The heat shrink helps keep all the dust in one place.

[EEVblog]

With Dave selling so many around the world, there is little incentive for Brymen to release this meter in their usual market. So this one might become an EEVBlog exclusive.

It might by default.

I was surprised at the small number of their initial batch.

[Dr.Joe]

Mine looks OK, I opened it up this evening and had a quick look. I don't suppose 2 in over 1000 units sold is such a bad failure rate. I wonder if the second failed unit has exactly the same fault?

As i am the owner of the second faulty unit: yes it did. Just not as bad as the one Dave repaired. My inductor was still intact, just lifted from the lower pad (by about 0.25mm). I did resolder the joint, and the Meter is working again.

I didn't even have to desolder the Input Board.

[retiredcaps]

As i am the owner of the second faulty unit: yes it did. Just not as bad as the one Dave repaired. My inductor was still intact, just lifted from the lower pad (by about 0.25mm).

@Dr. Joe

I wonder if yours is in the same serial number range as the one in Dave's video.  If yes, there may be some other owners who will want to look at their unit especially if you don't have to desolder the input board.

[Dr.Joe]

I can't remember Dave showing the serial number of the one he repaired, but mine is 161140685

[Kleinstein]

As the root cause seems to be the board layout with large form factor SMT parts soldered over a position with high mechanical load (or more accurate strain), the problem will not be limited to a few units. I am afraid all units with this board layout problem have a weak spot there.

I would consider this a point that should be addressed at the next board revision. It might be even a reason to start a new board version.

Some may fail early if the soldering is not the very best quality - failing solder may be the better case as resoldering can make it work again. With good solder joints the inductor itself can break - just like in the unit Dave showed.
 

[QuantumLogic]

I can't remember Dave showing the serial number of the one he repaired, but mine is 161140685

The one Dave repaired was 161140533.

[Flipflop]

The photo that Kean posted shows that those 2 inductors are almost the only thing holding the board together at that point.

I wonder if the BM257 and BM257s are the same in this area?

[Lightages]

The photo that Kean posted shows that those 2 inductors are almost the only thing holding the board together at that point.

I wonder if the BM257 and BM257s are the same in this area?

The BM257(s) does not have components in the same place on the circuit board. I suspect that there won't be any kind of failure in the same way if it does turn out to be a mechanical stress problem on the components in the BM235.

[blacksheeplogic]

It's good to see the honesty with a possible issue in these.

Also, to address a point on one of the threads regarding failure of the mid-tier brand verse a traditional top shelf (Fluke for example). I have had to returned two Fluke meters out of the box. A Fluke 28II  and a Fluke 1507. The Fluke 28II was replaced, and the Fluke 1507 was serviced and returned to me. I can accept that the Fluke 28II failed after factory calibration/shipping, but the Fluke 1507 should not have passed QA as it was a fault with the molding.

[jitter]

As the root cause seems to be the board layout with large form factor SMT parts soldered over a position with high mechanical load (or more accurate strain), the problem will not be limited to a few units. I am afraid all units with this board layout problem have a weak spot there.

Agreed. When looking at the photo Kean took of a board outside the case, I couldn't help but think "what were they thinking". And they did it knowing full well that it might not be that strong, they tried to counter it a bit by not having the slots under L2 and L3 sit in line with eachother, but they are offset.

I would consider this a point that should be addressed at the next board revision. It might be even a reason to start a new board version.

Indeed, especially if dropping the meter, or inserting a fuse leads to problems in that area.

Some may fail early if the soldering is not the very best quality - failing solder may be the better case as resoldering can make it work again. With good solder joints the inductor itself can break - just like in the unit Dave showed.

I think now that maybe Dave's conclusions on the inductor were wrong. He speaks of black goo under the inductor, and indeed when you look at the remains of the inductor at first it looks like black goo. But if you look closer you can see the light shine through a piece of it sticking out from the underside of the inductor (first attachment, it's clear not black). There's also quite clearly residue on the good end cap of the inductor, which I previously called contamination

I think I should stand corrected on my remarks that these are good solder joints. They are still good in the sense that there has formed a meniscus, and the slightly gritty nature of the joint is inherent to some alloys of leadfree solder. But there's also some yellowish residue covering the joint that you can see at 15:35 when Dave scrapes it off the joint with his pointer (second attachment).
I have seen this residue form on the boards we produce as well, but they should not form in such a thick layer under that inductor, but rather puddle around the pads like you can see at 12:28 around the unpopulated transistor pads.

I don't think the solder profile was wrong, but perhaps the solder paste was past its use by date, or something like that. And in the weakest area of the board this will be the first where it will lead to problems. So perhaps the inductor was fine when it came from the tape after all, but it was damaged by strain later on.

[coppice]

Why are there heatshrinks around PTC1 and PTC2 at 10:30? Sorry if it was explained in another video, I have missed it.

Thick heat shrink over PTCs and VDRs is a common practice. If you push these parts too hard they can fail explosively and rip right through a case. Restraining them at source greatl reduces the damage they can do.

They also make a mess.

Quite a few PTC were turned to dust in the meter robustness testing thread. The heat shrink helps keep all the dust in one place.

When the option is sharp objects punching through even tough polycarbonate cases, possibly taking an eye out, its legal liability which drives those heat shrink sleeves. Who cares if the shards are neatly gathered?

[Fungus]

Another option, which wouldn't change the PCB, only the mould

They have mould inside them? Interesting.

What it for? Does it absorb the inevitable crumbs of food that fall inside when you're connecting/disconnecting the leads?

[Len]

Another option, which wouldn't change the PCB, only the mould

They have mould inside them? Interesting.

What it for? Does it absorb the inevitable crumbs of food that fall inside when you're connecting/disconnecting the leads?

Don't get too excited. He wasn't talking about fungus, Fungus.

[Kleinstein]

There may be a good reason to start with relatively small batches - so there might be the option to get an more robust version in the future. The slots below the inductors make things mechanically worse, not better. Still wondering of these slots are really needed, as the inductors are more like for ESD purpose.

I would guess a board revision might be less trouble than changing the mold for the case. Some changes may be possible, but some are very expensive.

[Flipflop]

Still wondering of these slots are really needed, as the inductors are more like for ESD purpose.

I was thinking this also as the input/blast protect (apart from fuses) seems to be on the little board with the input jacks.

I'm a little surprised that Dave didn't pick up on this 'weakness' with this meter as he is usually very thorough doing his meter teardowns.

I was also surprised to see that the current jacks are split in 2 halves to act as switches for the 'beepjacks' feature as in some of the previous meter videos Dave frowned upon jack sockets with splits.

[rch]

Another option, which wouldn't change the PCB, only the mould

They have mould inside them? Interesting.

What it for? Does it absorb the inevitable crumbs of food that fall inside when you're connecting/disconnecting the leads?

'Mould' is the correct UK (?Australian) spelling.   'Mold' is a town in Wales, FWIW.

[EEVblog]

Brymen have investigated it, report attached.

[Kean]

Brymen have investigated it, report attached.

So they're basically saying it was likely damaged during assembly then, most likely during fuse insertion?  I agree.

I was flexing the main PCB with the terminal PCB disconnected, and I see that the red wire would provide support as it is quite stiff.  But if the fuse is inserted before adding the terminal PCB and gaining that extra strength, then I can easily see this happening to 2 in 1000 units.

An excellent response by Brymen in my opinion. 

[rrinker]

 

Clearly they are on top of things. I am very impressed with the quality feel of mine, and it's good to see the company has solid procedures in place to handle and correct small issues that may crop up.