Author Topic: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR  (Read 22679 times)

0 Members and 1 Guest are viewing this topic.

Offline EEVblog

  • Administrator
  • *****
  • Posts: 29943
  • Country: au
    • EEVblog
EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« on: May 28, 2016, 08:11:48 am »
Dave unboxes, tests, inspects and repairs the first returned faulty EEVBlog BM235 Multimeter.
What will the be the fault?

 

Offline Barny

  • Frequent Contributor
  • **
  • Posts: 285
  • Country: at
  • I'm from Austria, not Australia ;)
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #1 on: May 28, 2016, 08:39:45 am »
Finally the repair-curse is broken.

Excelent video.
Thanks for showing.
« Last Edit: May 28, 2016, 09:42:55 am by Barny »
 

Offline Wytnucls

  • Super Contributor
  • ***
  • Posts: 2836
  • Country: be
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #2 on: May 28, 2016, 09:43:48 am »
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. :)
 

Offline alper.y

  • Contributor
  • Posts: 13
  • Country: tr
    • Personal
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #3 on: May 28, 2016, 09:56:39 am »
Why are there heatshrinks around PTC1 and PTC2 at 10:30? Sorry if it was explained in another video, I have missed it.
 

Offline daqq

  • Super Contributor
  • ***
  • Posts: 1591
  • Country: sk
    • My site
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #4 on: May 28, 2016, 10:52:03 am »
You can get that from stress from some minor board bending. Note the big shunt 'resistor' near to it - this was hand soldered probably into the board and the cooling, thermal expansion etc. could cause long term stress to small SMD ceramic components.

See:

http://www.kemet.com/Lists/TechnicalArticles/Attachments/51/Technical%20Overview%20of%20Flex%20Mitigation%20Solutions.pdf
Believe it or not, pointy haired people do exist!
+++Divide By Cucumber Error. Please Reinstall Universe And Reboot +++
 

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: de
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #5 on: May 28, 2016, 11:37:49 am »
Solder wetting does not look very well on the other inductor, either.
Reflow temperature may be too low, or components solder contacts have quality issue.

The non broken side of the inductor also came off too easily, so bad solder wetting is very probable.

The case seems to be molded, so correct dry pack handling would be necessary.

If the package was left open too long, the component may have sucked in water, which cracks the case during reflow process.

In either way, high probability of two kinfd of series problem.

Frank
 

Offline bitwelder

  • Frequent Contributor
  • **
  • Posts: 821
  • Country: fi
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #6 on: May 28, 2016, 11:40:59 am »
Just curious: if a single SMD component on the tape-and-reel is a little 'wonky', would an average pick & place machine be able to detect the anomaly and discard it?
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 6335
  • Country: de
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #7 on: May 28, 2016, 12:20:14 pm »
Seeing the board, I am not surprised to see that part mechanical broken. Bending of the board gives a lot of stress to that component, as the board is really weak in that range. The connection to the input protection daughter board could give some stress to this area, e.g. when soldering that thick red ground wire, or just vibrations during transport.

So I would consider that one a weak spot in the design. So L3 might fail on other meters as well.

It's just poor placement of the inductor with a slit underneath at an already mechanical weak spot. 
 

Offline jitter

  • Frequent Contributor
  • **
  • Posts: 804
  • Country: nl
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #8 on: May 28, 2016, 12:36:00 pm »
Solder wetting does not look very well on the other inductor, either.
Reflow temperature may be too low, or components solder contacts have quality issue.

The non broken side of the inductor also came off too easily, so bad solder wetting is very probable.

To me the wetting looks fine because the so called solder meniscus looks exactly like it should be. Here's an example of how the solder meniscus should look, and the one in the Brymen meter looks fine too for Pb-free solder. (attachment, upper joint, lower joint may have been contaminated).



Bad solder wetting looks like this:



Quote
The case seems to be molded, so correct dry pack handling would be necessary.

If the package was left open too long, the component may have sucked in water, which cracks the case during reflow process.

In either way, high probability of two kinfd of series problem.

If that were the case, then the whole series of multimeters that was produced from the same lot of inductors would have shown problems. Assuming this happened, then this might be the first and more will follow eventually. But I doubt it and agree with Dave's asessment that this might be a one off bad inductor.

I have come across SMD components that were knocked off by carelessness. Sometimes the end caps remain in the solder, but often, and despite correct wetting, you're left with just solder. Especially SMD aluminium caps are prone to that. There's a thin intermetallic layer between the solder and the pads and the solder and the component leads. It is this layer that may be responsible for the tearing of an otherwise correct solder joint.
In this particular case, I' assuming that vibrations from transport may have broken the already weak inductor and then shaken loose the other pin as well on this intermetallic layer. Judging by the attachment, the goo from the inductor may have contaminated the solder without affecting wetting.

To me it looks to be a molding error. The goo under the inductor is probably not flux residue from the solder process. I think it's reasonable to assume that the inductor is molded from two part stuff (epoxy?) and that on this one the mixing didn't happen correctly.

Edit: looking at the attachment again, I can't help but think that the left inductor is too small to fit correctly on the pads. The end caps barely meet the pads! I think that according to the IPC610 standards, that's a no no.
« Last Edit: May 28, 2016, 12:53:35 pm by jitter »
 

Offline Kean

  • Supporter
  • ****
  • Posts: 916
  • Country: au
  • Embedded systems & IT consultant
    • Kean Electronics
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #9 on: May 28, 2016, 12:51:02 pm »
L2 seems to have a crazed pattern on it, so rather than a bad part, I suspect L2 & L3 both suffered a physical impact during assembly.
Dave, if you desolder L2, I wouldn't be surprised if it fell apart.
 

Offline jitter

  • Frequent Contributor
  • **
  • Posts: 804
  • Country: nl
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #10 on: May 28, 2016, 01:00:04 pm »
L2 seems to have a crazed pattern on it, so rather than a bad part, I suspect L2 & L3 both suffered a physical impact during assembly.
Dave, if you desolder L2, I wouldn't be surprised if it fell apart.

Now that you mention it...yeah, I think you're right. Attachment with a better view on L2, what are those dark streaks?
Would be interesting to see if that residue is also under that inductor... (and how it fits those pads).

« Last Edit: May 28, 2016, 01:05:20 pm by jitter »
 

Offline Kean

  • Supporter
  • ****
  • Posts: 916
  • Country: au
  • Embedded systems & IT consultant
    • Kean Electronics
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #11 on: May 28, 2016, 01:05:19 pm »
Now that you mention it...yeah, I think you're right. Attachment with a better view on L2, what are those dark streaks?

I'm just looking at L2 on one of my BM235's under a crappy magnifier (I'm working at the home bench, not the office) and it does appear to have some markings or pattern on it.  But in the video it looked like it had a chip out of it.
 

Offline ilanko

  • Contributor
  • Posts: 29
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #12 on: May 28, 2016, 01:45:52 pm »
Is that puppy is for sale ?
 

Offline retiredcaps

  • Super Contributor
  • ***
  • Posts: 3334
  • Country: ca
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #13 on: May 28, 2016, 03:21:57 pm »
Related thread on BM235 repair from a different customer ...

https://www.eevblog.com/forum/testgear/bm235-defective/
 

Offline craigh

  • Contributor
  • Posts: 10
  • Country: ca
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #14 on: May 28, 2016, 04:22:56 pm »
Hey Dave,

Sorry to hear about your ACL.  I'm a ski guide (take people into the backcountry, not ski fields) as well as an electronics engineer.  I blew my ACL a few years ago skiing.  Since it was a work related injury, I was fast-tracked into surgery. Recovery back to the level that I could ski for work took me a year and loads of physiotherapy (my age - 50+ - didn't help with a speedy recovery).  All good now, five years post-op and my knee works really well (thinking of having the other knee done  :)).

The best advice I got from my fellow ski guides who injured their knees as well (it's a common injury) was to work out as much as you can to improve the strength of the injured leg before surgery.  Makes recovery much quicker.

Cheers mate and good luck,

Craig.
 

Offline gardner

  • Regular Contributor
  • *
  • Posts: 124
  • Country: ca
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #15 on: May 28, 2016, 04:40:09 pm »
It seems that L2 and L3 bridge an incredibly narrow and flimsy section of the board.  It's narrow to begin with and with the routed gaps underneath is barely there.  With the large HRC fuse clip nearby, this part of the board should be expected to handle a lot of stress/force when the fuse is installed or pulled out.
--- Gardner
 

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: de
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #16 on: May 28, 2016, 04:45:27 pm »
Just curious: if a single SMD component on the tape-and-reel is a little 'wonky', would an average pick & place machine be able to detect the anomaly and discard it?

The pick n place machine itself usually does / can not detect t&r or component errors.
It just places the component.
Due to the surface tension of the fluid solder, most placement errors were corrected by self-alignment.

Later, residual errors like this one were detected by AOI.

Frank
 

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: de
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #17 on: May 28, 2016, 05:32:12 pm »
Solder wetting does not look very well on the other inductor, either.
Reflow temperature may be too low, or components solder contacts have quality issue.

The non broken side of the inductor also came off too easily, so bad solder wetting is very probable.

To me the wetting looks fine because the so called solder meniscus looks exactly like it should be. Here's an example of how the solder meniscus should look, and the one in the Brymen meter looks fine too for Pb-free solder. (attachment, upper joint, lower joint may have been contaminated).



Bad solder wetting looks like this:



Quote
The case seems to be molded, so correct dry pack handling would be necessary.

If the package was left open too long, the component may have sucked in water, which cracks the case during reflow process.

In either way, high probability of two kinfd of series problem.

If that were the case, then the whole series of multimeters that was produced from the same lot of inductors would have shown problems. Assuming this happened, then this might be the first and more will follow eventually. But I doubt it and agree with Dave's asessment that this might be a one off bad inductor.

I have come across SMD components that were knocked off by carelessness. Sometimes the end caps remain in the solder, but often, and despite correct wetting, you're left with just solder. Especially SMD aluminium caps are prone to that. There's a thin intermetallic layer between the solder and the pads and the solder and the component leads. It is this layer that may be responsible for the tearing of an otherwise correct solder joint.
In this particular case, I' assuming that vibrations from transport may have broken the already weak inductor and then shaken loose the other pin as well on this intermetallic layer. Judging by the attachment, the goo from the inductor may have contaminated the solder without affecting wetting.

To me it looks to be a molding error. The goo under the inductor is probably not flux residue from the solder process. I think it's reasonable to assume that the inductor is molded from two part stuff (epoxy?) and that on this one the mixing didn't happen correctly.

Edit: looking at the attachment again, I can't help but think that the left inductor is too small to fit correctly on the pads. The end caps barely meet the pads! I think that according to the IPC610 standards, that's a no no.

Exactly, the solder meniscus SHOULD look like in your GOOD example, but to me, it does not look like that at all, actually.


The video is not so good, so I tried an enlarged view with contrast improvement, and all the right sides of the components, where the light's shining on, all look really crappy and irregular.
Especially, the right side of the other inductor looks like your BAD case photo.

Maybe, a better photo with good illumination may show different, but here, the solder did not build a meniscus at all.

You cannot ignore, how easily Dave could bend away the inductor on the left side, and if you look closely, you still can see the rectangular imprint of the component inside the solder blob.. so it really hasn't been wetted correctly.. that's the only possible explanation.



And yes indeed, that can really be a series problem @Brymen, with their reflow soldering and maybe with dry pack handling (probably not a component problem)

1/1000 is 1000ppm already, that's much too much, as undetected soldering errors (especially when using AOI) must be in the sub ppm range, nowadays.. believe me, I got a lot of experience from automotive mass production.

Frank

PS: your own photo, showing both inductors side-by-side shows even much better, how bad both are soldered, you can clearly see the lack of the meniscus, and the black spots indicate bad solder junction.. should shine completely.



Here's a more contrasty version.. all junctions look crappy.
Maybe, that area of the PCB was in a thermal shadow, i.e. had seen lower temperatures than other parts.
« Last Edit: May 28, 2016, 06:05:16 pm by Dr. Frank »
 

Offline jitter

  • Frequent Contributor
  • **
  • Posts: 804
  • Country: nl
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #18 on: May 28, 2016, 06:04:14 pm »
I'm sorry to disagree. Those solder joints look perfectly acceptable for leadfree solder.
The existence of that shape meniscus means wetting did occur as it should have. If it hadn't, it would not have taken that shape. I agree that the joints look a bit iffy, maybe gritty, but that's normal for certain leadfree alloys.

The inductor was already shaken loose when Dave touched it for the very first time. And as I wrote, the goo on the inductor probably contaminated that joint.

To only mystery left, to me that is, are the black streaks on L2.
« Last Edit: May 28, 2016, 06:10:28 pm by jitter »
 

Online Dr. Frank

  • Super Contributor
  • ***
  • Posts: 1716
  • Country: de
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #19 on: May 28, 2016, 06:10:20 pm »

To only mystery left, to me that is, are the black streaks on L2.

Sign of bad soldering?
 

Offline jitter

  • Frequent Contributor
  • **
  • Posts: 804
  • Country: nl
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #20 on: May 28, 2016, 06:12:49 pm »
Bad soldering shouldn't affect the body of a component. It also doesn't look like overheating. I think I agree with Kean on a mishap. Perhaps something was dropped on the board, maybe someone slipped with a screwdriver.
If it's mechanical stress as a result of the design of the board, I guess we'll be hearing about this problem a lot more in the future.

I hope Dave will remove L2 as well and see if it falls apart and if there's also goo under that one.

Below: the right joint is what I expect to see on the boards we produce using the leadfree solder as SAC is the alloy we use. Looks bad, right? Well, it is a good joint. Lead containing joint on the left, the sight we are (were) used to, but which must not be used as a benchmark when judging SAC alloy joints.
« Last Edit: May 29, 2016, 04:52:16 am by jitter »
 

Offline station240

  • Supporter
  • ****
  • Posts: 857
  • Country: au
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #21 on: May 28, 2016, 08:10:30 pm »
Perhaps the crack is the result of the heat during reflow soldering, that is thermal shock.
As the component was broken then it moved before the solder cooled down.
 

Offline imidis

  • Frequent Contributor
  • **
  • Posts: 427
  • Country: ca
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #22 on: May 28, 2016, 09:14:27 pm »
Yeah, he should probably spot check a few to see. Were these meters from the current batch? Definitely will check out mine when it gets here.
Gone for good
 

Offline German_EE

  • Super Contributor
  • ***
  • Posts: 2284
  • Country: de
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #23 on: May 28, 2016, 09:57:08 pm »
Spotted it on the video about two seconds before Dave did, although when I saw the inductor at an angle I thought that it might be making contact at only one end. I reckon that this unit should be auctioned off as a 'special' as it's been worked on by the man himself.
Should you find yourself in a chronically leaking boat, energy devoted to changing vessels is likely to be more productive than energy devoted to patching leaks.

Warren Buffett
 

Offline Flipflop

  • Contributor
  • Posts: 17
  • Country: gb
Re: EEVblog #884 - EEVBlog BM235 Multimeter REPAIR
« Reply #24 on: May 28, 2016, 10:39:42 pm »
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?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf