Engine Management Units

[MJC]

In article <nce05k$n7j$2@dont-email.me>, curd@notformail.com says...

Nowadays
all enamel copper seems to be clear, but back in the day the enamel was a
sort of deep plum shade - which is what I see here. I'll do some searches
for pictures of TW and see if some match what I see on this board....

I didn't know modern enamel had gone clear. Sounds like a recipe for
confusion! Even if all enamel is now also the sort that turns into flux
if you try to solder the wire...

Mike.

 

[legg]

On Thu, 17 Mar 2016 10:08:49 -0000 (UTC), Cursitor Doom
<curd@notformail.com> wrote:

On Wed, 16 Mar 2016 22:50:06 -0500, legg wrote:

Looked to me more like dye concentrating as the curing epoxy puckered.

So not likely to cause a fault then?

Must have been pretty strong stuff, to strip colour out of the mask,
but no, not a conductive path.

Did you ever reset this thing, while it was still in the vehicle?

No. I think if the fix were that simple the service specialists who
swapped the unit over wouldn't have gone to the trouble of sourcing a
replacement. They charged me next to nothing for the job so I trust them.

?? Just charged you for the ECU, that they happened to have lying
around? Very obliging...

If you trusted them from past work, that's another thing entirely.

RL

 

[Cursitor Doom]

I can't find any pictures that resemble these things on the web.
Following a closer look with a stereoscope, there are outbreaks of these
growths in at *least* a dozen places on this board. Here's two more
sharper pictures with the problem areas outlined:

https://www.flickr.com/photos/128859641@N02/25236055623/in/dateposted-
public/

https://www.flickr.com/photos/128859641@N02/25836586336/in/dateposted-
public/

All the pictures I've seen on the net show whiskers as silver in colour
but these are dark, metallic plum. They typically spring from vias in a
kind of branching, fern-like formation and certainly reach out to touch
other traces on the board.

Anyone seen anything like this before? Ideas on how to proceed?

 

[Cursitor Doom]

On Thu, 17 Mar 2016 10:56:37 -0700, DaveC wrote:

Choose one of these, try acetone on a Q-tip cotton ear bud to dissolve
the coating (try to not disturb the “thing” that lies below the
coating--if,
in fact it is below the coating). Get a “burnishing brush” with
fiberglass bristles (used to clean oxidation from copper, etc.) and
start gently rubbing the site. I think you’ll quickly discover if it’s a
surface growth, or a crack or other phenomenon.

Good luck.

Thanks. They're definitely not cracks, they're fibres of some sort.

 

[Cursitor Doom]

On Thu, 17 Mar 2016 19:57:11 +0100, Look165 wrote:

> This type of caps is known for drying inside with time.

Yes, I know. But these are not faulty. I tested them with one of these
meters:

http://www.peakelec.co.uk/acatalog/jz_esr70.html

Current most likely possibility is shorting between PCB traces due to
growths of some kind:

https://www.flickr.com/photos/128859641@N02/25836586336/in/dateposted-
public/

 

[DaveC]

All the pictures I've seen on the net show whiskers as silver in colour
but these are dark, metallic plum. They typically spring from vias in a
kind of branching, fern-like formation and certainly reach out to touch
other traces on the board.

Cracks?
> Anyone seen anything like this before? Ideas on how to proceed?

Choose one of these, try acetone on a Q-tip cotton ear bud to dissolve the
coating (try to not disturb the “thing” that lies below the coating--if,
in fact it is below the coating). Get a “burnishing brush” with
fiberglass bristles (used to clean oxidation from copper, etc.) and start
gently rubbing the site. I think you’ll quickly discover if it’s a
surface growth, or a crack or other phenomenon.

Good luck.

 

" wonnder how you can determine an ESR visually !

And technically, it's not so easy to measure this parameter. "

With a square wave that has a fast enough slew rate for the vertical parts of the scope trace to disappear. The square wave comes from a known impedance and is high enough frequency so Xc is minimal. Measure the amplitude of the parts you can't see and then it is simple ohm's law.

I use 400 mV 1,000 Hz with a 360 ohm impedance which does not make the numbers easy but it makes for quick troubleshooting.

 

[Cursitor Doom]

On Thu, 17 Mar 2016 16:26:54 -0500, Jon Elson wrote:

Tin can definitely turn this "plum" shade.

Jon

In that case I wonder how I would go about measuring the resistance of
these things? Maybe I could lift one off with a piece of sticky tape if
they're not too fragile. Can't think of a better way!

 

[Look165]

This type of caps is known for drying inside with time.


Cursitor Doom a ĂŠcrit :

Hi guys,

I have an EMU from a Land Rover which just suddenly failed one day
without warning as I was driving down the road. I was wondering if
there's any possibility of fixing it, given that schematics are AFAIK
unavailable due to commercial confidentiality. Of course there are some
basic checks that can be done without a diagram, but take a look at this
thing and tell me what I could try doing with it. I've carefully
inspected it close up but can find *nothing* at all visibly wrong with it:

http://tinyurl.com/hvk5eac

BTW, the whole board has been thickly sprayed with clear varnish which
accounts for all the bright dots which appear on some of the parts -
they're not indicative of any failure.

Any suggestions?

 

[Cursitor Doom]

On Thu, 17 Mar 2016 19:25:50 -0500, legg wrote:

On Thu, 17 Mar 2016 19:00:52 -0000 (UTC), Cursitor Doom
curd@notformail.com> wrote:

https://www.flickr.com/photos/128859641@N02/25836586336/in/dateposted-
public/

You're going to have to get more localized images of individual sites,
using a minimum x10 magnification, and enhanced/varying side-lighting.
A 'star' group, or part group should occupy a full image. Manual marking
should not interfere and should use line widths <<smaller than the image
details themselves.

If possible, you should label or datestamp each individually for
reference.

The captive metalic detritus on the body of one resistor did not resolve
any better in the later image. If possible, angled images can be
informative, when monocular, and diffused lighting will reduce glare.

If you do post more detailed images to the same flikr thread, they can
be downloaded and examined by IPC Technet group members, with comments
at

http://listserv.ipc.org/scripts/wa.exe?A1=ind1603&L=TechNet

RL

Thanks for that. Yes, I would have liked to have posted images direct
from the stereoscope which is just in its own class for this kind of
thing and shows *so* much more detail, but I don't have a suitable
adaptor. I'll check out your link tomorrow as it's getting very late here
now...

 

[Jon Elson]

Cursitor Doom wrote:

Thank you, Jon. An interesting suggestion! But I was under the impression
that:

1. TW growth happened solely inside chips and so was normally invisible.

Absolutely not! I had some Xilinx chips where the leadframe was tinned
before the IC was mounted and encapsulated, then the leads were bent. This
put strain into the tin plating, and whiskers grew from that. Xilinx'
solution was to REALLY FRY the chips when reflowing the board, which I did
not think highly of! They seem to have changed their plating to reduce the
whisker growth. These were easily seen with a microscope, and were clearly
sprouting out of the leads RIGHT at the bends.

Lots of reports indicate tin whiskers on circuit boards, connectors, IC
leads, etc. Moisture exposure seems to be an important part of the
conditions where they grow.

> 2. tin whiskers were invariably silver in colour.
Yes, they should start out that way. But, oxidized Tin can take on a
VARIETY of colors. Most of these I'm more accustomed to see where the Tin
has been heated (as when soldering) but I have definitely seen Tin turn all
shades of purple. Also, the conformal coating may have chemicals that
caused the tin whiskers to change color.

These I have here are like the old kind of enamel copper wire. Nowadays
all enamel copper seems to be clear, but back in the day the enamel was a
sort of deep plum shade - which is what I see here. I'll do some searches
for pictures of TW and see if some match what I see on this board....
Tin can definitely turn this "plum" shade.

Jon

 

[Jon Elson]

Cursitor Doom wrote:

I can't find any pictures that resemble these things on the web.
Following a closer look with a stereoscope, there are outbreaks of these
growths in at *least* a dozen places on this board. Here's two more
sharper pictures with the problem areas outlined:

https://www.flickr.com/photos/128859641@N02/25236055623/in/dateposted-
public/

https://www.flickr.com/photos/128859641@N02/25836586336/in/dateposted-
public/

All the pictures I've seen on the net show whiskers as silver in colour
but these are dark, metallic plum. They typically spring from vias in a
kind of branching, fern-like formation and certainly reach out to touch
other traces on the board.

Anyone seen anything like this before? Ideas on how to proceed?

These do NOT look like the whiskers I have encountered before. But, they
could be some sort of conductive stuff that has grown between nets on the
board. One possibility is they are fibers from a brush that either cleaned
the board or applied the conformal coat. But, they could still be trouble,
maybe they start out non-conductive but become conductive when moisture is
absorbed. Since you say the sprout from vias, maybe the board had some
contaminant left in the vias that slowly creeps out under heat and
electrical field. Anyway, not much to lose by trying to remove the
conformal coat at these spots and see if the EMU comes back to life.
If it does, then you need to get some new conformal coat and re-seal those
spots.

Jon

 

[legg]

On Thu, 17 Mar 2016 19:00:52 -0000 (UTC), Cursitor Doom
<curd@notformail.com> wrote:

https://www.flickr.com/photos/128859641@N02/25836586336/in/dateposted-
public/

You're going to have to get more localized images of individual sites,
using a minimum x10 magnification, and enhanced/varying side-lighting.
A 'star' group, or part group should occupy a full image. Manual
marking should not interfere and should use line widths <<smaller than
the image details themselves.

If possible, you should label or datestamp each individually for
reference.

The captive metalic detritus on the body of one resistor did not
resolve any better in the later image. If possible, angled images can
be informative, when monocular, and diffused lighting will reduce
glare.

If you do post more detailed images to the same flikr thread, they can
be downloaded and examined by IPC Technet group members, with comments
at

http://listserv.ipc.org/scripts/wa.exe?A1=ind1603&L=TechNet

RL

 

[Jon Elson]

Cursitor Doom wrote:

On Thu, 17 Mar 2016 16:26:54 -0500, Jon Elson wrote:

Tin can definitely turn this "plum" shade.

Jon

In that case I wonder how I would go about measuring the resistance of
these things? Maybe I could lift one off with a piece of sticky tape if
they're not too fragile. Can't think of a better way!

I'm still thinking these were left behind by some kind of brush. But, they
could still be trouble after years of exposure to <something>.

Are these on top of the conformal coat or underneath/within it? If they are
in the coat, then it will be hard to extract them.

Jon

 

[legg]

On Thu, 17 Mar 2016 17:24:00 -0000 (UTC), Cursitor Doom
<curd@notformail.com> wrote:

I can't find any pictures that resemble these things on the web.
Following a closer look with a stereoscope, there are outbreaks of these
growths in at *least* a dozen places on this board. Here's two more
sharper pictures with the problem areas outlined:

https://www.flickr.com/photos/128859641@N02/25236055623/in/dateposted-
public/

https://www.flickr.com/photos/128859641@N02/25836586336/in/dateposted-
public/

All the pictures I've seen on the net show whiskers as silver in colour
but these are dark, metallic plum. They typically spring from vias in a
kind of branching, fern-like formation and certainly reach out to touch
other traces on the board.

Anyone seen anything like this before? Ideas on how to proceed?

Early technet response suggests chemical contamination in vias, but
would require SEM EDX to confirm. This contamination could difuse in
the conformal coating cure stress lines (ie the pucker).

I don't see vias involved in every instance - some sites seem to
develop at pure surface mount pad sites.

There's an image of copper dendrites on Steve Zeva's gallery

http://stevezeva.homestead.com/

http://stevezeva.homestead.com/Dendrite_1.jpg

These tend to be inter-node rather than centrally expanding.

A close examination of tin whiskers should reveal a characteristic
three dimensional spike/thread detail. Not likely here.

RL

 

[Cursitor Doom]

On Thu, 17 Mar 2016 21:36:17 -0500, Jon Elson wrote:

They seem to emerge from under it and spread along the surface, but I
can't confirm that yet as I've resisted disturbing them so far in case
they disintegrate and defy further analysis.
It's a good question, though, and I expect to be able to report back on
this later today. I'll sacrifice one of the patches and find out.

 

[Cursitor Doom]

On Fri, 18 Mar 2016 00:07:03 -0500, legg wrote:

Early technet response suggests chemical contamination in vias, but
would require SEM EDX to confirm. This contamination could difuse in the
conformal coating cure stress lines (ie the pucker).

Can you post a url for this site? I searched but got about a million
microsoft support links instead.

"SEM EDX"? Is that something to do with electron microscopy? I have an
admittedly huge collection of test gear here, but not a single electron
microscope.

I don't see vias involved in every instance - some sites seem to develop
at pure surface mount pad sites.

True, but they're in the minority. Maybe that's relevant; maybe not. The
we don't have a large enough sample here to say.

There's an image of copper dendrites on Steve Zeva's gallery

http://stevezeva.homestead.com/

http://stevezeva.homestead.com/Dendrite_1.jpg

These tend to be inter-node rather than centrally expanding.

Yes, still no precise matches.

A close examination of tin whiskers should reveal a characteristic three
dimensional spike/thread detail. Not likely here.

Yes, TW seem to come out at right angles to the plain; these are surface
creepers.

 

[Cursitor Doom]

They seem to emerge from under it and spread along the surface, but I
can't confirm that yet as I've resisted disturbing them so far in case
they disintegrate and defy further analysis.

Right, basically what I said above is correct. There seems to have been
some sort of breakdown of the clear coating which has enabled these
things to get a foothold. They then break out and traverse along the
board *on top* of the coating. So it's doubtful they could cause any
short circuits unless by sheer chance the tip of one of them hits another
spot of coating that's broken down above a trace - or a chip pin. When
touched with a scalpel blade they break apart readily into shards and can
be blown away by an air duster can.
The coating has lost its integrity and could probably be scrubbed away
with a toothbrush. I personally prefer to use clear varnish for my own
boards. It hasn't the depth that this coating has, but at least when it's
set, it's set for good.
I wonder what the reverse side looks like? Perhaps there's no coating at
all, in which case who knows what state it's in.

 

[Cursitor Doom]

On Fri, 18 Mar 2016 10:37:46 -0500, legg wrote:

Technet link was previously posted.

http://listserv.ipc.org/scripts/wa.exe?A1=ind1603&L=TechNet

Thanks, I hadn't spotted that link.

> Question is, what is your repair plan?

If there is some formal defintion of this term, I don't know it. But my
plans with this unit are to spend as little time as possible on it and if
it can't be fixed reasonably quickly, to get rid of it. Items like this
are way off-beat for me and I have a pile of other stuff more in my line
awaiting my attention in due course.

You don't seem to have any indication of the unit's present state other
than sticking it back in the car and seeing if the car starts. No fb on
xtal replacement yet.......

I'll get there. I just have to fit this EMU in between other jobs with
higher priorities, plus avoiding the usual divorce threats.

> The blemishes you're chasing may have nothing to do with actual fault.

I agree entirely.

The spatter pattern is also suggestive of the impact from an external
source, such as a flicked paint brush.

Nope. If you could see it through the stereoscope, you'd agree
straightaway. I shall have to post to the photo groups for adaptor ideas
to make it possible to post much better quality macro shots online than
the ones I'm pissing everyone off with at present.

The single metalic entrapment pictured should also flick off with a
knife blade. But without some before and after passive probing, you'll
never know if this did anything.

Easier said than done (see my update posted 2.10pm local here).

 

[Gunther Heiko Hagen]

On Fri, 18 Mar 2016 18:13:46 +0000, Mike Tomlinson wrote:

En el artĂ­culo <37cc548b-c211-4846-a1df-54c46f865865@googlegroups.com>,
pfjw@aol.com <pfjw@aol.com> escribiĂł:

why do the British drink warm beer?

So we can actually taste it. Have you drunk warm American beer? It's
icy cold to hide the fact that it tastes bloody awful. QED.

They don't appear to use hops for flavouring in N. America and Australia,
so it's not really beer at all but 'malt liquor' to be more accurate.
Young British men who don't know any better seem quite happy to drink
such monstrous concoctions as Coors, Bud and Fosters because they don't
know any better and fall for the flashy advertising and glitzy looking
dispensers. The best beers in the world - proper beers - are brewed by
the Belgians, the Czechs and the very best of all of course - the Germans.