brand new PCB with variable resistance between bus lines

INCREDIBLE!!
the problem on a new series PCB, the prototype worked well but in the
last version there was some resistance or
short circuit in the bus lines wich hangued the cpu and possibly
destroyed the XCS05 (with strange white or black substance going out
of the gnd or vcc pins).
With a 5V supply between the line and GND we eliminated the shorts
but
appeared after few minutes. We try to measure the resistance of the
lines with a tester and found to be lower and lower, accelerating the
decrease if we heated the PCB. Anybody knows the explanation of
that ?
thank you

Narcis Nadal

On Jan 29, 1:40=A0am, nna...@terra.es wrote:
> INCREDIBLE!!
> the problem on a new series PCB, the prototype worked well but in the
> last version there was some resistance or
> short circuit in the bus lines wich hangued the cpu and possibly
> destroyed the XCS05 (with strange white or black substance going out
> of the gnd or vcc pins).
> With a 5V supply between the line and GND we eliminated the shorts
> but
> appeared after few minutes. We try to measure the resistance of the
> lines with a tester and found to be lower and lower, accelerating the
> decrease if we heated the PCB. Anybody knows the explanation of
> that ?
> thank you
>
> Narcis Nadal

This is the most bizarre thing I have heard of regarding a PCB.  How
do you know the short is on the board and not one of the parts?  I
have ***no*** idea how a chip would "ooze" back and white foam from
the power pins.  Have you checked any bare PWBs?  If the problem is
the PWB you should try working with a bare PWB and do tests on that.

Rick

On Wed, 28 Jan 2009 22:40:47 -0800 (PST) nnadal@terra.es wrote in Message
id: <e592c0ab-5be0-4b78-b4de-f88cc107ce20@r38g2000vbi.googlegroups.com>:

>With a 5V supply between the line and GND we eliminated the shorts
>but
>appeared after few minutes. We try to measure the resistance of the
>lines with a tester and found to be lower and lower, accelerating the
>decrease if we heated the PCB. Anybody knows the explanation of
>that ?
>thank you

If you're "blowing" shorts with a 5V power supply, That's *not* a good
idea. I found out a number of years ago that the shorts can grow back
after hours or days of operation. I would guess that the act of blowing
the short leaves a bit of carbon between the traces, which is conductive.

On Jan 28, 11:40 pm, nna...@terra.es wrote:
> INCREDIBLE!!
> the problem on a new series PCB, the prototype worked well but in the
> last version there was some resistance or
> short circuit in the bus lines wich hangued the cpu and possibly
> destroyed the XCS05 (with strange white or black substance going out
> of the gnd or vcc pins).
> With a 5V supply between the line and GND we eliminated the shorts
> but
> appeared after few minutes. We try to measure the resistance of the
> lines with a tester and found to be lower and lower, accelerating the
> decrease if we heated the PCB. Anybody knows the explanation of
> that ?
> thank you
>
> Narcis Nadal

I have seen something like this before.

In my case, the PCB laminate was slightly porous, and the tin plating
oozed into cracks.  I would set me bench supply to 5V, and up the
current limit until I burned them out.  I would need to repeat this
ever other day until I got new PCB's in.

AL

On 29/01/2009 nnadal@terra.es wrote:

> 
> INCREDIBLE!!
> the problem on a new series PCB, the prototype worked well but in the
> last version there was some resistance or
> short circuit in the bus lines wich hangued the cpu and possibly
> destroyed the XCS05 (with strange white or black substance going out
> of the gnd or vcc pins).
> With a 5V supply between the line and GND we eliminated the shorts
> but
> appeared after few minutes. We try to measure the resistance of the
> lines with a tester and found to be lower and lower, accelerating the
> decrease if we heated the PCB. Anybody knows the explanation of
> that ?
> thank you
> 
> Narcis Nadal

Ah, the joys of RoHS. You probably have tin whiskers. See here:

http://www.siliconfareast.com/whiskers.htm

-- 
John B

On Jan 29, 7:40=A0pm, nna...@terra.es wrote:
> INCREDIBLE!!
> the problem on a new series PCB, the prototype worked well but in the
> last version there was some resistance or
> short circuit in the bus lines wich hangued the cpu and possibly
> destroyed the XCS05 (with strange white or black substance going out
> of the gnd or vcc pins).
> With a 5V supply between the line and GND we eliminated the shorts
> but
> appeared after few minutes. We try to measure the resistance of the
> lines with a tester and found to be lower and lower, accelerating the
> decrease if we heated the PCB. Anybody knows the explanation of
> that ?
> thank you
>
> Narcis Nadal

Rodent urine ?

-jg

On Jan 30, 5:47=A0am, "John B" <spamj_baraclo...@blockerzetnet.co.uk>
wrote:
> Ah, the joys of RoHS. You probably have tin whiskers. See here:
>
> http://www.siliconfareast.com/whiskers.htm

We have struck some field reliability issues we suspect is tin
whiskers/ fretting corrosion on  PLCC sockets. - the joys of RoHS
indeed.... :(

-jg

LittleAlex wrote:
> On Jan 28, 11:40 pm, nna...@terra.es wrote:
>> INCREDIBLE!!
>> the problem on a new series PCB, the prototype worked well but in the
>> last version there was some resistance or
>> short circuit in the bus lines wich hangued the cpu and possibly
>> destroyed the XCS05 (with strange white or black substance going out
>> of the gnd or vcc pins).
>> With a 5V supply between the line and GND we eliminated the shorts
>> but
>> appeared after few minutes. We try to measure the resistance of the
>> lines with a tester and found to be lower and lower, accelerating the
>> decrease if we heated the PCB. Anybody knows the explanation of
>> that ?
>> thank you
>>
>> Narcis Nadal
> 
> I have seen something like this before.
> 
> In my case, the PCB laminate was slightly porous, and the tin plating
> oozed into cracks.  I would set me bench supply to 5V, and up the
> current limit until I burned them out.  I would need to repeat this
> ever other day until I got new PCB's in.
> 

Wouldn't it work better with a higher voltage supply (with a large 
current capability) to more permanently blow out the shorts?  It would 
also make it easier to see where the problems are (look for the black 
marks!).

David Brown schrieb:
> LittleAlex wrote:

...

>> In my case, the PCB laminate was slightly porous, and the tin plating
>> oozed into cracks.  I would set me bench supply to 5V, and up the
>> current limit until I burned them out.  I would need to repeat this
>> ever other day until I got new PCB's in.
>>
> 
> Wouldn't it work better with a higher voltage supply (with a large
> current capability) to more permanently blow out the shorts?  It would
> also make it easier to see where the problems are (look for the black
> marks!).

In order to remove the short circuits on my homebrew PCBs I used an old
2200uF Capacitor and his maximum rated voltage (~30V).

Falk

On Thu, 29 Jan 2009 23:01:08 +0100, Falk Willberg
<Faweglassenlk@falk-willberg.de> wrote:

>In order to remove the short circuits on my homebrew PCBs I used an old
>2200uF Capacitor and his maximum rated voltage (~30V).

When I worked in an old (Strowger) telephone exchange, the 50v bus
bars were huge, and we did everything on live equipment. When
a pice of wirewrap (remember when you wire-wrapped with Lindstrom
pliers) wire fell into the frame, we just changed the rack fuses
for something unmentionably large.

The day a spanner fell across the bus bars was memorable. A single
cell of the exchange battery was a cube about 5 feet in each
dimension!

Stephen


-- 
Stephen Pelc, stephenXXX@mpeforth.com
MicroProcessor Engineering Ltd - More Real, Less Time
133 Hill Lane, Southampton SO15 5AF, England
tel: +44 (0)23 8063 1441, fax: +44 (0)23 8033 9691
web: http://www.mpeforth.com - free VFX Forth downloads