On 17/12/2010 00:34, peter green wrote:
> Oli Glaser wrote:
>> On 16/12/2010 21:12, Nathan Nottingham wrote:
>>
>>> I can't offer as much as others in contributing to the list, but hopefu=
lly I can brighten someone's day with my (self-inflicted) misery...
>>>
>>> My day has gone something like this:
>>>
>>> 1.) Removed a ~1000 ball (0.6mm) BGA GPU
>>> 2.) Cleaned old solder off board and chip
>>> 3.) Used a stencil to replace lead-free solder with leaded solder
>>> 4.) Aligned BGA back into place and started to heat gun it
>>> 6.) Decided it could use more flux so reached for the nearest syringe
>>> 7.) Injected a bunch of Chipquik SMD 291AX solder paste between the chi=
p and the board
>>>
>>>
>>>>  From now on, I'm storing the solder paste very far from my flux.  :)
>>>>
>>> - Nate
>>>
>> Ouch! I understand how that could be a tad depressing.. :-)
>> It's so easy to make silly mistakes when reworking or populating boards
>> though, especially if you're a bit tired - I managed to put a SMD
>> tantalum the wrong way round the other day, fired the board up and
>> couldn't work out why nothing was working, even got the hot air gun and
>> removed a PIC and a few other components, before working out what it
>> was. Sorted it out, then did exactly the same thing again later on with
>> another one on the same board.. :-)
>> All good fun..
>>
> Hmm, my experiance when screwing up with tantalums (e.g. putting them in
> backwards or overheating them during installation) is that you know
> about it from the bang and in some cases, fireworks display and then the
> massive stink left afterwards.

Yes, had that before.. :-)
When prototyping/populating boards I often use my bench supply initially=20
with current limit turned on, so in the event of this kind of thing less=20
or no damage is done (I usually set it for about twice the expected=20
consumption so I can tell if something is wrong) Of course you have to=20
allow for initial current surges and the fact some things (e.g.=20
switchers) may not start properly with too much limiting, it's just a=20
very basic initial precaution.
The problem with that here was that the cap didn't quite fail completely=20
(no fireworks or short) which made it a bit difficult to pin down (no=20
bang with the second one either - no current limit this time but on a=20
low current LDO) Mind you, it shouldn't have been really, but that was=20
kind of my point, in that when tired or not concentrating you can easily=20
make silly mistakes, or spend ages solving a simple problem.. :-)

In my experience, 90% of errors like these happen when I'm tired, so=20
usually at the first sign of them I call it a day, or at least have a=20
break (neither of which I did above, which quite probably led to the=20
further foolishness... :-)  )
  I'm a fan of "sleeping on a problem", letting the subconscious go to=20
work on it - amazing how very often something that baffles you last=20
thing is solved first thing the next day.



--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.