From: Humphreys, David (david.humphreys_at_honeywell.com)
Date: Fri Jul 09 2004 - 06:14:50 PDT
> However, if you were
> working with
> SMD youŽll probably be able to unsolder chips even w/o hot
> air gun.
Then you are asking for trouble.
> IŽve
> used a bigger soldering iron w/ about 100W and put a big blob
> of solder onto
> each row of pins. If all is heated up and melted you have
> only to turn the
> PCB quickly and the SMD will fall off.
Physically, I don't doubt that you can remove chips this way but you WILL damage them internally. This I guarantee.
> This worked very well
> for SMDŽs up to
> 2x 24 and even for 4x8. ThereŽs a danger that the unsoldered
> SMD is damaged
> by heat if you took to long
And that's the point. Heating up one side results in uneven heat flow to the substrate. This expands in an uneven way as
a result and stresses the silicon that is bonded to it and which forms the circuit inside. As the silicon has a
different temperature coefficient of expansion, it will change size at a different rate than the substrate. Something's
got to give. It usually is the silicon that forms micro cracks and destroys the chip. It doesn't matter how good a
solderer you are you cannot control the heating of the pins evenly with this method.
> but between 5 to maximum 8s
> should be okay (see
> the data sheets for what soldering time is specified).
The time specified in the data sheets refers to soldering in a professional environment. It has no relevance here using
this brute-force method.
> To
> short soldering
> time is bad as well. Heating up a soldering joint for less
> than 2s will
> produce unreliable contacts because the hot tin cannot
> diffund into the PCB
> pad surface.
Same as above.
>
> The next hurdle is to be shure that there are no tin shorts
> between the
> pins. The best bet is to use enough flux and solder removal
> stranded wire.
> And finally washing the PCB in an ultrasonic bath with
> alcohol is sort of
> mandatory.
>
> Marco
Look, the professional way of removing SMD components is by using an SMD rework station. This is a temperature
controlled hot air gun combined with a hot air bath.
This is used to slowly bring the chip AND the board up to temperature. Then, a low melting point solder alloy is melted
on all the pins. This forms an alloy with the existing solder on the board and brings the combined melting point down to
a level way below the melting point of ordinary solder. Then the hot air gun with a suitable nozzle, is applied to the
chip and when the solder reflows, it is pulled from the board. The low temperature reduces the risk of chip damage as
does the even heat soak before hand. It also guards against pads lifting off the board due to excessive heating.
Putting chips on the board is pretty much the reverse. The key is even heating. You cannot get that with a soldering
iron no matter how good you are at soldering.
Even the specialty tips that allow simultaneous heating of pins only solve a part of the problem. You cannot control
heat flow with any degree of accuracy with this method.
For those of you who have desoldered chip in this 'Radioshaft - You've got questions, we've got blank stares' way and
have been 'successful' it is more down to luck than skill and it is very likely that chip integrity has been
compromised.
The effects are random so it is impossible to predict how they have been compromised but, believe me, they have.
I understand that few people have this kind of equipment but that's what is needed to do the job correctly. Anything
else is sub-standard.
Sorry for the lecture but it is hard to just sit back and let folklore turn into reality. No offence is meant to anyone.
I just hate to see bad methods perpetuated. Please feel free to use whatever method you wish to remove your SMD
components. At least you now have some science to base your decision on.
Regards,
David Humphreys (PCBman)
-- This is the NewtonTalk list - http://www.newtontalk.net/ for all inquiries Official Newton FAQ: http://www.chuma.org/newton/faq/ WikiWikiNewt for all kinds of articles: http://tools.unna.org/wikiwikinewt/
This archive was generated by hypermail 2.1.5 : Fri Jul 16 2004 - 08:03:20 PDT