Re: gEDA-user: Tin pest

[Karel Kulhavy <clock@xxxxxxxxxxxxx>]

On Sat, Feb 04, 2006 at 11:22:38AM -0500, Bob Paddock wrote:
> On Saturday 04 February 2006 09:31 am, Karel Kulhavy wrote:
> 
> > The pest takes 1.5 years of freezing temperature to happen.
> 
> If done wrong (right?) you can get dendrite growth within a few seconds.
> It is a cool thing to watch on a microscope as they seem almost alive
> as they branch out like small trees, hence them name 'Dendrites'
> 
> Dentrites have always been a problem with ultra-high impedance circuitry,
> not with just Lead Free.  Your photo diode section might be prone to dendrite 
> problems.  Stress to the builders of your project that ultra clean boards are a must.
> 
> What Are Dendrites:
> http://www.rpi.edu/locker/56/000756/dendrite.html
> 
> Google for 'dendrites solder' and you get a lot of hits.
> then Google for 'dendrites high impedance'.
> 
> This page explains "Dendrites" are NOT "Whiskers" and gives a good overview
> of the whisker problem:
> 
> http://nepp.nasa.gov/WHISKER/background/index.htm
> 
> I do take exception to this advice at that page:
> 
> "Conformal Coat or foam encapsulation over the whisker prone surface appears to
> be beneficial but the limitations are not completely understood."
> 
> There is a significant limitation to Conformal Coating.  I used Conformal Coating 
> a lot in the coal mines to keep the caustic dust of the boards.
> 
> Since conformal coating is not a hermetic seal what real happens is
> the impurities in the water are kept away from the circuit, but the water
> itself reaches the traces, eventually.  Since the water is now fairly devoid of 
> contaminates the water acts more like a dielectric insulator.  You never
> notice in a digital circuit, but unless debugging is an obsession don't let it
> get near a RF tuning circuit or a high impedance sensor circuit.

Good is also putting 100m of UTP cable on a flat roof. If it rains, the
cable lies in ponds of water and the RF energy which travels partially
outside gets lost in the water. Packetloss at 100Mbps results. We had
this problem.

The solution is to use shielded cable.

CL<
> 
> 
> -- 
>                           http://www.softwaresafety.net/
>  http://www.unusualresearch.com/ http://www.bpaddock.com/
>                           http://www.designer-iii.com/