PJ wrote:
>> In a ckt like this you will have transients all over the place,
>> and to be most safe you also need protection all over the place.
>> The kontaktor sensor pin has absolutely no protection in your
>> ckt, and it connects to a 28" long antenna that runs straight
>> out to where the transients are being generated. I think this
>> is asking for trouble.
> hey, what can I say - some like living on the edge ;-)
>
>Would I also need protection on the other wire to
>the 'kontaktor sensor' the one that is connected to ground ?
>

Patrick,

Good question, hard to answer. There are probably several
people with ideas about how to handle these problems on piclist.
Hopefully, some others will submit theirs, too.

I guess any time you have wires going out to an area of
noise, then all of them can bring that noise into the ckt.

The worst offenders are probably the wires that go
directly to the cpu pins, for obvious reasons, so those
should definitely have some protection. I would use at a
bare minimum a series R with a cap at the cpu pin. There
are other methods too - transzorbs, LC filters, ferrite
beads, etc.

I would imagine that most designers would start with the
minimum they are comfortable with, and then increase that
if warranted, after testing the unit in the field. I used to
consult with a company that sold a microprocessor-controlled
icemaker and it was a terrible environment, with heavy
inductive switching loads. The machines had numerous
latchups. It took several years of different arrangements
before we pretty much nailed it down. But of course,
everything that you put in there costs $$$$.

Regarding the gnd line to the kontaktor sensor, that will
certainly bring in noise too. However, that pin terminates
on the pcb at the header, and not directly at the cpu.

A good pcb layout for this type of application might be
to have very heavy ground traces providing low impedance
pathways for conducting noise off the board and away from
the cpu. Gnd pins to external devices will connect here.
Also, by using various filters "directly" at the insertion
point on the pcb, you can help steer noise signals to the
gnd traces and away from the cpu.

This sort of thing goes on and on and on, and one of
the tricks to doing engineering design, I guess, is to
determine what is appropriate in any given situation.
Since every situation is different, you try to identify
the different problem areas, and use what your experience
and knowledge best dictates. I think, in electronics,
there is no one perfect answer. Many things work to
some extent, ultimately you get a pcb that's acceptable.

hope this helps,
- dan michaels
www.oricomtech.com
========================

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads