notes below.

Rick Thompson wrote:

>> You remember in the movie "Michael" when the angel Michael said "you 
>> can never eat too much sugar"?
>
>
> Excellent example.

I love that movie. I thought I'd die when the angel Michael walked down 
the stairs scratching his b- - - - -.

>
>> Well you can never have too much GND plane, and that includes copper 
>> pours.
>>
>> BUT... unconnected copper pour fragments will actually hurt, not 
>> help, an EMI situation. Make sure those
>> pieces are actually GND.
>
>
> Okay.  Thankfully, pcb design software is really good at removing free 
> copper automatically.

Yes, I know. I was thinking about simple layout programs. Actually I 
only own expensive ones (Accel, OrCad)
so I have no experience with Eagle, etc.

>
>
>> Did you put tiny caps on any long runs? Even caps as small as 5pF 
>> seems to stop a PIC from capturing noise
>> through a long trace. Just a thought.
>
>
> No kidding?  Please tell me at which end of the run is best, or right 
> in the middle if it matters?  Makes sense to put it right before the 
> input pin. The longest trace on my board is only about 2 1/2" so 
> unless you recommend otherwise I'll probably leave them off.


No point with just 2.5". yes, near the PIC input. Think of it as killing 
RF at a receiver. At what point will the cap do the most good? Nearest the
input itself.

>
>> 4Mhz is so slow that you won't radiate anything either. HINT: if you 
>> can't hear any noise, chances are great that
>> the PIC won't capture any, either.
>
>
> I think what my real problem is that on the previous revision I made a 
> bonehead compromize and left the crystal unprotected & close to the 
> edge of the board.  That's fixed on the new revision and I'm trying to 
> include any other noise trick I can think of, within reason.  Copper 
> pour is free.

You should go one step further, and unless precision is REALLY 
important, replace the crystal with a ceramic resonator. Crystals are 
very sensitive to shock damage, and worse, when it occurs, is literally 
impossible to troubleshoot, as the result is so similar to a firmware 
fault. I've been a hero to some pretty big companies when I suggested 
this simple, simple change. The makers of ceramic resonators (especially 
MuRata) have so dramatically improved their products that MuRata sells a 
0.1% resonator now! Other than timekeeping, precise measurement or USB, 
most applications can tolerate 1% easily.

>
> Lots to be learned here.

I learn something new every day.

>
>
>> You are on the right track.
>
>
> Pun intended???  :o)

sure!

Placing PWR/GND on top bottom: I tried it once or twice, and test people 
killed me. never did it again, although it would seems it would work well.
Seems they wanted to be able to easily cut traces and jumper things for 
ECNs (Engineering change notices).

>
>
> Thanks for all your help, Bob.
>
No problem. Pass it on.

> Rick
>
>
>


-- 
Note: To protect our network,
attachments must be sent to
attach@engineer.cotse.net .
1-866-263-5745 USA/Canada
http://beam.to/azengineer

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