At 06:11 PM 10/3/2005 +0100, you wrote:
>Bob,
>
>On Sun, 02 Oct 2005 18:06:19 -0700, Bob Axtell wrote:
> >...<
> > Purchasing 1000 PCBs that are not right is very costly. The chance
> > of you getting the layout correct the first time is impossible; most
> > commercial designs have 7-8 versions before committing to a
> > volume purchase.
>
>I can well believe this - and in that case, why do a lot of the board 
>houses have the so-called Prototype
>minimum of 5 pieces?  I know Olimex will do one-offs, but most of the 
>others won't, which seems a huge waste
>of money.  If you discover problems only "using" one board at each 
>iteration, you have 7 * 4 = 28 wasted
>boards!

The cost of making 1 board or 5 boards typically wouldn't be much different,
and 4 or 5 is not an unusual number to want, if not need, especially if more
than one physical location (often I do work for people 3-12 time zones
distant), or more than one firmware developer is involved, or if field
testing is required.

It's much better to have one or two spares available than to be short one!

The expected number of revisions to go through is partly a matter of company
culture-- lately, I've been taking the attitude that you should *expect* zero
errors, particularly when the EDA software is familiar, and any errors that
do crop up are an opportunity to improve the processes to weed them out
entirely.

Of course, it's always likely that spec's will change (this does not
count as an error), but taking the time to get errors down to approaching
zero the first time enforces design rigor that has beneficial effects in 
subtle
ways. OTOH, it takes a bit longer design time.

And another OTOH, in some cases, depending on who you are dealing with,
they don't really start thinking about things clearly until they see a physical
prototype, so getting to the first one quickly is very important.

Personally, I've ordered 1,000 multilayer boards without making even one
prototype, and also ordered a quantity of boards made and assembled without 
ever
seeing the bare boards. A bit nerve wracking in both cases, but they worked
out pretty well. I'd rather not do that on a first iteration, but it is
possible.

>Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


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