RE: [PROTEL EDA USERS]: Database management for different models of the same board

[lloyd.good@ps.ge.com]

We deal with this problem all the time. We have several products that use
the same pcb with different populations. What we have found is that your
first solution ends up with revisions making it through to some schematics
but not all, which is disastrous. We now use a new product numbering scheme.
The main schematic and pcb are called for example a XX-5000. Now any product
that uses this pcb and schematic would have a variant of this product
number. One build would be a XX-5001 which means the BOM includes the
XX-5000 with a particular build. We have a part number for a "Not populated"
component. In the schematic we fill in the different part fields with the
company part number and use multiple part fields to generated different BOMs
for the different products.
Eg. XX-5000 would use part field 1 - a complete build with all parts
	XX-5001 uses part field 2 - some of the same parts now have the
"Not-populated" number.
	XX-5002 etc....
This seems to keep the integrity of the schematic and pcb together. Any
revisions affect all products.
Its working for us.
Cheers,
Lloyd Good


-----Original Message-----
From: HxEngr@aol.com [mailto:HxEngr@aol.com]
Sent: Wednesday, September 06, 2000 6:33 AM
To: Multiple recipients of list proteledausers
Subject: [PROTEL EDA USERS]: Database management for different models of
the same board


I'm sure others have wrestled with this issue as well, but I've never seen a

really good solution. I have a customer who is designing a very complex 
board, which can be stuffed in several different variants to produce 
different models of a product. The PCB is identical in all cases, but the
BOM 
(perhaps 50 of about 800 parts) changes depending upon the exact model being

built. In some cases values change; in other cases specific parts are left 
off. Only one type of PCB is being fabricated in any case for this family of

products.

One approach is to generate a complete schematic for each model. This way 
everyone has a schematic to work from for the build, troubleshooting, field 
service, etc., and doesn't have to think about all the possible differences.

The opposite extreme is to make one schematic and just flag those parts
whose 
values can change (possibly to "None"), and make a separate appendix to the 
BOM listing the specific parts which are used on each model, either as a 
complete BOM for each model, or as one master BOM which is used on every 
model, and a separate BOM for each individual model which shows the parts 
which must be added to the baseline for that model.

The first approach appears to be easier in the long run, because you need 
only deal mentally with one model at a time. However, it is more work up 
front, and (to me, at least) the worst feature is the danger of divergence
as 
time goes on, to the point that the schematic might not even match the PCB 
any more, and it will become difficult to check for such divergence. The 
second approach is easier up front, but does require a bit more headwork
down 
the road to look up the exact part which should be on a particular model,
but 
guarantees that the schematic and PCB match.

To give a sense of scale, we're not talking about GM here - this unit will 
see production in the hundreds or thousands of units, not 100,000. And the 
parts which change are just minor adjustments in values, to accomodate 
changes in operating voltages, etc., plus a few which are depopulated where 
certain features of the electronics are not needed.

I bet somebody (probably quite a few somebodies) on this list have dealt
with 
similar problems and have clever solutions. Any that you'd like to share?

Steve Hendrix