Re: gEDA-user: 4538 symbole on the 4000 library

[sdb@xxxxxxxxxx (Stuart Brorson)]

Hi 

> I've taken a look at that kind of error:
> "order of pinnumber in slotdef does not match pinseq order"

A word of caution:

There are situations where pinnumber != pinseq.  In particular, my
spice-sdb netlister triggers off of pinseq in order to know which
order to emit pins for SPICE netlisting.  This is because SPICE
.subckts require that the nets be listed in a particular order to
connect them correctly.  This order is not the same as the numbering
of the pins on the package.

For example, the classic op-amp has pins numbered on an 8-pin package
(e.g. DIP-8 or SO-8) as follows:

1 -- NC (or bias)
2 -- Vin-
3 -- Vin+
4 -- VEE
5 -- NC
6 -- Vout
7 -- VCC
8 -- NC (or bias)

However, here's how the SPICE .subckt numbers the nets on an OP177,
which is a typical el-cheapo 8 pin op-amp:

* Node assignments
*               non-inverting input
*               | inverting input
*               | | positive supply
*               | | |  negative supply
*               | | |  |  output
*               | | |  |  |
.SUBCKT OP177   1 2 99 50 39

The reason I made spice-sdb trigger off of pinseq was that I
understood that PCB used only pinnumber to number the pins, and it
ignored pinseq.

Therefore, as long as PCB only uses pinnumber, please leave the pinseq
alone, unless you know for a fact that it must be renumbered to fix
some other netlister.

Stuart