[sdiy] Eagle analogue & digital groundplanes

[Sarah Thompson]

It's not *that* hard to get that many components into the space -- it would
have been easier if I could have used BGAs, but they are generally not
preferred for space applications so I had to get a waiver for the FPGA. It
basically means being REALLY careful about placement. If I could have used
standard vias it would have been easier too, but 4 routing layers and a
really fine pitch helps a lot. Looking at the board under a microscope
(sorry, I can't post photos!), there's not really much sign of any
variation in trace widths, even though the LVDS pairs are so fine that they
look like a single trace usually to the naked eye. The board house actually
could go even finer, believe it or not.

The main trick is that is *must* be assembled by a robot to stand any
reasonable chance of working first time. We had the assembly house do
flying probe tests after the board was populated to make sure it was
kosher, but the boards were all 100% tested before assembly too.

I'm a bit spoiled by being based in Silicon Valley -- it does make finding
that kind of crazy capability a lot easier.

Sarah

On Fri, Nov 27, 2015 at 1:52 AM, john slee <indigoid at oldcorollas.org> wrote:

> On 27 November 2015 at 20:30, Roman Sowa <modular at go2.pl> wrote:
>
>> Can't imagine 500 components on 2.5''sq. People call that alien
>> technology, and who has access to real alien technology if not your
>> employer...
>
>
> +1, I would love to see this. I'd make it my desktop background.
> Particularly if it was a nice white PCB!
>
> John
>



-- 
[s]
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://synth-diy.org/pipermail/synth-diy/attachments/20151129/42985f94/attachment.htm>