[Arm-netbook] Schematic and PCB layout CAD files

[Luke Kenneth Casson Leighton]

On Sun, Jun 16, 2019 at 5:14 PM Paul Boddie <paul at boddie.org.uk> wrote:
>
> On Wednesday 12. June 2019 02.09.12 Luke Kenneth Casson Leighton wrote:
> >
> > added here
> > http://rhombus-tech.net/pcmcia_sources/
>
> I have updated this page with some more details and some remarks about the
> physical constraints involved. If such remarks belong elsewhere, feel free to
> move them.

 no, great idea.

> It was interesting to work through some of the calculations. The 0.4mm offset
> of one of the header parts seems like a curious figure until one does the
> calculations and determines that it must be intended for boards of 1.2mm
> thickness (or less) with components mounted on the upper side.

yyep.  so the female header on the Card clearly has to be at a fixed
height relative to the plastic and the metal case, which means that
the offsetting of the SMT pins on the header will actually raise or
lower the PCB that is connected to it.

therefore, get this wrong, and the PCB will be so low that you can't
get any components on one of the two sides.

> Indeed, headers employing such an offset might conceivably be interesting when
> considering sockets on the "non-EOMA68" edge of a computer card, since these
> headers would maximise the space available to components on the upper side of
> a board, presumably permitting top-mount sockets to be used.

 interestingly, there's nothing to stop you putting the card inside
the PCB at a slight angle, such that whilst at the PCMCIA header end
there is plenty of room (relatively speaking) on either side of the
PCB, and at the other end the PCB becomes flush with the floor of the
metal case.

> I don't know
> whether this would have influenced the choice of sockets in the current batch
> of boards, but I wonder if it might be influential or useful to consider for
> any subsequent production.

 the issue is that even if the 1.2mm PCB is flush with the 5.0mm floor
(leaving no room for underside components for about... 35 possibly
even as much as 40mm (half) the Card, which would hugely complicate
the design, the connectors are still around 3.3mm in height (USB-OTG,
Micro-HDMI Type D)

 5.0 - 0.1 (thickness of the metal) - 1.2 (PCB thickness) = 3.7mm

 3.7 - 3.3 (height of the connector) = 0.4mm

 minus another 0.1mm for the top case shield.

 so that leaves only 0.3mm clearance, which means that the connectors
will poke out in a completely asymmetric way, which is very ugly.

 the mid-mount option gave a reasonably symmetric above- and below-
gap that doesn't look quite so ugly.

 plus, having 50% of the PCB impossible to put components on the
BOTTOM side means that the PMIC area would need redesigning (again),
spreading out the components to fit the ones that normally go on
BOTTOM.


> I hope this was helpful, anyway.

yes, very.

> Paul
>
> P.S. I have also been working on KiCad footprints for some of the parts I have
> found, in case anyone is still interested in such things.

 superb.  if you send me an ssh key i can arrange to add you to the
repo that i started 6 years ago:
 http://git.rhombus-tech.net/?p=eoma.git;a=summary

l.