[gta02-core] Use for hole left by GPS connector - shields

Rask Ingemann Lambertsen ccc94453 at vip.cybercity.dk
Sun Jun 28 19:51:02 CEST 2009

On Sat, Jun 27, 2009 at 08:22:48PM -0300, Werner Almesberger wrote:
> Rask Ingemann Lambertsen wrote:
> >    I used to work at a place which did shielding solutions based on a
> > photo-chemical process. To make a shielding can, you cut a sheet[1] of an
> > appropriate metal to the desired shape and then you bend it.
> Thanks for the explanation ! That's all already quite complicated. I
> was thinking more of just putting a sheet into a CNC mill, cutting it
> to size, and bending manually, perhaps over a customized support. But
> perhaps that's too optimistic.

   I didn't think of cutting with a CNC mill, but I suppose it will work
fine. For sure you should be able to try out a new shape faster that way.
But creating the support for bending the can is the "interesting" part of
it. You want to use a hard material (e.g. tool steel) for the support, but
that will do your CNC drill bit no good at all if you try to make the
support that way. We did do some using photo-chemical milling, but we didn't
have to bend a structure as tall as a can wall. It was a few tenths of a mm,
so in range for e.g. the dimpling for the snaplock mechanism to keep the lid
in place.

   Maybe we can look at the two likely situations differently:

1) Mass production of a few thousand cans. OM found something that worked.
Do we know how much it cost? From the PCB photos[1][2] it looks like the
shape of the cans was not changed from the GTA01 to the GTA02 in spite of
substantial PCB changes, suggesting tooling costs were an issue.

2) Making 10-20 prototypes. Then you can live with tools that wear out
quickly, 100% inspection, parts that need rework, low yield, material
choices suboptimum for mass production, etc. Nobody will expect it to look

   Perhaps I should pay a visit to <URL:http://www.etch.dk/>. I could say
hello to Lars Traberg - we worked at the same place until a little over two
years ago - and I could bring my GTA02 so we can talk about possible can
shapes, alloys, material thicknesses, ballpark figures for costs and such.

[1] https://wiki.openmoko.org/wiki/Image:Gta01b_v3_bottom.jpg
[2] https://wiki.openmoko.org/wiki/Image:Gta02a5_pcba_cs.JPG

> >    So if we're considering a new shielding can, I'd say go for an existing
> > design. For example, Digikey has a product family called "RF shields" in the
> > category "RF and RFID". They don't appear to ship in bent form, but maybe
> > shielding cans are available prebent from somewhere else.
> You mean these ?
> http://media.digikey.com/pdf/Data Sheets/Fotofab PDFs/Standard RF Shields SPECS.pdf
> They claim that the shields can be bent manually.

   On the photos, it looks like they've etched partially into the material
to create a weak line to aid in bending the cans. They're using 0.381 mm
brass, which is quite thick. I was thinking along the lines of 0.15 mm MA3M
stainless steel, although come to think of it, I've not seen it used for
shielding cans, only shielding gaskets. Cans were usually nickel silver. The
choice of material determines e.g. heat conductivity, shielding properties,
weight, mechanical properties, solderability and corrosion resistance.

   How are the shields soldered to the PCB at the factory? You pretty much have
to solder the components first and the shields afterwards, don't you? You
somehow apply a lot of heat to the shields without unsoldering the
components outside the shield?

> Unfortunately, the
> GTA02 shape needs rather non-rectangular cans, and we would want them
> to be removable.

   Maybe the can was just made as large as possible to not limit the GPS
part design too much. If the gta02-core GPS part will fit in a rectangular
can, then there's nothing to stop us from using a rectangular can, is there?

Rask Ingemann Lambertsen
Danish law requires addresses in e-mail to be logged and stored for a year

More information about the gta02-core mailing list