Test points - a possible future direction

Werner Almesberger werner at openmoko.org
Sat Jul 19 21:58:42 CEST 2008

A while ago, I posted a suggestion to one of our internal lists, for
making test points a bit more useful. This is for future products,
hopefully GTA03 and beyond.

The basic idea is to route some of the signals that currently go to
test points to an FPC connector instead, and to make a small FPC
that can be soldered to whatever cabling is needed for testing or for
communication. The assumption is that this FPC can be made cheaply,
so developers who want to connect there could get them in packs of ten
units or so.

Such an FPC would have the following advantages over the current test

- compact: takes up less board space (although some of our test points
  occupy areas that wouldn't be used for anything else).

- removable: external cabling can be easily removed and re-attached,
  so the device can also be used for non-development purposes.

- modular (1): it's also possible to have different sets of external
  cabling, which can be swapped easily.

- modular (2): the same external wiring can be attached to different

- off-line soldering: the external cabling can be soldered and tested
  without getting near the Neo, thereby lowering the risk of accidently
  damaging it.

- open: 3rd parties could easily make variations of the FPC, e.g., with
  a different spacing or geometry, or even with some circuits residing
  on the FPC.

Of course, there are also some disadvantages:

- current limit: those FPC connectors can only handle about 1 mA and
  have a high contact resistance, so they would not be suitable for
  exporting power to the external circuit or even for powering the

- longer signal paths: test points can be placed right next to the
  source of the signal, while signals may have to be routed over a
  long distance for a central connector.

- more exposed signals: some of the test points are currently inside
  a shield. Exporting them to a connector would also require proper
  EMI and ESD protection. (Well, the signals we currently export on
  the debug connector of GTA01 and GTA02 have neither ... :-( )

- more components: FPC connector and that special FPC.

I would suggest to use a 39 position 0.3mm FPC like for the debug
board. That way, the same FPC can be used also there, and we don't
necessarily have to export any signals already available there.

Here's an example drawing of what I have in mind:

A 50mil spacing of the solder pads would make it easy to attach a
standard ribbon cable, and it's still a fairly generous spacing
for manual soldering. If this should turn out to be woefully
inadequate, someone else can always make a better design.

Not sure if the connector should face into the board or away from
it. Into the board would allow for more compact assemblies, while
away from the board would make it easier to access the contacts.
Perhaps the best place would be on top, facing away from the board.
That way, the FPC could be bent around the edge and below the board,
if necessary.

Joerg, you already said that you liked that idea. Perhaps you could
discuss this a bit with our hardware team during your next visit ?

- Werner

More information about the hardware mailing list