I haven’t checked prices on pogo pins recently, however it would be neat to have the various in-circuit serial programmers have optional pogo pin attachment instead of requiring headers, with the pogo pins compatible with current target boards with unpopulated thru-hole ISP headers. Then, just “press… program… go!”. Optionally the thru-holes could be eliminated on some boards since the pogo pins only need pads. Alignment can be done by using extra-long male header on the programmer side which slips into the target board off-center thru holes, so pogo pins are placed properly. This assumes that pogo pins are relatively cheap enough to not severely increase the cost of the programmer (though needles might do just as well).
Needles don’t really work very well. I was hacking a rainbow iKey a while back and made an eeprom reading jig with sewing needles. In theory, a good idea - but in practice, it’s very hard to maintain consistent contact with something that inflexible. Pogo pins would be much better…
signal7:
Needles don’t really work very well. I was hacking a rainbow iKey a while back and made an eeprom reading jig with sewing needles. In theory, a good idea - but in practice, it’s very hard to maintain consistent contact with something that inflexible. Pogo pins would be much better…
Hmm. Maybe some good mechanical design would fix that. The needles would likely need mounting on a separate springloaded board, a single fat spring might do. Unless pogo pins are cheap enough not to bother.
there are many flavors, but I’ve used http://search.digikey.com/scripts/DkSea … =ED5069-ND to some success. there are other ones i’ve used that are much longer (2 inches long with maybe a half inch travel) but i haven’t a clue what they were. (hurray for ‘stuff’ in the lab)
even better, the ICSP pads could just be routed out a bit from the PICs pads and not require a header footprint.
By the time you are done spring loading a needle, I think you’ll find pogos to be much much easier.
Philba:
even better, the ICSP pads could just be routed out a bit from the PICs pads and not require a header footprint.By the time you are done spring loading a needle, I think you’ll find pogos to be much much easier.
Another option could be to use the target PCB as a card edge, either on single side with a row of contacts or double sided board with contacts on both sides. Then the programmer contacts slides on/off the edge of the target board for ICSP. Although this would only be useful for bare boards and not so useful for boards mounted in a case and is physically larger than using pogo’s, it is still easier than sliding a header block on/off.
I still like the pogo pin idea though.
I recently saw a nice solution made by Tag-Connect ( http://www.tag-connect.com/ ). The TC2030-MCP-NL looks promising. It has a very minimal footprint, uses guide holes to ensure proper orientation, and looks well made (compared to some other pogo pin programmers I’ve seen). A little pricey at $30, but for quick in-circuit programming, probably worth it. My only wish is that they would also offer it with an IDC header for use with Atmel programmers.