Are you using in-circuit programming?
If so you may be able to get a number of different "tries" on a single OTP
chip. It depends on the size of your code & how much of the available
program space you use. You can even re-assign subroutine addresses so only
part of your code is changed but you need to be careful about page
switching. Just remember that programming will only clear bits - not set
them.
I've been getting 7 tries out of a 16C73A recently - but now my code is >
0x200 it's complicated things somewhat.

The microchip document TB013 gives a quick example of this.

Richard

> -----Original Message-----
> From: Dave Johnson [mailto:djohnson@SIRIUS.COM]
> Sent: Saturday, September 04, 1999 7:22 AM
> To: PICLIST@MITVMA.MIT.EDU
> Subject: OTP Development strategies
>
>
> I'm working on a project where the target board is tiny, using surface
> mount parts (including the PIC in a TQFP). But of course I need to
> reprogram the PIC umpteen times during development. What's the usual
> strategy here? It's impractical to breadboard this project so that I
> could use a windowed part, but of course I don't want to
> throw away a PIC
> every time I need to fix a firmware bug, not to mention removing and
> replacing the part from the board.
>
> I suppose the ideal solution would be an ICE, but I can't afford it.
> Maybe some sort of adapter to use the windowed part?
>
> Any suggestions/strategies are appreciated.
>
> Dave Johnson
>