[PicForth] Re: more on flash-write failure

[easlab at absamail.co.za]

David McNab wrote:
> [snip]
> What I'm seeing with my test prog is:
> - I do a flash-write, then do a flash-read to verify
> - the flash-read most often produces the original data, not the
>    data I've just written (although about 1 in 5 times, it produces
>    the data just written)
> - after a reset, I'm seeing the data my test prog has written

If it's inconsistent, that points to some parameter(s) at their 
threshold, instead of being safetly inside the guaranteed range.
Could be voltages, but more likely to be times.
eg.  if the device is guaranteed to 'do the function' provided the
signal is applied for minimum 8 <time units>, then if the time 
allowed is only 6 time units, you would get such inconsistent results.

> [snip]
> I'm using the bootloader from 
> www.microchipc.com, and between a couple of hacks to the bootloader, 
> plus a hacked-up picforth include file, I'm able to compile picforth 
> progs to a .hex format which will download and run fine via this 
> bootloader.
> 
Kevin L. Pauba wrote:
] This is interesting to me.  Any chance that you could report the details 
] on the PicForth Wiki?

Indeed. 
Or if you don't want to do wiki, just a plain text description here ?
And it may be good to 'find out' what the problem was.
Sometimes a work-around just allows it to come back and bite you
later, when it's more painful.

== Chris Glur.