PIC 16F628 programming problems?
Mitchell Mlinar
I have a home-brew David Tait-based programmer that has been working
well for years on 16C73, 16F873 and 16F876. However, there have been
periodic failures on a new PIC chip, the 16F628 -- catastrophic
failures where the chip is no longer programmable.
Some details:
- 16F628 programming (other than memory size) is being treated the
same as 16F873 and 16F876 (i.e. flash style and the same programming
commands)
- I use hi-voltage programming with the voltage set precisely at 13V
(to be compatible with 16C series); I do this because I tend to fully
utilize all pins and cannot afford to dedicate one to LVP (although on
the 628, that can be resolved by going to internal MCLR and using that
pin instead)
- sometimes a 628 will program, but often it aborts partway through
programming; reprogramming sometimes fixes that
- when a 628 fails for good, it fails the SAME way: I only read 3FFF
and the hi-voltage line into the chip is shorted to ground
I have tried several 628 chips from different batches bought from
Digikey. 16C73 and 16F87X still program 100% successfully, so it is
not the hardware or the software, it must be something quirky in the
628 which demands special attention. I have programmed 2 good 628
but also have 4 fried 628 on my hands at this point...
I suspect I need to switch to low-voltage programming for these babies
which will hopefully save me some chips :-), but would like to
understand the problem here. Any ideas?
-Mitch
Joe Kupcha
with my EPIC programmer made by MicroEngineering Labs. I called them and
they new about it and had a fix. I needed to cut a trace and add a jumper
on the ZIFF socket. I believe that this switched the low voltage
programming input pin (RB.4)from being connected to Vdd to being connected
to Vss. Its been awhile and I may have it backwards but there was
definitely a change to my programmer when I switched to a device which
supported LVP.
Hope this helps,
Joe Kupcha.
Mitchell Mlinar <mml...@gte.net> wrote in message
news:ni9rut44u4up576jf...@4ax.com...
Joe Kupcha
saying its only an input. I just looked and here is a snippet from the spec
sheet. "RA5 is a Schmitt Trigger input only and has no output drivers."
Hope this helps,
Joe
Mitchell Mlinar <mml...@gte.net> wrote in message
news:ni9rut44u4up576jf...@4ax.com...
Mitchell Mlinar
Joe pointed out, you cannot set RA.5 as output. I missed that line in
the doc...
So, after I set RA.5 as input, the chip no longer goes goofy but runs
properly, but that did not solve the HiV programming problem.
BTW, for the others who point out issues with the LVP pin, those have
been around since the F87x. When my programmer is in HiV programming
mode, the LVP pin is pulled to ground at all times. As someone
pointed out, you really only need do this the first time (if you set
the configuration word for HiV programming only), but the programmer
just does it as a matter of course. 5-wire programming is a
non-sequeter.
I am afraid to try further with HiV given I have LVP mode -- just a
lot safer...
-Mitch
On Tue, 13 Nov 2001 13:38:56 -0500, "Joe Kupcha" <cup...@znet.com>
wrote:
Tim H.
I'm not sure if this will be any help, but here goes....
I followed this thread a little bit and was under the impression that
you can't use HVP mode on them (16F628s) when they're fresh. So, I got
my batch of 5 yesterday. I too have a Tait-style programmer (w/out
inverters) but wrote my own software. So, I plop the 16F628 chip in the
socket and program like normal, it worked w/out any problems.
Programmed/re-programmed several times, still worked. So, given the
programming specs, I think that HVP is always available. They come from
the factory with the LVP bit set, yet I was able to program in HVP mode.
Btw, the LVP pin is floating in my programmer. So far it seems like the
16F628s are drop-in replacements. Nice to see Microchip keep the same
algorithm throughout their chips too.
Regards,
Tim Hamel