Searching \ for 'Single stepping a PIC16F84 programmer' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/microchip/devprogs.htm?key=programmer
Search entire site for: 'Single stepping a PIC16F84 programmer'.

Truncated match.
PICList Thread
'Single stepping a PIC16F84 programmer'
1999\03\10@042916 by Jxrn Lind-Nielsen

flavicon
face
Hi,

I have just assembled Steve Marchant's PP84 PIC16F84 programmer
(http://www.ccc.nottingham.ac.uk/~cczsteve/pic84.html) and (no surprise
here I guess) it only works on selected (read: old) parallel ports.

Assuming that the programmer actually works and the programming software
(which I have written myself) is correct - would it then be possible to
single step the software, using a debugger, and still program the PIC
correct? It is, after all, a clocked programming, so I would expect it
to work.

Another question - can somebody point me to some links about the
noice/grounding/whatever problems related to the parallel ports?

-- Thanx, Jorn Lind-Nielsen

1999\03\10@114858 by Marc

flavicon
face
> I have just assembled Steve Marchant's PP84 PIC16F84 programmer
> (http://www.ccc.nottingham.ac.uk/~cczsteve/pic84.html) and (no surprise
> here I guess) it only works on selected (read: old) parallel ports.

I have not seen this programmer. But if it does not have an external power
supply, it is not a good one.

Make sure that 13V programming voltage is _stiff_ and has _quick_ rise
time. If it is slow, the PIC might start executing instructions while
MCLR travels through 2-7V on its way to 13V.  Executing instructions
increments PC, which is used as address pointer during programming. All
words in the PIC will be off by 1 or 2. This happens especially with
RC or external clock setup, because those startup with no delay.

I made a programmer once with the 14V/0V decision made by an NPN and
pullup to 13V. It didn't work reliable for that very reason: the rise
time was just too slow.

> Assuming that the programmer actually works and the programming software
> (which I have written myself) is correct - would it then be possible to
> single step the software, using a debugger, and still program the PIC
> correct? It is, after all, a clocked programming, so I would expect it
> to work.

With the F84, yes.

> Another question - can somebody point me to some links about the
> noice/grounding/whatever problems related to the parallel ports?

If you suspect ground loops, battery power the PIC circuit as well as
the programmer. If it suddenly works, you have located the problem.

1999\03\13@050631 by paulb

flavicon
face
Marc wrote:

> Make sure that 13V programming voltage is _stiff_ and has _quick_ rise
> time. If it is slow, the PIC might start executing instructions while
> MCLR travels through 2-7V on its way to 13V.  Executing instructions
> increments PC, which is used as address pointer during programming.
> I made a programmer once with the 14V/0V decision made by an NPN and
> pullup to 13V. It didn't work reliable for that very reason: the rise
> time was just too slow.

 Ahem!  What you mention applies to ISP programmers.  It does *not*
AFAIK apply to stand-alone programmers as - they have no crystal or RC
to oscillate!  (Unlike AVR programming.)

 No oscillator - no clocking - no execution of instructions/ no
increment of PC.

 If you want to be sure, provide a link (across the capacitor) to
ground OSC1 whilst programming.

 If you have no crystal or OSC1 is grounded and it still malfunctions,
look elsewhere than the MCLR risetime.
--
 Cheers,
       Paul B.

1999\03\16@033656 by Jxrn Lind-Nielsen

flavicon
face
Thanks to the people who helped me get my programmer working.

> Assuming that the programmer actually works and the programming software
> (which I have written myself) is correct - would it then be possible to
> single step the software, using a debugger, and still program the PIC
> correct? It is, after all, a clocked programming, so I would expect it
> to work.

The answer to my own question is now:
  It is possible to single step the programmer - but it fails more
often than not.


- J¿rn Lind-Nielsen

More... (looser matching)
- Last day of these posts
- In 1999 , 2000 only
- Today
- New search...