David Tait wrote:

> The Maplin programmer is more or less the same as the ASCII sketch in
> the file pic84pgm.zip (all programs mentioned are in my PIC archive).

  OK, got that.  Diagram is fine.

> It is a kit and I'm sure some people have difficulties because they
> have not built the thing correctly.

  I'm sure that would be right!  That's why they don't warranty kits!

> If my designs are really "dodgy" and causing people lots of
> frustration it would pain me a lot.

  Hard to tell, isn't it?  Especially given the point above.

  My point about part substitution is that firstly, I'm uncomfortable
with 4066 gates being used to provide programming current; in the case
of the 16F84, this relates not to Vpp but to Vdd which is specc'ed at
50mA maximum which the 4066 is most certainly not designed to supply.  I
realise this is by no means usual, but ... that's the spec.

  I'm not even sure a 74HC4066 is intended to supply this current when
HCMOS gates in general are not.  I'm much more comfortable with Don's
style of circuit using PNP transistors (which are probably cheaper
as well!).  To be quite honest, I can't figure why David picked the 4066
in the first place, but then I have had some "fun" using a 4051 to drive
LEDs many, many years ago which put me off using analog gates in this
manner.

  Secondly, the description of the gate used as "TTL O/C buffers,
74xx06, or 74xx07" worries me.  Only the 7406/ 7407/ 7416/ 7417 and the
Texas Instruments brand of 74LS06/ 74LS07 are specc'ed to operate at
12 volts; other brands of LS gates, all HC gates and *all* 74xx05 are
specifically *not* designed for this application.  They may perhaps
work, but the design sheets specifically say they *won't*!

> Nowadays I can only deal with a fraction of the mail that my
> programmer stuff and web pages attract (which gives me a permanent
> guilty feeling) but at one time I spent a lot of time trying to help
> people get their programmers working.

  No surprises there.  But it's a bit of a challenge for others to do
so!

> On balance I think my stuff is still useful but perhaps I'm wrong.

  Quite so, and by no means.

  I still want to know what ICs Andy has in his kit, what results he
had on the "commissioning" test (setting the port to the appropriate
values and testing the voltages at the programming socket, using the
appropriate 100 ohm "dummy load" across pins 5 and 14), and those tests
passed, exactly what could and could not be done with the device.

  In fact, http://www.dontronics.com/dt001cha.html#testing gives a good
outline of the steps, even though this is a different version.
--
  Cheers,
        Paul B.