gnupic: DIY USB programmer ?
Subject:
Re: DIY USB programmer ?
From:
David Willmore ####@####.####
Date:
13 Jan 2005 18:44:26 +0000
Message-Id: <200501131843.j0DIhmQH011850@localhost.localdomain>
> > Exactly. The 556 can be used in two places in the circuit--once for the
> > bootstrapper and once in a HV power supply switcher. That keeps the
> > micro from having to walk, chew gum, pat its head and rub its belly all
> > at once. ;)
>
> You'll have to change up some parts to get it going as I'm sure that the
> frequency for the HV will have to be higher than the delays for the
> monostables.
I was thinking of having two different places on the board for the chip to
fit in. It would cost one socket, but the bootstrap 'option' for the kit
would be the socket for the 556; an E9 shell; and the handful of caps,
resistors, and diodes. From a company that just sells boards and simple
kit parts, they can sell it with the bootstrap option--as they are unlikely
to have PIC programming ability. But places that can program the PIC will
just not supply the 'bootstrap' parts.
For those who just want to build the think from digikey parts, they can so
with just a bare board--buying the bootstrap parts themselves.
So, don't worry about the circuit around the 556, we'll move the whole
chip. :)
> I took a 5 minute look at Pikdev. It'll require adding a new type of programmer
> separate from the parallel and serial programmers defined now. Fortunately
> the class only has a handful of definitions. I didn't see how to connect the
> type to the class though. It should be too difficult to pull off.
I thought it would fit in pretty easily. You'd need to make a new programmer
dialog for it, but it would only have to have a pulldown for serial port as an
option--maybe we'd need some calibration stuff in there too, huh?
> I think I'm going to adopt pikdev for my programmers since it has 18F support
> and works with the Trivial programmers.
I've been using it for use with your THVP for a while now and I've had great
luck. For those PICs that it didn't support, I found that it wasn't that
hard to add it. Of course, I have to finish submitting that... Sorry Alain...
> > :) Well, for my part, I'm going to try to finish up some Pikdev chip
> > support code (sorry Alain) and watch Byron work his magic! :)
>
> I hope to get a chance to work on it some more between now and the weekend.
> I'm continuing to think on the verify problem that I outlined in my
> preliminary results post. I realize that I need to isolate the PIC Program
> Data pin from both the incoming TX circuit (which I was going to do anyway)
> and the outgoing RX circuit. So it'll take a couple more transistors.
NPNs if they don't have any special requirements are damn near free, so don't
feel bad. Just follow these kit building guidelines: If you have to use an
expensive part, use it. If you can avoid it by using a few common parts,
do so. If you've already used a part using it again is fine.
Having 'kitted' a few hundred kits, can safely say that the worst kits to
prepare are the analog ones where you have *one* of this odd value, *one*
of that, *one* of some other identical value, but 1% tolerance, etc. I love
digital designs that are like "C1-C50 .1uF momo cap, R1-R18 10K 5% 1/8W,
R19-R40 1K 5% 1/8W" Woooo Hooo! That's three parts as I count them. :)
> I'm starting to think that a modular design may be possible without too
> much effort. That way you can build the design with or without verification.
I'm always leary of capacitors as they have the largest tolerance specs of
any common electrical part. Resistors you can get in any arbitrary tolerance.
Inductors tend to be available in better than 10% tolerances. But caps,
sheesh. Common 'lytics are +70%/-20% tolerance! Do a timing circuit with
them, not without calibration. Now, expensive mylar caps can be 1%, but
they're hard to find and they're not cheap.
That's just my take on kitting. Maybe someone else does it differently--and
they'd like to interject with their design guidelines.... :)
Keep up the good work, Byron!
Cheers,
David
