[dorkbotpdx-blabber] Designing a Parallel-to-USB Converter

Wed Feb 25 21:10:09 EST 2009

I was actually questioning the gate at all. Strobe the data in with a  
pin change interrupt.

On Feb 25, 2009, at 4:02 PM, Greg Grunest wrote:

> Nice schematics.  Looks like a cool project.
>
> A couple of questions / suggestions:
>
> 1) I see an inverter on the strobe line of the parallel port.  I  
> don't see
> it on the eagle layout.  Maybe I'm looking at the wrong one.
>
> 2) You're using an edge triggered device (the 374) for a parallel port
> application. You said you had it prototyped.  Try dropping in a 373  
> instead
> and take out the inverter (74LS04) on the strobe line.  The 373 is  
> level
> triggered.  If for some reason you run long distances with your  
> parallel
> cable or through a parallel switchbox the increased impedance in the  
> cable
> can cause the strobe signal to rise too slow causing edge triggered  
> devices
> (the ls374) to kind of "hang" in the middle and oscillate creating a  
> huge
> amount of system noise and rendering the data in the latch  
> unusable.  If you
> try the ls373 instead, it should work and not require the inverter  
> (and
> doesn't have the noise issue).  All your software should be the same.
> (I know from experience that a 373 will work through 50 ft of  
> parallel cable
> with a 3-way switchbox in the middle and a 374 wont.)  Really  
> though, I'd
> only change this if you have a distance problem or if you need to  
> get rid of
> an inverter.
>
> 3) Depending on what your final application is, some software (none  
> I ever
> wrote.  BAD PROGRAMMER, BAD!) *will* actually use the NACK and BUSY  
> signals.
> You might tie them into unused ports on the AVR chip and that way in  
> the
> future you could implement them in software if needed.
>
> 4) No support for bidirectional parallel ports??  SHAME!  :)
>
> -greg
>
>
> -----Original Message-----
> From: dorkbotpdx-blabber-bounces at dorkbot.org
> [mailto:dorkbotpdx-blabber-bounces at dorkbot.org] On Behalf Of Tim Hatch
> Sent: Wednesday, February 25, 2009 3:29 PM
> To: dorkbotpdx-blabber at dorkbot.org
> Subject: [dorkbotpdx-blabber] Designing a Parallel-to-USB Converter
>
> Hi dorkbotters,
>
> This is my first post on the list, I usually just lurk because I'm
> actually in Santa Barbara.  I was wondering if anyone could offer some
> hints on designing an AT90USB-based peripheral.  I need it to be able
> to input parallel (yes, ieee-1284) and re-send the data over usb.  I
> have a proof of concept using a Stickduino along with some TTL logic
> that works great.  I've begun laying out a surface mount version with
> an AT90USB162, but want to get it right the first time.  Eagle files
> and image of schematic:
>
> http://timhatch.com/ex/2009/02/parallel_usb/
>
> It's loosely based on the AVRopendous and Benito, which in turn appear
> based on AT90USBKEY.  I have a little experience with surface-mount,
> but some parts of EE are a little murky, as I'm trained in software.
> This is essentially a weekend project that my job now wants to be able
> to use, but I'd rather learn the why's so I can do this again for
> other projects.  The office has also agreed to make the result of my
> learning experience open, so anyone can use it.
>
> The second revision has the approximate outline I want to fit inside
> (the giant drills are holes for mounting screws), and is basically the
> same schematic (I think I just yanked out a few LEDs for space
> reasons).  It's actually more like the 8th version, as I keep ripping
> up all the traces and moving stuff around.  If you haven't already
> clicked the link, I have screenshots there that might be helpful for
> following along.
>
> Goal
> ------
>
> Read 8 bits of input on the falling edge of the strobe line, send to
> the host over usb.  (I have this much working already on a Stickduino
> which uses FT232R, with the same inverter/latch combination using DIP
> parts).  This is of course easiest if a whole port is available, but
> on the Stickduino PC6 acts funky (always LOW), and I'm hoping that
> PORTB on the AT90USB162 behaves.  I plan to make it run at 16MHz, for
> no other reason than I know the code I have now works at 16MHz.
>
> Questions for everyone
> -------------------------------
>
> (Note, I won't be offended at all if you just answer the question with
> a link...)
>
> Q: What kind of logic goes into choosing a crystal/oscillator?
> I found the AWSCR-16.00MTD-T which already has caps built-in (making
> fewer joints for me to use), but I don't see any other boards using it
> so I'm wondering if it's inappropriate here.  As far as I can tell,
> the only SMT oscillator used for other AT90 boards is digikey
> 644-1132-1-ND or related, and all the Arduino boards use through-hole
> HC49's or ceramic oscillators.
>
> Q: What kind of logic goes into deciding whether to use the built-in
> 8MHz oscillator?
> Is it just accuracy, with the 10% tolerance mentioned in the  
> datasheet?
>
> Q: Are there any special rules relating to trace length of USB lines,
> xtal lines, or the giant cap next to the AT90?
> I'm trying to keep them as short as possible, but wondering if they
> shouldn't come near each other, for example.  Is it a problem to run D
> + on one layer and D- on another?
>
> Q: Am I doing the best thing with the tac switches?
> The switches I have on there are huge, given that I'm trying to fit it
> inside a DB25 hood.  I notice Paul S has done something cool with a
> BAT54C diode on the Teensy[2] -- I can see it would save space, and if
> you're only using HWB for bootloader you don't lose anything.  It
> appears that most of these tiny boards use Digikey 401-1427-1-ND for
> their buttons (which isn't in any Eagle libraries I've found)?
>
> Q: Is the only difference between the AT90USB162 and AT90USB82 the
> flash size?
> My current code compiles to 1700 bytes or so, do you think it's worth
> switching to the -82 even though I have zero experience with it?  What
> would the reasons be (power? a few cents cheaper?)
>
> And lastly, I'm planning to use a through-hole USB Mini-B connector,
> and solder wires directly for a captive cord.  Do you foresee any
> problems with that?
>
> Tim
>
> [1]: My day job is doing some Python-based software for Nutricate -
> their website is http://nutricate.com/ but be warned there's lots of
> Flash.
> [2]: http://www.pjrc.com/teensy/schematic.html
> _______________________________________________
> dorkbotpdx-blabber mailing list
> dorkbotpdx-blabber at dorkbot.org
> http://music.columbia.edu/mailman/listinfo/dorkbotpdx-blabber
>
> _______________________________________________
> dorkbotpdx-blabber mailing list
> dorkbotpdx-blabber at dorkbot.org
> http://music.columbia.edu/mailman/listinfo/dorkbotpdx-blabber

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://music.columbia.edu/pipermail/dorkbotpdx-blabber/attachments/20090225/962ed780/attachment-0001.html