From: Spiro Trikaliotis on
Hello,

petela(a)mudplace.org wrote:

> I'm an embedded software engineer by trade, so I'm quite comfortable
> writing PIC/AVR/8051/etc code, and have been considering some of the
> newer (and cheap!) USB microcontrollers out there. So I've started
> working on a project to do the USB<-->CBM cable. As part of my
> background search, there is reference to the XU1541 cable, but all the
> links I have found are broken. Does this cable still exist? Are
> there archives of the work that was done?

See here:
http://www.trikaliotis.net/xu1541

Note that the links are not (yet) working. However, the sources and the
schematics (Eagle format) are in the OpenCBM CVS.

> I saw some discussion of integrating the XU1541 into OpenCBM. I've
> downloaded OpenCBM, and it appears to be a toolset plus a low-level
> driver. I haven't explored it much, but it appears to bit-bang the
> parallel port via the X1541 (et al) cable.

Right.

Note that the integration into OpenCBM is done. In fact, the only part
still missing is the installation, which needs some manual steps at the
moment.

I am currently working on this. As soon as I am not ill anymore, I hope
to be able to finish it. Thus, some testers would be good (then).

> Not to stir up the USB vs parallel port debate, I wanted to announce
> my own work on a USB cable. I have created a website and blog (my
> first ever, so don't expect much) documenting my ideas and the
> project. I'm open to ideas, and help. Check out my blog at
> http://mudplace.org/?cat=u1541 for more information.
>
> I think the biggest barrier to the USB version of the cable is cost.
> X1541 cables are very inexpensive and easy to make. A USB solution
> requires several more dollars (probably close to $30 in parts in small
> quantity purchases).

IIRC, the XU1541 cost less than EUR 15,- - that is, ignoring the
soldering, only the parts.

The important part for the XU1541 which makes it so cheap is that the
USB protocol is performed completely in software.


> Anyways, if anybody is interested, I'd be glad to hear from them.

If you want a solution and to integrate it into OpenCBM, please

1. Use the current CVS version of OpenCBM, and

2. contact me so we can work together on this.

Regards,
Spiro.

--
Spiro R. Trikaliotis http://opencbm.sf.net/
http://www.trikaliotis.net/ http://www.viceteam.org/
From: Dombo on
petela(a)mudplace.org schreef:
> Not to stir up the USB vs parallel port debate, I wanted to announce
> my own work on a USB cable. I have created a website and blog (my
> first ever, so don't expect much) documenting my ideas and the
> project. I'm open to ideas, and help. Check out my blog at
> http://mudplace.org/?cat=u1541 for more information.

I've been playing with that idea for sometime. I have look at AVR based
solutions, and I'm convinced it is doable. Unfortunately I have already
too much stuff ordered that I never got around to put to good use (due
to a structural lack of time).

> I think the biggest barrier to the USB version of the cable is cost.

Maybe it isn't that bad; today I saw a kit for �8,- which could, with
very few additional parts, turned into a 1541-to-USB converter:
http://www.samenkopen.net/action_product/878690/767990

With AVR it is pretty easy to have cycle exact timing, and they are
pretty fast too (compared to a 6502). I believe it is possible to speed
up the transfer between the AVR and the 1541 with a dedicated transfer
protocol.

Good luck with your project.
From: Sam Gillett on

"redrumloa" wrote ...

> Ignore any noise, options are good. And yes, legacy devices like
> parallel ports are getting scarce. My recent motherboard has a built
> in floppy adapter, but it doesn't even work! All legacy devices are
> going bye-bye :-/

Last time I checked you could get a USB 3.5" floppy drive. Commodore was
ahead of their time with an external, serial port drive when they came out
with the 1581. ;-)
--
Best regards,

Sam Gillett

Change is inevitable,
except from vending machines!



From: Suudy on
Oops. Google chopped off the bottom when I was reading, so I missed
part of this message. My response is below.

On Jan 21, 10:43 am, Jim Brain <br...(a)jbrain.com> wrote:
> However, instead of re-inventing the wheel, join the OpenCBM-devel list
> (https://lists.trikaliotis.net/listinfo/opencbm-devel) and see if anyone
> has the firmware.  It was an AVR-based solution, with soft USB.  If you
> want to use a newer AVRUSB device, I am sure you can ditch the soft USB.
>   A nice bonus is that XU1541 support is built into the OpenCBM
> routines, so no integration costs.

I'll do that. In my musings on my blog, I was considering the
implemention of the protocol between PC and the cable. And OpenCBM
seemed a good choice. And already existing integration into a toolset
would be a huge bonus. No front-end work would be required, just the
back-end. And I could steal a lot from the XU1541.

It sounds like dropping the soft USB would be a big help in terms of
performance. It looks like the ATmega8 and AT90USB use the same
processor core, so there won't be a big jump in internal performance.
However, offloading all the USB stuff into hardware, the additional
RAM and flash, and that the AT90USB can be run up to 16MHz help a
lot. Also, the AT90USB supports external memory, so it is possible to
stuff up to an additional 64K of external SRAM to help.
From: Simon Scott on
On Mon, 21 Jan 2008 18:43:11 +0000, Jim Brain wrote:

> It does exist, but political infighting or something made the original
> author grow uninterested in the project. At the beginning of the year,

Political in-fighting concerning the production of a cable to connect an
antique disk drive to a PC?

Now I have, literally, heard everything.