Re: Resending: Repairing a SFD-1001 (8250/LP, 8050, 4040, 3040, 2040)

From: Wolfgang Moser <womo_at_news.trikaliotis.net>
Date: Sun, 14 Aug 2011 23:45:03 +0200
Message-ID: <j29fgi$aev$1@vs5413.trikaliotis.net>
Hello Steve,

Steve Gray schrieb:
> Excellent email. Thanks for re-sending!

thanks.

> A while ago I discussed creating a replacement RRIOT with Jim Brain. We came to 
> the conclusion that it would be possible to create an adapter with 6532 chip 
> that is about the same size as the socket by using a small surface mount FLASH 
> ROM underneath, thereby making it universal to any drive.

There is no need to create an adapter that small. I found a way to
create a layout that is nothing else than a direct 8250LP adapter
replacement. By shrink wrapping the replacement a bit, it now also
perfectly fits into a SFD-1001 board. A friend already verified that the
adapter also fits into a 8050 mainboard, where there is plenty of free
space around the RRIOT chip and no ceramic capcitors eggting too high.
Until now I don't see a need for creating a SMD version of the adapter,
if not costs are the reason.

> A flash rom would make 
> dip switches unnecessary and the whole thing a lot smaller.

I already used FlashROM, but nevertheless used DIP switches do easily
change different preprogrammed ROM banks. To have the FlashROM
programmed in-place this would required a bit more glue-logic and pretty
much programming efforts to have a reliable flashing tool that's able to
recognise all the different Flash types.
Currently my PC based Flasher programs the chip regardless of its type,
it virtually knows thousands of types and I'm not in the need of reading
all the datasheets and constantly extending a tool for havin new chips
supported.

> Unfortunately that's 
> about where it ended, since demand for such an adapter was unknown and Jim's 
> time is understandably limited.

Maybe I can jump in here. But my time is also limited so please don't
expect some boards made before chrismas this year. But a little
prototype run of profesionally manufactured PCBs would be feasable and
well suited to actually test for true demand.

> Old Atari 2600 consoles can be a source of 6532 chips if you don't mind 
> de-soldering.

Yeah, but that would be much too costly. For one, Ataris are not _that_
cheap on eBay even if defective and then think about of all the work of
getting the chips pulled. I'm very sure that some professional or
semi-professional electronics guy would be able to source bigger amounts
of 6532 chips from some of these China "New Old Stock" retailers.

> Anyway, I have at least one SFD unit that is waiting for a replacement RRIOT, so 
> I'm very interested in your solution. I have most of the commodore IEEE drives 
> here if I can help with dumping the code. I also have 3 or 4 space SFD 
> motherboards missing many chips, but there might be some usable chips left.

Same here. As I wrote, I got a couple of boards and replacement parts
from a huge inheritance sales that took place on eBay Germany some years
ago. Also many SFD boards with missing parts, only the DOS ROM and the
GCR ROM were left in.

> It's true that the 100tpi mechanisms are also impossible to find now. I could 
> use a few myself. Due to the different "standards" any replacement mech will 
> have 96tpi so you'd have to look at a mechanical modification to make them 
> usable. The difference is very slight so perhap all that is needed is a 
> new wheel for the stepper motor.  I suppose if you really wanted you could use a 
> 96tpi mech as-is and at least you could have a working unit, but you would 
> sacrifice read/write compatibility with actual 100tpi formatted disks.

I will write up another follow-up article on this special topic.
Currently I believe that the 100tpi-vs.-96tpi mechanism issue is fixable
by replacing the 96tpi stepper motor with the 100tpi one from the
SFD/8250LP drive. Of course this requires a mechnism that makes use of
the same or at least a very similar stepper motor type and some
mechanical preparation.

Personally I'm really concerned about the R/W head itself, its
resistances and impedances of the two read and write coils as well as
the two erasing coils of the tunnel erasing head (most floppy R/W heads
are constructed as tunnel-erasing heads).

When measuring the resistances of the four coils and not checking the
impedance I cannot find any other floppy mechanism, neither from Quad
Density drives nor from High Density or Dual Density (not Double
Density) drives that would match the coil resistances of the Matsushita
JU-570/-2 drive.


What I would be most interested in is, if it is easily possible to adapt
the analog electronics of the SFD/8250LP to compensate for other
resistance and/or impedance values? How is this done? What needs to be
changed?


Womo

-- 
  ------ to obtain more infos about me, look up the page ------
    ------ http://www.wmsr.de | wm (at) wmsr (dot) de  ------

       Message was sent through the cbm-hackers mailing list
Received on 2011-08-14 23:00:22

Archive generated by hypermail 2.2.0.