On Apr 6 2005, 0:08, Dan Williams wrote:
> It made me laugh and it's on topic.
>
> item=5181057334
Looks like I'm sitting on a potential fortune, then :-)
--
Pete Peter Turnbull
Network Manager
University of York
>Subject: Re: rx01 w/o controller board
> From: Pete Turnbull <pete at dunnington.u-net.com>
>Your host adapter is little more than a parallel interface; it talks
>over a 40-way ribbon cable (but less than half the pins are used) to
>one of the boards on the RX02 unit which has a state machine to execute
>various commands, and that board in turn connects via another ribbon
>cable to the lower board, which has the amplifiers and so forth on it.
> The head cables plug into this lower board.
The state machine is a simple but programable non-von cpu that can do
the needed tasks to execute complex command to seek, read or write
a sector. The RX02 is a 2901based bitslice that does more as it
has to do the needed tricks for DD format.
An aside for CPU hackers, the core cpu of the RX01 processor is remarably
simple and can execute 5 distinct microinstructions with a few variations.
A varient of the same logical core was used to build the LA34 printer and
LA120 and a similar design is in the LP26 DAFU. It was a fun project
years ago building one and tweeking it to do a simple 8bit cpu.
>AFAIK the drive mechanisms for an RX01 and RX02 are the same; only the
>boards differ.
Correct. The drive and power suppies are interchangable, The two
boards are different for the 01/02 though mechanically interchangable.
Allison
Well, I read through the digest replies and don't agree with all of them
... so :):
Re: the floppy cable. The twist on a HD cable is closer to the edge than
the floppy cable. It is obvious if you hold the two of them together
which is which. The floppy twist starts at pin 10 while the HD twist
starts about pin 6 or so (don't have one in front of me.) A floppy cable
will work okay but *don't* use the end with the twist!
Make sure the data cable is connected to the correct connector if the
card supports two drives (IIRC one 40 pin connector and two 20 pin
connectors.)
A terminator is also required at the last drive in the chain. Assuming
you are using just one drive, not having the terminator installed would
probably cause the drive not to work properly.
If you are using a HD that originally had a twist, the jumper select
will most likely be set to drive 1. Using a straight cable will cause
the drive to appear as drive D instead of C ... I don't remember what
problems that will cause. So if using a straight cable, make sure the
drive select is at 0 (of 0-3) or 1 (of 1-4).
If you are using an 8-bit card with the bios enabled, make sure in the
386 setup that there is NO HD installed (the bios will take care of it.)
If you have the bios disabled, then go ahead and set the 386 HD setup to
Type 2 or 4? (can't remember, 615 cyl, 4 heads, 17 sectors) but you
*will* have to do a low level format to get the drive to work.
If you are trying to save files on the HD AND it was installed with an
8-bit controller, you don't have any choices but to use the 8-bit card
bios to access the data AND an identical controller to the one that was
used to low level format the drive. Fred can probably comment on whether
it has to be the same type of controller, or the same controller for
this to work. If you have any idea of what the original machine it came
out of was, that would help :).
Finally, the clicking you hear on the drive is not a good sign. I would
power up the drive with no cables connected (except of course the power
cable) and if the clicking still continues, the drive is probably bad.
If it is important enough, you could also send the drive to a data
recovery service and leave it to them to deal with it.
Marvin
P.S. - I think this is an example where top posting makes everything a
lot easier to follow :)!
> OK, I posted about this a few weeks ago and now I have some more
> information.
>
> The drives in question are a ST-225 20MB (615/4/17) that came out of an
> XT (so would have been using a 8-bit controller) and a ST-238R 30MB
> (615/4/26) of unknown origin.
>
> I don't have or even know what the original controllers were, the
> machines were gutted/given away years ago by the owners.
>
> I have a 386 testbed which I'm using, and plan to laplink the files over
> to another machine once they are accessible.
>
> The only card in the 386 is a VGA card. Turbo is off.
>
> I have the following controllers:
> - WD1002A-27X 8-bit RLL
> - Adaptec ACB-2370A S2 16-bit RLL/floppy
> - WD1002SV-SR2 16-bit RLL/floppy
> - Everex EV-346 16-bit MFM/floppy
>
> I also have two Miniscribe 8438 30MB RLL drives. However, they were
> used with the 27X, which means the geometry is ambiguous depending on
> how the jumpers were set when it was formatted.
>
> Now, I'm either doing something wrong or all these drives are dead.
>
> I am using a floppy cable with a twist in it and a 20-pin data cable.
> All of the drives spin up (the Miniscribes needed a little coaxing) and
> sound "healthy" as they dance their little self-test jigs. Here's where
> the trouble starts.
>
> I tried the ST-238R with all three of the RLL controllers. With the 27X
> and with no hard drive entered in the NVRAM setup, since the BIOS can't
> be disabled, it runs the BIOS, which can't figure out what's going on
> and returns a 1701 POST code.
>
> With both of the 16-bit RLL controllers, 615/4/26 entered in the NVRAM
> and the controller BIOS disabled, the drive makes a repetitive seeking
> sound like an ECC retry about twice a second, until eventually the BIOS
> gives up and returns Drive C: failure. What this seems to indicate to
> me is that the drive was formatted with different geometry, or the
> tracks have drifted so bad it can't get its bearings.
>
> I also tried the Miniscribe drives with the 27X for kicks. Everything
> sounds normal, including the long growl that I remember that
> drive/controller combo doing during POST. But, a 1701 is returned,
> which isn't normal - this drive *used* to work with this exact
> controller.
>
> The only MFM controller I have is the Everex, so I tried the ST-225 with
> that one, entering 615/4/17 in the NVRAM and disabling its BIOS.
> Unfortunately, I get mostly the same behavior as with the ST-238R;
> except that with this drive, the retry clicks are about a second or two
> apart. Eventually the BIOS gives up anyway.
>
> I'm fresh out of ideas at this point. Maybe I have a bad cable? Do I
> need to find the exact controller the drives were paired with - what
> were the most common ones for each drive? Is the 386 the problem, too
> fast for these cards?
>
> --
> Ryan Underwood, <nemesis at icequake.net>
>
river wrote on Tue, 5 Apr 2005
>Does anyone else here build old stuff like this? Or do you build your own debug
>and monitor software? Any experimenters/builders here?
My answer is yes to all the above. I have many old processors and plan to put them
all to use someday. I've done something similar to your SC/MP project with an
Intel 4004. ( http://webpages.charter.net/bkotaska/mcs4_micro.htm )
Latest project is a redesign of the HP Programmable ROM Module for the HP-85
( http://webpages.charter.net/bkotaska/prm85.htm ).
Anything to make old computers (and chips) more useful!
Bill
Found another roll of 4.25" x 400' Black thermal paper
and some DC100A tapes including an HP-85 Standard pac,
00085-13001.
Will bundle with the other rolls.
Collect or postage from the UK.
Lee.
Send instant messages to your online friends http://uk.messenger.yahoo.com
>Subject: Re: rx01 w/o controller board
> From: Brad Parker <brad at heeltoe.com>
>
>
>Allison wrote:
>>
>>Brad did you buy a bare drive?
>
>yes, apparently I did :-)
>
>>If so you need the cab, power supply and both logic cards.
>
>So, it seems I need the 2 logic cards. All that for a floppy :-)
>
>thanks!
Well err, yes! DEC did things a bit different as the RX01 dates
to before the era of FDC chips. So what they did was encapsulate
both the drive electronics and a full floppy data interface into
the system. By doing that they also simplified stuff by way of
consolidation that was both expensive at the time and power hungry.
What that amounts to is the RX01 box is two dives and the interface
and all that's needed in the system is a fancy {but cheap to design
and build} parallel port to interface it to the bus. Very handy as
the RX01 could be used with PDP-8, PDP11 both Unibus and Q-bus.
As a result a nominal Sa800 drive is not even a close analoge to
the complete RX01.
So having the bare drive is like buying a car, well ok, the frame
and transmission for one.
Generally RX01 and RX02 drives sans the bus interface card are
not hard to find. Don't ship it though, they weigh in around 50+
pounds! Whats a good replacement is DSD880s, those are RX01/02
work alike boxes that are a contoller, floppy and some models
include a hard disk from 10 to 30mb in one case 6" high. Even
then you need both the complete box and the bus interface card
to match.
Allison
OK, I posted about this a few weeks ago and now I have some more
information.
The drives in question are a ST-225 20MB (615/4/17) that came out of an
XT (so would have been using a 8-bit controller) and a ST-238R 30MB
(615/4/26) of unknown origin.
I don't have or even know what the original controllers were, the
machines were gutted/given away years ago by the owners.
I have a 386 testbed which I'm using, and plan to laplink the files over
to another machine once they are accessible.
The only card in the 386 is a VGA card. Turbo is off.
I have the following controllers:
- WD1002A-27X 8-bit RLL
- Adaptec ACB-2370A S2 16-bit RLL/floppy
- WD1002SV-SR2 16-bit RLL/floppy
- Everex EV-346 16-bit MFM/floppy
I also have two Miniscribe 8438 30MB RLL drives. However, they were
used with the 27X, which means the geometry is ambiguous depending on
how the jumpers were set when it was formatted.
Now, I'm either doing something wrong or all these drives are dead.
I am using a floppy cable with a twist in it and a 20-pin data cable.
All of the drives spin up (the Miniscribes needed a little coaxing) and
sound "healthy" as they dance their little self-test jigs. Here's where
the trouble starts.
I tried the ST-238R with all three of the RLL controllers. With the 27X
and with no hard drive entered in the NVRAM setup, since the BIOS can't
be disabled, it runs the BIOS, which can't figure out what's going on
and returns a 1701 POST code.
With both of the 16-bit RLL controllers, 615/4/26 entered in the NVRAM
and the controller BIOS disabled, the drive makes a repetitive seeking
sound like an ECC retry about twice a second, until eventually the BIOS
gives up and returns Drive C: failure. What this seems to indicate to
me is that the drive was formatted with different geometry, or the
tracks have drifted so bad it can't get its bearings.
I also tried the Miniscribe drives with the 27X for kicks. Everything
sounds normal, including the long growl that I remember that
drive/controller combo doing during POST. But, a 1701 is returned,
which isn't normal - this drive *used* to work with this exact
controller.
The only MFM controller I have is the Everex, so I tried the ST-225 with
that one, entering 615/4/17 in the NVRAM and disabling its BIOS.
Unfortunately, I get mostly the same behavior as with the ST-238R;
except that with this drive, the retry clicks are about a second or two
apart. Eventually the BIOS gives up anyway.
I'm fresh out of ideas at this point. Maybe I have a bad cable? Do I
need to find the exact controller the drives were paired with - what
were the most common ones for each drive? Is the 386 the problem, too
fast for these cards?
--
Ryan Underwood, <nemesis at icequake.net>
>
>Subject: Re: The SC/MP is finally alive!
> From: "river" <river at zip.com.au>
> Date: Wed, 06 Apr 2005 07:44:02 +1000
> To: <cctalk at classiccmp.org>
>
>Hi,
>
>Yes, for it's time (1976) the SC/MP had the logic that allowed it to share
>the bus with other processors etc. The 8080 chip could also do this, but
>this device required 3 separate power supplies and you needed three chips
>to get the CPU working properly. The 6800 also offered similar multiprocessing
>abilities and, like the SC/MP was a single chip system, but I'm not sure if it
>came out before or after the SC/MP.
Commnet:
The biggest differnce between the SC/MP bus intnerface was that
the sc/mp didn't assume it was the bust master. So the bus
interface pins did a priority resolution with other potential
bus masters.
All of the other cpus (8080, 8085, z80, 6502, 6800 etal)
assumed they were the bus master and you took the bus only
after being granted DMA access.
While the difference is a small one it's notable in that all
of it's peers of the time were different.
>The SC/MP I (ISPA/500) was a PMOS device and required +5 and -7 volt power
>rails. The SC/MP II (ISPA/600) was NMOS and required only +5 rail (same as the
>6800), it also had three of it's CPU control lines inverted (as compared to the
>SC/MP I) and though it could take 4 times the speed of the clock of the SC/MP I,
>due to internal clocking it ran only twice as fast.
>
>What made the SC/MP popular back in those days was the cost. It was about a
>quarter of the cost of the 8080 and 6800 processors.
The cost factor was insignificant by 1978. The CPU was only part of a
systems cost.
>Finally, it's good to see some other old SC/MP dudes around and also others who like
>to build and program their systems from scratch.
I started a SC/MP cpu in TTL and Bit slice (2901) and then
abandoned it for a design that expanded it in some obvious
ways.
the sc/mp was in the same class for cpu hackers as the 1802,
6100 (PDP8 in cmos), and a few others.
A significant part of my collection and expeimentation is with SBC
(Single or Small Board Computers). I have SBCs for 1802, SC/MP 8a/500,
Nibble basic 8073 sc/mp, 6800, 6809, 6502 Kim-1, 6100 Intersil sampler,
IMSAI IMP-48 an 8035 SBC, Ti9900 Technico Superstarter board,
NEC TK80 (8080), several 8085, Z80, Z280 and Z8001 sbc of my
design, and T-11 (a 40 pin dip pdp-11) homebrew. The only one
I haven't played with is the 2650, 16032, 68000... yet.
Allison
>I gave Chris maybe 8 ~ 12 unique models IIRC. Sometime later he had
>a crash of storage space... I have almost no idea how many he kept.
I actually found users for most of them. Many of them were repeats of the
IIci, and LC 1, 2, 3 series. A number didn't work, and were combined into
functioning machines. All the 68040 machines (mostly Performa 63x) that I
didn't keep were given to new homes, as well as most of the LC models.
The IIci's were surprisingly hard to place and in the end most were
stripped for parts and traded off in pieces.
I did keep a few that were the "rarer" models (like the IIvx and MacTV,
as well as one each of the LC versions, IIci, and IIsi... I also kept
some of the 68040's, in fact, I currently run my mail and web servers off
them, the mail servers are running the Mac OS, and the web servers are
Mac68k NetBSD).
I'm no where near having one of every model however. At one point I
wanted to make such a collection, but alas, it will be some time before I
have that kind of storage space available to me again.
>Somebody just gave me my first G3. Very good looking machine.
>Even has that prominent DEC chip (21154 PCI) to look at. Can't wait
>to load something on it.
Which model? I can check to see if I have the Apple restore CD for it.
-chris
<http://www.mythtech.net>
