5 Jun
2007
5 Jun
'07
11:39 a.m.
I've been checking through the stuff I have around here and I've decided
that I have a few surplus PC mobos that I'll be getting rid of (via a
certain auction site). Before I do I thought I'd try them out with
TESTFDC
and see what it says.
I have three 5.25" drives (of currently unknown functionality) which I
could use. One is a DEC RX33, one is a Tandon TM75-8 and the last one
is a Sankyo (or that might just be the motor manufacturer :-)). I also
have whatever is sitting in my main machine, but I'm not keen to
drag that out just foir fun! I also have a few 1.44MB 3.5" drives, but
they've not been used in quite some time, so heaven knows what state
they are in.
So what do I need to do to extract the maximum useful information
about these motherboards (aside, that is, from checking that they
are not already on the list :-))? And exactly what media will I need?
Should I test both 3.5" and 5.25" (I've not tried the 300/360 mod yet;
I may do in the future, but for now I'd just like to get some
testing done)?
Anything to watch out for?
Any particular version of DOS recommended (I have 6.2? kicking around,
would FreeDOS do if that is easier to find)?
Thanks
Antonio
6 Jun
6 Jun
6:34 a.m.
I have three 5.25" drives (of currently unknown
functionality) which I
could use.
Drives of known functionality will usually give more meaningful TESTFDC
results.
One is a DEC RX33, one is a Tandon TM75-8 and the last
one
is a Sankyo (or that might just be the motor manufacturer :-)). I also
have whatever is sitting in my main machine, but I'm not keen to
drag that out just foir fun! I also have a few 1.44MB 3.5" drives, but
they've not been used in quite some time, so heaven knows what state
they are in.
I would recommend testing the drives on a known good system first.
Unless you know the drives are good, you can't trust the TESTFDC
results (A failure could be indicative of a bad drive, not a FDC
limitation).
So what do I need to do to extract the maximum useful
information
about these motherboards (aside, that is, from checking that they
are not already on the list :-))? And exactly what media will I need?
Should I test both 3.5" and 5.25" (I've not tried the 300/360 mod yet;
I may do in the future, but for now I'd just like to get some
testing done)?
To completely test the FDC, you need to check Single, Double and
Double/128 at each of 250, 300 and 500 kbps rates. These are
represented in a standard PC at:
250 = 5.25 low-density drive only, 3.5" low-density diskette
300 = 5.25 Low-density diskette in high-density drive
500 = 5.25" or 2.5" high-density diskette
So to check all possible data rates without a 300/360 modded drive,
you need to run two tests - either the 5.25" LD and HD drives, or a
5.25" HD and 3.5" drive.
Anything to watch out for?
As noted above, make sure your drives are working - Ideally they should
have recently passed the tests you are about to run when installed in
a system with a known good FDC.
Any particular version of DOS recommended (I have 6.2?
kicking around,
would FreeDOS do if that is easier to find)?
It shouldn't matter - TESTFDC talks to the hardware directly. Just don't
have any TSR's or other resident software loaded which might touch the
FDC (best to run under bare DOS).
Dave
--
dave06a (at) Dave Dunfield
dunfield (dot) Firmware development services & tools: www.dunfield.com
com Collector of vintage computing equipment:
http://www.classiccmp.org/dunfield/index.html
10:02 a.m.
Dave Dunfield wrote:
I would recommend testing the drives on a known good
system first.
Unless you know the drives are good, you can't trust the TESTFDC
results (A failure could be indicative of a bad drive, not a FDC
limitation).
What's "good enough" testing? As a minimum I'll boot DOS,
format a floppy with the drive and put DOS onto the floppy, then
boot the floppy. Is that enough or are there oddball quirks that
might be better exposed by something else?
So to check all possible data rates without a 300/360
modded drive,
you need to run two tests - either the 5.25" LD and HD drives, or a
5.25" HD and 3.5" drive.
Good. I should be able to rustle up at least one of those combinations.
Thanks
Antonio
10:26 a.m.
Antonio Carlini wrote:
Dave Dunfield wrote:
A few of the PC motherboards I found were highly dependent on the drives used
as to what tests they'd pass. Mitsumi drives seemed to cause a lot of
problems, and I had one board which refused to pass tests with any 3.5" drive
despite passing the same tests with any 5.25" drive.
So yes, there are quirks (through no fault of testfdc) :-(
I would recommend testing the drives on a known
good system first.
Unless you know the drives are good, you can't trust the TESTFDC
results (A failure could be indicative of a bad drive, not a FDC
limitation).
What's "good enough" testing? As a minimum I'll boot DOS,
format a floppy with the drive and put DOS onto the floppy, then
boot the floppy. Is that enough or are there oddball quirks that
might be better exposed by something else?
12:25 p.m.
I would recommend testing the drives on a known good
system first.
Unless you know the drives are good, you can't trust the TESTFDC
results (A failure could be indicative of a bad drive, not a FDC
limitation). 
12:16 p.m.
There are some NEC and Mitsubishi 3.5" drives that have a fair number
of additional signals on the odd-numbered connector pin side--usually
used on CNC and some lab equipment on 9801-series architectures.
When replacing these, we've found that any old 3.5" drive won't do--
an exact replacement is necessary.
I'm surprised no one mentioned 3.5" DSED testing.
Cheers,
Chuck
4:17 p.m.
There are some NEC and Mitsubishi 3.5" drives
that have a fair number
of additional signals on the odd-numbered connector pin side--usually
used on CNC and some lab equipment on 9801-series architectures.
When replacing these, we've found that any old 3.5" drive won't do--
an exact replacement is necessary.
I guess I should have been more specific... the information I posted
was specifically about the setup you are likely to find in PC drives.
There are lots of variations in possible drive types if you open up
to "anything", but I didn't (and still don't) feel that discussing all
the things you are not likely to see on a PC would make the water any
less muddy for those guys who are trying to figure out what they need
to do to characterize their PC FDC.
I'm surprised no one mentioned 3.5" DSED
testing.
Neither TESTFDC nor IMD support ED, simply because I don't have an ED
compatible drive - I don't think I've ever run across one. If I can't
test it, I can't reliably implement it (anyone want to send me a drive? :-)
I also don't find ED very relevant to archiving vintage computer diskette
images - are there any systems in need of such attention that use ED as
their standard disk format?
Dave
--
dave06a (at) Dave Dunfield
dunfield (dot) Firmware development services & tools: www.dunfield.com
com Collector of vintage computing equipment:
http://www.classiccmp.org/dunfield/index.html
4:03 p.m.
On 6 Jun 2007 at 18:17, Dave Dunfield wrote:
I also don't find ED very relevant to archiving
vintage computer diskette
images - are there any systems in need of such attention that use ED as
their standard disk format?
I do have a few IBM PS/2 distros on ED. I am aware of a few
companies selling into the PS/2 market for whom 2.88M was the
distribution medium. The Teac drive that will handles these in
either media-sense or host-specified mode is the FD-235J.
IBM used (IIRC) host-side density control on much of the PS/2 line,
so you *could* format anything from 720K to 2.88M on a plain-Jane
DS2D disk. Maybe the result wasn't entirely reliable, but it could
be done.
The "right" media is a barium ferrite formulation and has the density
aperture closer to the bottom edge of the disk than do the DSHD
diskettes. If you stick one of these things in a 1.44MB-only drive,
it appears as a DS2D.
As far as "PC drives", those Mitsubishi/NEC 360RPM drives ran MS-DOS
2.10 on their host system (the flavor is sometimes called "DOS-V").
Diskettes were generally formatted as 8x2x1024x77 or 8x2x1024x80;
some of the CP/M formats used 256-byte sectors.
Even Windows NT running on a 9801 supported that format--as does (in
read-only mode) Windows XP/2000 today. As far as Microsoft back in
the day was concerned, the 9801 platform was a PC, just not a
"Western" PC. At one time, the 9801 platform owned something like
70% of the PC market in Japan.
Hope I got this all correct.
Cheers,
Chuck
5:11 p.m.
Fred Cisin wrote:
On Wed, 6 Jun 2007, Dave Dunfield wrote:
The *early* NeXT standard disk format was 256MB magneto-optical.
ED stiffies came later.
-Frank McConnell
I also don't find ED very relevant to
archiving vintage computer diskette
images - are there any systems in need of such attention that use ED as
their standard disk format?
early NeXT? 
5:45 p.m.
On Wed, 6 Jun 2007, Frank McConnell wrote:
The *early* NeXT standard disk format was 256MB
magneto-optical.
ED stiffies came later.
Ah, yes.
Thanks for the correction.
NeXT called the ED's "4Meg", which was the unformmated capacity.
IBM called them 2.88M, which is the formatted capacity when using a bogus
definition of megabyte.
(80 * 2 * 36 * 512)
The dividing line between regular and obscure systems is not always clear.
Some people claim that dual speed 1.2 was as common as 360RPM.
Outside of the USA, 360RPM 3.5" wasn't as rare.
2.8M was popular in some niche markets, along with 20M floptical, LS120,
Kodak/Drivetec 6? M, 12M? 5.25", Weltec 180RPM 1.2M with 250K data
transfer rate, etc.
'course some of those drives used on PC's were not a "normal" SA400
style
interface, including parallel port interfaces, and some used SCSI.
And, of course, there are some with special harware that hardly ever were
on PC's, such as hard-sectored, 2.5", etc.
If this ever does make it to a website write-up, it should cover the
"normal" ones, and have a paragraph listing the obscure and bizarre.
That way, people with "non-standard" setups won't be excluded.
--
Grumpy Ol' Fred cisin at xenosoft.com
9 Jun
9 Jun
9:55 a.m.
I've found a working 5.25" 1.2MB drive (Mitsubishi MF504C-310MP) and
a 3.5" 1.44MB drive (NEC FD1231T). Both drives can successfully
format LD and HD media. I can fill disks and copy stuff off
with no apparent issues.
I'm booting DOS 6.22 off a 20MB partition on a 3GB drive.
Dave Dunfield wrote:
To completely test the FDC, you need to check Single,
Double and
Double/128 at each of 250, 300 and 500 kbps rates. These are
represented in a standard PC at:
250 = 5.25 low-density drive only, 3.5" low-density diskette
300 = 5.25 Low-density diskette in high-density drive
500 = 5.25" or 2.5" high-density diskette
Before I go through a bunch of motherboards and find I've
made an elementary mistake, I want to check that what I'm seeing
is OK for a "typical" modern motherboard. FWIW this is a
Gigabyte GA-7DXR+ and it takes Athlon/Athlon XP CPUs, the
Super I/O controller is a VT82C686B ... just so you have an idea
of what I've started with.
I have the 5.25" drive as A: and the 3.5 drive as B: (and the BIOS
has been told what they are).
When I test the 3.5" drive I see a pass at 250k DD (with a 720K disk)
and a pass at 500k DD (with a 1.44MB disk) and the other four cases
fail. I think that means it can read/write the 720K and 1.44MB PC-style
diskettes, so it doesn't buy me anything "odd" but the results do at
least
correspond with expected reality.
When I test the 5.25" drive I see a pass at 300k DD (with a 360K disk)
but with the 1.2MB disk I see three failures. I was expecting 500k DD
to pass with a 1.2MB disk ... am I wrong? I've tried twice with the
same results and I've reformatted the 1.2MB disk afterwards and in each
case it formats with no errors.
Are my expectations in error or am I slipping up somewhere?
BTW: I'm not likely to find a fully-passing motherboard in
the set of "modern" ones that I'm testing now. Any ideas
how far back in time I need to go to find something that
I can use as the basis of a "media archiving" machine?
Most of the "PPP PPP PPP" ones are P1 systems. Although
I have a bunch of older motherboards and processors,
I'm not sure that I have "matching pairs" (or the docs
to determine such, where separate components are
involved). Ideally I'll find an add-in Adaptec board
or similar, but, in case I don't, I'd like to minimise
my search :-)
Thanks
Antonio
6407
days inactive
6411
days old
11 comments
6 participants
participants (6)
-
arcarlini＠iee.org -
cclist＠sydex.com -
cisin＠xenosoft.com -
dave06a＠dunfield.com -
fmc＠reanimators.org -
julesrichardsonuk＠yahoo.co.uk