Intel FDCs

[William Beech]

I found an Intel manual with the schematics of the early CDC FDC.  I
believe this is the design of the isbc-201 and 202.  In 9800885A, they
talk about one controller board and different interface boards for SD FM
and DD M2FM.  I don't have the iSBC 201 and 202.  Could someone that has
both controllers compare them and possibly figure out the difference? Is
it just the microcode in the ROM set?  Does anyone have the capability
to read series 3000 ROMs?

There are Intel schematics in 9800425B for the DD controller. 9800422A
has schematics for one controller board and 2 interface boards (SD and
DD).  9800420A has the DD channel and interface boards marked as
SBC-202.  9800349B has the schematics of the controller and interface
boards (SD) and the label on the drawings contains "CDC" who designed
the original controller.

Thoughts?

Bill

[roger arrick]

Some info you might already have:

There are 2 different multibus FD systems used on early MDS systems
called SBC-201 for single density, and SBC-202 for double density.
Each was comprised of 2 boards - an interface board and a channel board.

Single density FD channel board, Intel PWA 1000467 same as DD except ROMSs

Single density FD Interface board, Intel PWA 1000603

Double density FD channel board, Intel PWA 1000467 but different ROMs than SD.

Double density FD interface, Intel PWA 1001036

SBC201 was at 88H and SBC202 was at 78H so they can both reside in a system at the same time,
and ISIS-II assigns them with DD first if it exists.

--  Roger Arrick -- Tyler, Texas, USA -- Ro...@Arrick.com --

---

From: intel-...@googlegroups.com <intel-...@googlegroups.com> on behalf of William Beech <nj...@nj7p.org>
Sent: Monday, December 16, 2024 12:50 PM
To: intel-devsys <intel-...@googlegroups.com>
Subject: intel-devsys Intel FDCs

 

--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/7c234d58-ca7f-41fd-9749-c0dd60a71505%40nj7p.org.

[Herbert Johnson]

https://www.retrotechnology.com/herbs_stuff/intel_FDC_field_guide.txt

This document includes some info from Roger as below; and some info
asBill Beech mentions. The document is referenced from this Web page

https://www.retrotechnology.com/herbs_stuff/m2fm.html#isis1

and all comes from 2018 discussion on these controllers. I'll add
additional information and any corrections as I see more or am informed
about more information. - regards Herb

On 12/16/2024 2:30 PM, roger arrick wrote:
> Some info you might already have:
>
> There are 2 different multibus FD systems used on early MDS systems
> called SBC-201 for single density, and SBC-202 for double density.
> Each was comprised of 2 boards - an interface board and a channel board.
>
> Single density FD channel board, Intel PWA 1000467 same as DD except ROMSs
> Single density FD Interface board, Intel PWA 1000603
>
> Double density FD channel board, Intel PWA 1000467 but different ROMs
> than SD.
> Double density FD interface, Intel PWA 1001036
>
> SBC201 was at 88H and SBC202 was at 78H so they can both reside in a
> system at the same time,
> and ISIS-II assigns them with DD first if it exists.
>
> --  Roger Arrick -- Tyler, Texas, USA -- Ro...@Arrick.com --
>

> ------------------------------------------------------------------------
> *From:* intel-...@googlegroups.com <intel-...@googlegroups.com> on

> behalf of William Beech <nj...@nj7p.org>

> *Sent:* Monday, December 16, 2024 12:50 PM
> *To:* intel-devsys <intel-...@googlegroups.com>
> *Subject:* intel-devsys Intel FDCs

> I found an Intel manual with the schematics of the early CDC FDC.  I
> believe this is the design of the isbc-201 and 202.  In 9800885A, they
> talk about one controller board and different interface boards for SD FM
> and DD M2FM.  I don't have the iSBC 201 and 202.  Could someone that has
> both controllers compare them and possibly figure out the difference? Is
> it just the microcode in the ROM set?  Does anyone have the capability
> to read series 3000 ROMs?
>
> There are Intel schematics in 9800425B for the DD controller. 9800422A
> has schematics for one controller board and 2 interface boards (SD and
> DD).  9800420A has the DD channel and interface boards marked as
> SBC-202.  9800349B has the schematics of the controller and interface
> boards (SD) and the label on the drawings contains "CDC" who designed
> the original controller.
>
> Thoughts?
>
> Bill
>
> --

--
Herb Johnson, New Jersey USA
http://www.retrotechnology.com or .net
preserve and restore 1970's personal computing
email: hjohnson @ retrotechnology dot com
or try later at herbjohnson @ comcast dot net

[scott baker]

I have a double density board set on the way — channel board and controller, with the little bridge board that connects the two of them. Was working when the previous owner removed it. A couple of questions…

1) what do people usually do for a floppy cable, if you don’t have a stock cable? The pin out of the 100-pin connector s weird. I was thinking of just using a spare s-100 connector and hand wiring it to a 34-pin pin out…

2) did I read correctly that someone here was using 3.5” drives with an 8” controller?if so, have any tips? I do have a few 8” drives around, but they’re assembled into other projects.

Thanks, Scott

[roger arrick]

I saw that DD board set on Ebay.

The cables are difficult.  You'll have to build several.

There is a cable going from the interface board (100 pin card edge) to the back of the CPU (37 pin DSub).

Then a cable to go from the CPU (37 pin DSub) to  the drive box (50 pin telco).

Then an internal cable to the 50 pin card edges of the disk drives, or header connectors if the old CDC drive.
For the old CDC drives, the ones with the little lever, there was a paddle board on each drive.

These cables are all in the online docs.
My recollection is that the single density system has a slightly different pinout than double density.

I started with a 50 pin M-F telco cable from ebay.  It came with 25 twisted pair wire which I undid and rewired.
Make sure not to twist signals together, twist signals with grounds.
You'll need a DC37 male connector and hood, and a DC37 chassis-mount female solder cup connector.
And you'll need a 100 pin .1" card edge connector to go to the interface card.
The channel card and the interface card both have 100 pin card edges at the top, but the drive cable just connects to the interface card.

--  Roger Arrick -- Tyler, Texas, USA -- Ro...@Arrick.com --

---

From: intel-...@googlegroups.com <intel-...@googlegroups.com> on behalf of scott baker <smb...@gmail.com>
Sent: Saturday, December 21, 2024 1:54 PM
To: intel-devsys <intel-...@googlegroups.com>
Subject: Re: intel-devsys Intel FDCs

 

--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/b0b44e0b-3358-4dc4-ab6b-17ea872b28edn%40googlegroups.com.

[Vale, Martyn]

 

Hi Scott,

 

See attached a solution originally designed by Richard Main which I have reproduced and enhanced.

 

All cables are relatively simple ribbon cables and you can use it with both 8” and suitable 3 ½” Drives, you can even use both sides of the 3 ½” drive giving you more drives (well 4 with two drives).

 

All the design gerbers etc. are on the group Google under MartynV. Bill can give you access if you haven’t already.

 

I don’t think the Double Sided Board files are up to date, but if you decide to go forward with this method let me know and I can update and help out if required.

 

Bye

Martyn.

 

 

From: intel-...@googlegroups.com <intel-...@googlegroups.com> On Behalf Of scott baker
Sent: 21 December 2024 19:55
To: intel-devsys <intel-...@googlegroups.com>
Subject: Re: intel-devsys Intel FDCs

 

I have a double density board set on the way — channel board and controller, with the little bridge board that connects the two of them. Was working when the previous owner removed it. A couple of questions…

--

You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/b0b44e0b-3358-4dc4-ab6b-17ea872b28edn%40googlegroups.com.

[scott baker]

Thanks Martyn, this looks perfect! I'll batch them up with my next order. Sometimes I think I single-handedly keep JLCPCB and DHL in business... :)

I either don't have access to the files section, or don't know where to look. Bill, can you set me up?

Thanks,

Scott

[William Beech]

Scott,

They are on my Google drive under MartnV/PCBs, I believe.  Scott, do you have access to it?

Intel did got through some weird boards to get from the interface board to the drive.  Since I use the ZX-200A controller board, it is a standard Shugart interface.

Bill

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/4caab0a6-b77e-412c-bcc1-72951e15c77en%40googlegroups.com.

[scott baker]

I don't think I've ever accessed your google Drive, Bill -- can you send me the link?

Scott

[Vale, Martyn]

 

 

Hi Scott,

 

I’ve uploaded the latest Double Sided – Motor Control Board design files, although you don’t need this board I’d highly recommend it as it will avoid wear on your 3 ½” floppies from constant rotation of the motor (remember 8” drives have a head load solenoid unlike 3 ½”  drives) it will also give you access to both sides of the floppy.

 

You’ll need:

 

MartynV/PCBs/SA800

MartynV/PCBs/MDS-202-II

MartynV/PCBs/Double Sided – Motor Control V2

 

You need this connector for connection to the Intel Floppy Drive Adapter PCB:

 

https://www.ebay.co.uk/itm/253179642597

 

This is the ONLY connector know to work as the edge connector on the Intel PCB is so shallow.

 

Google Drive Link is here:

 

https://drive.google.com/drive/folders/147raTizgJ_T5oUZQLDu2UiusVrE4MPui

 

Thanks

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/d0680f47-0676-4062-9881-a535729e1565n%40googlegroups.com.

[scott baker]

Thanks Martyn, and thanks also Bill who sent me the google drive invite. Looks like I got some stuff to order!

For some reason I had just assumed the FDC connector was the same pitch as a 100-pin S-100 connector. Unfortunately, clearly not the case, which is too bad because I actually have some nice shallow S-100 connectors.

Scott

[scott baker]

I was looking at the hardware reference manual for the SBC-201/SBC-202 and it looks like the interface board needs -10V on P1-77/P1-78. It promptly runs this into a 7905 to get -5V. I was note aware of -10V being part of the multibus spec. I'm thinking about just taking my -12V supply and dropping it through a couple of diodes to get the voltage within limit. But I was curious, those of you with existing multibus systems do you typically have a -10V supply?

Scott

[William Beech]

Scott,

If you use the 8008 or 8080 or 2704/8 EPROMs, you need the -5 VDC on pin 9/10 on the multibus connectors. I  used a 7905 regulator to provide that voltage from the -12 VDC.  I use ATX power supplies to run both my MDS-800 (MDS-I) and MDS-II.  With more modern boards you only need the -12 VDC.

Bill

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/742833e2-90f4-465b-a706-33d7c6ef4cacn%40googlegroups.com.

[scott baker]

Hmm, I have two different schematics -- one from 98-422A "Intellec Double Density Diskette Operating System Hardware Manual" and the other from 98700420A "SBC 202 Double Density Diskette Controller Hardware Reference Manual".

98-422A has the 7905 attached to pins 77/78 and 98700420A is pulling -5V from pins 9/10. Both the snippets are from sheet 1 on the double density interface board. It looks like the -5V comes in to play on sheet 6 of the interface board.

I suppose the right thing to do is to just look at the board and see what's connected to those pins.

Scott

[scott baker]

Looks like the answer is "both". My SBC-202 Interface Board is connected to -5V on pins 9/10. It also has an unpopulated 7905 footprint connected to pins 77/78.

[scott baker]

Filed under "things I should have checked a week ago..."

My OEM P2 Coupler as well as Martyn's P2 Coupler pcboards are 0.700" center-to-center on the edge card connectors. My card cage / backplane is 0.600" center-to-center.

It's baby steps on this FDC project...

Scott

[Vale, Martyn]

Hi Scott,

It sounds like you have an MDS-800 P2 coupler the same as my design which are wider. You need a MDS-ll/III coupler or you could narrow my design.

Martyn.

On 2 Jan 2025, at 07:19, scott baker <smb...@gmail.com> wrote:



To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/611f79a2-52d9-48ef-9d12-33baabe5e094n%40googlegroups.com.

[scott baker]

Thanks Martyn,

I've done a layout for an 0.600" coupler and will submit it with my next board order. Simple enough board, and cheap enough. It's the shipping on PCBs that has got out of hand lately, so I usually wait and batch things up.

In the meantime, there was a guy one state over from me that was selling IDC edge card connectors that I *think* I can just crimp onto a ribbon cable and muddle my way through that way. The connectors were cheap, and they'll get here in a couple days, so I'm going to give that a shot.

I built up all your other boards -- SA800, adapter, drive motor board, etc. All that went just fine.

Scott

[scott baker]

Okay, I got my P2 coupler soldered up and working. I got all of Martyn's boards built and wired up. My drive light comes on at just the right time, tries to read a sector, and promptly declares a disk error!

I take it the issue is that I'm trying to use my Flashfloppy Gotek, which supports MFM, and the SBC 202 boards expect M2FM. This seems to be confirmed by looking at the schematics, which detail a bunch of jumpers that are defaulted to M2FM and need to be cut and resoldered to do MFM. I'd rather not alter the board (nor do I know if cutting and flipping all those jumpers would even work, or whether software changes are also needed).

I've got plenty of actual 3.5" drives, but how do I go about getting an image put onto a M2FM disk? I have a greaseweazle.

Scott

[scott baker]

Turns out initializing a disk was super easy, barely an inconvenience.

I just stuck IMG2MDS in ROM and wrote a short loader to copy it to RAM and execute it. It would actually be pretty easy to make it straightforward rommable, I think there's just a few initialized data structures one would have to be careful. Really awesome utility, Martyn! The Heathkit H8 folks have a similar scheme to initialize disks on a system over serial.

Scott

[Herbert Johnson]

I suggest, as a matter of prudence, having someone else read your
(8-inch? 3.5 inch?) M2FM disks as soon as you can, to make sure you are
producing disks *anyone else* with proper equipment (preferably
different equipment but start with the same) can read. Same with reading
their disks on your setup. Reading your own disks is not enough to
confirm your results are consistent and reliable.

Put it this way: it would be annoying to discover an incompatibility
months, years from now, after many disks were written. Or not long after
you sell those 202 cards, or the buyer discovers the problem. Or when
you need to recover from backups, or change floppy drives (or change
your floppy-replacement-widget).

M2MF makes the problem worse, there's so few M2FM Intel users today. And
those using 3.5" diskettes with M2FM?

>> The Heathkit H8 folks have a similar scheme to initialize disks on a
>> system over serial.

Dave Dunfield has similar schemes for Northstar Horizon (S-100 hard
sectored), some other system too. And Apple II's have audio files that
download in the cassette port, that create and populate a bootable
diskette.

I may have to do that for a SWTPC 6800 I'll bring up this winter, the
floppy controller is pretty ancient but single-density FM. But I'll try
someone else's boot disks, and try to read/write them myself on 20th
century PC's.

Regards Herb Johnson

On 1/10/2025 11:24 PM, scott baker wrote:
> Turns out initializing a disk was super easy, barely an inconvenience.
>
> I just stuck IMG2MDS in ROM and wrote a short loader to copy it to RAM
> and execute it. It would actually be pretty easy to make it
> straightforward rommable, I think there's just a few initialized data
> structures one would have to be careful. Really awesome utility, Martyn!
> The Heathkit H8 folks have a similar scheme to initialize disks on a
> system over serial.
>
> Scott
>
> On Friday, 10 January 2025 at 18:10:35 UTC-8 scott baker wrote:
>
> Okay, I got my P2 coupler soldered up and working. I got all of
> Martyn's boards built and wired up. My drive light comes on at just
> the right time, tries to read a sector, and promptly declares a disk
> error!
>
> I take it the issue is that I'm trying to use my Flashfloppy Gotek,
> which supports MFM, and the SBC 202 boards expect M2FM. This seems
> to be confirmed by looking at the schematics, which detail a bunch
> of jumpers that are defaulted to M2FM and need to be cut and
> resoldered to do MFM. I'd rather not alter the board (nor do I know
> if cutting and flipping all those jumpers would even work, or
> whether software changes are also needed).
>
> I've got plenty of actual 3.5" drives, but how do I go about getting
> an image put onto a M2FM disk? I have a greaseweazle.
>
> Scott

[William Beech]

Herb and Martyn,

If I incorporate Martyn's BOOT.PRN into the MDS-800 and MDS-II ROMs,
would this allow a cold start of the MDS without available disks to boot
up?  Jack, Richard, and I messed with a device that would allow this to
happen with a Hobbytronics USB stick reader.  This seems SO MUCH easier!

Herb, I agree, we need to make sure the images we hold will work on the
hardware and SIMH simulators.

I have written a SIMH simulator for the SWTPC 6800 and it has several
OS's that work with it.

Bill

[Vale, Martyn]

 

Hi Scott,

 

Glad you found IMG2MDS helpful.

 

If you look here:

 

https://drive.google.com/drive/folders/1jec93O0Ao_BMaW8JCPeTzyDgS87iL2YX

 

https://www.youtube.com/watch?v=uCHvWm5drHc&t=195s

 

There’s a img2mds.img which contains the disk based version of IMG2MDS which you can create further disks as necessary without the necessity to go through the boot process. Normally once I generate a working disk I then do an SCP image of it with Greaseweasle which I then use as a backup to quickly recreate the disk if it gets corrupted, which is easily done when you are experimenting.

 

One of these days I’ll get round to speeding it up a bit !

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/51feca3a-2fc2-4110-a729-46cee092aa49n%40googlegroups.com.

[Vale, Martyn]

Hi Bill,

 

Yes, I don't see why that would be a problem and would definitely be preferable to punching in the code by hand. I would be fairly easy to re ORG the code to run anywhere you wanted it to reside.

 

Talking of the USB device I did design a set of PCB's to use this on the MDS-800 as below:

 

 

Unfortunately it doesn’t fit the MDS-II/III, so my plan eventually is to redesign it to simply plug in to one of the MDS-II/III ports directly without the need to mess around inserting a board in to the 8251 socket on the IPC and in fact of course the 8251 is soldered in on the IPB which makes this impossible anyway.

 

Bye

Martyn.

 

-----Original Message-----
From: intel-...@googlegroups.com <intel-...@googlegroups.com> On Behalf Of William Beech
Sent: 11 January 2025 05:18
To: intel-...@googlegroups.com
Subject: Re: intel-devsys Intel FDCs

 

⚠ Caution: External sender

--

You received this message because you are subscribed to the Google Groups "intel-devsys" group.

To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.

To view this discussion visit https://eur01.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgroups.google.com%2Fd%2Fmsgid%2Fintel-devsys%2F0c787817-b551-4a8b-9d5a-c23e749ff157%2540nj7p.info&data=05%7C02%7Cm.vale%40ucl.ac.uk%7C136c0b701fda4191d78d08dd31ff7409%7C1faf88fea9984c5b93c9210a11d9a5c2%7C0%7C0%7C638721695415031745%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C0%7C%7C%7C&sdata=bpv%2F4nuFpHjCpXIj3NZc6Ah95BJYObxW%2BbcO3hlC%2BT4%3D&reserved=0.

[Herbert Johnson]

Bill, I'll check out your SIMH/6800 work, thanks. Also, glad you
mentioned the Hobbytronics USB/flash drive product. I updated my ISIS
Web page to point to their current version, it's a little cheaper I
think and a different processor but apparently still available.

The point of my post, was to encourage use of original mylar/iron oxide
media and drives to confirm the M2FM results were reproducible and for
archival and exchange purposes.

My interests are in vintage hardware preservation and restoration.
modern solid-state microcontroller replacements are fine but they don't
need Herb Johnson's support, they have plenty of support. Even so, M2FM
Intel format is such a marginal use case, those modern substitutes (and
simulators) may develop issues few would see or debug. But I don't want
to distract from the subjects of this thread, just issue a cautionary
note.

regards Herb

> --
> You received this message because you are subscribed to the Google Groups "intel-devsys" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.

> To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/0c787817-b551-4a8b-9d5a-c23e749ff157%40nj7p.info.

[scott baker]

Herb,

I understand the point about the disks not necessarily being readable elsewhere -- sometimes on my other computers even ones with reliable controllers, disks written in one drive might not even read in another drive due to alignment issue. All this old floppy hardware is getting pretty sketchy, and the 3.5" media was never great in the first place and several decades later seems even worse.

That said, unfortunately I was only able to image 2 disks with IMG2MDS last night. My remaining half-dozen attempts have all failed at some point during the xmodem upload. It probably spit out a "write error" message, but I can't see because the terminal program's xmodem is obscuring any output. It's entirely possible that I've just run into a bad run of floppies. I'm hesitant to try rewriting the two successful disks, because they are still readable and useful. Three different disks have now failed to write at exactly 86%, so I wonder if it is some component that has drifted in the controller, rather than bad disks. Nevertheless, it's been a fun project so far. I've always been fascinated by the bit slice cpu chipset, and this is the first project where I've had one.

Martyn, one of the two images I burned was the IMG2MDS image (it was the first image I tried), though I suspect that unless this version uses ISIS system calls, the binaries on it won't work for me. I'm running everything from an 80/24A board and I had to reassemble IMG2MDS with different port numbers for the 8251

Scott

[William Beech]

Reviewing Martyn's code on the Google Drive, I see it is more complex
than that, but still doable.  I would need to implement img2mds or
imgm2mds in the boot ROMs.  There is some space in boot loader part of
the ROM.

Bill

[William Beech]

Martyn,

What I will do is to start with the latest MDS-II ROM.  I have made one modification to bring up both serial ports at 9800 baud.  Let me see if I can shoehorn in your code into the the two open blocks of the ROM.

Bill

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/DU2PR01MB793667D0248687DFDB461118C11D2%40DU2PR01MB7936.eurprd01.prod.exchangelabs.com.

[William Beech]

Martyn,

An ISIS-II disk uses different skews on different tracks.  Specifically "14555555555555555555555555555555555555555555555555555555555555555555555555555" from track 0 to track 76.  This is included in the disk header at 0c80-0cff in a disk dump.  I see that is in your disk image file, as well.  Do you incorporate that in your software?

Thanks!

Bill

On 1/11/2025 5:49 AM, Vale, Martyn wrote:

Hi Bill,

 

Yes, I don't see why that would be a problem and would definitely be preferable to punching in the code by hand. I would be fairly easy to re ORG the code to run anywhere you wanted it to reside.

 

Talking of the USB device I did design a set of PCB's to use this on the MDS-800 as below:

 

c

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/DU2PR01MB793667D0248687DFDB461118C11D2%40DU2PR01MB7936.eurprd01.prod.exchangelabs.com.

[Herbert Johnson]

On 1/11/2025 12:33 PM, scott baker wrote:
> Herb,
>
> I understand the point about the disks not necessarily being readable
> elsewhere -- sometimes on my other computers even ones with reliable
> controllers, disks written in one drive might not even read in another
> drive due to alignment issue. All this old floppy hardware is getting
> pretty sketchy, and the 3.5" media was never great in the first place
> and several decades later seems even worse.

I often hear about sketchy hardware and unreliable disk media. And yet,
floppy controllers continue to work half a century later. I read disks
from the 1970's. And, media of quality *was* produced in the era, and is
*still* readable today.

As for "alignment", well, that happens but that can be tested out. I'm
*very* reluctant to align floppy drives because more failures are due to
heads and dirt and controllers. But if you determine your drive is
misaligned, then fix it or shelve it.

What happens on controllers is that components drift - capacitors and
resistors, and particularly one-shot monostables, a common design dodge.
Those need replacement, even the IC's. This might happen on drives too
but that's hard to determine practically.

Floppy drives suffer from age, particularly heads through physical wear
and dirt. Mechanical parts "drift" also - plastic read/write head mechs
have long arms, plastics distort with age, QED. Some brands, models of
drives survive better than others, exposure to temperature and humidity
matter.

Verify your drives in their intended use, to be sure the drives are not
the weak link. Marginal results suggest you check rotation rates, get
within 1% if you can.

As for diskette media, 0) make sure your floppy drive heads are CLEAN.
And stay clean. 1) choose name brands like 3M, Dysan, BASF, Verbatum and
a few others. 2) examine media by eye for mold and scratches. 3) format
and test before use - for 3.5's use a desktop PC.

Some USB 3.5" floppy drives support both high-density 1.44M and double
density 720K, so both can be tested. Don't use high-density media to
write at double-density flux rates, that media is not designed for that
purpose. Just as DD media will fail when formated HD.

There's a number of these and other "secrets" to 21st century use of
floppy drives and media. I try to keep them here:

https://www.retrotechnology.com/herbs_stuff/drive.html

all these things sound tedious, to people who expect things to run
forever. Like Star Trek episodes where centuries-old-computers still
run when found - hah! Engineering is where expectations and theory meet
the real world. I call these "engineering problems", they are fun enough
for me.

> That said, unfortunately I was only able to image 2 disks with IMG2MDS
> last night. My remaining half-dozen attempts have all failed at some
> point during the xmodem upload. It probably spit out a "write error"
> message, but I can't see because the terminal program's xmodem is
> obscuring any output.

parallel another terminal to the Intel system's serial output to capture
whatever XMODEM loses.

> It's entirely possible that I've just run into a
> bad run of floppies. I'm hesitant to try rewriting the two successful
> disks, because they are still readable and useful. Three different disks
> have now failed to write at exactly 86%, so I wonder if it is some
> component that has drifted in the controller, rather than bad disks.

I referenced brand and physical condition of floppies. Find more
floppies, clean the heads. Always good to clean the heads, but don't
bust them in the process. 3.5" head mechs are pretty fragile.

86% is diagnostic, dig deeper. "eyes and ears, Watson". Do those
failures appear near the end of the diskette copy? Recall the
fundamentals. Diskette tracks start from zero at the outermost
diameter/radius; the last tracks are innermost, the smallest diskette
diameter. That's the densest flux-per-inch to read or write (same track
data less track length). If your floppy drive has low signals (dirty
heads, wrong media density, possibly rotation rate) then that's where
the errors will occur.

(also: some 8-inch drives need a "track 34" signal to tell it "tracks
are getting dense, increase write signals". Some drives don't. I'm
losing track of whether you are writing 8-inch or 3.5-inch.)

Also, use of high-data-rate formats (I'm carefully not saying "high
density" but that's one example) on low density media will classically
fail on the innermost tracks. Why? again: the double-density media
won't support the higher density flux changes demanded on those inner
tracks. This is engineering in action.

> Nevertheless, it's been a fun project so far. I've always been
> fascinated by the bit slice cpu chipset, and this is the first project
> where I've had one.

It's still fun. Failures are diagnostic. If you aren't failing, what are
you learning that's *new* to you? ;) Bill Beech tells me privately
"Remember, this is a) supposed to be fun, b) keep our minds healthy!" I
think I can share that.

Regards Herb

[scott baker]

I just prepped a second drive, and did an IDISK and COPY * from the first to the second without issue. The disk is only about 25% full so it's not an exhaustive tests.

Now I've swapped the two drives and am trying the other one out with IMG2MDS.

The media I'm using is a mix of Verbatim, TDK, 3M and Sony. All HD. The verbatim were the initial two disks that succeeded last night.

The drives I have are all Samsung SFD-321B. These had to be modified for 360 RPM operation, basically bridge a couple pads and it will function like a Teac FD235HG. I probably should verify that they are indeed running at 360 RPM as expected; I assume that can be done by putting a scope on trk0. I vaguely remember frustration with these drives with the H8 and swapping them around thinking there were drive issues, but I never conclusively got to the bottom of it.

Herb, what is your recommended procedure for cleaning heads?

Scott

[mark.p...@btinternet.com]

Bill

The disk formatting is a little more complex. The Isis.lab  three characters after the label itself give the interleave for track 0,1,2+. There is however an inter track skew for tracks 3 onwards. 

The decompiled code in my GitHub repo have all the details as it is slightly different for SD and DD and the version of OS

Mark

Regards

Mark Ogden

---

From: intel-...@googlegroups.com <intel-...@googlegroups.com> on behalf of William Beech <nj...@nj7p.info>
Sent: Saturday, January 11, 2025 7:22 pm
To: intel-...@googlegroups.com <intel-...@googlegroups.com>

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/eb0b1a79-7333-498e-a92b-75290f20140f%40nj7p.info.

[scott baker]

Initialized 2 more disks in the second drive just fine, using IMG2MDS. I have at least stuck a "suspected bad" label on the original drive so it doesn't end up back in rotation again. Would still be good to diagnose it / clean the heads / whatever else.

Scott

[Herbert Johnson]

Scott is making progress!

>> Herb, what is your recommended procedure for cleaning heads?

With caution, use cotton swaps and isopropyl alcohol 90%. Be real
careful not to bend anything. brush along the head-mech screw to wipe
out the head gap; brush along the track to cleand adjacent to the head
gap, good can pile up.

Apple Macintosh (sony) 3.5 inch floppy drives have a copper plate to
tension the head, that's easily bent. for larger drives the heads are
sprung and dimensions are much larger.

I also use cleaning diskettes and alcohol. But *caution* on cleaning
single-sided drives! The other side is a felt/cloth pad - you'll abrade
the pad with a cleaning disk! Put a barrier on the pad side of the
cleaning disk. But i don't like single-side drives because of pad wear,
which can cause failures and track misalignment.

Al Kossow says, don't use alcohol on dirty diskette media. He's right
BUT good media won't be degraded (much) by it and dirt is worse than
chemical degradation. (If the coating cleans off, try better disks.)
There's other cleaning solutions for diskettes but it's
hairdresser-chemicals a little obscure to obtain. Check my Web page for
discussion of this.

> scope on trk0.

no, scope on the index hole signal, measure the pulse frequency. But
better is a digital counter. There's cheapie frequency-counter widgets
on Amazon.

Better is a program to time the signal - many floppy testing programs
will perform that measurement. Since you want 1-2 percent a 'scope for
timing won't be good enough unless the scope can measure frequency.

https://jope.fi/drives/SAMSUNG-SFD321B-070103.pdf

high density 1.6M option for that drive. NOte that the recording density
(I called it flux density) for 1.6Mb is intermediate to single and
double density. Check to see if you need to toggle the "density select"
signal. For practical use, a little toggle switch may be advised. This
is not work I've done much so verify my considerations.

Modern users of 8-to-3.5 use of these drives will know better. I *think*
you want to use high-density diskettes for that "1.6M" scheme? While
8-inch disks are iron oxide, they have more track length than 3.5 inch
disks, thus higher-density media is advised. again I lack
practical/recent experience here.

Regards Herb

On 1/11/2025 3:19 PM, scott baker wrote:
> I just prepped a second drive, and did an IDISK and COPY * from the
> first to the second without issue. The disk is only about 25% full so
> it's not an exhaustive tests.
>
> Now I've swapped the two drives and am trying the other one out with
> IMG2MDS.
>
> The media I'm using is a mix of Verbatim, TDK, 3M and Sony. All HD. The
> verbatim were the initial two disks that succeeded last night.
>
> The drives I have are all Samsung SFD-321B. These had to be modified for
> 360 RPM operation, basically bridge a couple pads and it will function
> like a Teac FD235HG. I probably should verify that they are indeed
> running at 360 RPM as expected; I assume that can be done by putting a
> scope on trk0. I vaguely remember frustration with these drives with the
> H8 and swapping them around thinking there were drive issues, but I
> never conclusively got to the bottom of it.
>
> Herb, what is your recommended procedure for cleaning heads?
>
> Scott

> Initialized 2 more disks in the second drive just fine, using IMG2MDS. I have at least stuck a "suspected bad" label on the original drive so it doesn't end up back in rotation again. Would still be good to diagnose it / clean the heads / whatever else.
>

[Herbert Johnson]

On 1/11/2025 4:00 PM, Herbert Johnson wrote:> There's other cleaning

solutions for diskettes but it's
> hairdresser-chemicals a little obscure to obtain

You can use clean or distilled water with drops of dish detergent and
cotton swaps, to wipe dirt from diskette surfaces. It's not as solvent
as isopropyl alcohol, and you need more drying time. Don't use hot-air
to dry the diskettes, err, they melt.

Regards Herb

[roger arrick]

As for 8" floppy disks, they are incredibly durable.

In 1980 I left an 8" ISIS-II disk in my '68 Mustang during the Texas heat - it warped the jacket.
I cut the edge and slipped out the media, put it in another jacket and it read fine.

I've recovered 80% of my 8" floppy disks that have been stored in a Texas attic for 40 years, it gets 130°+ up there and down to the teens.

If you have one with a lot of dirt, I'd have to see it, I'd try cutting the jacket edge to remove the media and rinse it in warm water, maybe help off some of the debri with a cotton ball, reinstall in the jacket and see what it does.  

Make sure the felt pad is good on your drive.  And look at the heads and clean them with alcohol and QTip.  I've had some media flake off onto the head and stick, then it grinds grooves in every disk after that.  So EVERY time you pull a disk out of a drive, look at it for grooves, that's an indication of a dirty head or a bad felt pad.

The index timing pulse on Shugart SA80X drives often needs to be adjusted.  It's in the shugart maintenance manual and requires a scope to measure the pulse width.  Easy adjustment and doesn't require an alignment disk.

--  Roger Arrick -- Tyler, Texas, USA -- Ro...@Arrick.com --

---

From: intel-...@googlegroups.com <intel-...@googlegroups.com> on behalf of Herbert Johnson <hjoh...@retrotechnology.com>
Sent: Saturday, January 11, 2025 3:19 PM
To: intel-...@googlegroups.com <intel-...@googlegroups.com>; scott baker <smb...@gmail.com>

Subject: Re: intel-devsys Intel FDCs

--
You received this message because you are subscribed to the Google Groups "intel-devsys" group.
To unsubscribe from this group and stop receiving emails from it, send an email to intel-devsys...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/f98f7dc5-7af9-4f68-bb80-adc75819aa7e%40retrotechnology.com.

[Vale, Martyn]

Hi Bill,

 

At the time I wrote the software I couldn’t find a reference to the correct skew, but Mark pointed this out to me at a later date.

 

I think from memory it uses a fixed skew of 2 as this appeared to give the best performance for booting at least. Once you have one functioning floppy you can format a fresh one with ISIS and then use IMG2MDS to create a floppy on that disk whilst saying No to the format disk option in IMG2MDS at this point you will have a true ISIS skewed floppy.

 

Thanks

Martyn.

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/eb0b1a79-7333-498e-a92b-75290f20140f%40nj7p.info.

[William Beech]

Martyn,

Indeed, that is a good answer.  I learned from Mark the difference between interleave and skew, as well.

The need to be able to create a disk image on a hardware system directly from a file on another machine is something we needed from the beginning.  And not just on the ISIS machines.  I am also going to make this work on the CP/M-80 machines and the 6800 Flex machines.

Bill

To view this discussion visit https://groups.google.com/d/msgid/intel-devsys/DU2PR01MB793685E046484654CF9989D7C11E2%40DU2PR01MB7936.eurprd01.prod.exchangelabs.com.