On Thu, Feb 1, 2018 at 1:08 PM, Paul Berger via cctalk < cctalk at classiccmp.org> wrote: This is true. I have the pinouts at home if anyone wants them, I can post them when I get home from work. However: I will note that the YE-DATA 8" drives have heads that are rough on some 8" media, particularly those prone to shedding. The heads are also a pain to get to to clean after such shedding has taken place. I've stopped using mine for 8" archival for this reason. The Shugart 850/851s are much nicer in this regard (Al suggested them to me awhile back) -- easier on disks, much easier to clean. - Josh

On 2/1/2018 2:52 PM, Ali wrote: Here's what I have, from notes I took about 10 years ago when I built a cable after beeping things out.? The YE-DATA drive itself uses the standard Shugart interface, IIRC. The below table lists the 50-pin edge connector pins of interest, the 37-pin d-sub on the back of the drive unit, and the 34-pin PC floppy cable.? I make no guarantees, but it worked for me: 50-pin edge??? ?? 37-pin Dsub??? ??? 34-pin edge -----------??? ?? -----------??? ??? ----------- 1??? ???? ??? ??? 20??? ??? ? ? ? ? ??? GND 2??? ??? ??? ???? 2??? ? ??? ? ?? ? ??? ????? (TG43) 9??? ???? ??? ??? 25??? ? ??? ??? ? ??? GND 10??? ??? ??? ??? 7??? ? ??? ? ?? ? ??? N/C 11??? ??? ??? ??? 37??? ?? ??? ? ?? ??? GND 12?? ? ?? ??? ??? 19?? ?? ??? ? ??? ??? 34??? (Ready) 13??? ?? ? ?? ??? 36? ??? ??? ? ??? ??? GND 14??? ?? ?? ? ??? 18?? ?? ??? ? ??? ??? 32??? (Side1) 17??? ??? ? ??? ? 28? ??? ??? ? ??? ??? GND 18??? ?? ?? ? ??? 10? ??? ??? ? ??? ??? 16??? (Motor ON/Load) 19??? ? ??? ? ??? 24? ??? ??? ? ??? ??? GND 20??? ? ??? ? ??? 6?? ?? ??? ?? ??? ??? 8??? (Index) 21??? ?? ?? ? ??? 27? ??? ??? ? ??? ??? GND 22??? ?? ?? ? ??? 9?? ?? ?? ? ? ??? ??? N/C 25??? ??? ?? ? ?? 26? ??? ??? ? ??? ??? GND 26??? ??? ? ??? ? 8?? ?? ? ?? ? ??? ??? 12??? (DS1->DS0) 33??? ?? ? ?? ??? 29? ??? ??? ? ??? ??? GND 34??? ? ??? ? ??? 11?? ?? ??? ? ??? ??? 18??? (Direction) 35??? ??? ??? ? ? 30?? ?? ??? ? ??? ??? GND 36??? ??? ? ??? ? 12?? ?? ??? ? ??? ??? 20??? (Step) 37??? ??? ?? ?? ? 31? ??? ??? ? ??? ??? GND 38??? ?? ? ?? ??? 13? ??? ??? ? ??? ??? 22??? (Write Data) 39?? ?? ? ??? ??? 32? ??? ??? ? ??? ??? GND 40?? ?? ? ??? ??? 14? ??? ??? ? ??? ??? 24??? (Write Gate) 41??? ?? ? ?? ??? 33? ??? ??? ? ??? ??? GND 42?? ?? ? ??? ??? 15? ??? ??? ? ??? ??? 26??? (Track 0) 43?? ?? ? ??? ??? 34?? ?? ??? ? ??? ??? GND 44??? ?? ? ?? ??? 16?? ?? ??? ? ??? ??? 28??? (Write Protect) 45??? ?? ? ?? ??? 35?? ?? ??? ? ??? ??? GND 46??? ?? ? ?? ??? 17?? ?? ??? ? ??? ??? 30??? (Read Data)

On 02/01/2018 02:08 PM, Paul Berger via cctalk wrote: The drives may be standard, but I am curious how standard the format of what's on the disk is. How likely is it that a PC OS from the last ~30 years will understand how to read what's on the disk? Or, I'm guessing it's likely that archival programs don't need to understand how to read the disk format for raw disk imaging. -- Grant. . . . unix || die

On Thu, 1 Feb 2018, Grant Taylor via cctalk wrote: DOS/Windoze will have absolutely no idea how to understand what is on the disk. It is a different file system structure, and DOS/Windoze didn't deal with that. THAT is why there existed programs such as 22Disk, XenoCopy, Uniform, etc. to interpret the file system of disks from CP/M, P-System, TRS-DOS, etc. that had different file system structures, but similar recording methods. There were somewhere around 2500 mutually incompatible floppy formats. (XenoCopy included capability for 400 of them) If the physical track structure is "standard" MFM, then it is possible to read sectors from them. That means that Imagedisk, Teledisk, or a disk file transfer program that has implemented THAT format can read them. If the physical track structure is different, even if using the same disk drive, such as hard-sector, MFM without similar sector headers, MMFM (FM sector headers, MFM data), or GCR (Apple, Commodore, Sirius/Victor, etc.), then you can not read sectors using a PC FDC chip. Nevertheless, flux transition, such as Kryoflux, Cat Weasel, Option Board, can still capture the raw flux transitions, and maybe make sense out of them. -- Grumpy Ol' Fred cisin at xenosoft.com

On Fri, 2 Feb 2018, Terry Stewart wrote: But, you have to tread very carefully. Because SOMETIMES [now rarely], they could mean the PRE-5150 definition where "IBM compatible" meant 3740/SSSD 8" with 128 byte sectors. 5150 really was a danger to itself and/or others; it totally destroyed the meaning ot "IBM compatible" disks.

On Thu, 1 Feb 2018, Ali via cctalk wrote: If it is a 50 pin edge connector, then it is LIKELY that it follows the Shugart 800/850 pinout. BUT, it is still a good idea to find and check the pinout of the specific model drive, since some added additional signals and/or repurposed pins for other purposes. And, check the drive specific pinout for power connections! 5.25" drives had a standardized power connection. 8" drives do NOT have standardized power connections. Not all used the same power nor connector nor pinout as Shugart. In addition to 5VDC, you may need 24V or 12V, and maybe even -5VDC, or even AC. If it IS SA800/850, then FPADAP, http://www.dbit.com/fdadap.html is an easy way to handle the cabling. The website say that they have 5 in stock, so order immediately if you need it! If you are reading ONLY, and not writing to the drive, then you do not need the TG43 signal and circuitry, in which case it is trivial to make your own cable. The majority of the 5.25" (SA400) signals are even in the same order, so you can start with a 34 pin flat cable, peel out a small handful of the wires, and crimp them in a 50 pin IDC edge connector. YMMV.

I second that. I've got one and works just as it should. Terry (Tez)

On Thu, 1 Feb 2018, Chuck Guzis via cctalk wrote: for reading, it doesn't do anything other than cabling. It means using a flat 34 cable and a flat 50 cable, without the trivial effort of making your own 34 to 50 cable. THAT is the part that my homemade cables don't do. But, it was extremely rare that I needed to WRITE 8". When I did have to, I tried to write to fresh empty disks, so that I could stay mostly on lower numbered tracks.

"Current" versions ("since 2009") have two small 7-segment LEDs that display current track number. THAT might be handy sometime, although most such times are amenable to software methods. Analagous example: When installing the beta version of Windows 3.10, it would hit an unrecoverable error at a certain point. But, SMARTDRV wouldn't permit anything other then futile further retries (since SMARTDRV had already reported SUCCESS), or abort the installation, at which time, it destroyed all indication of where on the disk the problem was. Neither SSTOR nor Spinrite could find the consistent bad spot. If I had known cylinder and head, I could have trivially mapped out the bad spot. I should have written a TSR to be able to interrupt and find out! (Without SMARTDRV, I could have done an "Ignore" and renamed the specific file that failed into "BADSECS", and written another copy of that file to another location.) I told MS Win31 beta support that write-cacheing of SMARTDRV (which at that point could not be turned off!) was capable of creating disasters that would ultimately force recalls (DOS 6.00 to 6.20 replacement); they said that it was a HARDWARE problem, not THEIR problem, and that it was not the responsibility of the OS to find, work around, nor recover gracefully from hardware problems.

Of course if you want to get fancy w/ the FDADP: www.ibm51xx.net ;) -Ali

So would I Fred, but I have not found 34 to DC37 adapters readily. The SCSI stuff is out there for easy purchasing - at least for now. Plus many people have drives in cases already w/ 50 pin connectors on them (I have a couple and that is what that cable is being used for) so a DC37 to DC37 cable is no bueno... I have been toying with making a bunch of brackets for odd connectors (I haven't seen a bracket that has holes cut for DC37 for example readily available for purchase). A friend of mine has access to an industrial laser cutter and has told me he would be happy to cut brackets in quantity for free.... -Ali

Well, I think they're overpriced, but here you go: http://www.jdr.com/product_p/bkt-d37.htm JDR sold all of their ISA prototype boards with a DC-37 bracket also. --Chuck

On 02/01/2018 07:17 PM, Fred Cisin via cctalk wrote: IN my case, all of my drives are terminated in DC37M connectors and my controllers have DC37F. Half-height and 3.5" drives are housed in old IBM 4569 boxes (we bought 50 of the things back in the day) and the 8" and FH 5.25" have their own external boxes with PSU. So, no matter what I want to do, it's a matter of plugging something in and turning it on. Need a 100 tpi drive? Just grab it off the shelf and plug it in. That extents to not only PC floppy controllers, but things such as Catweasels (I fit them with a buffer board ending in a DC37F on a bracket) and even my WD-based controllers and my MCU setups. It streamlines the process greatly. --Chuck

On 02/01/2018 07:53 PM, Fred Cisin via cctalk wrote: Yeah, I've got an old ABCD manual printer switch with the "centronics" connectors that swaps all 36 lines, but it always seemed like more trouble than it was worth, so it resides in a storage cabinet. I use a rack-mounted PC with a couple of DC37s on the front panel, so swapping boxes is pretty simple. I've seen some bizarre stuff, but the one that takes the cake was a boot floppy from a PLC. Said PLC was equipped with two floppy drives--one for booting only and the other for user data. The boot floppy had a very oddball track spacing; something like 120 tpi; the user floppy was a normal 135 tpi. This was measured with developer (Kyread) and a microscope. I just couldn't believe that a manufacturer would go to that length. I guess the PLC manufacturer didn't want their software to be copied. Needless to say, replacement boot floppy drives were near unobtanium and stupid expensive when you could find one. --Chuck

On 02/01/2018 09:39 PM, Fred Cisin via cctalk wrote: Other than a brief encounter with an SA400 (single-sided 48 tpi. 35 cylinders), the Micropolis 100 tpi drives were my first real run in with 5.25" drives. Great drives, heavy, built like a tank--and expensive. I've got a late-model 96 tpi Micropolis that illustrates their NIH mindset--the PCB pivots as you close the door latch. Still used the precision-ground leadscrew positioner. Buffered seek; something like 4 steps per cylinder, if you can believe it. ISTR that if you stick a formatted 100 tpi ddisk in a 96 tpi drive and do a read ID, the first cylinder reads as 6. That is, the 100 tpi disk is offset a bit more toward the outside of the disk. You can read a few cylinders after that, before the misregistration causes things to get weird. --Chuck

On Thu, 1 Feb 2018, Bill Gunshannon via cctalk wrote: I agree. I always made my own cables, but I did mention [ABOVE] that FDADAP is the easy way to handle the cabling. First cable that I made, I used a solderless bread-board, and looking at both pinouts, started placing wires. Once I saw what I was ending up with, subsequent cables were a flat 34 pin cable, that I peeled out a few wires, and very carefully held them in place to crimp on a 50 pin connector.

On Thu, Feb 1, 2018 at 7:04 PM, Chuck Guzis via cctalk < cctalk at classiccmp.org> wrote: Well, great, but then what do you do when you want to read and write 80 (decimal) byte MFM sectors? :-)