Possible PUTR bug?

[Fred Cisin]

On Sat, 11 May 2019, Douglas Taylor via cctalk wrote:
> Finding a PC that supports the 5-1/4" floppy drive is difficult, the BIOS or 
> FDC chips only support 3-1/2" floppies in many late model PC's.  It appeared 
> only a few of the older PC's that supported the 5-1/4" drives could actually 
> change the spindle speed so you could read/write RX50 format.

{Written in a hurry; making a list of all of the errors is left as an 
exercise for the reader]

You can fool the computer in software. 
Most of those problems are with DOS/Windoze.
There is little or no problem with the FDC (other than as detailed in #3 
below). 
The main change that you are referring to was that some companies 
cut corners and stopped supporting a 300K bits per second data transfer 
rate that had been needed to use "360K" floppies in SOME early "1.2M" 
drives.   Don't use THOSE "1.2M" drives on THOSE computers for reading 
5.25" 96TPI "720K"/"800K" floppies.

1) the computer can't tell the difference between a "720K" 3.5" drive and 
a "720K" 5.25" drive, such as Tandon TM100-4, Teac 55F, Mitsubishi 4853, 
or Shugart/Panasonic/Matsushita 465.  My favorite was the 465, since it 
didn't have the "NEED for index" mentioned in #3 below.
The methods of being able to figure out which kind of drive it REALLY is 
are hardly never implemented in PCs.  (cf. undocumented algorithms to 
identify which processor is present)
If you are concerned about the menus in the CMOS SETUP or the choices in 
the DOS FORMAT program, LIE TO IT!  (Tell it that the tax increase is 
TEMPORARY, that the check is in the mail, that the current software 
release is completely bug-free)

When 3.5" drives and disks came out (also 3" and 3.25"), they were usable 
in ALL of the 5150s, 5160s, and 5170s.  With 5170s, you just had to LIE 
to the CMOS SETUP and tell the BIOS that your "720K" 3.5" drive was a 
"360K" 5.25" drive. 
FDC and BIOS couldn't tell the difference.  But, now, you need to lie in 
the other directions.
DOS support of "720K" was finally included in certain OEM versions of 
MS-DOS 2.11; and then got full support in MS/PC-DOS 3.20.  For use of the 
DOS formats of those drives, you were instructed to use DRIVER.SYS, or 
the partially undocumented DRIVPARM (present, but not documented in 
PC-DOS, because it did not work with some "real" IBM BIOS'es and would 
give an "UNRECOGNIZED" error, although the same boot disk would work with 
after-market BIOS!)


2) There did exist briefly, some "1.2M" 5.25" drives that could only run 
at 360 RPM, and required a 300K bits per second data transfer rate for 
handling "360K" and "720K" floppies in the "1.2M" drive.  Support for that 
data transfer rate might not be present in cut-corner "modern" PCs.
On those machines, you need to either use a "720K" 5.25" drive, or, if you 
need to use a "1.2M" drive, it needs to be one that can run at 300 RPM. 
Check the jumper settings.


3) However, SOME "400K" and "800K" floppies do still have a problem. 
WD style disk controllers, such as the 179x series and some other custom 
disk controllers are capable of writing closer to the index pulse than the 
NEC 765 style FDCs can handle.  The NEC style needs a little time after 
the index pulse before it can read.  Some compare that to "flash 
blindness", where a human can not see anything immediately after a photo 
flash.

There are numerous kludges to get around that.  On most drives (NOT 
including most TEAC 55), you can successfully read those diskettes by 
physically blocking the index hole of the floppy, using a SOLIDLY applied 
write-protect tab. (Do I really need to tell you to make sure that it 
won't fall off in the drive?)

OR, by fashioning a floppy drive data cable with a SWITCH in series with 
the INDEX signal.

Both of those approaches have a very minor problem that many unrelated 
disk errors will be mis-reported as being "DRIVE NOT READY" (error code 
80h), since the FDC doesn't see an index pulse.

Some people have had some success by slightly reducing the motor speed 
(nominally 300 RPM) of the drive!

If you still have access to the machine whose disks you want to read, you 
can FORMAT a disk with an adequate index gap (using software on your PC), 
take THAT disk back to the alien machine, copy the files onto it, and then 
bring it back to the PC to read.  That also works for a few other minor 
incompatabilities, such as formats that use an incorrect value in any of 
the sector header fields.

--
Grumpy Ol' Fred     		cisin at xenosoft.com
XenoSoft                        http://www.xenosoft.com