Current source for RX50 media?

Fred Cisin cisin at
Sat Jun 25 11:18:37 CDT 2016

On Sat, 25 Jun 2016, Warner Losh wrote:
> I was looking at the wrong row in my tables. The RX-50 uses the
> 300K transfer rate, while the 1.2MB IBM-AT floppy uses 500k. The
> rotation rate is different, but I'm having trouble locating that. Sorry
> for the possibly bad info. I added support for RX-50 to the FreeBSD
> driver (preliminary, mostly works for reading, writing has issues and
> formatting doesn't work) several years ago, so I needed to know...

Your info is good.

The RX-50 and the IBM "360K" when in their correct drives, spin at 300 
RPM, with a data transfer rate of 250K bits per second.   The "360K" is 
40track/48tpi; RX-50 is 80track/96tpi

The IBM "1.2M" diskettes (and 8"DD) spin at 360 RPM, with a data transfer 
rate of 500K bits per second.   80track/96tpi

It gets more squirrelly when you want to use a "360K" disk (or RX-50) in a 
"1.2M" drive!
You can either change the drive speed to 300RPM (on drives that support 
both speeds) and use a 250K data transfer rate, OR (on drives that don't 
change speed) leave the speed at 360 RPM, and change the data transfer 
rate to 300K bits per second to compensate for the different rotational 
speed.  (It can get confusing because "300" is both the rotational speed 
of the "360K" and RX-50, AND is the data transfer rate when compensating 
for 360RPM, AND is the Oersted (coercivity) of the disks for "360K" and 
RX-50.  "360" is both the rotational speed of "1.2M" and 8", AND is the 
polpular nake for IBM's DSDD 5150/5160 disks)

Since the RX-50 is DD with 80 track/96tpi  (what the marketing people 
called "QUAD density"), a "360K" drive won't work.
A "1.2M" drive is 80 track/96tpi, but is more commonly used at 360 RPM and 
density that the marketing people called "high" (which is what the 
marketing people were; you won't believe what the Superbrain/Intertec 
people called DSDD80track!)

The RIGHT drive for RX-50 would be a 5.25" "720K" drive, which weren't 
commonly used on PC.  Such as Tandon TM100-4, Shugart/Matsushita 465, or 
Teac 55F.
("1.2M drives are #475 or 55G.  55FG is a "1.2M" drive that is explicitly 
intended to also be able to do "720K".  Some "1.2M" drives can switch 
between 300RPM and 360RPM, some are stuck at 360RPM and need a data 
transfer rate change to compensate)

Because the "1.2M" drive needed to also be able to read "360K" diskettes, 
provision was made.   40 track V 80 track can be read just fine by using 
every other track, BUT because the track width is different a "1.2M" drive 
does not adequately cover/erase a wide ("360K" track that it tries to 
write over.
(1/3mm track width at 1/2mm spacing V 1/6mm track width at 1/4mm spacing)
Need an analogy?  Two bicyles can ride on the tracks made by a car, but 
the bicycle tire-prints won't obliterate the car tire-prints.

In addition, for "360K" in a "1.2M" drive, either the rotational speed or 
the data transfer rate needs to change.

SO, . . .
to to do RX-50 in a "1.2M" drive, we need the 80track/96tpi track spacing, 
but the "360K" RPM/data transfer rate.

Proper:          300RPM  with   250K data transfer rate,  96tpi.
"1.2M" kludge:   360RPM  with   300K data transfer rate,  96tpi

(Note: the Tandon TM100-4 is an excellent drive for this, but the TM100-4M 
is 100tpi instead of 96!)

RX-50 and "360K" diskettes both have a coercivity of 300 Oersted.  BUT, 
"360K" is tested at 48tpi.  If there are flaws between the tracks, they 
won't be spotted in testing, but will show up for RX-50!  So, ideally, 
you want disks tested for 96tpi (marketing "quad density"), or GOOD disks 
without flaws, even between the tracks.  In theory, any good "360K" 
diskette should do, but some folk have encountered errors on disks not 
tested for 96tpi.

"1.2M" diskettes have a coercivity of 600 Oersted.  When written by a 
drive expecting 300 Oersted, data longevity may be shortened.  (Roytype 
"HD" with TRS80 had an average data retention of less than an hour)  Do 
NOT use "HD" ("1.2M") diskettes for any RX-50 data that you would like to 

Grumpy Ol' Fred     		cisin at

More information about the cctech mailing list