reading ancient floppy formats

[D. Hugh Redelmeier]

| From: Allan Swanepoel <allanice001 at gmail.com>

| firstly, it could be that the floppies are stuffed.

Sure.  In fact, the reason I'm trying to read them now is to rescue
them.

| These are among the
| first magnetic disks I know of.

Really?  The IBM RAMAC from the 1950s was the first magnetic disk I
know of.  The oldest one I used much were IBM 1311 drives and disk
packs from the early 1960s.

| Any electronic interference could make these
| go haywire.

Yeah, even the earth's magnetic field is supposed to be able to do
that.

| How close are they to your cellphone?

Not close.  CRT's are probably worse.  But closed magnetic fields
apparently drop off with the cube of the distance.

| secondly, there could be a blown circuit on the controller board. this is
| capable of giving bad reads, as is a damaged cable.

Sure.  But this drive appears to be working during very-slight testing
with an HD disk.

QD is a funny beast.  The write signals are just like DD, not HD.  But
the track density (and thus write gap size) are like HD.

I think that HD drives can read QD disks -- after all, they can read
DD disks.  But the controller settings are not normal.

| and lastly, WTF are you doing with a machine that belongs in a museum next
| to the the freaking ark???

I'm trying to rehabilitate my museum before it is too late.

My NABU's hard drive is toast and not many ST-506-style disks pass my
way.
  http://en.wikipedia.org/wiki/ST-506
Worse: the boot ROM and UNIX have wired-in drive geometry.

BTW, the Nabu has an 5MHz 8086 CPU and 1M of RAM (homebrew upgrade
from 512K RAM).  It supports 4 users with 7th edition UNIX.  In fact,
there is more memory than required: I used 512k as a RAMDISK for /tmp.
Networking?  Sure, it supported uucp :-)

Boot time?  Roughly the same as Linux now.  Sad.

Google found this
  http://www.classiccmp.org/dunfield/nabu/n1600a.jpg
  http://www.classiccmp.org/dunfield/nabu/n1600b.jpg

| From: Allan Swanepoel <allanice001 at gmail.com>

| > On Sun, Sep 20, 2009 at 10:55 PM, D. Hugh Redelmeier <hugh at mimosa.com>wrote:
| 
| >> The whole block is 0.  I don't know if that is expected (I don't
| >> remember how the machine that wrote it laid down a 7th edition file
| >> system on a floppy).  The machine was a Nabu 1600, not a PC clone (it
| >> predates PC clones).

| btw, if at first you don,t succeed, ask google
| http://minnie.tuhs.org/UnixTree/V7/

Yeah.  Henry Spencer (credit at the bottom of that page) shared a
cubicle with me in grad school in the 6th edition days and we're still
friends.

The file system on this particular floppy has some tar files (if the
handwritten label is correct).  I was refering to things like the inode
struct itself, the array of them, the number of reserved blocks at the
start, etc.

Plan A is to capture the raw bytes on the floppies and later (possibly
much later) dissect them to get the data.