Subject: Re: 8" disk drive project - maybe 3.5" project too???
From: Grant Stockly <grant at stockly.com>
Date: Tue, 06 Nov 2007 11:32:34 -0900
To: "General Discussion: On-Topic Posts Only" <cctech at
classiccmp.org>
Write trim erase is a signal to the head assembly
(driver) that turns on
the erase segment of the head for erasing the area either side of the head.
It's also called tunnel erase. This gives some tolerence to off track reading
as the intertrack gaps are "clean". Tunnel erase is still part of all
floppy drives but is activated by Write-enable and you never "hear" of it.
In 1976 all this was new and magical until Sugart came out with the much
simplified 5.25 SA400.
From reading the information on the Internet I have pretty much
guessed that much, that it was done automatically on new drives.
Any
comments or ideas on the idea? Is it worth trying? I've got a
very weird tarbell card that formats and uses 3.5" disks as a 70k
mini floppy. I guess anything is possible. : )
Myself I'd persue something using a 5.25 or 3.5" drive with a current
softsector interface to the drive and enough CPU smarts to fake looking
like the Altair interface which was dumb as a rock and depended on the
8080 to do most everything.
That is what I was thinking about too. I was thinking about a
controller that hot wired into the existing drive card set. Drive 1
would activate the fake 3.5" disk or flash memory and drive 0 would
activate the 8" drive for example. That would allow for easy
migration from 8" to a different format.
If you can that works.
It would also be nice to have a media readable on a PC.
A 3.5" disk
with the fake hard sector formatting wouldn't be possible...
Thats impostant as most users will find the availble on media offerings
a bit thin is you go with hard sector (faked or real).
What do you think about drive RPM? Should I be able to
fake the 32
hard sectors at any rate I want as long as its as fast or slower than
the 8" drive?
I won't touch that. It's one of many warrens and burrows to dive into.
The 3.5" disk should be writing the bits at
whatever
bit rate is requested, right? The only reason I want to do the 3.5"
disk interface with faking 32 hard sectors is because it would be a
fast way to get going. Hardly any hardware development required.
Drive rpm is always a problem. Even for soft sector the data sep has
to deal with that. Faked hard sector is just causing additional pain.
Take it from someone that uses a NS* hard sector system where the media
and the controller are far more common and pretty reliable and it's a
major pain. Media is hard to get and hard sector always was a closed
gate to transfering software. I felt the nonportable media was always
a major pain and handicap. I still maintain a "stock NS* Horizon"
as an artifact but due to media availability and portability my
second Horizon that I do use a lot has a softsector controller and
hard disk as primary storage.
I'd suggest two things:
Do complete Altair disk and drives and 1 floppy with it and make
additional media the users problem like in 1976, 1980, 1985 and
more so now.
Or a completely new controller (non replica) that acts like the Altair
part (runs unaltered altair basic and all) but is normal softsector
using easier to get 5.25 or 3.5 media with portability (and PC as a
disk generator) in mind. This would make the system more easily shown
and used. Also the user can than have CP/M as well as it was also
historically important to any Altair user of the time.
Like the 8800BT I have the few that had the Altair contoller usually added
a Tarbell or other later on to gain access to the larger software market.
Those that started with the base Altair and were slower to go to floppy
often didn't even bother with the altair disk and use NS*, Tarbel or one
of a handful of others with NS*dos or CP/M as the OS platform. It would
be interesting to see how many Altairs (all versions of the 8800) wer sold
and then how many disk controllers. I suspect based on my experince of
the time it will be something like 20:1 likely greater. The reason was
at that time by 1977 the cost differential was fairly great to have
Altair floppy disk over brand-x and CP/M was seen a cleary the up and
comming OS. It also was a factor that long shipping delays by MITS
at times and problems experinced by some tended to push some more toward
the after market add ons. My experience, I built and tested the first
Altair FDC sold on LI for a company.
Just my opinion.
Allison
Thanks!
Grant