Northstar Advantage disk problem...PLS help!
Andrea
yesterday I turned on my Advantage after some months and there is this
strange new problem with the disk drive: it doesn't format disks
anymore....when it starts writing it says T4 D1 S0000 that according to
the manual means No INDEX PULSE: wrong type of disk or bad seated disk.
If I format an already written disk it goes straight to the end. It
works. I tried to copy files from HD to the disk drive on a formatted
disk and it works, it writes correctly. I tried also the diagnostic disk
and it asks to put a blank disk in drive 1, I do it but it keeps asking
it again...just like if there was no disk in the drive. I did a crazy
thing, I replaced the drive with the one from a TRS80 model 3 and I get
all the same errors, the funny thing is that the drive works, it can
read the Advantage disks...but still it doesn't format unformatted or
unwritten disks. So I believe the problem could be in a ROM or in a
logic IC....Anyone can help me?
By the way I found out I have the Xterm program and I'd like to send my
disks to the PC to create disk images...I think I need to build a cable
and some software on the PC side, can anyone help me out?
It would be great to make images of the programs and system disks for
future use...
Sorry for the long message.
Thanks in advantage to everyone that will help me.
Bye
Andrea
nos...@nouce.bellatlantic.net
wrote:
I'll bet your using the wrong media. NS* Advantage is hardsector
media not PC soft sector. The differnce is the number of sector holes
11 for NS and 1 for the softsector stuff.
Allison
Andrea
> I'll bet your using the wrong media. NS* Advantage is hardsector
> media not PC soft sector. The differnce is the number of sector holes
> 11 for NS and 1 for the softsector stuff.
>
> Allison
You are so right!
This morning I checked all disks and they are soft sectored Apple II or
Commodore disks.
Later I found another box with disks I had used with the advantage and
they are all hard sectored and they work fine.
Thanks for the info, I didn't know about this differences in 5,25 disks.
Bye
Andrea
nos...@nouce.bellatlantic.net
wrote:
Great to hear success. It was my first guess as I have Both NS*
Horizon and Advantage so I'm familiar with the disk issues.
Good luck,
Allison
Lee Hart
>> I'll bet your using the wrong media. NS* Advantage is hardsector
>> holes (11 for NS and 1 for the softsector stuff).
Andrea wrote:
> You are so right! This morning I checked all disks and they are
> soft sectored Apple II or Commodore disks. Later I found another
> box with disks I had used with the Advantage and they are all
> hard sectored and they work fine.
Just an aside. Heath H8 and H89 computers used these same
10-hard-sectored disks with their hard-sector controller. At one time I
looked into rewriting the BIOS to use soft-sector disks in the
hard-sector controller. I played with it enough to see that it looked
workable.
The stock Heath BIOS has ten 256-byte single-density sectors per track.
What I did was to increase the sector size to 3k bytes, so each track
had was just one large sector. It required a large sector buffer (3k of
precious RAM!), but read/write was considerably faster. It also
increased disk capacity from 40 x (10 x 256) = 100k to 40 x 3k = 120k
bytes.
I used a single byte checksum per sector (same as the stock system). The
only drawback I found in limited testing was that the error rate was
higher. I thought about rewriting it to use a 16-bit CRC check (which
would have to be calculated in software), but never got around to it. it
wasn't hard enough to get hard-sector disks at the time to bother
pursuing.
Maybe now that hard-sector disks are harder to get, it would be worth
re-thinking this idea? Would the Northstar controller also be capable of
such a trick?
--
*BE* the change that you wish to see in the world.
-- Mahatma Gandhi
--
Lee A. Hart, 814 8th Ave N, Sartell MN 56377, leeahart_at_earthlink.net
nos...@nouce.bellatlantic.net
wrote:
Lee,
I have looked at it. Answer is not really. With a mod or two it may.
The problem is the sector counter is also reset by the passage of the
sector hole. If that doesnt happen the whole sector timing is off.
The logic used by the NS* controller is quite primitive but effective
and there lies the difficulty in doing soft sectoring with it.
Allison
Barry Watzman
auxilliary board that lies between the drive and the controller that
"generates" the sector pulses and simulates a hard sector disk from a
soft-sector disk.
And yet another idea is to build a "sector hole punch"
There are various solutions, none ideal or perfect.
nos...@nouce.bellatlantic.net
<Watzma...@neo.rr.com> wrote:
>Another idea that has been proposed is to build a small circuit on an
>auxilliary board that lies between the drive and the controller that
>"generates" the sector pulses and simulates a hard sector disk from a
>soft-sector disk.
I tried that years back.
The setup was a clocked osc and counter chain that was reset off the
softsector index. The clock was scaled to match nominal rpm and a
good direct drive (teac FD55B) was used) this also worked. It assumed
the rotation rate was exact and didn't drift much per rotation.
It worked meaning it formatted and read and wrote media
with passable error rate, though not as good as the real thing.
Also it didn't work at all (no logic to differentiate) with hardsector
as there were too many "index" pulses. Also I'd only tried it with
the SD controller not the later DD controller.
Allison
Barry Watzman
exact actual rotation rate, instead of assuming the nominally correct
300rpm. Certainly possible (more so now than "then"), but not by any
means effortless.
Dave Dunfield
>>>auxilliary board that lies between the drive and the controller that
>>>"generates" the sector pulses and simulates a hard sector disk from a
>>>soft-sector disk.
>>
>> I tried that years back.
>> The setup was a clocked osc and counter chain that was reset off the
>> softsector index. The clock was scaled to match nominal rpm and a
>> good direct drive (teac FD55B) was used) this also worked. It assumed
>> the rotation rate was exact and didn't drift much per rotation.
>>
>> It worked meaning it formatted and read and wrote media
>> with passable error rate, though not as good as the real thing.
>> Also it didn't work at all (no logic to differentiate) with hardsector
>> as there were too many "index" pulses. Also I'd only tried it with
>> the SD controller not the later DD controller.
>A better approach would be to use the index pulses to determine the
>exact actual rotation rate, instead of assuming the nominally correct
>300rpm. Certainly possible (more so now than "then"), but not by any
>means effortless.
Shouldn't be all that tough - I've been planning just such a device for
when I "run out" of hard-sector media ... so far I haven't had to build
it... but the day is coming!
With a little single-chip micro, you could fairly easily build the device
to:
- Generate 'n' sector pulses evenly spaced between the physical
index pulses. ('n'=10 for NorthStar).
- Time the interval between index to index pulse on an ongoing basis
and adjust the intervals for the sector pulses accordingly.
- Remember the timing for up to 4 drives, so that once you have accessed
a drive, you can begin generating sector pulses after the first occurance
of Index (otherwise you would need two revolutions to determine the speed
- three to do it reliably).
- Recognize when an index pulse occurs "too soon" after motor-on, indicating
that the drive may not be up to speed yet and wait for the next one.
- Recognize a hard-sector disk and pass the pulses transparently, so that
you can use either media without having to switch.
I'd be suprised if someone hasn't already built this. If someone really needs
it and it doesn't already exist, I can try and move it up my "to do" list for
winter projects.
Regards,
Dave
--
Dunfield Development Systems http://www.dunfield.com
Low cost software development tools for embedded systems
Software/firmware development services Fax:613-256-5821
nos...@nouce.bellatlantic.net
<Watzma...@neo.rr.com> wrote:
>A better approach would be to use the index pulses to determine the
>exact actual rotation rate, instead of assuming the nominally correct
>300rpm. Certainly possible (more so now than "then"), but not by any
>means effortless.
>
I considered a PLL locked to the index pulse and the hardware to do
that was a PITA then and now. *If* I wer to do it today I'd take a
PIC and programit to do the software synthesis of a PLL and provide
the correct pulses. It could also shut itself off if the media was
really 10sector too.
Allison
nos...@nouce.bellatlantic.net
Dave.D...@use.techsupport.link.on.my.website (Dave Dunfield)
wrote:
>
>Shouldn't be all that tough - I've been planning just such a device for
>when I "run out" of hard-sector media ... so far I haven't had to build
>it... but the day is coming!
>
>With a little single-chip micro, you could fairly easily build the device
>to:
>
>- Generate 'n' sector pulses evenly spaced between the physical
> index pulses. ('n'=10 for NorthStar).
>
>- Time the interval between index to index pulse on an ongoing basis
> and adjust the intervals for the sector pulses accordingly.
>
>- Remember the timing for up to 4 drives, so that once you have accessed
> a drive, you can begin generating sector pulses after the first occurance
> of Index (otherwise you would need two revolutions to determine the speed
> - three to do it reliably).
>
>- Recognize when an index pulse occurs "too soon" after motor-on, indicating
> that the drive may not be up to speed yet and wait for the next one.
>
>- Recognize a hard-sector disk and pass the pulses transparently, so that
> you can use either media without having to switch.
>
>I'd be suprised if someone hasn't already built this. If someone really needs
>it and it doesn't already exist, I can try and move it up my "to do" list for
>winter projects.
One suggestion. Use drives like the TEAC or Toshibas that use a
ceramic resonator or crystal clock to establish the motor speed.
Those by far are the most stable. The worst were any of the belt
drives (MPI, SA4xx, Tandon) as the motor uses a tach drive and
DC servo that drifts with time and temperature. The SA400s were
by far the worst for motor speed variations and bit jitter due to
sloppy spindle bearings.
Allison
Lee Hart
> Shouldn't be all that tough - I've been planning just such a device
> for when I "run out" of hard-sector media ... so far I haven't had
I did this way back around 1985. I used a phase locked loop to lock a
10x clock onto the index hole. This was just a couple simple old
4000-series CMOS chips. As Allison said, it worked but had a higher
error rate. The main problem seemed to be that floppies with a lot of
friction in drives that had weak rotation motors could have too large a
speed variation.
On read, I suppose the way to do it is to use the read data itself to
tell you the rotational speed. Use this with the PLL to generate the
fake index hole pulses.
--
Never doubt that the work of a small group of thoughtful, committed
citizens can change the world. Indeed, it's the only thing that ever
has! -- Margaret Mead
nos...@nouce.bellatlantic.net
wrote:
>On read, I suppose the way to do it is to use the read data itself to
>tell you the rotational speed. Use this with the PLL to generate the
>fake index hole pulses.
Aww, you had to give the secret away. I tried that, it works.
there are a few tricks that have to be applied as you will get
irregular clocks from the gaps but thats the trick.
I found the easiest is to use a really good drive that hold speed.
Then clocking off the index is reliable enough.
Allison
Dave Dunfield
>> for when I "run out" of hard-sector media ... so far I haven't had
>> to build it... I'd be suprised if someone hasn't already built this.
>I did this way back around 1985. I used a phase locked loop to lock a
>10x clock onto the index hole. This was just a couple simple old
>4000-series CMOS chips. As Allison said, it worked but had a higher
>error rate. The main problem seemed to be that floppies with a lot of
>friction in drives that had weak rotation motors could have too large a
>speed variation.
I have observed this - I find that more modern decent quality drives
with large "flat" direct drive solid-state motors tend to work a lot better
than the older SA-400 type pully system. Also, one would have to avoid
"bad" floppies - again, I find that newer disks tend to work a lot better.
I suppose one could add an optical encoder ring to the flywheel on the
bottom of the drive and track the speed to a much finer degree through
the entire revolution - for both read and write, however my initial tests
indicated that with a decent drive and disks, this would not be required -
but as I as noted earlier - I haven't actually had to build it yet, so this is
all conjecture.