lynchaj wrote:


> If anyone has one or more of these hard disk drives they would like to
> sell me, please contact me or reply here.

ST-412 were common in early IBM PCs. That should be the easiest way to
find one.
ST-506 can be found in old Apple profile 5mb hard disks, I also have one
inside my Northstar Advantage, AFAIK the ST-506 was the first 5,25" hard
drive ever produced.

Andrea

--

http://myretrocomputing.altervista.org

lynchaj wrote:
> Hi,
>
> I am restoring a Vector Graphic computer and would like to add a hard
> disk drive. The controller only supports two kinds of hard disk
> drives, the ST506 and ST412. The HD controller is hard coded to
> support these DRIVES not just a compatible interface and relies on a
> special signal present only in these actual drives.
>
> If anyone has one or more of these hard disk drives they would like to
> sell me, please contact me or reply here.
>
> Thanks!
>
> Andrew Lynch

Hi Andrew,

look at Ebay Item number: 270199570627.
I'm NOT the seller, just found this!

Happy new year,

Uwe.

>In case you haven't noticed them, 4drives.com (also an ebay store)

Awww, they're not related to the old old shareware of the same name
(supported a secondary hard drive controller under DOS).

On Dec 30 2007, 11:02 am, lynchaj wrote:
> Hi,
>
> I am restoring a Vector Graphic computer and would like to add a hard
> disk drive. The controller only supports two kinds of hard disk
> drives, the ST506 and ST412. The HD controller is hard coded to
> support these DRIVES not just a compatible interface and relies on a
> special signal present only in these actual drives.
>
> If anyone has one or more of these hard disk drives they would like to
> sell me, please contact me or reply here.
>
> Thanks!
>
> Andrew Lynch


Hi All
From another forum, I found that Andrew truly needed a ST506 or
ST412.
It seems that the system he is connecting to needs to have a specific
hardware modification done to the drive that was only originally done
on these specific drives.
There may be other drives that would work but each would require a
little more analysing of the drives control board.
It seems he needs to create a continuous Index signal, that is not
gated by select.
At one time, the ST506s were easy to get. They'd modified MSDOS
such that it would no longer work with the 5Meg drives. Many early
machines were upgraded with ST225s or ST251s and ST506s were in
piles at local surplus shops. Today they are quite rare.
Dwight

dkelvey@hotmail.com wrote:

(snip)

> From another forum, I found that Andrew truly needed a ST506 or
> ST412.
> It seems that the system he is connecting to needs to have a specific
> hardware modification done to the drive that was only originally done
> on these specific drives.
> There may be other drives that would work but each would require a
> little more analysing of the drives control board.
> It seems he needs to create a continuous Index signal, that is not
> gated by select.

The index signal is on the 34 pin cable, and is shared between
drives connected to the same controller. A continuous index signal
would only make sense with a one drive system. In that case,
ground the drive select 0 line and it should generate continuous
index pulses. The controllers I knew kept the last drive select
line low until another was selected, but maybe not all do that.

http://en.wikipedia.org/wiki/ST506

-- glen

On Jan 9, 11:17 am, glen herrmannsfeldt wrote:
> dkel...@hotmail.com wrote:
>
> (snip)
>
> > From another forum, I found that Andrew truly needed a ST506 or
> > ST412.
> > It seems that the system he is connecting to needs to have a specific
> > hardware modification done to the drive that was only originally done
> > on these specific drives.
> > There may be other drives that would work but each would require a
> > little more analysing of the drives control board.
> > It seems he needs to create a continuous Index signal, that is not
> > gated by select.
>
> The index signal is on the 34 pin cable, and is shared between
> drives connected to the same controller. A continuous index signal
> would only make sense with a one drive system.

This is true but that is a requirement of the system he has.

In that case,
> ground the drive select 0 line and it should generate continuous
> index pulses. The controllers I knew kept the last drive select
> line low until another was selected, but maybe not all do that.

You are assuning that the system in question parallels all the Index
signals. This may not be true for this particular system. It may
not be clear that the system was even intended to support multiple
drives either.
Remember, the Index signal could be split out as an independent
signal from each drive. It is just a wire.
Dwight


>
> http://en.wikipedia.org/wiki/ST506
>
> -- glen

>

dkelvey@hotmail.com wrote:

> On Jan 9, 11:17 am, glen herrmannsfeldt wrote:

(I wrote)
>>The index signal is on the 34 pin cable, and is shared between
>>drives connected to the same controller. A continuous index signal
>>would only make sense with a one drive system.

> This is true but that is a requirement of the system he has.

>> In that case,

>>ground the drive select 0 line and it should generate continuous
>>index pulses. The controllers I knew kept the last drive select
>>line low until another was selected, but maybe not all do that.

> You are assuning that the system in question parallels all the Index
> signals. This may not be true for this particular system. It may
> not be clear that the system was even intended to support multiple
> drives either.

The interface assumes it by having two selct lines. If a system
doesn't support two drives, it should ground at least one of the
select lines. Whether or not the system supports more than one,
there is only one index line on the cable. (I checked to be sure
it wasn't on the 20 pin cable.)

I might see the problem, though. Pin 2 is Reduced Write Current
for the ST506, and ignored for the ST412. For drives with more than
eight heads, it is a head select. The controller will ground this
for certain cylinders, which will not work for a drive with more
than eight heads, or one that checks the line anyway. In that case,
disconnect this line.

ST506/ST412 have at most four heads, so the controller will probably
only work with four. It may also limit the cylinders. See:

http://www.bitsavers.org/pdf/seagate...nual_Apr82.pdf

-- glen


> Remember, the Index signal could be split out as an independent
> signal from each drive. It is just a wire.

On Jan 9, 5:47*pm, glen herrmannsfeldt wrote:
---snip---
>
> The interface assumes it by having two selct lines. *If a system
> doesn't support two drives, it should ground at least one of the
> select lines. * Whether or not the system supports more than one,
> there is only one index line on the cable. *(I checked to be sure
> it wasn't on the 20 pin cable.)
>
Hi Glen
You are still missing the point. This isn't a PC. It may have
multiple 34 pin cables. There is no reason it has to parallel
the 34 pin cables. It is not a standard controller and has no
reason to do any particular method of getting the control
signals to the drives.
Think a little outside the box for a minute.
Dwight

dkelvey@hotmail.com wrote:

> You are still missing the point. This isn't a PC. It may have
> multiple 34 pin cables. There is no reason it has to parallel
> the 34 pin cables. It is not a standard controller and has no
> reason to do any particular method of getting the control
> signals to the drives.
> Think a little outside the box for a minute.

PC doesn't have anything to do with it.

The 34 pin cable is part of the ST506/ST412 standard.
(Well, it started the standard.)

Those pins exist on the cable whether it is connected to
one or four drives. Yes, it could be designed for one, but
it still has to pull one of the drive select pins low, and
when it does the index pin should be active. If no drive
select is low, the index pin must not be active.
It looks like stock drives come with all four drive select
pins connected.

The signals have to arrive at the drive on the 34 pin
and 20 pin cables, or it isn't ST506/ST412.

As I said, though, there is a difference between the
ST506 and later drives, in that the ST506 uses pin 2
as a Reduced Write Current pin. The ST412 ignores it,
but later drives use it for a fourth head select line.

The controller should pull it low for cylinders 128
and higher, which will fail for a drive that uses it
for head select.

-- glen

On Jan 9, 10:56*pm, "dkel...@hotmail.com" wrote:
> On Jan 9, 5:47*pm, glen herrmannsfeldt wrote:
> ---snip---
>
> > The interface assumes it by having two selct lines. *If a system
> > doesn't support two drives, it should ground at least one of the
> > select lines. * Whether or not the system supports more than one,
> > there is only one index line on the cable. *(I checked to be sure
> > it wasn't on the 20 pin cable.)
>
> Hi Glen
> *You are still missing the point. This isn't a PC. It may have
> multiple 34 pin cables. There is no reason it has to parallel
> the 34 pin cables. It is not a standard controller and has no
> reason to do any particular method of getting the control
> signals to the drives.
> *Think a little outside the box for a minute.
> Dwight
-----------

Still getting into arguments, eh Dwight ;-)

I should probably let Andrew respond, but what the heck...

You're both missing some details:

The particular controller in question does indeed only support one
hard disk but it's a combined FD/HD controller and some of the signals
(e.g. Step, Dir, HS0, Trk0, WG) are shared by the HD and the FD, so a
permanently selected HD would presumably not work even though there is
only one and the index signals are separate. In fact, the HD is
largely treated as just another floppy with a few unique signals and
separate cables as far as the interface is concerned.

On the other hand, the differences between the 412 and 506 don't
matter because the controller knows about those two specific drives
and can handle them differently as required.

In any case, as discussed on another list, modifying a drive to issue
a constant index would be trivial, but since the parameters are hard-
coded in the controller any substituted drive would have to match the
406 or 512 pretty closely in terms of geometry, step rates, etc.

mike

On Jan 10, 1:42*am, MikeS wrote:
> On Jan 9, 10:56*pm, "dkel...@hotmail.com" wrote:
>
>---snip---
> -----------
>
> Still getting into arguments, eh Dwight ;-)
>
> I should probably let Andrew respond, but what the heck...
>
> You're both missing some details:
>
>
Hi Mike
Thanks for the clarification.
My problem with Glen is that he almost insist that one
couldn't make a machine with multiple 34 pin cables.
There is no physical or electronic reason why this couldn't
be done. It would be like saying that Apple could create
their own disk format because others were using IBMs
format on the disk drives. The cable is just a bunch of
wires. An engineer can do anything he or she wants with
them to solve a particular problem. I could even believe
that they might remove the index signal pair at each
connector to create a separate one for each drive.
The twisted select line used in PC was not in the original
spec for floppy drives but became standard usage.
Who would have thought that one could make drive select
3 the select for the second drive.
Dwight

dkelvey@hotmail.com wrote:

(snip)

> My problem with Glen is that he almost insist that one
> couldn't make a machine with multiple 34 pin cables.

No, I wasn't trying to say that. I was saying that in that
case, as far as the cable and drive are concerned it is
a one drive system.

> There is no physical or electronic reason why this couldn't
> be done. It would be like saying that Apple could create
> their own disk format because others were using IBMs
> format on the disk drives. The cable is just a bunch of
> wires. An engineer can do anything he or she wants with
> them to solve a particular problem. I could even believe
> that they might remove the index signal pair at each
> connector to create a separate one for each drive.

They could, but it is only active for a selected drive.
That has to be true if you allow more than one drive on
the cable, and ST506/ST412 did.

> The twisted select line used in PC was not in the original
> spec for floppy drives but became standard usage.
> Who would have thought that one could make drive select
> 3 the select for the second drive.

True, but the INDEX pin is still active only for
the selected drive.

-- glen

On Jan 10, 10:34*am, "dkel...@hotmail.com"
wrote:
> On Jan 10, 1:42*am, MikeS wrote:> On Jan 9, 10:56*pm, "dkel...@hotmail.com" wrote:
>
> >---snip---
> > -----------
>
> > Still getting into arguments, eh Dwight ;-)
>
> > I should probably let Andrew respond, but what the heck...
>
> > You're both missing some details:
>
> Hi Mike
> *Thanks for the clarification.
> *My problem with Glen is that he almost insist that one
> couldn't make a machine with multiple 34 pin cables.
> There is no physical or electronic reason why this couldn't
> be done. It would be like saying that Apple could create
> their own disk format because others were using IBMs
> format on the disk drives. The cable is just a bunch of
> wires. An engineer can do anything he or she wants with
> them to solve a particular problem. I could even believe
> that they might remove the index signal pair at each
> connector to create a separate one for each drive.
> *The twisted select line used in PC was not in the original
> spec for floppy drives but became standard usage.
> *Who would have thought that one could make drive select
> 3 the select for the second drive.
> Dwight
--------
Well, you're arguing over an essentially irrelevant hypothetical
situation; I've certainly never seen an HD controller with separate
control cables for each drive, although of course in theory it is
possible.

But as a matter of fact this particular controller *is* close to what
you're talking about, albeit with a hard disk on one cable and
floppies on the other; the main controller logic handles both the
floppies and the hard disk and the common signals are brought out on
both 34pin cables while the unique signals like /INDEX are brought out
separately on their respective cable. That's how it can run the HD
with a constant index (presumably to keep the PLL in sync while the HD
is deselected) without interfering with the FDC signal.

And Glen, yes, I imagine that another drive could probably be used
but, knowing Andrew, I suspect he'd prefer if possible to at least
start with the original drives that the controller specifies before
exploring substitutes.

Having just spent quite a bit of time trying to replace some old
Tandon and Micropolis 100TPI floppy drives with 96TPI 1.2Mb units, I
was surprised how many subtle differences there can be among drives
that at first glance should be more or less compatible. Hard disks
would probably not be quite as bad, but nevertheless issues like
buffered seek, step rates (even if higher) etc. may well cause
compatibility problems with such an old controller, especially since
it's SSI TTL and doesn't use any industry-standard controller chips.

mike

MikeS wrote:
(snip)

> And Glen, yes, I imagine that another drive could probably be used
> but, knowing Andrew, I suspect he'd prefer if possible to at least
> start with the original drives that the controller specifies before
> exploring substitutes.

They might be pretty hard to find, though.

Slightly later Seagate models might be somewhat easier to find, and
one would hope not so different. As far as I remember, the 20MB
drives were much more popular from the beginning of the PC and XT years.

-- glen

> From another forum, I found that Andrew truly needed a ST506 or
> ST412.
> It seems that the system he is connecting to needs to have a specific
> hardware modification done to the drive that was only originally done
> on these specific drives.

Another tidbit of info that wasn't mentioned here, but is quite
relevant to the discussion, is that the system in question,
specifically a Vector Graphic S-100 bus machine with a VEDMCS
controller, didn't support more than one hard drive at a time.

De

On Jan 11, 6:17*pm, glen herrmannsfeldt wrote:
> Dennis Boone wrote:
> > *> *From another forum, I found that Andrew truly needed a ST506 or
> > *> ST412.
> > *> It seems that the system he is connecting to needs to have a specific
> > *> hardware modification done to the drive that was only originally done
> > *> on these specific drives.
> > Another tidbit of info that wasn't mentioned here, but is quite
> > relevant to the discussion, is that the system in question,
> > specifically a Vector Graphic S-100 bus machine with a VEDMCS
> > controller, didn't support more than one hard drive at a time.
>
> It seems that the ST506 and ST412 do have a jumper option that makes
> them appear selected without any select lines going low. *As described,
> the only difference is that the LED on the front does not come on.
>
> As I previously noted, drives with more than eight heads will use the
> pin marked "reduced write current" as a head select. *That will likely
> cause problems unless pin 2 is cut. *(Or put tape over the pin before
> installing the connector.)
>
> -- glen


Hi Glen,

Thanks for the replies and helpful information. Thanks also to
Dwight, Dennis, and Mike and others for their contributions. I do
appreciate your help!

I agree that substituting a later model interface compatible hard disk
like a ST225 is possible *in theory* and is what I would like to do
eventually. However, since the Vector Graphic system I am restoring is
quite old its VEDMCS hard disk controller has some peculiarities which
require special drive configuration.

As Mike said, I would like to get the default configuration working
before trying any modified configurations due to subtle
incompatibilities in drives. At this point the whole discussion is
entirely theoretical so I am looking for the lowest risk solution
possible before cutting in real hardware.

As I have come to discover, the Vector Graphic VEDMCS hard disk
controller supports only a single hard drive at a time -- either a
ST-506 or a ST-412. It requires a constant /INDEX signal which is
accomplished by doing a "cut and jumper" on the ST-506/ST-412 drive
controller board (not the hard drive controller -- the drive's PCB is
modified) which brings the constant /INDEX signal out to the interface
on pin 20 of the 34 pin edge connector.

Essentially, this is accomplished by isolating the NAND output driver
gate for /INDEX, tying the "OUTPUT ENABLE" signal high so that INDEX
is passed though as /INDEX constantly. The reason, I am told, is the
VEDMCS needs the signal for its PLL to sync before being able access
the hard disk. I believe this practically ensures the VEDMCS will
only be able to support a single hard disk since it is effectively
"tied" to single hard disk signal.

I have since found the ST-506 schematics and an ST-506 hard drive and
will be putting this theory to the test in the near future. More
information as it comes available!

Thank you very much for your help and I do appreciate you taking the
time to reply to my posting.

Andrew Lynch

lynchaj wrote:

(snip)

> I agree that substituting a later model interface compatible hard disk
> like a ST225 is possible *in theory* and is what I would like to do
> eventually. However, since the Vector Graphic system I am restoring is
> quite old its VEDMCS hard disk controller has some peculiarities which
> require special drive configuration.

> As Mike said, I would like to get the default configuration working
> before trying any modified configurations due to subtle
> incompatibilities in drives. At this point the whole discussion is
> entirely theoretical so I am looking for the lowest risk solution
> possible before cutting in real hardware.

> As I have come to discover, the Vector Graphic VEDMCS hard disk
> controller supports only a single hard drive at a time -- either a
> ST-506 or a ST-412. It requires a constant /INDEX signal which is
> accomplished by doing a "cut and jumper" on the ST-506/ST-412 drive
> controller board (not the hard drive controller -- the drive's PCB is
> modified) which brings the constant /INDEX signal out to the interface
> on pin 20 of the 34 pin edge connector.

Just wondering, is the floppy side hard sectored? The only reason
I can think of for a PLL on the index line is for indexing sectors.
I once thought about using a PLL on index to generate hard sector
index pulses with a soft sector drive. And yes, it would take a
while to lock, so would need a continuous index.

> Essentially, this is accomplished by isolating the NAND output driver
> gate for /INDEX, tying the "OUTPUT ENABLE" signal high so that INDEX
> is passed though as /INDEX constantly. The reason, I am told, is the
> VEDMCS needs the signal for its PLL to sync before being able access
> the hard disk. I believe this practically ensures the VEDMCS will
> only be able to support a single hard disk since it is effectively
> "tied" to single hard disk signal.

I suppose the other signals are shared with the floppy, and so
must stay controlled with the select line. In that case I would
put a board in between with 7438's and 7414's to control the lines
that needed to be controlled, and pass index through.

-- glen

On Jan 12, 11:49*pm, glen herrmannsfeldt
wrote:
> lynchaj wrote:
>
> (snip)
>
>
>
>
>
> > I agree that substituting a later model interface compatible hard disk
> > like a ST225 is possible *in theory* and is what I would like to do
> > eventually. However, since the Vector Graphic system I am restoring is
> > quite old its VEDMCS hard disk controller has some peculiarities which
> > require special drive configuration.
> > As Mike said, I would like to get the default configuration working
> > before trying any modified configurations due to subtle
> > incompatibilities in drives. *At this point the whole discussion is
> > entirely theoretical so I am looking for the lowest risk solution
> > possible before cutting in real hardware.
> > As I have come to discover, the Vector Graphic VEDMCS hard disk
> > controller supports only a single hard drive at a time -- either a
> > ST-506 or a ST-412. *It requires a constant /INDEX signal which is
> > accomplished by doing a "cut and jumper" on the ST-506/ST-412 drive
> > controller board (not the hard drive controller -- the drive's PCB is
> > modified) which brings the constant /INDEX signal out to the interface
> > on pin 20 of the 34 pin edge connector.
>
> Just wondering, is the floppy side hard sectored? *The only reason
> I can think of for a PLL on the index line is for indexing sectors.
> I once thought about using a PLL on index to generate hard sector
> index pulses with a soft sector drive. *And yes, it would take a
> while to lock, so would need a continuous index.
>

Yes, the Vector Graphics FDCs are all hard sectored for 16 sector
floppy disks.


> > Essentially, this is accomplished by isolating the NAND output driver
> > gate for /INDEX, tying the "OUTPUT ENABLE" signal high so that INDEX
> > is passed though as /INDEX constantly. *The reason, I am told, is the
> > VEDMCS needs the signal for its PLL to sync before being able access
> > the hard disk. *I believe this practically ensures the VEDMCS will
> > only be able to support a single hard disk since it is effectively
> > "tied" to single hard disk signal.
>
> I suppose the other signals are shared with the floppy, and so
> must stay controlled with the select line. *In that case I would
> put a board in between with 7438's and 7414's to control the lines
> that needed to be controlled, and pass index through.
>
> -- glen- Hide quoted text -
>
> - Show quoted text -

Thanks!

Andrew Lynch

On Jan 12, 11:49*pm, glen herrmannsfeldt
wrote:


> Just wondering, is the floppy side hard sectored? *The only reason
> I can think of for a PLL on the index line is for indexing sectors.
> I once thought about using a PLL on index to generate hard sector
> index pulses with a soft sector drive. *And yes, it would take a
> while to lock, so would need a continuous index.
---
Ummm... I don't follow you there; yes, as Andrew says the FD is hard
sectored, but how is that relevant to the HD index?

Why would a PLL not be required for the HD? This would be news for the
designers of this controller, since as it happens the HD index signal
goes directly to a 4046 PLL, while the FD index first goes through a
discriminator to separate the sector pulses from the index pulse
before the two are multiplexed onto a common bus.

mike




>
> > Essentially, this is accomplished by isolating the NAND output driver
> > gate for /INDEX, tying the "OUTPUT ENABLE" signal high so that INDEX
> > is passed though as /INDEX constantly. *The reason, I am told, is the
> > VEDMCS needs the signal for its PLL to sync before being able access
> > the hard disk. *I believe this practically ensures the VEDMCS will
> > only be able to support a single hard disk since it is effectively
> > "tied" to single hard disk signal.
>
> I suppose the other signals are shared with the floppy, and so
> must stay controlled with the select line. *In that case I would
> put a board in between with 7438's and 7414's to control the lines
> that needed to be controlled, and pass index through.
>
> -- glen- Hide quoted text -
>
> - Show quoted text -