16 Jun
2015
16 Jun
'15
8:25 a.m.
As long as we're talking philosophy, what do y'all think about emulating the TU58
drive, vs. emulating the TU58 *tape*?
I cannot properly express my opinion of that tape cartridge design even if I violate list
rules about use of profanity. But the drive itself isn't all that bad, aside from not
having enough motors to manipulate a well-designed tape cartridge mechanism.
Emulating the whole tape drive is pretty easy since it helpfully interfaces over a plain
old asynchronous serial port. But replacing the whole drive with an emulator, or worse yet
tethering its computer to a modern computer, leaves a bit of an empty feeling if one likes
their vintage machines to be original. There's something missing when you don't
hear the drive whirring, and the system boot completes within a modern attention span.
But what about emulating the tape cartridge, instead? Imagine a gizmo in the form factor
of a TU58 cartridge, containing a wheel for the capstan roller to engage, but connected to
an encoder instead of the ******* ************ ***** ** **** belt drive of an original
cartridge? Where the tape would normally be exposed, there is instead a magnetic head
which rests against the tape drive head like in one of those gizmos for injecting line
level audio into an audio cassette drive. It might need an external power source, but for
the sake of argument, let's pretend that a suitable rechargeable battery can be
embedded. Maybe it has an SD card slot on the rear, or maybe it looks just like a real
TU58 cartridge when inserted, and you swap the whole thing to change tapes (this is open
for discussion).
Would this be more or less acceptable in terms of keeping the system as close to original
as possible, vs. unplugging the original drive and plugging in a drive emulator?
No, I'm not going to build the thing. I'll just build my TU58 drive emulator to
fit in the cartridge slot but plug into the computer in place of the original drive, with
the cables snaked through the original drive mechanism. And I'll feel a little bit
dirty, but the thing will work reliably and will be easy to implement. I'm just
curious about the philosophical implications of my silly cartridge emulator idea.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
16 Jun
16 Jun
9:10 a.m.
As long as we're talking philosophy, what do y'all think about emulating the TU58
drive, vs. emulating the TU58 > *tape*?
I cannot properly express my opinion of that tape cartridge design even if I violate list
rules about use of
profanity. But the drive itself isn't all that bad, aside from not having enough
motors to manipulate a
well-designed tape cartridge mechanism.
The tape cartridge (along with QIC cartridges, etc) is an interesting design. Of course
the general problem is
to run the tape at constant linear speed past the head while winding it off one reel onto
the other. In general
there have been 3 solutions to this.
One is to drive the tape itself using a capstan and maybe a pinch roller. Then have the
takeup spool
driven too fast with a slipping clutch in-line. And a weak brake on the supply spool. This
of course is what
was done in audio recorders (reel to reel and cassette), Video recorders (ditto), etc.
Another is to simply drive the takeup spool but to measure the speed of the outside of the
tape wound on
it. thus keeping a constant tape speed. I had forgotten this one until I realised
that's how the Cipher F880
works. Or maybe to run the tape round a rotating guide that has a tacho disk on it.
I've not seen that used,
but I bet it was somewhere.
The last is to drive the outside of the tape wound on the spools. Of course this keeps
constant tape speed.
Mechanically it is by far the simplest. All you need is one motor driving a belt that is
tensioned against the
outside of the tape spools. This is how the TU58, etc, work. The problem is due to the
physical properties
of the tape and belt, which leads to tape wear and damage and a short life.
There are at least 2 types of TU58 cartridge (DEC called them 'metal base' and
'plastic base' I think. Is
one any better than the other?
Incidentally, given the fact that a constant motor speed -> constant tape speed, it
should be possible to
make a device to put the timing track on a blank tape for the TU58. Has anyone done that?
But what about emulating the tape cartridge, instead?
Imagine a gizmo in the form factor of a TU58 cartridge,
containing a wheel for the capstan roller to engage, but connected to an encoder instead
of the *******
************ ***** ** **** belt drive of an original cartridge? Where the tape would
normally be exposed, there is
Now that I want to see!
Even a module containing coupling head and encoder disk that slots into a TU58 drive and
which links
to an external box of electronics containing the SD card, etc, would be interesting.
You would, of course, not know which track it was reading, so you would have to output 2
blocks, one on
each track, at once. And how would you detect it was writing? Look for an extra signal at
the coupling head
or something?
-tony
9:21 a.m.
On Jun 16, 2015, at 12:10 PM, tony duell <ard at
p850ug1.demon.co.uk> wrote:
Another solution is to avoid the problem entirely by not requiring constant linear speed.
That?s what DECtape (the real one) does.
As long as we're talking philosophy, what do y'all think about emulating the TU58
drive, vs. emulating the TU58 > *tape*?
I cannot properly express my opinion of that tape cartridge design even if I violate list
rules about use of
profanity. But the drive itself isn't all that bad, aside from not having enough
motors to manipulate a
well-designed tape cartridge mechanism.
The tape cartridge (along with QIC cartridges, etc) is an interesting design. Of course
the general problem is
to run the tape at constant linear speed past the head while winding it off one reel onto
the other. In general
there have been 3 solutions to this.
One is to drive the tape itself using a capstan and maybe a pinch roller. Then have the
takeup spool
driven too fast with a slipping clutch in-line. And a weak brake on the supply spool.
This of course is what
was done in audio recorders (reel to reel and cassette), Video recorders (ditto), etc.
Or a variant: drive the takeup reel with a controlled torque motor and no slip clutch. I
think professional grade tape decks may have been built that way.
paul
9:51 a.m.
Mark, I've thought of that for my HP 9845, too. Sure would be nice to
fabricate something that's flexible hardware that can be programmed for the
peculiarities of various implementations. I'll put it on my list of things
to do once I'm done with my dissertation. :-)
On Tue, Jun 16, 2015 at 4:21 PM, Paul Koning <paulkoning at comcast.net> wrote:
design. Of
course the general problem is
--
Ian S. King, MSIS, MSCS, Ph.D. Candidate
The Information School <http://ischool.uw.edu>
Archivist, Voices From the Rwanda Tribunal <http://tribunalvoices.org>
Value Sensitive Design Research Lab <http://vsdesign.org>
University of Washington
There is an old Vulcan saying: "Only Nixon could go to China."
On Jun 16, 2015, at 12:10 PM, tony duell <ard
at p850ug1.demon.co.uk>
wrote:
>
> As long as we're talking philosophy, what do y'all think about
emulating the TU58 drive, vs. emulating the TU58 > *tape*?
>
> I cannot properly express my opinion of that tape cartridge design even
if I
violate list rules about use of
> profanity. But the drive itself isn't all
that bad, aside from not
having enough motors to manipulate a
well-designed tape cartridge mechanism.
The tape cartridge (along with QIC cartridges, etc) is an interesting to run the tape at constant linear speed past the
head while winding it
off one reel onto the other. In general
there have been 3 solutions to this.
Another solution is to avoid the problem entirely by not requiring
constant linear speed. That?s what DECtape (the real one) does.
One is to drive the tape itself using a capstan and maybe a pinch
roller. Then
have the takeup spool
driven too fast with a slipping clutch in-line.
And a weak brake on the
supply spool. This of course is what
was done in audio recorders (reel to reel and
cassette), Video recorders
(ditto), etc.
Or a variant: drive the takeup reel with a controlled torque motor and no
slip clutch. I think professional grade tape decks may have been built
that way.
paul
10:31 a.m.
On 2015-06-16 18:21, Paul Koning wrote:
And yet another (possibly the most common one on computers) is to have a
small drive wheel that pulls the tape at constant speed across the
heads, and then have some other construction that drives the tape reels
depending on tape tension or length. Think vacuum columns or spring
loaded arms.
And then we have drives like the TU80/TU81 which do not seem to fit into
any of the mentioned categories. I'm not entirely sure how they work,
but I think they are similar to the DECtape in this sense.
Johnny
On Jun 16, 2015, at 12:10 PM, tony duell <ard
at p850ug1.demon.co.uk> wrote:
Another solution is to avoid the problem entirely by not requiring constant linear speed.
That?s what DECtape (the real one) does.
As long as we're talking philosophy, what do y'all think about emulating the TU58
drive, vs. emulating the TU58 > *tape*?
I cannot properly express my opinion of that tape cartridge design even if I violate list
rules about use of
profanity. But the drive itself isn't all that bad, aside from not having enough
motors to manipulate a
well-designed tape cartridge mechanism.
The tape cartridge (along with QIC cartridges, etc) is an interesting design. Of course
the general problem is
to run the tape at constant linear speed past the head while winding it off one reel onto
the other. In general
there have been 3 solutions to this.
10:51 a.m.
And yet another (possibly the most common one on
computers) is to have a
small drive wheel that pulls the tape at constant speed across the
heads, and then have some other construction that drives the tape reels
depending on tape tension or length. Think vacuum columns or spring
loaded arms.
That is closely related to the capstan and weakly-driven take up spool I think.
And then we have drives like the TU80/TU81 which do
not seem to fit into
any of the mentioned categories. I'm not entirely sure how they work,
but I think they are similar to the DECtape in this sense.
I don't know. I can think of several ways it could work....
1) There is a tacho on an arm that rests against the takeup spool (as in the Cipher F880).
The control
system drives the spools to keep that speed constant.
2) Similar, but the tacho is a rotating tape guide that the tape goes round
3) It assumes the takeup spool is intially empty. It then drives forwards, and measures
the
speed of the supply spool. From that, you can work out the diameter of the supply spool,
and
thus I guess the amount of tape. You can then (a) determine where you are on the tape by
the
relative speed of the 2 spools, this gives you the amount of tape on the takeup spool and
thus
how fast you need to turn that to get the right speed.
One of my VHS video recorders does something similar to work out how much recording space
is
left on the cassette.
-tony
11:16 a.m.
On 2015-06-16 19:51, tony duell wrote:
I would disagree. There is no slipping clutch. Instead, the motors (on
both sides) are driven by the uptake available. When there is too
little, the motors run to give out some more tape. When there is too
much, the motors run to pick up the excess.
So, motors can actually run in both directions, as needed. They are
actually controlled by a pretty dumb closed loop system.
The actual tape movement as such, is all done by the small wheel next to
the head, which just runs the tape past the head.
I can also think of several. I just don't know for sure what it does.
And yet another (possibly the most common one on
computers) is to have a
small drive wheel that pulls the tape at constant speed across the
heads, and then have some other construction that drives the tape reels
depending on tape tension or length. Think vacuum columns or spring
loaded arms.
That is closely related to the capstan and weakly-driven take up spool I think. And then we
have drives like the TU80/TU81 which do not seem to fit into
any of the mentioned categories. I'm not entirely sure how they work,
but I think they are similar to the DECtape in this sense.
I don't know. I can think of several ways it could work.... 1) There is a tacho on an arm that rests against the
takeup spool (as in the Cipher F880). The control
system drives the spools to keep that speed constant.
There can definitely be a tacho on the motor, but since you do not know
how much tape is on the reel, this will actually not tell you the tape
speed. But there is no tacho on the tape itself, unless my memory fails me.
2) Similar, but the tacho is a rotating tape guide
that the tape goes round
See above.
3) It assumes the takeup spool is intially empty. It
then drives forwards, and measures the
speed of the supply spool. From that, you can work out the diameter of the supply spool,
and
thus I guess the amount of tape. You can then (a) determine where you are on the tape by
the
relative speed of the 2 spools, this gives you the amount of tape on the takeup spool and
thus
how fast you need to turn that to get the right speed.
That would be very hard, as you do not know the thickness of the tape,
nor the diameter of the tape reel center.
I've been wondering if there is some sensor of tape tension/pressure at
the heads, and this will allow the drive to figure out how much
faster/slower the reels must run, relative to each other, in order to
keep the tape tensioned. Then you can figure out tape speed across the
heads (if you care) by just observing flux changes.
One of my VHS video recorders does something similar
to work out how much recording space is
left on the cassette.
That can't be very precise... :-)
Johnny
12:20 p.m.
And yet another (possibly the most common one on
computers) is to have a
small drive wheel that pulls the tape at constant speed across the
heads, and then have some other construction that drives the tape reels
depending on tape tension or length. Think vacuum columns or spring
loaded arms.
That is closely related to the capstan and weakly-driven take up spool I think. The actual tape movement as such, is all done by the
small wheel next to
the head, which just runs the tape past the head.
Which is essentailly the same as the capstan in an audio tape recorder, albeit the
computer
drive doesn't have a pinch roller
I've been wondering if there is some sensor of
tape tension/pressure at
the heads, and this will allow the drive to figure out how much
faster/slower the reels must run, relative to each other, in order to
keep the tape tensioned. Then you can figure out tape speed across the
heads (if you care) by just observing flux changes.
Only if there is something on the tape. These computer tape drives could surely record on
a
totally blank tape and get the right number of bits per inch. So the thing can't use
the data rate
at the head as a speed measurement.
One of my VHS
video recorders does something similar to work out how much recording space is
left on the cassette.
That can't be very precise... :-)
12:31 p.m.
On Jun 16, 2015, at 3:20 PM, tony duell <ard at
p850ug1.demon.co.uk> wrote:
...
Most don?t. Some do. I remember using IBM tape drives on a 360 Model 44 that were
amazingly badly designed.
Part 1: two capstans spinning all the time, in opposite directions. Solenoid-activated
pinch rollers would press the tape against the capstan to set it moving. The same
solenoid-activated pinch roller moved in the opposite direction would press the tape
against a non-moving ?capstan" to stop it.
Part 2: vacuum columns ? no surprise there. But why on earth would anyone build a vacuum
column that has the oxide side of the tape facing OUTward, rubbing the oxide against the
column side walls?
Amazingly enough, those drives did work reasonably reliably, but I would never want such a
beast in my shop.
paul
The actual tape movement as such, is all done by
the small wheel next to
the head, which just runs the tape past the head.
Which is essentailly the same as the capstan in an audio tape recorder, albeit the
computer
drive doesn't have a pinch roller
12:40 p.m.
On Jun 16, 2015, at 12:31 , Paul Koning <paulkoning
at comcast.net> wrote:
Part 1: two capstans spinning all the time, in opposite directions. Solenoid-activated
pinch rollers would press the tape against the capstan to set it moving.
That's also how the magnetic card is moved in and out in the Mag Card II typewriter.
There's a pair of counter-rotating capstan wheels below the card, and a pair of
solenoid-actuated pinch rollers opposite them, above the card. The head is moved from
track to track with a solenoid-actuated escapement that lets the head move one track at a
time under spring pressure, and then yet another clutch lets a free-running roller grab a
cable to pull the head back to the first track.
I have one waiting for my attention. I think the job will be much more mechanical than
electronic!
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
12:46 p.m.
> On Jun 16, 2015, at 12:31 , Paul Koning
<paulkoning at comcast.net> wrote:
> Part 1: two capstans spinning all the time, in opposite directions.
Solenoid-activated pinch rollers would
> press the tape against the capstan to set it moving.
That's also how the magnetic card is moved in and
out in the Mag Card II typewriter. There's a pair of counter-
rotating capstan wheels below the card, and a pair of solenoid-actuated pinch rollers
opposite them, above the
And how the paper tape is moved in a Trend HSR500 reader. 2 capstans, one each side of the
read head
and 2 solenoid-operated pinch rollers under them. And an electromagnet and armature as a
brake.
For something even madder, look at the design of the original Radio Shack 'Line
Printer' which was
actually a Centronics something-pr-other (733?). This thing (which is not a line printer
at all) has a belt
running across the chassis with a motor continuously driving it (a shaded pole motor I
think). Solenoids
on the printhead carriage grab the top or bottom run of the belt depending on which way
the carriage
is to move.
-tony
12:52 p.m.
On Jun 16, 2015, at 12:46 , tony duell <ard at
p850ug1.demon.co.uk> wrote:
For something even madder, look at the design of the original Radio Shack 'Line
Printer' which was
actually a Centronics something-pr-other (733?). This thing (which is not a line printer
at all) has a belt
running across the chassis with a motor continuously driving it (a shaded pole motor I
think). Solenoids
on the printhead carriage grab the top or bottom run of the belt depending on which way
the carriage
is to move.
Is that the same screwball printer that has a metal platen spinning behind the paper with
horizontal raised lines on it, and a single vertical striker in the printhead that strikes
at the moment when the platen and striker intersect at the desired X/Y location of the dot
to be printed? I've considered buying one just to hear what it sounds like.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
1 p.m.
[Centronics belt-grabbing printer]
Is that the same screwball printer that has a metal
platen spinning behind the paper with horizontal raised lines on
it, and a single vertical striker in the printhead that strikes at the moment when the
platen and striker intersect at
the desired X/Y location of the dot to be printed? I've considered buying one just to
hear what it sounds like.
No. That sounds like a Seikosha (?spel) Unihammer printer. Radio Shack sold it as the
DMP110, I think
Amstrad and Commodore sold them too (probably with different firmware). From what I
remember
(I have a DMP110, but haven't used it for many years) they produce a fairly
high-pitched buzz.
Then of course there is the Olivetti sparkjet printer. A really crazy thing that blasts
solid toner from a
rod of same by producing a spark from said rod to a fixed electrode. The toner doesn't
all end up
on the electrode, some goes on the paper. Or at least that#s the idea. Setting one up is
an interesting
job (I turned a pointed pin to make things a little easier) and the quality is
non-existant even when it is
working properly.
-tony
1:54 p.m.
On 2015-Jun-16, at 12:31 PM, Paul Koning wrote:
Once ran across a Honeywell drive (7-track) that had a 'vacuum capstan'. The
capstan wheel had a vacuum path through the axle and then transferred the vacuum out to
holes around the perimeter. So rather than a pinch roller or rubber frictional surface on
the capstan to grip the tape, it was a vacuum holding the tape to the smooth metal surface
of the rotating capstan. Or was this common on higher-end vac-column drives?
On Jun 16,
2015, at 3:20 PM, tony duell <ard at p850ug1.demon.co.uk> wrote:
...
Most don?t. Some do. I remember using IBM tape drives on a 360 Model 44 that were
amazingly badly designed. The actual tape movement as such, is all done by
the small wheel next to
the head, which just runs the tape past the head.
Which is essentailly the same as the capstan in an audio tape recorder, albeit the
computer
drive doesn't have a pinch roller 
2:30 p.m.
On 06/16/2015 01:54 PM, Brent Hilpert wrote:
Once ran across a Honeywell drive (7-track) that had a
'vacuum
capstan'. The capstan wheel had a vacuum path through the axle and
then transferred the vacuum out to holes around the perimeter. So
rather than a pinch roller or rubber frictional surface on the
capstan to grip the tape, it was a vacuum holding the tape to the
smooth metal surface of the rotating capstan. Or was this common on
higher-end vac-column drives?
Two counter-rotating capstans on the CDC 60x, 65x, 66x drives at least.
The vacuum/pressure was switched by means of a voice-coil valve. IBM
seemed wedded to the pinch-roller model.
I can recall having a discussion about who had the "sloppy write/careful
read" versus the "careful write/sloppy read" technology, but don't
recall who claimed which. I do remember hours spent with various tapes,
developer (e.g. Magna-see) and a loupe looking at patterns on problem tapes.
--Chuck
12:48 p.m.
Johnny Billquist wrote:
Then you can figure out tape speed across the
heads (if you care) by just observing flux changes.
Tony Duell wrote:
Only if there is something on the tape. These computer
tape drives could surely record on a
totally blank tape and get the right number of bits per inch. So the thing can't use
the data rate
at the head as a speed measurement.
Not for something that can write a blank tape (e.g. a 7-track or
9-track), but there were other tape systems that required a formatted
tape, and servoed the reel motors to the flux transition rate.
Obviously the factory had to use a drive with a different tape speed
control system to format the tape.
5:54 p.m.
On Jun 16, 2015, at 3:48 PM, Eric Smith <spacewar
at gmail.com> wrote:
Johnny Billquist wrote:
DECtape has a timing track, but that can be written on site and simply runs the tape
takeup motor at its design speed, as far as I can tell. The reel dimensions of DECtape
ensure that the speed doesn?t change all that much (min : max diameter ratio is rather
modest).
paul
Then you can figure out tape speed across the
heads (if you care) by just observing flux changes.
Tony Duell wrote:
Only if there is something on the tape. These
computer tape drives could surely record on a
totally blank tape and get the right number of bits per inch. So the thing can't use
the data rate
at the head as a speed measurement.
Not for something that can write a blank tape (e.g. a 7-track or
9-track), but there were other tape systems that required a formatted
tape, and servoed the reel motors to the flux transition rate.
Obviously the factory had to use a drive with a different tape speed
control system to format the tape.
2:17 p.m.
On 2015-06-16 21:20, tony duell wrote:
I would disagree. There is no slipping clutch. Instead, the motors (on
True. I meant it was a system where the tape speed was determined by the
capstan, and the reels are driven to keep the tape round up, not all over the
machine room floor :-)
Ok. Then I would agree. :-)
Right.
Ah... Good point. I didn't even think about writing. Stupid of me. :-)
Anyway, it seems the drive do some clever computations to figure out how
to keep the speed constant.
Johnny
--
Johnny Billquist || "I'm on a bus
|| on a psychedelic trip
email: bqt at softjar.se || Reading murder books
pdp is alive! || tryin' to stay hip" - B. Idol
And yet another (possibly the most common one on
computers) is to have a
small drive wheel that pulls the tape at constant speed across the
heads, and then have some other construction that drives the tape reels
depending on tape tension or length. Think vacuum columns or spring
loaded arms.
That is closely related to the capstan and weakly-driven take up spool I think. The actual
tape movement as such, is all done by the small wheel next to
the head, which just runs the tape past the head.
Which is essentailly the same as the capstan in an audio tape recorder, albeit the
computer
drive doesn't have a pinch roller I've been
wondering if there is some sensor of tape tension/pressure at
the heads, and this will allow the drive to figure out how much
faster/slower the reels must run, relative to each other, in order to
keep the tape tensioned. Then you can figure out tape speed across the
heads (if you care) by just observing flux changes.
Only if there is something on the tape. These computer tape drives could surely record on
a
totally blank tape and get the right number of bits per inch. So the thing can't use
the data rate
at the head as a speed measurement. 
11:08 a.m.
-----Oorspronkelijk bericht-----
From: Johnny Billquist
Sent: Tuesday, June 16, 2015 7:31 PM
To: cctalk at classiccmp.org
Subject: Re: On the Emulation of TU58s
[... snip ...]
And then we have drives like the TU80/TU81 which do not seem to fit into
any of the mentioned categories. I'm not entirely sure how they work,
but I think they are similar to the DECtape in this sense.
Johnny
-----
The round (approx 2" diameter) "bearings" of the TU80 do not rotate.
Instead, the metal has tiny holes. Air pressure pushes the tape away
from the bearings. As the tape moves, due to movement
pressure, the
tape comes closer to the air bearing. That is measured. To get the
air pressure back to some sort of equilibrium the pickup or supply
reel speed is adjusted.
That is how I understand the mechanism ... now writing this down, I
think I should read up the Theory of Operation chapter ...
- Henk
11:20 a.m.
On 06/16/2015 11:08 AM, Henk Gooijen wrote:
The round (approx 2" diameter)
"bearings" of the TU80 do not rotate.
Instead, the metal has tiny holes. Air pressure pushes the tape away
from the bearings. As the tape moves, due to movement pressure, the
tape comes closer to the air bearing. That is measured. To get the
air pressure back to some sort of equilibrium the pickup or supply
reel speed is adjusted.
That is how I understand the mechanism ... now writing this down, I
think I should read up the Theory of Operation chapter ...
Pretty much the standard CDC vacuum-column design since the 1960s. The
TU80 is, after all, a CDC product. Two counter-rotating capstans--tape
movement is determined by applying positive or negative pressure to
ether or both (i.e. positive pressure on both results in the tape "floating"
--Chuck
11:22 a.m.
On 2015-06-16 20:20, Chuck Guzis wrote:
On 06/16/2015 11:08 AM, Henk Gooijen wrote:
Uh... The TU80/TU81 do not have vacuum columns...
Johnny
The round (approx 2" diameter)
"bearings" of the TU80 do not rotate.
Instead, the metal has tiny holes. Air pressure pushes the tape away
from the bearings. As the tape moves, due to movement pressure, the
tape comes closer to the air bearing. That is measured. To get the
air pressure back to some sort of equilibrium the pickup or supply
reel speed is adjusted.
That is how I understand the mechanism ... now writing this down, I
think I should read up the Theory of Operation chapter ...
Pretty much the standard CDC vacuum-column design since the 1960s. The
TU80 is, after all, a CDC product. Two counter-rotating capstans--tape
movement is determined by applying positive or negative pressure to
ether or both (i.e. positive pressure on both results in the tape
"floating"
11:23 a.m.
On Jun 16, 2015, at 11:22 , Johnny Billquist <bqt
at update.uu.se> wrote:
Or rotating capstans...
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
Pretty much the standard CDC vacuum-column design since the 1960s. The
TU80 is, after all, a CDC product. Two counter-rotating capstans--tape
movement is determined by applying positive or negative pressure to
ether or both (i.e. positive pressure on both results in the tape
"floating"
Uh... The TU80/TU81 do not have vacuum columns...
11:30 a.m.
On 2015-06-16 20:23, Mark J. Blair wrote:
Bu the way. For the vacuum column drives that I know
(TU16/TU45/TU77/TU78), tape movement is not determined by vacuum as such.
Each column have about 8 pressure switches. When 4 have low pressure,
and 4 have atmospheric pressure, the drive is happey, and the reels are
resting. When the point of underpressure/pressure moves down, the reel
on that side starts moving to pick up more tape, thus moving the point
back towards neutral. When the point moves up, the reel starts moving to
give out more tape, to move the point back down again. The more switches
triggered in either direction, the faster the reel will move.
The movement of the tape across the head is controlled by a third motor,
which have a smaller wheel on it. The tape lies against this reel, but
it's all done by friction. This wheel pretty much runs the tape across
the head at constant speed, and this movement of the tape will also
cause the points in the two columns to move, which in turn will trigger
action as described in the previous section.
Not sure where the counter-rotating capstans would fit into this either...
And there is only negative pressure, and it is applied to the tape at
all time.
Johnny
On Jun 16, 2015, at 11:22 , Johnny Billquist
<bqt at update.uu.se> wrote:
Or rotating capstans...
Pretty much the standard CDC vacuum-column design since the 1960s. The
TU80 is, after all, a CDC product. Two counter-rotating capstans--tape
movement is determined by applying positive or negative pressure to
ether or both (i.e. positive pressure on both results in the tape
"floating"
Uh... The TU80/TU81 do not have vacuum columns...
11:32 a.m.
On 2015-06-16 20:28, Chuck Guzis wrote:
On 06/16/2015 11:22 AM, Johnny Billquist wrote:
But I can't see any relationship between vacuum column tape drives and
how the TU80/TU81 works. You'll have to explain this to me in way more
detail, because right now, you have totally lost me. :-)
Johnny
Uh... The TU80/TU81 do not have vacuum columns...
I didn't say that they did, Johnny. I said that the TU80 hails back to
the vacuum-column CDC drive design of the 1960s. At least that's my
understanding.
11:30 a.m.
-----Oorspronkelijk bericht-----
From: Johnny Billquist
Sent: Tuesday, June 16, 2015 8:22 PM
To: cctalk at classiccmp.org
Subject: Re: On the Emulation of TU58s
On 2015-06-16 20:20, Chuck Guzis wrote:
On 06/16/2015 11:08 AM, Henk Gooijen wrote:
Uh... The TU80/TU81 do not have vacuum columns...
Johnny
-----
well, the tiny space between the tape and the air bearing could
be regarded a vacuum column, but it's a very short one :-)
The two air bearings are immediately above and below the R/W
and erase head. And they do *not* rotate. The tape glides on
an air cushion over the bearing.
The round (approx 2" diameter)
"bearings" of the TU80 do not rotate.
Instead, the metal has tiny holes. Air pressure pushes the tape away
from the bearings. As the tape moves, due to movement pressure, the
tape comes closer to the air bearing. That is measured. To get the
air pressure back to some sort of equilibrium the pickup or supply
reel speed is adjusted.
That is how I understand the mechanism ... now writing this down, I
think I should read up the Theory of Operation chapter ...
Pretty much the standard CDC vacuum-column design since the 1960s. The
TU80 is, after all, a CDC product. Two counter-rotating capstans--tape
movement is determined by applying positive or negative pressure to
ether or both (i.e. positive pressure on both results in the tape
"floating"
11:34 a.m.
On 2015-06-16 20:30, Henk Gooijen wrote:
-----Oorspronkelijk bericht----- From: Johnny
Billquist Sent: Tuesday,
June 16, 2015 8:22 PM To: cctalk at classiccmp.org Subject: Re: On the
Emulation of TU58s
On 2015-06-16 20:20, Chuck Guzis wrote:
Hmm, I guess it could. But it is a very different construct compared to
a vacuum column drive. In this case we're talking about overpressure to
act as a cushion, and not underpressure to act as a takeup/buffer mechanism.
Johnny
On 06/16/2015 11:08 AM, Henk Gooijen wrote:
Uh... The TU80/TU81 do not have vacuum columns...
Johnny
-----
well, the tiny space between the tape and the air bearing could
be regarded a vacuum column, but it's a very short one :-)
The two air bearings are immediately above and below the R/W
and erase head. And they do *not* rotate. The tape glides on
an air cushion over the bearing. The round (approx 2" diameter)
"bearings" of the TU80 do not rotate.
Instead, the metal has tiny holes. Air pressure pushes the tape away
from the bearings. As the tape moves, due to movement pressure, the
tape comes closer to the air bearing. That is measured. To get the
air pressure back to some sort of equilibrium the pickup or supply
reel speed is adjusted.
That is how I understand the mechanism ... now writing this down, I
think I should read up the Theory of Operation chapter ...
Pretty much the standard CDC vacuum-column design since the 1960s. The
TU80 is, after all, a CDC product. Two counter-rotating capstans--tape
movement is determined by applying positive or negative pressure to
ether or both (i.e. positive pressure on both results in the tape
"floating"
11:50 a.m.
On 2015-06-16 20:34, Johnny Billquist wrote:
On 2015-06-16 20:30, Henk Gooijen wrote:
So I decided to read up on the TU81 operation, and it does work by
creating an overpressure to act as a bearing before and after the head.
The overpressure is also used to detect tape tension so that a constant
tension can be maintained.
The supply reel is controlled by the air pressure sensors, so that
tension is maintained, while the pickup reel is directly controlling the
speed. Both reels have tachos as well.
There is no capstan, but constant linear speed can be maintained by the
fact that the takeup reel is of a known size, and you have the tachos,
and you essentially can estimate good enough how much tape is on the
takeup reel.
(Lots of computations in here... :-) )
Johnny
-----Oorspronkelijk bericht----- From: Johnny
Billquist Sent: Tuesday,
June 16, 2015 8:22 PM To: cctalk at classiccmp.org Subject: Re: On the
Emulation of TU58s
On 2015-06-16 20:20, Chuck Guzis wrote:
Hmm, I guess it could. But it is a very different construct compared to
a vacuum column drive. In this case we're talking about overpressure to
act as a cushion, and not underpressure to act as a takeup/buffer
mechanism. On 06/16/2015 11:08 AM, Henk Gooijen wrote:
Uh... The TU80/TU81 do not have vacuum columns...
Johnny
-----
well, the tiny space between the tape and the air bearing could
be regarded a vacuum column, but it's a very short one :-)
The two air bearings are immediately above and below the R/W
and erase head. And they do *not* rotate. The tape glides on
an air cushion over the bearing. The round (approx 2" diameter)
"bearings" of the TU80 do not rotate.
Instead, the metal has tiny holes. Air pressure pushes the tape away
from the bearings. As the tape moves, due to movement pressure, the
tape comes closer to the air bearing. That is measured. To get the
air pressure back to some sort of equilibrium the pickup or supply
reel speed is adjusted.
That is how I understand the mechanism ... now writing this down, I
think I should read up the Theory of Operation chapter ...
Pretty much the standard CDC vacuum-column design since the 1960s. The
TU80 is, after all, a CDC product. Two counter-rotating capstans--tape
movement is determined by applying positive or negative pressure to
ether or both (i.e. positive pressure on both results in the tape
"floating"
1:05 p.m.
On 06/16/2015 11:30 AM, Henk Gooijen wrote:
well, the tiny space between the tape and the air
bearing could
be regarded a vacuum column, but it's a very short one :-)
The two air bearings are immediately above and below the R/W
and erase head. And they do *not* rotate. The tape glides on
an air cushion over the bearing.
Yeah, a case of mental dyspepsia.
So do any post-1980s 1/2" open-reel tape drives employ vacuum columns?
I was thinking of DG drive (about half-rack size) that I believe had
them, but I"m not sure of the date of manufacture.
Does this apparent lack of start-stop on a dime drives have anything to
do with the decline in tape as a working medium (as opposed to a backup
or distribution medium)? This is obvious when one looks at an OS such
as UNIX or PRIMOS; the tape was treated as pretty much an archival device.
Used to be that either Flores or (later) Knuth on sort/merge operations
was a standard on many a programmer's bookshelf. Not so much, I
gather, in the last couple of decades?
--Chuck
10:44 a.m.
[Driving tape]
They were. Or at least some semi-professional audio recorders were built like that. The
Revox G36 (736) and Brennell Mk 5 among them (I happen to have both). Actually, the
Philips V2000 video recorders (home machines,
not professional but quite the best of the home systems) did that. They had a pair of DC
motors to directly
drive the 2 spools along with the capstan and head drum motors. No slipping clutches. Back
tension was
provided not by a light brake on the supply spool but by a suitable current through the
supply spool motor
in those machines.
-tony
Another solution is to avoid the problem entirely by
not requiring constant linear speed. That?s what DECtape
(the real one) does.
Sure. A number of tape drives were built that way, the HP9865 (and thus the built-in tape
drive on the
HP9830) is another example.
One is to drive the tape itself using a capstan
and maybe a pinch roller. Then have the takeup spool
driven too fast with a slipping clutch in-line. And a weak brake on the supply spool.
This of course is what
was done in audio recorders (reel to reel and cassette), Video recorders (ditto), etc.
Or a variant: drive the takeup reel with a controlled torque motor and no slip clutch. I
think professional grade > tape decks may have been built that way.
10:48 a.m.
On Jun 16, 2015, at 09:10 , tony duell <ard at
p850ug1.demon.co.uk> wrote:
You would, of course, not know which track it was reading, so you would have to output 2
blocks, one on
each track, at once. And how would you detect it was writing? Look for an extra signal at
the coupling head
or something?
That seems to me like it would be the trickiest part. The TU58 schematic appears to
indicate that there are separate erase head gaps (not on a separate head like in audio
cassette drives), so maybe those could be monitored to detect when a track is being
written? Of course, that probably means that a custom 4-gap head would need to be made
with gaps matching the two read/write and two erase gaps (and a concave surface, too,
rather than recycling some audio tape head with suitable geometry to line up with just the
read/write gaps.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
10:54 a.m.
That seems to me like it would be the trickiest part.
The TU58 schematic appears to indicate that there are
I think so. Of course you could cheat by making it a read-only tape. But I don't like
that.
separate erase head gaps (not on a separate head like
in audio cassette drives), so maybe those could be
I am pretty sure there is just one physical head on the TU58 drive chassis. So if there is
a separate erase
field, it is a second gap in that head.
monitored to detect when a track is being written? Of
course, that probably means that a custom 4-gap head
would need to be made with gaps matching the two read/write and two erase gaps (and a
concave surface,
too, rather than recycling some audio tape head with suitable geometry to line up with
just the read/write gaps.
Ouch. Good luck in getting that, and even more luck in getting it positioned properly!
-tony
11:18 a.m.
On Jun 16, 2015, at 10:54 , tony duell <ard at
p850ug1.demon.co.uk> wrote:
I am pretty sure there is just one physical head on the TU58 drive chassis. So if there
is a separate erase
field, it is a second gap in that head.
Correct. Single head with multiple gaps.
Ouch. Good luck in getting that, and even more luck in getting it positioned properly!
Yup! That's why I'm presenting this as a philosophical debate, rather than
something even crazy old me would think of building! :)
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
1:02 p.m.
On Jun 16, 2015, at 09:10 , tony duell <ard at
p850ug1.demon.co.uk> wrote:
Incidentally, given the fact that a constant motor speed -> constant tape speed, it
should be possible to
make a device to put the timing track on a blank tape for the TU58. Has anyone done that?
There's no timing track in the TU58 scheme, but there are magnetic BOT and EOT markers
that have to be written so that the drive can identify the tape ends. I assume there are
also block headers much like on most floppy disks, but I haven't gotten that deep into
the formatting yet.
The tape includes the BOT and EOT sensing punched holes, and the cartridge includes the
angled mirror behind the tape to allow the holes to be sensed with a right-angled optical
path. But the TU58-XA drive mechanism does not include the optical sensor that would be
needed to sense tape ends on an unformatted tape. I don't know if this was meant as a
way to further cost-reduce the already mechanically simple tape drive mechanism, or if DEC
did that deliberately to make sure that non-DEC DC100/DC150 cartridges could not be be
formatted in the field, so that users would be stuck buying preformatted cartridges from
DEC.
If new cartridges with brand new, un-decayed belts could be manufactured, then it should
be possible to hack up a TU58-XA mechanism for formatting them. I think there may be a
little hole in the plastic casting of the drive where one of the optical sensors might be
glued in place, if I recall correctly.
Oh yeah, the metal vs. plastic base cartridges were also mentioned in this thread.
I've only encountered the metal ones so far. Based on the manual pictures, I think the
plastic ones use a shorter belt with a simpler path.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
3476
days inactive
3477
days old
33 comments
9 participants
participants (9)
-
ard＠p850ug1.demon.co.uk -
bqt＠update.uu.se -
cclist＠sydex.com -
henk.gooijen＠hotmail.com -
hilpert＠cs.ubc.ca -
isking＠uw.edu -
nf6x＠nf6x.net -
paulkoning＠comcast.net -
spacewar＠gmail.com