27 Jan
2015
27 Jan
'15
9:04 a.m.
I have a bunch of DECTAPE II cartridges from which I want to try recovering as much data
as possible, including the console cartridges for my VAX-11/730, none of which have
managed to boot the machine. Inspired by the various floppy disk imagers I have, my
thinking has wandered in the direction of building an imaging device using a TU58-XA
mechanism and my own drive electronics. I happen to have a few TU58-XA mechanisms sitting
about to experiment with. The imagined imager would sample both tracks simultaneously with
ADCs so that I could post-process the data repeatedly from a single physical pass past the
heads. Maybe I'd even hack in the missing optical end-of-tape sensor rather than
relying on the written end-of-tape signals. This scheme might let me recover data from
tapes which confuse the normal TU58 drive electronics (say, because of corrupted sector
marks).
With that in mind, I dug into a box of acquired tapes to select a sacrificial dummy for
experimentation, and picked out one labeled "NFG" by the previous owner,
figuring I had nothing to lose with that one. Inspection revealed that the drive belt had
broken and was stuck to the tape, and as expected, it peeled the oxide right off when I
removed it. Is there any way to remove pieces of stuck drive belt from these tapes without
harming the underlying oxide?
Next, I moved on to trying to replace the drive belt with a Plastiband brand elastic band,
as I've previously seen discussed for tape cartridges with this style of mechanism
(including the larger QIC cartridges). I used the 2-1/8" size, and found that it
seemed to be of suitable diameter to stretch around the required path. However, I have had
no luck getting it to stay in place when the tape is moved. It quickly jumps off and gets
tangled when I drive the tape, and I suspect that it's because these Plastibands are a
bit too narrow to properly ride the crowned drive wheel and idler rollers.
Let me interject that I have nothing good to say about this belt-driven cartridge design.
I thought it was a bad design the first time I encountered one, and nothing has changed my
opinion since then!
Now I finally get to the main topic I'd like to discuss: "Destructive"
imaging of DECTAPE II media. By eye, the DECTAPE II tape media looks very close to the
same width as normal audio cassette tape. What if I built an imaging drive in which I
remove the tape from any DECTAPE II cart to be imaged rather than trying to use the
original &(#$%*$ belt drive system? I wonder whether there's any prior art for a
scheme like this.
One idea would be to transplant the tape media into a cassette tape housing, but I'm
not yet sure whether that might offer any advantages over building an ad-hoc open reel
system, or even a reel-less system since any given tape would only be run past the heads a
small number of times in one event, and then might be discarded once any remaining data is
extracted.
I have doubts that a cassette transport's pinch roller and capstan system would work
well for this scheme, since the normal tape speed of a DECTAPE II is over 15x the tape
speed of an audio cassette. So, I'd probably need to fashion a different sort of
capstan system, possibly negating any advantage of using cassette tape housings. I'm
not sure whether it would be better to pull the tape with a capstan and pinch roller drive
vs. pulling it with a take-up reel hub and providing a free-spinning capstan with an
optical encoder to provide tape speed feedback. Either scheme would allow me to move the
tape past the head at constant speed, but I'm not sure if one scheme might be easier
to build than the other.
Next, the imaginary device would need appropriate heads. I might salvage the head from a
TU58-XA, though that could be a bit challenging since the heads are epoxied in place after
adjustment. Or, maybe I would get lucky and discover that the heads from an auto-reversing
audio tape deck (which I believe have four gaps in order to play stereo tapes in either
direction without flipping them) might have gaps in the right places for reading both
tracks of DECTAPE II media? Audio cassette tape heads have a few potential advantages,
including not being epoxied in place and having a tape alignment guide welded on one
side.
What do you folks think about this silliness?
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
27 Jan
27 Jan
9:22 a.m.
I think it is a great start! But if you're doing destructive tape
recovering, why not using a 15x slower mech and using a slower (and cheaper)
ADC?
Or use a band from a known good cartridge, at least to recover the data.
Or you can use a reel-to-reel mech :)
---
Enviado do meu Apple IIGS (pq eu sou chique)
Meu site: http://www.tabalabs.com.br
Meu blog: http://tabajara-labs.blogspot.com
----- Original Message -----
From: "Mark J. Blair" <nf6x at nf6x.net>
To: "General Discussion: On-Topic and Off-Topic Posts"
<cctalk at classiccmp.org>
Sent: Tuesday, January 27, 2015 3:04 PM
Subject: Destructive Imaging of DECTAPE II Media
I have a bunch of DECTAPE II cartridges from which I want to try recovering
as much data as possible, including the console cartridges for my
VAX-11/730, none of which have managed to boot the machine. Inspired by the
various floppy disk imagers I have, my thinking has wandered in the
direction of building an imaging device using a TU58-XA mechanism and my own
drive electronics. I happen to have a few TU58-XA mechanisms sitting about
to experiment with. The imagined imager would sample both tracks
simultaneously with ADCs so that I could post-process the data repeatedly
from a single physical pass past the heads. Maybe
I'd even hack in the
missing optical end-of-tape sensor rather than relying on
the written
end-of-tape signals. This scheme might let me recover data from tapes which
confuse the normal TU58 drive electronics (say, because of corrupted sector
marks).
With that in mind, I dug into a box of acquired tapes to select a
sacrificial dummy for experimentation, and picked out one labeled "NFG" by
the previous owner, figuring I had nothing to lose with that one. Inspection
revealed that the drive belt had broken and was stuck to the tape, and as
expected, it peeled the oxide right off when I removed it. Is there any way
to remove pieces of stuck drive belt from these tapes without harming the
underlying oxide?
Next, I moved on to trying to replace the drive belt with a Plastiband brand
elastic band, as I've previously seen discussed for tape cartridges with
this style of mechanism (including the larger QIC cartridges). I used the
2-1/8" size, and found that it seemed to be of suitable diameter to stretch
around the required path. However, I have had no luck getting it to stay in
place when the tape is moved. It quickly jumps off and gets tangled when I
drive the tape, and I suspect that it's because these Plastibands are a bit
too narrow to properly ride the crowned drive wheel and idler rollers.
Let me interject that I have nothing good to say about this belt-driven
cartridge design. I thought it was a bad design the first time I encountered
one, and nothing has changed my opinion since then!
Now I finally get to the main topic I'd like to discuss: "Destructive"
imaging of DECTAPE II media. By eye, the DECTAPE II tape media looks very
close to the same width as normal audio cassette tape. What if I built an
imaging drive in which I remove the tape from any DECTAPE II cart to be
imaged rather than trying to use the original &(#$%*$ belt drive system? I
wonder whether there's any prior art for a scheme like this.
One idea would be to transplant the tape media into a cassette tape housing,
but I'm not yet sure whether that might offer any advantages over building
an ad-hoc open reel system, or even a reel-less system since any given tape
would only be run past the heads a small number of times in one event, and
then might be discarded once any remaining data is extracted.
I have doubts that a cassette transport's pinch roller and capstan system
would work well for this scheme, since the normal tape speed of a DECTAPE II
is over 15x the tape speed of an audio cassette. So, I'd probably need to
fashion a different sort of capstan system, possibly negating any advantage
of using cassette tape housings. I'm not sure whether it would be better to
pull the tape with a capstan and pinch roller drive vs. pulling it with a
take-up reel hub and providing a free-spinning capstan with an optical
encoder to provide tape speed feedback. Either scheme would allow me to move
the tape past the head at constant speed, but I'm not sure if one scheme
might be easier to build than the other.
Next, the imaginary device would need appropriate heads. I might salvage the
head from a TU58-XA, though that could be a bit challenging since the heads
are epoxied in place after adjustment. Or, maybe I would get lucky and
discover that the heads from an auto-reversing audio tape deck (which I
believe have four gaps in order to play stereo tapes in either direction
without flipping them) might have gaps in the right places for reading both
tracks of DECTAPE II media? Audio cassette tape heads have a few potential
advantages, including not being epoxied in place and having a tape alignment
guide welded on one side.
What do you folks think about this silliness?
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
9:46 a.m.
On 1/27/15 9:22 AM, Alexandre Souza wrote:
why not using a 15x slower mech and using a slower (and cheaper) ADC?
Because the amplitude from the read head is a function of tape speed.
Or use a band from a known good cartridge, at
least to recover the data.
I was going to suggest that. Any DC-100 sized cart can be used as a donor, and you
only need a couple of them.
There was a post about saving HP tapes by doing something to them (heating them?)
but now I can't find the post. Chuck was trying to find something to keep the hole
in the oxide from interfering with the EOT/BOT sensing, but I don't know if he found
something that worked reliably. You need something that is opaque to IR.
I also thought someone already imaged the 730 carts, but I can't find those either.
Maybe they were for the 750.
I have 100+ TU58s that I need to get to this year and already replaced the rollers with
new ones made from norprene tubing.
11:22 a.m.
On Jan 27, 2015, at 09:46 , Al Kossow <aek at
bitsavers.org> wrote:
On 1/27/15 9:22 AM, Alexandre Souza wrote:
Right. I don't necessarily need to run the tape at 30 ips like a real TU58, but I
probably will need to run it somewhere in that ballpark.
why not using a 15x slower mech and using a slower (and cheaper) ADC?
Because the amplitude from the read head is a function of tape speed. Or use a band from a known good cartridge, at
least to recover the data.
I was going to suggest that. Any DC-100 sized cart can be used as a donor, and you
only need a couple of them. There was a post about saving HP tapes by doing
something to them (heating them?)
but now I can't find the post.
I thought that I had read something like that, but I also don't recall when/where. A
baking cycle might be beneficial just like it is for other old magnetic media, but I
don't recall whether baking was also helpful in releasing stuck drive belt from the
oxide.
Chuck was trying to find something to keep the hole
in the oxide from interfering with the EOT/BOT sensing, but I don't know if he found
something that worked reliably. You need something that is opaque to IR.
I wish that DEC had included unpopulated provisions for optical end of tape sensing in the
TU58-XA rather than entirely cost-cutting that sensor out of the design. I see that the
cartridges still include the BOT/EOT holes and the right-angle mirror; presumably, the
drives that they used to preformat the tapes in the factory used those for BOT/EOT sensing
prior to writing the magnetic BOT/EOT symbols.
I also thought someone already imaged the 730 carts,
but I can't find those either.
Maybe they were for the 750.
I have found some 730 cart images online, and at the moment I expect that my only way of
booting my 730 will involve using those with tu58em. I'd still like to image the carts
that came with my machine, though, which may be different versions.
I have 100+ TU58s that I need to get to this year and
already replaced the rollers with
new ones made from norprene tubing.
I've already replaced the drive roller rubber in a few of my drives with black latex
rubber tubing. The stuff I used is essentially the same as the yellow tubing used for
wrist rockets and water weenies, except it's black, and of a suitable diameter to fit
the aluminum roller hub and have about the right outside diameter when installed. To cut
it, I stuck a short length on a piece of plastic rod, chucked the rod in a Harbor Freight
mini lathe, and cut pieces with a surgical scalpel while the chuck was spinning. I simply
cut many more than I needed and selected the ones that turned out the best, as there was a
lot of variation in width from my eyeball-cutting, and variation in squareness of the
edges depending on how I held the scalpel. I probably threw away five pieces for every one
that I selected, but the end results look very neat.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
12:31 p.m.
On Tuesday (01/27/2015 at 11:22AM -0800), Mark J. Blair wrote:
I rebuilt several TU58s in a similar fashion and had all nice spiffy
rollers and each tape I stuck in there promptly broke the band after it
had moved less than an inch. In the end, out of dozens of tapes, I think
I had two that worked. Even NOS tapes, still wrapped in cellophane,
would fail when unwrapped and put in the drive for the first time.
I have 100+ TU58s that I need to get to this year
and already replaced the rollers with
new ones made from norprene tubing.
I've already replaced the drive roller rubber in a few of my drives with black latex
rubber tubing. The stuff I used is essentially the same as the yellow tubing used for
wrist rockets and water weenies, except it's black, and of a suitable diameter to fit
the aluminum roller hub and have about the right outside diameter when installed. To cut
it, I stuck a short length on a piece of plastic rod, chucked the rod in a Harbor Freight
mini lathe, and cut pieces with a surgical scalpel while the chuck was spinning. I simply
cut many more than I needed and selected the ones that turned out the best, as there was a
lot of variation in width from my eyeball-cutting, and variation in squareness of the
edges depending on how I held the scalpel. I probably threw away five pieces for every one
that I selected, but the end results look very neat. the original &(#$%*$ belt drive system?
I know what that word is.
I think we should take the tape out of the cartridges and build a thing
that pulls it through a small pool of Magna-See (do they still make that
stuff? Is that what it was called?) and then image them with a camera and
decode the data from the video. Maybe even counting the bits by hand will
be easier than trying to read those &(#$%*$ tapes in any sort of drive.
In any case, I'm totally in support of any DC100 destructive methods.
Using them instead of clay pigeons also comes to mind. :-)
Chris
--
Chris Elmquist
1:32 p.m.
On Jan 27, 2015, at 12:31 , Chris Elmquist <chrise
at pobox.com> wrote:
I think we should take the tape out of the cartridges and build a thing
that pulls it through a small pool of Magna-See (do they still make that
stuff? Is that what it was called?) and then image them with a camera and
decode the data from the video. Maybe even counting the bits by hand will
be easier than trying to read those &(#$%*$ tapes in any sort of drive.
I think that reading them with magnetic heads will be much easier than trying to read them
optically, but I wouldn't mind having some of that magnetic visualization fluid for
experimental use anyway.
I've found the track dimensions of the DECTAPE II in a specification on Bitsavers, and
I found dimensions of various tracks formats used on audio cassette tapes on another site.
I'm about to compare them to see if there's any potential compatibility, since
audio cassette tape heads generally have a built-in tape alignment guide that might make
things easier mechanically vs. the TU58-XA head (assuming I didn't use a TU58-XA
mechanism). I'll follow up with whatever I find. I did determine that both tapes are
the same width: 0.15" or 3.81mm.
In any case, I'm totally in support of any DC100
destructive methods.
Using them instead of clay pigeons also comes to mind. :-)
Once the data is recovered, of course. :)
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
28 Jan
28 Jan
1:51 a.m.
On Tue, 27 Jan 2015, Chris Elmquist wrote:
rollers and each tape I stuck in there promptly broke
the band after it
had moved less than an inch. In the end, out of dozens of tapes, I think
The problem is that the band loses its elasticity and gets longer. The
black PU based bands can be made functional again with the method that I
described in a post some months ago (hot water). You may try this method;
actually you must use this method with all tapes that haven't been used
for some time. My experience even with NOS QIC cartridges from the 90s is
that all such bands have that problem nowadays.
Christian
8:14 a.m.
On 1/28/15 1:51 AM, Christian Corti wrote:
The problem is that the band loses its elasticity and
gets longer. The black PU based bands can be made functional again with the method that I
described in a post some months ago (hot water). You may
try this method; actually you must use this method with all tapes that haven't been
used for some time. My experience even with NOS QIC cartridges from the 90s is that all
such bands have that problem
nowadays.
When I was working on this a couple of months ago, I had hundreds of old DC450/600/6150
carts to experiment on and I
sent out a message about the results. Short answer is bands from 300/450 and 600
aren't worth messing with. 6150 are
a slightly different formulation and are also a bit tighter. This seems to be true for all
of the larger capacity
carts beyond 150Mb as well. The 6150 bands either do not require the warm water treatment
or should be rehydrated at
significantly lower temperature/time. Since I disassemble and replace the bands when I
bake the tapes, it isn't a big
deal to change them out. I also posted a rather long post on how to put the bands on.
Like I said, I have a big batch of TU58 tapes that I need to process, so I'll
experiment with different brands of DC-100
sized tapes when I do it. I'm in the middle of setting up new lab space at CHM right
now, and as soon as that's done I'll
bump up the TU58 recovery in the queue. I know I have 730 tapes in the piles to do.
29 Jan
29 Jan
1:53 a.m.
On Wed, 28 Jan 2015, Al Kossow wrote:
When I was working on this a couple of months ago, I
had hundreds of old
DC450/600/6150 carts to experiment on and I
sent out a message about the results. Short answer is bands from 300/450 and
600 aren't worth messing with. 6150 are
That's more or less true. I have many old 300/600 tapes with with belts
that broke during or shortly after the treatment. Therefore I use belts
from much newer (about one decade) NOS 600 tapes
(thanksfully I have many
dozens of them). They need the hot water treatment
beforehand, too. But at
least they work reliably after that.
If you are archiving such tapes you don't need a new belt for each
cartridge. I simply transfer the belt from the archived cartridge to the
next one and keep it in a parts box for a future archiving session.
The 6150 bands either do not require the warm water
treatment or should
be rehydrated at significantly lower temperature/time. Since I
I had very few 6150 that needed the treatment, too. But in general, you're
right.
Like I said, I have a big batch of TU58 tapes that I
need to process, so I'll
experiment with different brands of DC-100
sized tapes when I do it. I'm in the middle of setting up new lab space at
I have many TU58 tapes, but I have never used a TU58 so far. But I
also have a lot of DC100 tapes from HP terminals, and they surely have
the sticky and worn belt problem. And because DC100 tapes are so rare and
*** expensive, I try anything to keep the original belt going.
Christian
31 Jan
31 Jan
10:39 a.m.
Poking around on eBay, I see that there were other brands (Teac, Wangtek, Everex...) of
streaming tape backup drives that look like they may have used those audio-cassette-like
tapes with the notch on the back edge, much like the TU60. Maybe one of those transports
could be adapted for this wacky scheme? I might buy one of the cheaper ones and some
matching tapes to experiment with.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
10:47 a.m.
On 01/31/2015 10:39 AM, Mark J. Blair wrote:
Poking around on eBay, I see that there were other
brands (Teac,
Wangtek, Everex...) of streaming tape backup drives that look like
they may have used those audio-cassette-like tapes with the notch on
the back edge, much like the TU60. Maybe one of those transports
could be adapted for this wacky scheme? I might buy one of the
cheaper ones and some matching tapes to experiment with.
Are you thinking of the DC2000/2100 type drives? I think that they're a
bit too large. How about, say, an Irwin drive using DC1000 minicarts?
--Chuck
11 a.m.
On Jan 31, 2015, at 10:47, Chuck Guzis <cclist at
sydex.com> wrote:
On 01/31/2015 10:39 AM, Mark J. Blair wrote:
No, I'm thinking of things like Teac MT-2ST drives, using audio-cassette-like CT-600H
cartridges. Some random googling suggests that those sorts of drives may move the tape at
around 90 ips, so hopefully they don't use capstan and pinch roller drive that's
designed for slow audio tape movement.
I just ordered the cheapest one of those drives listed, and some matching tapes, to
experiment with. Maybe it'll provide a better transport than the data logger tape
transport that I also have on the way? And either of them might be easier to hack for this
project than a regular audio cassette transport that's designed for much slower tape
speed.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
Poking around on eBay, I see that there were
other brands (Teac,
Wangtek, Everex...) of streaming tape backup drives that look like
they may have used those audio-cassette-like tapes with the notch on
the back edge, much like the TU60. Maybe one of those transports
could be adapted for this wacky scheme? I might buy one of the
cheaper ones and some matching tapes to experiment with.
Are you thinking of the DC2000/2100 type drives? I think that they're a bit too
large. How about, say, an Irwin drive using DC1000 minicarts?
11:31 a.m.
On Jan 31, 2015, at 11:18, Al Kossow <aek at
bitsavers.org> wrote:
On 1/31/15 11:00 AM, Mark J. Blair wrote:
That sounds very promising. I'll look for the maintenance manual. I think that about
the only thing that would be better would be a system where I could lift the reels and
tape out of a DECTAPE II cart without needing to transfer the media linearly or do any
splicing, but that would take a lot more effort. For me, the mechanical part would be a
lot harder than the electronic and software parts.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
hopefully they don't use capstan and pinch roller drive
There is a maintenance manual on bitsavers
It is servo based with an encoder attached to a roller in contact with the tape.
The two reels are motor driven.
1:08 p.m.
On 1/31/15 11:31 AM, Mark J. Blair wrote:
DC-100 reels float on two pins, so at a minimum you'll have to find one of the good
quality cassette tapes
that had actual reels inside and transfer the tape to those.
I just picked up a TEAC cassette mechanism at a surplus place, and the part of the tach
that touches the
tape is mostly metal with a channel cut into it with what looks like something rubber-ish
where the tape
would touch at the bottom. That seems a little strange so I'll pull it apart and take
pictures of the mechanism.
On Jan 31, 2015, at 11:18, Al Kossow <aek at
bitsavers.org> wrote:
On 1/31/15 11:00 AM, Mark J. Blair wrote:
That sounds very promising. I'll look for the maintenance manual. I think that about
the only thing that would be better would be a system where I could lift the reels and
tape out of a DECTAPE II cart without needing to transfer the media linearly or do any
splicing, but that would take a lot more effort. For me, the mechanical part would be a
lot harder than the electronic and software parts.
hopefully they don't use capstan and pinch roller drive
There is a maintenance manual on bitsavers
It is servo based with an encoder attached to a roller in contact with the tape.
The two reels are motor driven. 
1 Feb
1 Feb
3:19 a.m.
Mark J. Blair wrote:
I've got three Verbatim ST-600 "High Density Streamer Cassette" and five
Maxell CS-600HD "High density data cassette back coated 183m/600ft"
audio-cassette-like tapes with the notch on the back edge. Most have data
written on them from around 1994 but one of each type is unused and still in
plastic wrapping. Another of the Verbatim tapes is missing the box.
If they are of any use, they are available for postage from Ireland.
Regards,
Peter Coghlan.
On Jan 31, 2015, at 10:47, Chuck Guzis <cclist
at sydex.com> wrote:
On 01/31/2015 10:39 AM, Mark J. Blair wrote:
No, I'm thinking of things like Teac MT-2ST drives, using audio-cassette-like
CT-600H cartridges. Some random googling suggests that those sorts of drives
may move the tape at around 90 ips, so hopefully they don't use capstan and
pinch roller drive that's designed for slow audio tape movement.
I just ordered the cheapest one of those drives listed, and some matching
tapes, to experiment with. Maybe it'll provide a better transport than the
data logger tape transport that I also have on the way? And either of them
might be easier to hack for this project than a regular audio cassette
transport that's designed for much slower tape speed.
Poking around on eBay, I see that there were
other brands (Teac,
Wangtek, Everex...) of streaming tape backup drives that look like
they may have used those audio-cassette-like tapes with the notch on
the back edge, much like the TU60. Maybe one of those transports
could be adapted for this wacky scheme? I might buy one of the
cheaper ones and some matching tapes to experiment with.
Are you thinking of the DC2000/2100 type drives? I think that they're a bit
too large. How about, say, an Irwin drive using DC1000 minicarts?
10:15 a.m.
On Feb 1, 2015, at 03:19, Peter Coghlan <cctalk at
beyondthepale.ie> wrote:
I've got three Verbatim ST-600 "High Density Streamer Cassette" and five
Maxell CS-600HD "High density data cassette back coated 183m/600ft"
audio-cassette-like tapes with the notch on the back edge. Most have data
written on them from around 1994 but one of each type is unused and still in
plastic wrapping. Another of the Verbatim tapes is missing the box.
If they are of any use, they are available for postage from Ireland.
I don't think I'll need them, but thank you for the offer!
I've received the Teac MR-30 data recorder tape transport that Bob Rosenbloom
suggested. It looks promising, though I also plan to experiment with a Teac MT-2ST
streamer transport that may work better. I made a video introducing TU58 innards and the
MR-30 transport, in case anybody is interested:
https://www.youtube.com/watch?v=jo4qfVl-Y-o
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
3612
days inactive
3617
days old
17 comments
7 participants
participants (7)
-
aek＠bitsavers.org -
alexandre.tabajara＠gmail.com -
cc＠informatik.uni-stuttgart.de -
cclist＠sydex.com -
cctalk＠beyondthepale.ie -
chrise＠pobox.com -
nf6x＠nf6x.net