29 May
2016
29 May
'16
1:03 p.m.
I have been fiddling with a TU58-EX device, dual TU58 drives in a small box.
The capstans is replaced. I used silicone tubing which I glued on and then
sanded down a bit. PVC tubing in a size that would fit seems to be
unavailable in Sweden.
The two capstans were a little bit different in diameter at first. Drive
one closer to 17 mm but drive zero around 16.5mm. Drive zero read 5 of 8
tapes (two more tapes had belt breakage). One tape gave "Invalid Directory"
in RT11 the two other gave "Error reading directory". On drive one just one
tape was readable.
So the decision was to get closer to the nominal 5/8" (which I read was the
OD in a post by Tony Duell). With both drives at 16mm drive zero still read
the same amount of tapes and drive one read the same tapes plus one more
which was not readable on drive zero.
Highly annoying. So I decided to read more on the TU58. The spec says that
the bit time is 41.2 us. When I measure I get reading of between 42 and 44
us. (Yes I should have measured before trimming the capstans). So now the
tape is too slow. Although the OD is slightly above the nominal.
Is silicone tubing too soft?
Having been working on a project to recover a tape from a Zilog S8000
machine together with AJ (http://mightyframe.blogspot.se/) I just thought
that it might be possible to read the TU58 and HP DC100 tapes with some
other hardware doing post processing in a regular Linux box.
Would it be possible to use a Floppy Tape (QIC-117) tape drive to read
them? It appears that the tapes are not identical in size. A DC1000 is 0.25
" while the DC100 is 0.15". The capstan position would also differ. Are
there other physical differences that I am not aware off?
Since the TU58 drive is not able to format a tape it could be useful to
also write a new tape with TU58 format. But I guess that there are
difference in coercivity between different tapes.
I have read that Rik Bos successfully converted HP85 drives to take DC1000
tapes by modifying the capstan and changing the write current.
Before I go ahead buying some old Colorado T1000 drive dirt cheap I just
like to ask if this project is doomed because of whatever reason.
/Mattis
29 May
29 May
2:10 p.m.
On 5/29/16 10:03 AM, Mattis Lind wrote:
Would it be possible to use a Floppy Tape (QIC-117)
tape drive to read
them?
No. The heads are movable on floppy tapes, and the format is completely different.
If you send me your address, I can send you a chunk of tubing that Brad Parker and I
have used to repair the drive wheels.
4:05 p.m.
2016-05-29 20:10 GMT+02:00 Al Kossow <aek at bitsavers.org>:
On 5/29/16 10:03 AM, Mattis Lind wrote:
Yes. I am aware of that. But that also means that the head can be moved
into the spot where the information is recorded. My quick reading of the
QIC-117 spec give that the one control the drive with a quite weird
interface consisting of the STEP, TRACK ZERO and INDEX lines.
The READ DATA is supposed to be the actual flux transitions. Likewise is
WRITE DATA.
What is not yet clear to me is whether that the QIC-117 drive is rather
dumb and leaves most work to the FDC or if it includes a lot of logic to
handle the reading process.
http://www.qic.org/html/standards/11x.x/qic117j.pdf
The QIC36 drive that was used to recover the information stored on the
Zilog S8000 tapes is completely stupid. The read data just reflected the
flux transitions which AJ recorded using a logic analyzer so that it was
possible to decode the MFM data stream in software. The tracks of course
also differed since the QIC36 drive was 9 track while the DEI drive used
for recording the S8000 tapes used four (fixed) heads. But nevertheless it
was possible to recover the entire tape thanks to the circuitry that AJ
devised to control the head position. Using a more modern SCSI drive it
will not be possible to read these tapes since they only handle GCR
encoding (as does QIC02 drives)
Could this kind of operation be done with a QIC-117. I.e. is the drive
stupid enough? Reading more of the spec may indicate that "segment" concept
might be the culprit that makes it non-feasible. It could be that the drive
keeps track of the segments somehow.
Of course there is always the possibility of "hacking" the drive. But then
it would help with schematics which is probably not available for these
quite "modern" things.
All input appreciated!
Would it be possible to use a Floppy Tape
(QIC-117) tape drive to read
them?
No. The heads are movable on floppy tapes, and the format is completely
different.
If you send me your address, I can send you a chunk of
tubing that Brad
Parker and I
have used to repair the drive wheels.
That would be very nice. Send you a message off-list.
/Mattis
2:34 p.m.
My first attempt to use silicone tubing to repair my TU58 rollers was also unsuccessful.
Maybe the material I used is too soft? Your experience suggests to me that I should hook
up an oscilloscope to measure bit timing, and then adjust the roller composition and
diameter to arrive close to the correct bit length.
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
3 p.m.
On 5/29/16 11:34 AM, Mark J. Blair wrote:
My first attempt to use silicone tubing to repair my
TU58 rollers was also unsuccessful. Maybe the material I used is too soft? Your experience
suggests to me that I should hook up an oscilloscope to measure bit timing, and then
adjust the roller composition and diameter to arrive close to the correct bit length.
dug this up from the cctlk archive
--
http://www.usplastic.com/catalog/item.aspx?itemid=23485
It just says "norprene A-60-F" the outside. It measures 3/8" I.D.
I think other's have said 7/16" i.d., but I could not find that. I know I bought
it from US Plastics and it's A-60-F.
You have to buy 10' but its not very expensive.
-brad
4:06 p.m.
-----Original Message-----
From: cctalk [mailto:cctalk-bounces at classiccmp.org] On Behalf Of Al Kossow
Sent: Sunday, May 29, 2016 3:00 PM
To: cctalk at classiccmp.org
Subject: Re: TU58 yet one more time.
On 5/29/16 11:34 AM, Mark J. Blair wrote:
dug this up from the cctlk archive
--
http://www.usplastic.com/catalog/item.aspx?itemid=23485
It just says "norprene A-60-F" the outside. It measures 3/8" I.D.
I think other's have said 7/16" i.d., but I could not find that. I know I
bought it from US Plastics and it's A-60-F.
You have to buy 10' but its not very expensive.
-brad
-----
"A-60-F" would be here?
http://www.usplastic.com/catalog/item.aspx?itemid=23886 $26.95 for 10 foot
length. Ouch!
3/8" ID x 5/8" OD x 1/8" Wall TygonR A-60-F Hot Food & Beverage Tubing
-----
My first attempt to use silicone tubing to repair my
TU58 rollers was also
unsuccessful. Maybe the material I used is too soft? Your
experience
suggests to me that I should hook up an oscilloscope to measure bit timing,
and then adjust the roller composition and diameter to arrive close to the
correct bit length.
4:12 p.m.
2016-05-29 20:34 GMT+02:00 Mark J. Blair <nf6x at nf6x.net>:
My first attempt to use silicone tubing to repair my
TU58 rollers was also
unsuccessful. Maybe the material I used is too soft? Your experience
suggests to me that I should hook up an oscilloscope to measure bit timing,
and then adjust the roller composition and diameter to arrive close to the
correct bit length.
It might be too soft. But the capstan doesn't look to flat when the tape is
inserted. So I am not sure. Pin 5 on the 8085 has the read data and pin 7
is the strobe data.
Another question is how sensitive it is to speed variations. Apparently it
is. Or my tapes are bad. But I cannot tell for sure since I have no working
reference.
The encoding scheme is rather simple. And decoding is done by some analogue
electronics doing integration. Since it is a fixed capacitor value used it
can be problematic.
/Mattis
--
Mark J. Blair, NF6X <nf6x at nf6x.net>
http://www.nf6x.net/
3009
days inactive
3009
days old
6 comments
4 participants
participants (4)
-
aek＠bitsavers.org -
mattislind＠gmail.com -
nf6x＠nf6x.net -
pbirkel＠gmail.com