Re: HP 7970B Capstan?
On 09/08/2017 07:24 AM, Jay Jaeger via cctalk wrote: > On an IBM 1410 (at least - but I suspect this was likely widely true), > tape marks were *always* EVEN parity, 0x0F (Bits 8421), even on an odd > parity tape. Oh, I know--but remember that I was reading a 7-track tape in a 9-track drive. With the 9-track "scramble" of tracks and the position of the tracks, it makes perfect sense. I'm reading data that isn't massaged by a formatter--this is the (deskewed) output of the read channels. There are a number of documents on bitsavers that explain the 7/9 track differences. As an aside, I know that not all 9 track 800 NRZI drive formatters included the CRC byte before the final LRCC. I'll have to keep an eye out for those. --Chuck
Re: HP 7970B Capstan?
On 9/7/2017 9:35 PM, Chuck Guzis via cctalk wrote: > On 09/07/2017 07:18 PM, Jon Elson wrote: > >> Ah, well, I can see why a 7 track tape won't read well on a 9-track drive! > > I was a bit puzzled at why a tapemark would read as 135 (hex). Sigh--at > least the parity is correct. > > --Chuck > > On an IBM 1410 (at least - but I suspect this was likely widely true), tape marks were *always* EVEN parity, 0x0F (Bits 8421), even on an odd parity tape. Learned about this one day when all of a sudden all of our FORTRAN compiles caused the machine to error stop, but the machine seemed OK otherwise. WTH?? So we called IBM, explained what was going on. The CE calmly opened the console panel, and turned on the Asterisk Insert switch. He explained about the tape mark, and the fact that the FORTRAN compiler happened to write its intermediate files in odd parity. With that switch off, the even parity tape mark read under odd parity was placed into core that way - with even parity, which was an invalid core character, which stopped the machine. With that switch on, the tape mark came into storage as an asterisk, and the machine was happy. (In either case, the end-of-file latch was set, of course. See IBM 1410 Principles of Operation, "Read or Write Tape with Word Marks" - in the -3 version, it is pages 87 and 88.) JRJ
Re: HP 7970B Capstan?
On 09/07/2017 07:18 PM, Jon Elson wrote: > Ah, well, I can see why a 7 track tape won't read well on a 9-track drive! I was a bit puzzled at why a tapemark would read as 135 (hex). Sigh--at least the parity is correct. --Chuck
Re: HP 7970B Capstan?
On 09/07/2017 02:24 PM, Chuck Guzis via cctalk wrote: On 09/07/2017 09:32 AM, Al Kossow via cctalk wrote: The problem you're fighting is things are just a little marginal at 800 ? As I just mentioned to Al in an offlist email, it turns out that the stack of tapes labeled "800 NRZ 9 track" are, in fact, 7 track. Kyread doesn't lie. Argh. Time to try out the 7 track stack that Al sent me. Ah, well, I can see why a 7 track tape won't read well on a 9-track drive! Jon
Re: HP 7970B Capstan?
On 09/07/2017 09:32 AM, Al Kossow via cctalk wrote: > The problem you're fighting is things are just a little marginal at 800 ? As I just mentioned to Al in an offlist email, it turns out that the stack of tapes labeled "800 NRZ 9 track" are, in fact, 7 track. Kyread doesn't lie. Argh. Time to try out the 7 track stack that Al sent me. --Chuck
Re: HP 7970B Capstan?
On 09/06/2017 10:09 PM, Chuck Guzis via cctalk wrote: Here's a thought--can skew be adjusted by writing,a block of data in a forward direction, then reading it in reverse? Nope, it will always show perfect skew. What you COULD do, is take the tape off the reel and turn it around. THEN, any skew error would be doubled, This would be pretty annoying, you'd have to write, turn around, check, adjust and then repeat with the write step. Jon
Re: HP 7970B Capstan?
On 9/6/17 9:20 PM, Chuck Guzis via cctalk wrote: > If you could spare it for a week or so, I'd be grateful. On the other > hand, my MCU setup for the drive works great--it reads a tape and > stashes the data as a .TAP file on an SD card. The problem you're fighting is things are just a little marginal at 800 ?
Re: HP 7970B Capstan?
On 09/06/2017 08:14 PM, Al Kossow via cctalk wrote: > I could mail you a skew tape. > > On 9/6/17 8:09 PM, Chuck Guzis via cctalk wrote: > >> I'm just scratching around looking for ideas. If you could spare it for a week or so, I'd be grateful. On the other hand, my MCU setup for the drive works great--it reads a tape and stashes the data as a .TAP file on an SD card. At the same time, the SD card appears as a USB storage device, so you don't need to remove it from the MCU to read it. Interaction is done via a serial connection. Thanks, Chuck
Re: HP 7970B Capstan?
I could mail you a skew tape. On 9/6/17 8:09 PM, Chuck Guzis via cctalk wrote: > I'm just scratching around looking for ideas.
Re: HP 7970B Capstan?
On 09/06/2017 06:50 PM, Jon Elson wrote: > On 09/06/2017 12:11 PM, Chuck Guzis via cctalk wrote: >> Will do--I've mentioned in another message that I'm preparing a test >> tape using the drive, so I should get some inkling of timing and >> adjustments pretty quickly. --Chuck > While that is good, an all-ones test tape made on ANOTHER drive that is > known to be in good calibration (both speed and skew) is very helpful in > standardizing drives. Here's a thought--can skew be adjusted by writing,a block of data in a forward direction, then reading it in reverse? And then writing in reverse and reading in the forward direction? Ideally, when all is right with the world, there will be no difference in the skew. I'm just scratching around looking for ideas. --Chuck
Re: HP 7970B Capstan?
On 09/06/2017 06:50 PM, Jon Elson wrote: > On 09/06/2017 12:11 PM, Chuck Guzis via cctalk wrote: >> Will do--I've mentioned in another message that I'm preparing a test >> tape using the drive, so I should get some inkling of timing and >> adjustments pretty quickly. --Chuck > While that is good, an all-ones test tape made on ANOTHER drive that is > known to be in good calibration (both speed and skew) is very helpful in > standardizing drives. I realize that, but I don't have another NRZI known-good drive to calibrate against. PE and GCR, sure, but not NRZI. I'll have to rely on luck for skew adjustment. I can probably work off the tach signal for speed. --Chuck
Re: HP 7970B Capstan?
On 09/06/2017 12:11 PM, Chuck Guzis via cctalk wrote: Will do--I've mentioned in another message that I'm preparing a test tape using the drive, so I should get some inkling of timing and adjustments pretty quickly. --Chuck While that is good, an all-ones test tape made on ANOTHER drive that is known to be in good calibration (both speed and skew) is very helpful in standardizing drives. Jon
Re: HP 7970B Capstan?
On 09/06/2017 09:41 AM, Jon Elson via cctalk wrote: > On 09/06/2017 11:03 AM, Al Kossow via cctalk wrote: >> >> On 9/5/17 11:09 PM, Chuck Guzis via cctalk wrote: >> >>> The read results have improved considerably. >> tension arm servo acting reasonably? >> >> what style arm sensor does your drive have? The same as shown in the 7970B/C service manual--a semicircular slotted metal disk attached to the sensor arm intersecting a photocell/lamp box. Looking at the drive casting, it's marked as a Rev F DHB, if that means anything. Just judging by the CE notes scribbled on the inside of the door, I'm guessing that this is a late model.The write-enable sensor assembly is different from that pictured in the service manual. The sensor "finger" is attached to an L-shaped flat arm with two bolts for adjustment. I did replace the microswitch, as the lever arm had broken off at some point. > OK, so one of the first checks is to put a scope on the tach signal > (should be a test point for that) and verify the velocity is a perfect > trapezoid. If there are regular spikes in the trace, that could > indicate a problem in the tachometer. If there are bobbles in the trace > at jerk points, then adjusting the damping on the servo loop will help. > > Then, you need a tape written at known density, and read that, looking > at a data track. The best is a tape with long records of all ones, > which puts transitions on all channels (in 800 BPI NRZI mode). Adjust > the right pot to set the data rate to the calculated value for that > drive's tape speed. > > Then, move the scope to the skew test point. The drive will have 9 FFs > (one per track) and these go to a resistor summing point. You should > see nearly a square wave when skew is perfect, but will get much more > slope with little steps on the flanks when skew is not good. The > all-ones tape is perfect for this adjustment. You will see the shape of > the signal fluctuate, as tape laying on the reel for some time will > stretch just a bit, and tend to "weave" across the tape guides. There > will be adjusting screws to set the head skew to minimize this. > > All the above applies mostly to 800 BPI NRZI. PE and GCR are far less > critical on this. Will do--I've mentioned in another message that I'm preparing a test tape using the drive, so I should get some inkling of timing and adjustments pretty quickly. --Chuck
Re: HP 7970B Capstan?
On 09/06/2017 11:03 AM, Al Kossow via cctalk wrote: On 9/5/17 11:09 PM, Chuck Guzis via cctalk wrote: The read results have improved considerably. tension arm servo acting reasonably? what style arm sensor does your drive have? OK, so one of the first checks is to put a scope on the tach signal (should be a test point for that) and verify the velocity is a perfect trapezoid. If there are regular spikes in the trace, that could indicate a problem in the tachometer. If there are bobbles in the trace at jerk points, then adjusting the damping on the servo loop will help. Then, you need a tape written at known density, and read that, looking at a data track. The best is a tape with long records of all ones, which puts transitions on all channels (in 800 BPI NRZI mode). Adjust the right pot to set the data rate to the calculated value for that drive's tape speed. Then, move the scope to the skew test point. The drive will have 9 FFs (one per track) and these go to a resistor summing point. You should see nearly a square wave when skew is perfect, but will get much more slope with little steps on the flanks when skew is not good. The all-ones tape is perfect for this adjustment. You will see the shape of the signal fluctuate, as tape laying on the reel for some time will stretch just a bit, and tend to "weave" across the tape guides. There will be adjusting screws to set the head skew to minimize this. All the above applies mostly to 800 BPI NRZI. PE and GCR are far less critical on this. Jon
Re: HP 7970B Capstan?
On 9/5/17 11:09 PM, Chuck Guzis via cctalk wrote: > The read results have improved considerably. tension arm servo acting reasonably? what style arm sensor does your drive have?
RE: HP 7970B Capstan?
> -Oorspronkelijk bericht- > Van: cctalk [mailto:cctalk-boun...@classiccmp.org] Namens Chuck Guzis via > cctalk > Verzonden: woensdag 6 september 2017 8:09 > Aan: CCtalk <cctalk@classiccmp.org> > Onderwerp: Re: HP 7970B Capstan? > > On 09/05/2017 10:47 PM, David Collins wrote: > > From my old HP CE days you could always get black off the capstan if > > you used alcohol or similar (never recommended). The key thing is it > > shouldn't be sticky/gooey. > > I noticed the issue by starting a read from load point and observing that the > bit > timing and data was all over the place until things came up to speed. I was > getting different data with each try. > > I pulled the capstan off, cleaned the powdery stuff off with isopropanol > and gave it a 30 minute soak in methyl salicylate and isopropanol. The > capstan comes out clean (no more shedding) and firm. > > I also noticed that the bits of foam on the inside of the capstan had > degraded to > goo. I replaced those with some new foam. > > The read results have improved considerably. Not perfect, but much better. > > --Chuck Chuck, The tacho is a combined motor generator unit, it could be a good idea to read the adjustment procedures for the capstan tacho. The behavior is electronic controlled and adjustable. You can find the manual at the HP Museum site: http://www.hpmuseum.net/document.php?hwfile=2807 7970CEAdjusment...etc shouldn't be difficult to find. The speed and transient behavior of the capstan should be checked and possibly adjusted. You should also check if the capstan turns easy without noise. -Rik
Re: HP 7970B Capstan?
On 09/05/2017 10:47 PM, David Collins wrote: > From my old HP CE days you could always get black off the capstan if > you used alcohol or similar (never recommended). The key thing is it > shouldn't be sticky/gooey. I noticed the issue by starting a read from load point and observing that the bit timing and data was all over the place until things came up to speed. I was getting different data with each try. I pulled the capstan off, cleaned the powdery stuff off with isopropanol and gave it a 30 minute soak in methyl salicylate and isopropanol. The capstan comes out clean (no more shedding) and firm. I also noticed that the bits of foam on the inside of the capstan had degraded to goo. I replaced those with some new foam. The read results have improved considerably. Not perfect, but much better. --Chuck
Re: HP 7970B Capstan?
From my old HP CE days you could always get black off the capstan if you used alcohol or similar (never recommended). The key thing is it shouldn't be sticky/gooey. The official cleaning solution when they were in active use was water and a low lint cloth. David Collins (Sent from out of office) > On 6 Sep 2017, at 3:18 pm, CuriousMarc via cctalk <cctalk@classiccmp.org> > wrote: > > All my capstans are feel rather hard (I think because the rubber is pretty > thin), and do not feel sticky either. The finish is matte. But they are not > powdery at all, they would not leave any residue rubbing your finger on them, > so something looks odd with yours. When the tape is threaded but the power is > off, they grab the tape well enough that turning the bottom reel by hand will > move the capstan easily, without any slippage. > Marc > > -Original Message- > From: cctalk [mailto:cctalk-boun...@classiccmp.org] On Behalf Of Chuck Guzis > via cctalk > Sent: Monday, September 04, 2017 9:17 PM > To: CCtalk > Subject: HP 7970B Capstan? > > I've got an HP 7970 tape drive here that's a little dicey on its reading > (haven't tried writing yet). One thing that I noticed today was that the > surface of the capstan is a bit powdery--wipe it with an alcohol swab and you > get a lot of black schmutz. > > It wasn't very soft and it occurred to me that it might be slipping, so I > cleaned it off and treated it with some methyl salicylate (oil of > wintergreen). It's a bit softer now, but I'm wondering, for those of you who > still have functioning drives, how soft or "sticky/grippy" is right. > > Any ideas? > > Thanks, > Chuck > > >
RE: HP 7970B Capstan?
All my capstans are feel rather hard (I think because the rubber is pretty thin), and do not feel sticky either. The finish is matte. But they are not powdery at all, they would not leave any residue rubbing your finger on them, so something looks odd with yours. When the tape is threaded but the power is off, they grab the tape well enough that turning the bottom reel by hand will move the capstan easily, without any slippage. Marc -Original Message- From: cctalk [mailto:cctalk-boun...@classiccmp.org] On Behalf Of Chuck Guzis via cctalk Sent: Monday, September 04, 2017 9:17 PM To: CCtalk Subject: HP 7970B Capstan? I've got an HP 7970 tape drive here that's a little dicey on its reading (haven't tried writing yet). One thing that I noticed today was that the surface of the capstan is a bit powdery--wipe it with an alcohol swab and you get a lot of black schmutz. It wasn't very soft and it occurred to me that it might be slipping, so I cleaned it off and treated it with some methyl salicylate (oil of wintergreen). It's a bit softer now, but I'm wondering, for those of you who still have functioning drives, how soft or "sticky/grippy" is right. Any ideas? Thanks, Chuck
Re: HP 7970B Capstan?
On 09/05/2017 10:55 AM, Jon Elson wrote: > Some spring-arm drives had really bad problems with tape slipping. I > know some Pertecs had circular grooves in the capstan to prevent an air > bearing effect. On one of our spring-arm drives, I cut thumbs out of > latex gloves and stretched them over the capstan. This TOTALLY removed > the slipping problem until the latex wore out. The 7970 capstan has those same grooves. I may resort to the latex trick if OoWG doesn't make it grippy enough. --Chuck
Re: HP 7970B Capstan?
On 09/04/2017 11:16 PM, Chuck Guzis via cctalk wrote: I've got an HP 7970 tape drive here that's a little dicey on its reading (haven't tried writing yet). One thing that I noticed today was that the surface of the capstan is a bit powdery--wipe it with an alcohol swab and you get a lot of black schmutz. It wasn't very soft and it occurred to me that it might be slipping, so I cleaned it off and treated it with some methyl salicylate (oil of wintergreen). It's a bit softer now, but I'm wondering, for those of you who still have functioning drives, how soft or "sticky/grippy" is right. Some spring-arm drives had really bad problems with tape slipping. I know some Pertecs had circular grooves in the capstan to prevent an air bearing effect. On one of our spring-arm drives, I cut thumbs out of latex gloves and stretched them over the capstan. This TOTALLY removed the slipping problem until the latex wore out. Jon
Re: HP 7970B Capstan?
On 09/04/2017 09:41 PM, Mike Loewen via cctalk wrote: > On Mon, 4 Sep 2017, Chuck Guzis via cctalk wrote: > > > >The rubber on my 7970E capstan is not tacky, but fairly firm. I've > used Techspray Rubber Roller Rejuvenator successfully on keypunch > rollers to make them a bit "grippier" - maybe it would help? > > https://www.techspray.com/roller-rejuvenator > Probably a related compound to oil of wintergreen. "Rubber Renue" was another such stuff--basically OoWG and Xylene Smells nice. The biker community seems to like it to refurb rubber parts, either mixed with xylene, or hot water or isopropanol. Worth a little more experimenting. The rest of the drive is near-pristine, not even any dust inside. I'm driving it with a STM32F407 MCU. --Chuck
Re: HP 7970B Capstan?
On Mon, 4 Sep 2017, Chuck Guzis via cctalk wrote: I've got an HP 7970 tape drive here that's a little dicey on its reading (haven't tried writing yet). One thing that I noticed today was that the surface of the capstan is a bit powdery--wipe it with an alcohol swab and you get a lot of black schmutz. It wasn't very soft and it occurred to me that it might be slipping, so I cleaned it off and treated it with some methyl salicylate (oil of wintergreen). It's a bit softer now, but I'm wondering, for those of you who still have functioning drives, how soft or "sticky/grippy" is right. Any ideas? The rubber on my 7970E capstan is not tacky, but fairly firm. I've used Techspray Rubber Roller Rejuvenator successfully on keypunch rollers to make them a bit "grippier" - maybe it would help? https://www.techspray.com/roller-rejuvenator Mike Loewen mloe...@cpumagic.scol.pa.us Old Technology http://q7.neurotica.com/Oldtech/
HP 7970B Capstan?
I've got an HP 7970 tape drive here that's a little dicey on its reading (haven't tried writing yet). One thing that I noticed today was that the surface of the capstan is a bit powdery--wipe it with an alcohol swab and you get a lot of black schmutz. It wasn't very soft and it occurred to me that it might be slipping, so I cleaned it off and treated it with some methyl salicylate (oil of wintergreen). It's a bit softer now, but I'm wondering, for those of you who still have functioning drives, how soft or "sticky/grippy" is right. Any ideas? Thanks, Chuck
