Schedule May 2006

• Present were Ron Williams and Bill Flora, with Ron Crane and Robert Garner arriving later.

• Ron Williams and Bill Flora retiming the circuit breakers [CBs] in the 1402 card reader trying to get rid of the reader checks. The "CBs" are not the replacements for fuzes in your house, but cam operated contact points to open and close a circuit to an accuracy of 1 degree of rotation - Very much like the "points" in the ignition system in older gasoline engines - and the timing is even more vital.

  That didn't work, so a wiring check showed a broken wire from column 71 of the read check brushes. Now fixed.

• Ron Williams was also working on the Instruction Overlap problem, which prevents the machine from (among other things) executing a loader reading in an object deck.

  He is now suspicious of TAU (Tape Adaptor Unit) signals.

• Ron Crane fixed more SMS cards. Ron Crane thinks we should swap out ALL the slow 2N1303's with the fast, 102-eqiuv silicon 2N2894's on the repaired TAU boards. (For a while we were replacing failed 102 type transistors with 2N1301 transistors, which turned out to be much slower.) About a dozen suspects.

We have an intercontinental project going. The 1401 system came from Germany. Not only do "spare" parts and help come in from Red Bluff, Ca, Atlanta, GA. , e-bay, Digikey [of Thief River Falls ;-), Minn] , ... - but now from Australia - just about the other end of the earth. Ron Williams and team have been straining failing eyesight trying to read the ALDs (IBM term for schematics) from Germany. - The Germaan ALDs are the results of many generations of copying - faint and bleary, just like our eyes :-(( Even Ron's big magnifying glass is weary. John Dean of Sidney, Australia has a much more readable set of 1401 ALDs with about the right options. He sent several scanned samples to Robert Garner. John found a copy place willing to scan the ALDs to CD or DVD for A$525 (Australian Dollars). Robert Garner has created a work statement and is arranging suitable funds to be transferred to John - and of course the work will be done to Al Kossow's specifications. The e-mail copy list is too long to list here, and includes Max Burnet of Australia who visited CHM and our 1401 a year ago :-)) Ain't we got fun :-))

Present were: Bob Erickson, Ed Thelen, Robert Garner, Frank King, Chuck Kantmann, Don Cull, Bill Flora, Grant Saviers, Ron Williams, Ron Crane. John Toole and Dave Dial (both of CHM) also dropped by.

• Bob Erickson and Frank King continued working on the 077 collator. The mechanical numerical comparitors are back out on the bench -

• Ron Williams, Bill Flora, Don Cull continued to struggle with instruction overlap.

• Grant Saviers brought in some more tools, and 100 more separators for the SMS drawers. He arranged tools in the main tool chest, and made/placed labels.

• Ed Thelen and Chuck Kantmann continued making labeling SMS drawer separators, and taking the unsorted SMS cards from the various boxes and sorting them into the SMS drawers. There are some types of SMS cards We have a large number (say 30+) of some types of SMS cards. As there will not be enough room in the 4 large SMS drawers for all the cards, some of the surplus cards are being returned to now partitioned and labeled boxes.

• Ron Crane fixed several obviously damaged cards from the spare SMS cards.

• (Allen Palmer (of the 729 tape drives) is off traveling again. Last time it was Paris. There are rumors of ?Tahiti? this time? Life must be tough in the fast lane. ;-))

• We are proposing that this fakey 1401 operator panel (on the 1401 in Visible Storage) be replaced with one of "our" spares - one from original purchase from Germany, one donated to "us". Note that someone drilled holes and inserted non-period yellow LEDs into the this front panel, and the white paper function labels are also bogus. A real 1401 operator's panel is on the right.

• This is the Visible Storage 1403 printer ribbon with the ends of the quick change print chain just visible on each side. The 1403 print chain on "our" machine is not the "quick change" type - the operator cannot change it. A Customer Engineer must do the more complex work.

Present were Glenn Lea, Frank King, Chuck Kantman, Don Cull, Don Luke, Bob Erickson, Ed Thelen, Ron Williams. Robert Garner and Ron Crane were there briefly.

• DonL, FrankK, BobE

  Electromechanicals

  Bob Erickson found another relay with an open magnet coil in the 077 collator that was preventing successful compares between the two input feeds. An application form to access an 077 in Visible Storage or elsewhere has been submitted.

• Chuck Kantmann and Ed Thelen completed labeling, and filling the four new drawers with SMS cards. Chuck Kantmann and Ed Thelen (with Ron Williams our customer observing intently) completed a count/inventory of the SMS cards in the four drawers and overflow boxes. And yes - the drawers are a little out of sequence :-(( After filling the drawers we discovered the bottom four drawers were *not* the same size. - When life is less pressing, we may swap the contents and labels of two drawers.

• The overflow boxes with new white labels reflecting contents :-))

• Ron Williams continued struggling with the instruction overlap "feature" of our 1401 - he did get the combination instruction "3" "Write and Read" of - read a card - print a line instruction going.

• a random scene - Chuck Kantmann and Ron Crane intently ???

• Frank King feels that he has worked himself out of the job of getting the 1403 printer going, :-))) and is getting organized to overhaul the 083 sorter that came in with the shipment from Germany.
  Ed Thelen Xeroxed some manuals selected by Frank, and while Xeroxing them,
  Len Shustek of CHM came by and suggested that the troublesome long fold out schematics could be copied at Kinkos - they have appropriate equipment.
  I went to Kinkos and copied two of the long fold out schematics.
  - Really neat - you can copy indefinite length (like say four feet long) fold out schematics for $4.75/sheet :-))
  - - Folks who have spent hours "stitching" together foldouts will really appreciate this option. -
  - and even enlarge a 1 foot high input by a factor of 2 giving a 24 inch high output.

• Glenn Lea removed the final power supply from the 729 tape drives for checking. This may not sound like much of a job, but you would not believe the dust, grit, grime, vacuuming, goop wiping involved - not to mention the 40ish wires to disconnect and label, and man-handling the 50+ pound unit out of the 729 -

• The current status of "our" 026 Key Punch - is - as I understand from Bob Erickson -

  The "interposers" in the punch part of the 026 are sticking, and these are read by the print logic to set up the character to print. The print trouble we were having was *not* a problem in the mis-alignment of the print platten vs. print wires. Bob says that cleaning these interposers is a 25 hour job for someone who is current in doing the job - maybe double the time for Bob who has not done that for 50 years. The probable 50ish person-hours to disassemble, work on, reassemble the 026 key punch is many work sessions where a working key punch would not be available. - And we need a working key punch to make check decks for the 1402 card reader, the 077 collator, and the upcoming 083 sorter - much less any software functions. Therefore, Bob would like one or more working key punches available to the 1401 restoration crew before attempting to clean/lubricate/... the 026 interposers and get the 026 print unit working. Bob thinks that if a working key punch is available, we should borrow it.

---

from Bob Feretich

Wednesday, May 17th, (Grant's shop): (Grant S., Jeff S., Ron C. (afternoon), & me)
729-V drive debug:
We disassembled and refurbished the 12 port switching relays from the drive. The pin wires were ultrasonically cleaned. Fixed contacts were blasted with glass beads. One relay was damaged (part of a coil pin broke off), but the damage doesn't seem to effect the relays operation. The damaged relay has been marked, just in case in the future... Once re-assembled, contact resistance was typically <0.2 Ohms. The contacts and wires look brand new.
Tape Analyzer Debug:
Jeff downloaded the Analog Board Test Loops that he and Dave created into the Analyzer's PIC. Grant and Jeff debugged the Analog Board's registers and pulse shaping circuits.

extra Saturday May 20th - TAU Team

TAU/729V Debug: (Ron C. and me):

Replaced a card in the 729 and one in the TAU. Now the drive successfully performs "Rewind" and "Rewind & Unload" operations issued from the TAU. Yeah!

I tried "Go Forward" and "Go Backward" operations from the CE Panel. The tape moves in the appropriate direction, but there is loud contact chatter. I looked at the "GO" signal, it is going active for about 15mSec at a ~20mSec period. This makes sense if short records are being read. I did not like the sound of the chatter, so I ceased debugging until I could meet with Allen and get his opinion on the situation.

I also noticed that the 729's "START" switch and the "Caspain Out" switches are intermittent. The START switch has to be rocked back and forth to activate. Sometimes the caspains are out, but the switch signal does not detect it. We may need to refurbish the switch contacts.

Ron Crane found multiple bad transistors per 729 card that we had marked for repair. So, we pulled all of the APF (+P to S converters) and APE (-N to S converters) from both Model 5 drives. These card receive signals from the TAU. Each drive has 3 APFs and 6 APEs. Each card has 6 transistors. We built a test jig for the cards.

The first 729-V (the one attached to the 1401) had many bad cards ( I think 7 of the 9 cards had faults.). We replaced 17 transistors and a diode. (Roughly 1/3rd of the transistors were bad.) The cards in this drive are very dusty. Some transistors were leaky. Some transistors had open junctions. These level converter cards are more sensitive to transistor degradation than logic cards, but that is a high percentage of bad transistors. Was this drive ever subjected to bad power conditions?

The other 729-V had only one bad card. We replaced a cracked diode on this card. (We also replaced a transistor, but we believe that the diode was the root problem.) The cards in this drive look new (very little dust & many metal film resistors in places where the first drive's.cards had carbon composition resistors.)

BTW: The second 729-V drive does not have the dual port feature (no relays).

Regards,
Bob

Grant, Allen, and I met today to discuss Tape/TAU debug plans.

1. Relays: One Mod 2 drive has the dual port feature (same types of relays as the Mod 5 that was just refurbished). The two Mod 2 and one Mod 4 drives use a lot of relay logic. Allen pulled three sample relays from the drives. The relays were in better shape than in the original Mod 5 drive, but each relay had at least one contact with resistance >100 Ohms. We decided that something needs to be done to improve them. Grant suggested a cycling fixture that would use an electric arc to clean the oxide on the contacts. We decided to take no action until Grant is back from his travels.
2. CP16: Early in the session we popped CP16 several times. (CP = circuit protector (aka circuit breaker)). Allen exercised the caspian in/out sense switches and the problem stopped occurring.
3. Chattering noise: Allen Identified the "chattering noise" that occurs during the "Go Forward" and "Go Backward" operations as the "normal noise the drive makes when transferring lots of short records".
4. The  drive seems to perform the TAU "Go Forward" and "Go Backward" operations successfully.
5. The  drive moves tape correctly for TAU "Write" operations, error indicators appear on the TAU panel. More work to do here.
6. Each of the two Mod 5 drives have 32 MX__ SMS cards. (By far, its the most heavily used card in the drive.) It would probably make sense to build a test bed for these cards and test them like we tested the APE and APF cards. MX is a quad 2-inp NAND. Only one transistor per gate.
   

   Ron, please take a look at the circuit for this card and suggest a way to check it out. I would like to know if the degraded transistors are limited to the interface level shifter cards, or if the problem is wide spread throughout the drive.
   

TAU/drive debug to continue Thursday.

Regards,
Bob

Canceled by Allen Palmer on Wednesday May 24th due to faulty 729 drive.

Well, maybe it wasn't?

TAU & 729-V drive debug:
Jeff and I came in Thursday morning to start a debug session. We were surprised to find that all of the MX cards had been removed from the drive. Since our debug plan for the morning required the drive, we postponed it to the following week. (An e-mail communication would have helped here.)

Ron C. came in around noon to complete evaluation of the MX cards. Out of 36 cards (144 transistors), I believe that he found about 4 bad cards. (Ron can provide exact numbers and symptoms.) Most significantly, the failure trend that we observed in the APE and APF cards did not continue with the MX cards. This leads me to believe that the transistor failures may have been specific to the batch of transistors. Did IBM stamp the transistors with a date/batch code?

Tape Channel Analyzer debug:
Dave, Jeff, and I held an Analyzer Spec. Review session in the afternoon. Dave is getting into the meat of the firmware, so hopefully the review enhanced the common understanding.

Next Thursday, we plan to reinstall the MX cards into the drive and continue with the debug of the Tape Write Operation.

I have seen the status declaring that the "Overlap Bug" has been shot. The diagnostic decks are on a self near the 1406 Core Storage box and the documentation is in a binder in the rack near the air conditioner. I can't come in on Wednesday, but I am available most of the day via phone (408-464-3286).

Regards,
Bob Feretich

• From Ron Mak - 1401 software lead

  Thanks, Ed, for opening up the 1401 room and spending some time with Prof. Chen Friday. Chen Yu teaches at Renmin University in Beijing and is one of the leading computer scientists in China today. Prof. Chen had a nice meeting afterwards with John Toole (CEO of the Computer History Museum). Prof. Chen gave John a brief history of Chinese computer technology. (I wish we had videotaped the meeting.) There may be some collaboration in the future that will get us some Chinese computer artifacts.

• Present were Ron Williams, Bob Erickson, Tim Coslet, Ron Crane. Jeff Stutzman came in about 3:00.

• As of 3:30 - Ron Williams announced OVERLAP IS WORKING. It had been stuck ON, and interfered with card reading by the load programs that were not expecting Overlap. The overlap "feature" has connections all through the 1401.

  Now, the Overlap feature is reset to "OFF" when the machine is powered on or the card load is depressed

  While chasing the problem maybe 6 bad cards have been found as bad, still with out fixing the OVERLAP problem.

  The implications are that we will be able to read cards successfully, and be able to read in a loader and object deck and execute the program.

  Tim Coslet said that the bad card has an open diode.

• I had come in Friday and worked on a 729 Mod II power supply. I reformed/verified that the capacitors could handle their nominal voltages - but also noted that the 140 volt power supply (unregulated) was emitting 160 volts when run from the ferroresonant transformer excited by 50 Hz !! Checking further I noted that one of the wires from the ferroresonant transformer to its resonating capacitors had been too close to a bleeder resistor and had the plastic insulation charred - and figured that the wire might be open decreasing the number of connected resonating capacitors from three to two - and maybe that was causing the incorrect output voltage.

• Tim Coslet worked on the above 729 Mod II power supply and noted that the big green 1,000 ohm bleeder resistor was open, and its wires were also damaged.

• Ron Crane tested 36 MX cards from the rebuilt (green dot identification) 729 Mod V tape drive. Three transistors on those cards had leakage greater than 2 ma. (rather bad) He is replacing them.

• Bob Erickson continued checking and replacing actuator coils in the comparitors of the 077 collator.

	... picture show a gash in the cover made by the motor belt that had been operated in the wrong pulley position for quite some time! (the "unused" inner track, closest to the front. and now we see why...) Allen Palmer says "The cover is the 'head read/write protection' cover. It protects the head up/down mechanism, the wiring to the erase head, lamps and photocells.
	just a little scary: Bill Flora's screw driver is pointing to the screw for adjusting the card-punch card thickness (as I recall... ;-)
	This is Glenn Lea removing a power suppy from a 729 Mod II tape drive.
	This is the complete stack of magnetic clutches on the shaft with the reel hub on the left. The geared clutch is energized when the reel needs to move slowly as in loading and unloading the vacuum columns. A separate motor turns the worm gear that drives this clutch.
	A stack of clutches in one of the Mod II 729 tape drives (before restoration ;-)

Jeff Stutzmann kindly provided these images and explaination

Regarding: 1401 Memory Overlap feature. Enclosed are 3 pictures to choose from showing the "Memory Overlap" feature of the 1401 working properly. You can get more details from Ron, but I think that what the scope is showing is the Card reader activating proper timing of a card read and the Memory Overlap feature pulse happening at the end, instead of showing between all other pulses. As you may know this was a big win for the team. It was one open diode. (also picture enclosed but hard to see the open cat hair in the diode.)

---

• Present were - oh golly, so many - Ron Williams, Frank King, Glen Lea, Bob Erickson, Don Luke, Chuck Kantmann, Ed Thelen - I'm sorry, I probably forgot a few - this senility is just killing me. Robert Garner was there in the morning.

• Bob Erickson found so many bad (open) magnet windings that he has decided not to restore the top four of 16 comparitor in each of the two units. I think the last count was 11 bad magnets, he feels that robbing so many from the good comparitor we took from Visible Storage would leave that one too bad. I must remember to ask Dag if we can rewind the bad electro-magnets.

• Ron Williams celebrated killing the Overlap problem by doing a card reader to printer copy of several diagnostic decks. Robert Garner and Ron determined that many of the diagnostic decks were for testing the tape units.

• Don Luke & Chuck Kantman started struggling with a bad "B" bit in the 8,000-11,999 range in memory. They don't think that is related to the known broken wire in the 1406 unit.

• Frank King is starting restoration of the 083 Sorter. After a lot of vacuuming and grime wiping the machine still looks poorly - lots of rust all over. He is doing an initial survey
  - pocket select works for about half of the pockets
  - frequent jams - maybe every 100 cards
  - some of the roller bearings in the card path are "lumpy"