Schedule July 2007

Wed July 04 - general

Ron Williams will be there - so its an official work day ;-))

• Present were: Ron Williams, Bob Erickson, Judith Haemmerle, Ed Thelen. The building was otherwise empty.

• Ed Thelen was in Tuesday - and the newly patched Multiple Big Print demo program had a strange bug, the text that this version stored away in core for reuse on for multiple guest printouts was getting overlayed if there was more than 4000 characters of text.

  I called Ron telling him of this bug, and he said there was still a problem involving the 8,000 memory boundary. Unfortunately, this did not exactly fit my bug as I started pouring text starting 3,000 up, and got a roll-over of the index register (X3=3,999) at memory location 7,000.

• So today Ron (and a reluctant me ;-) continued to trouble shoot the 8,000 boundary problem. Ron had noted that at particular locations, moving characters down from say 8201, 8200, ... except as shown in Ron working paper, the next address was 199 instead of 8119. But at other address boundaries, the machine did just fine, as shown 8101, 8100, 8099, 8098. OK what was the machine trying to tell us??? I figured that the 8K bank and the 2 hundreds bit has a problem subtracting :-))

• Ron said "right, and I think the problem is on this page" and soon guided my hand to the gate at the right of the paper clip. :-)) - With Ron around, life is so simple :-))) Note the red lines which Ron had already drawn tracing the problem - looks easy :-|

• Well, OK - now the hard part! What is the text in the box? and what is the meaning of the text??? We tease each other who owns Ron's magnifying equipment, two magnifying glasses. But now is the time to try to make out the hazy text. (The "originals" we received weren't any better. I had taken great pains to adjust the reproducing machine for best quality working copies for us to work with.)

• The text you and I are interested in: "01" of the "01AC" means first 1401 cabinet, "A" of the "01AC" means top box, "8" of the "8E17" (it is not a "B") means 8th gate "E" of the "8E17" means column E, counting from hinge out - - (note empty A column) "17" of the "8E17" (a 7 not a 1, it slants) is card, counting down

• The left image shows the inputs, there was no change in the output of the defective card.

  This is the card we used to replace the no output gate

  And here is the good output of the replacement card.

• And Ron can now do a Storage Scan with out getting alarms :-))

  Being continuously reset by this handy tooth pick :-))

  And Tim Coslet has another (CJWF) card to fix - this one looks easy :-))

  See how easy it is???

  ---

• Taking a break from the 1401, we now turn to Bob Erickson and Judith Haemmerle working on the 513 reproducing (and summary) punch. Bob and Judith had re-assembled the gear box, and hand cranked the machine. It did not work correctly - so the gear box cover comes off - again. Behind the Geneva gear, Bob found one of the clutch springs had been damaged :-((

  I'm not sure I have the story correct, but there was a question in the re-assembly about the position of one of the cams while trying to get it all together. and something about a pivot adjustment to be done through a hole in one of the gears?

  I have no idea if/how Bob "fixed?" the spring. Ron Williams said that Orchard Supply has a good stock of springs - but the did not make a trip.

• Here is (Bob's) copy of the IBM 514 Summary Punch - the physical punch mechanism is apparently identical with our 513. Looks as though I should make a shop copy !!

  and the outside of the gear case

• E-mailed to me later - here is the "last shot of the day". The gear box cover has been reinstalled - again - and hopefully the machine will/does hand crank over just fine. That big selenoid operated the clutch (inside the gear box

  ---

• Back to the 1401 - the repair of the strange 8,000 decrement problem did not fix the newly patched Multiple Big Print demo problem :-(( Darn, not our lucky day.

  Ron said that when you add or subtract something from an X Register, the X Register has to have a word mark in the high order position, as that is the B-Operand. (I had assumed that a word mark in either A or B operands terminated the operation.) So Ron helped me experiment. I put in a little three instruction through the front panel - I won't say how long to took to get my 10 thumbs working correctly.

  After some experimentation, we found that Ron was correct :-((
  Ya know, It just ain't fair !!

  So I patched the patched Multiple Big Print demo program to place word marks in the high characters of the X-Registers - but the roll-over problem in the program did not go away.
  The program works just fine if there are less than 4,000 characters of text - including blank lines and spaces.

Wed July 11 - general

• Present were Ron Williams, Bob Erickson, Judith Haemmerle , Joe Preston, Glenn Lea, Robert Garner.
  The following report is from Ron Williams

• 1401, there is a temporary "short/open" in/near the 01B2 gate - Ron figures the cable/paddle area. The symptom is that when you run diagnostic tests and physically rotate the gate (even just a little), the diagnostic test will do unexpected things. Normal vibration and manipulating the cabling doesn't cause a problem.

• 720 Mod V Tape, sick "A" bits. The previously reported failure to stop tape seems solved.

• On the 513 reproducing punch, Bob Erickson, Judith Haemmerle, Glenn Lea unmounted the motor generator, ...

• The third KeyPunch motor is still overheating - the output stacker just above it gets noticably warm, the motor itself gets too hot to handle. Joe Preston worked on the/a friction clutch - now adjusted to minimum, he also found a bearing not too good, ...

• Ron Williams wrote a little program to test Ed's suspected wrap-around of indexing at 4,000 characters. He started with index at 3950, indexing by 10, moving data, for or over 4000 indexing boundary - The machine seems to function correctly. Ed will re-examine his program :-|

Saturday July 14 - 2nd Saturday - Volunteer day - free food

• Present were Ron Williams, Bob Erickson, Allen Palmer, Judith Haemmerle, Tim Coslet, Ed Thelen, and Rafael Skodlar who had dropped off our team list. Mattthias Goerner (on his way to study Computer Science at Berkeley in the fall) joined us for the day. Mike Cheponis joined us about 3:00.

• Allen got right to work on the 729 tape problem. This is Allen with the scope he used while fixing 729s in the old days. (A year or so ago we had to replace a transformer in it - )

• There seemed to be at least two problems involved with the tape system or interface:
  First a problem that popped up and then disappeared, was that
  - card images copied from the card reader to magnetic tape
  - then read back from tape to the print buffer
  (got that?) appeared on the printer to have lots of leading blanks

  Just how can this happen? Allen insisted the blanks were not on the tape -
  and that he was not sending leading blanks back from the tape units analog amplifiers.
  and Ron couldn't figure how he was padding leading blanks into the printer buffer.

  Then the problem vanished - poof -

• Back to the original tape problem - some translation, cross coupling, ... of the A bit to/from tape.
  - we punched up some cards with numbers and letters not containing the A bit
  - examined the resulting printouts
  As Allen left for the day, he said that if we left the tape drive in it's current condition, we would not have any trouble
  - The tape drive was in the turned off condition !!
  - Don't you think Allen deserves the teasing we give him?
  - Of course, he gives as good as he gets ;-))

• Matthais Goerner (our student from Germany) seems very interest in all aspects of computing. After he visited with other team members, I introduced him to the 026 keypunch. That guy is quick. He was interested in the printing function of the print head, and quickly deduced the general form of the layout of the print plate. A real "quick study" :-))

• Ron Williams introduced Matthais to our current worry - If gate 01B2 (see Wed July 11) is more than half way open, then the handle lifted vertically just 1 cm - any program (even a branch to itself) running in the 1401 will halt in an illegal state - no Storage Address Register (STAR) is selected and the memory fetch is invalid - causing a Machine Check.

  Of course, one can say "don't open that gate". But of course, most of us have been bitten by "Murphy's Law" (Things will fail at the most inconvienent time!! Preferably during a VIP demo or a payroll run.) We want to get to the root of this thing before it gets to us!!

  Here is Matthais Goerner going after the problem. "We" were aware that relaxing the support rods would cause the problem, but Matthais started to worry about the hinge area where the wires were routed into the gate. There is considerable stress/flexture at this hinge.

• And in the mean time !!

  Bob Erickson, Judith Haemmerle, and Rafael Skodlar were reassembling the 513 reproducing punch - again. Here is Rafael doing some microsurgical prying while Bob (almost unseen) is trying to reassemble part of the frame assembly. Note the precision steel prybar ;-)) and supporting 2x4s. Serious complexity - how about working on your car's automatic transmission?

• Allen split about 3:00 claiming a number of "Honey-Dews". But he wanted the record to show that he did indeed "log" a previous problem. He had found this little-bitty contact broken in some great big plug, and replaced the contact with one from some spare plug.

  Details added ;-))

• Mike Cheponis talked with Ed about transistor failures, types, types vs. usage, ... We found Ron Williams' logs detailing much of the above - Mike will study. Mike also proposed that SMS board testing by automated using a PC via the very common 115 kilobit/sec serial port. Mike felt that this port will be accessable for a long time - and that the interface is simple. He says all operating systems support this port in a simple manner.

• Tim Coslet talked with Ed about transistors, history, sensitivity to faults. He said that during the 1620 restoration only about 10 to 15 transistors were replaced. We wondered if the difference in our experience was the moist storage rusting the casings breaking the hermetic seal?

  Tim mentioned the unpassivated junction edges of the older construction techniques. Like the alloy type used so much in our machine had a exposed ?base/emmiter/collector? junctions that could easily get contaminated.

• Its been a long busy day - note the IBM wall clock. Bob Erickson, Judith Haemmerle, and Rafael are still struggling with the 513 in the back corner. Ron Williams and Tim Coslet (Seated) have been struggling with the 1401 and repairing SMS cards respectively. Mike Cheponis has come in. After some rowdy comments about putting contact juice on the contacts, we see our little theater. Matthais is explaining to Mike what little we know about the flexure problem. (They were soon trying jumpers to see if that would help.)

Wed July 18 - general

Wed July 25 - general

Sat July 28 - 4th Sat