Schedule November 2008
<!FONT SIZE=-1> return to main 1401 Restoration Page
go to Team Bios
<!A HREF="#05">Wed Nov 5 - General, <!A HREF="#06">Thurs Nov 6 - tape team Wed Nov 12 - General, Thurs Nov 13 - tape team Wed Nov 19 - General, Thurs Nov 20 - tape team, Sat Nov 22 - 4th Sat Happy Thankgiving :-))
Wed November 05 - general
Thursday November 06 - Tape Team
Wednesday November 12 - General
- A bunch of folks were physically present - oh golly - Ron Williams, Bob Erickson, Bill Flora, Frank King, Glenn Lea, Joe Preston, me, Robert Garner, ... - too many for my short and long term memory -
The CT 1402 card reader rollers did not "exactly" match "some other" rollers :-(( Is Grant Saviers (our ace on-call machinist) going to get another call for help?
Ya know, Joe Preston doesn't seem to move very fast - but when he is around good things keep happening - One of the CT 729 Mod II tape drives now reads and writes reliabily - assuming good tape of course ;-))
Ya know - life ain't worth livin' without a little fun!! The huge collection of documentation and small artifacts got moved "off site" and the area is now empty for future exhibits. Stan Paddock just couldn't pass up the opportunity - Especially with Karen Kroslowitz and other pretty ladies nearby ;-))
- Glenn and I double checked the second 729 Mod IV power supply. Glenn is in the process of re-installing it after the previous failure.
Thursday November 13 - Tape Teamfrom Robert Feretich, Jeff Stutzman was also present
- Jeff is back from working 70 hour weeks - for mere money :-(( - how depressing ;-))
- When we powered on the CT machine today the TAU would not reset. The RDD latch was stuck on. This is a new failure. We replaced a bad card to fix it.
- Although data transfered correctly to and from emulated tape, we were still seeing data dependent R/W Reg VRC errors, and intermittent skew errors on Write Tape instructions.
- The R/W VRC error was being caused by improper ending of the write operation. Part of the TAU mistakenly thought it was in a Read operation and latched the LRC in the R/W Reg. (Parity checking LRCs is not valid.) Replacing a card fixed the problem.
- We believe the skew errors are also being caused by improper operation ending. The Read input amplifiers should be disabled about 600 uSecs after the LRC is received. They are left on for several milliseconds. During this period, noise on the Read Bus will cause skew errors. We haven't found the root cause yet.
- We revisited the TAU CE Panel "GO Down" rate bug. I installed a diode across the the output transistor (base to emitter) of the controlling single shot. This had a minuscule effect on the pulse width of the first pulse, but it was lengthened enough so that the continuous write operations were able to be run at the maximum rate reliably (today). The initial pulse width is still about 1 millisecond and it grows to about 10 milliseconds over a 50 millisecond period. (It should always be a consistent 10 mSec at maximum rate. I tried placing a resistor from the NB card's A-pin to Gnd. This resistor made the A-pin go to -6V instead of -12V when the input transistor was turned off. The resistor made all the pulses very short (~1 millisecond). This was the opposite of the effect that I was looking for, so I removed the resistor, but left the diode in the circuit. The diode should also protect the transistor's B-E junction. The current passing through it during the initial pulses was over several hundred milliamps.
Bob Feretich filling in the new ECO number ;-)) The previous date was June 1962. Jeff Stutzman signing off as - did he say "Executive VicePresident of High HankiPank"??
- Ron Crane bought some 10V 40mAmp (ES10) indicator lamps at HSC. He modified them to attach wire leads. They work excellently in the TAU CE panel. They are as bright as the brightest existing lamps.
- Ed Thelen came in and took some pictures of the 729 Emulator and the debug effort.
Joe Feng (of the RAMAC restoration group) came to visit. We got to talking about the method RAMAC used to fly heads, about more modern flying heads, data recovery, modern methods of encoding data to make modern systems with 10-6 surface error rates act as though they have 10-12 error rates, ... I swear to you that Joe is not describing "the one that got away"!!
Wednesday November 19 - GeneralFrom Stan Paddock:
Last week, Joe Preston and I did some work on the first tape drive on the DE machine - The 10 AMP circuit breaker for the 7.5 VDC supply tripped the "FUSE" light line without popping out. - It was also very hot. We measured the 7.5 VDC current at 8.75 AMPS.
Unable to find a matching circuit breaker, we installed a panel mount fuse holder with a 10 AMP fuse. While not the best solution, it has not been a problem since.
Since the DE TAU was repaired last Thursday, it gave me a chance to try out my new tape demonstration program today. My program reads in a parameter card that defines:
The program prints out:
- The size of the record to be written
- The number of records to be written per loop (rewind at the end of each loop)
- The number of loops to be executed.
- How many total records were written
- How many write errors were detected
Try 1; 8000 records written; 526 records with errors Ron Williams suggested that I do a single loop with 8000 records to get to cleaner tape. Try 2; 8000 records written; 8 records with errors. Found a tape head cleaning pac on Allen's desk and cleaned the tape head Try 3; 8000 records 2345 written; records with errors Decided to change the density from 800 BPI to 556 BPI Try 4: 8000 records written; 4323 records with errors. . Decided to go home before I got to 100 % error records Note: I intentionally did not attempt to recover from these errors in my code. I thought the information would be more meaningful without recovery.
Bill Worthington and I tried to get the high speed rewind to function. We found the light bulb for that light source was burned out. We replaced it with the one from the "Donor" machine. It too was burned out. We took the bulb from the second machine on the CT machine and got light. We aligned the light to shine on the photo detector. After all of this, the machine still does not want to go into high speed rewind. We did not want to proceed for fear that, with my luck, the machine would not rewind at all! :>)
Ron Williams, Bob Erickson and I decided that with Thanksgiving next week, we would skip next Wednesday.
San Jose, CA
from Robert Garner
"A sitting stone gets depressed?" ;-) Since the 088's card reader roller also had similar depressions, Bill's theory is that it got this way after sitting in non-operating equipment for 15+ years. Ron W. baked it in his oven and then rounded, retaining its original diameter. It's been reinstalled, but there appears to be a signaling problem (reader picker?) between 1401 and 1402 that Joe Preston searched for at 11/26 session.
Thursday November 20 - Tape Teamfrom Robert Feretich
TAU Debug Status - 11/20/08 (Sam S., Ron, W. and I)
Allen reported problems with the German 1401, so we focused on it today.
The German 1401 was hanging on Tape Write operations. The Write, Write Delay, Go, and Delay Counter bit 2 indicators were the only indicators on.
There are three phases to a Tape Write operation, tape acceleration (indicated by Write Delay), data transfer (indicated by Write Cond), and tape deceleration (indicated by Write Disconnect Delay - WDD). (The Read operation has three similar phases.) The indicators showed that the operation hung in the acceleration phase. During this phase, the TAU counts the Delay counter to about 144, then transitions to the data transfer phase. Since the transition was not occurring, we suspected that the Delay Counter was having a problem.
Sure enough, the Delay Counter was counting 1,2,1,2,1,2..., it never reached 3. Bit 2 if the counter was getting reset prematurely. We replaced a card to fix the problem.
The next time we tried to write, the TAU would hang waiting for the first data character to be echoed from the tape drive. (All Write operations are verified, by the Write data being wrapped back through the drives read head and sent back to the 1401.) We noticed that the density settings on the 1401/729 were set to 800 CPI. The oscillator card that supported reading this density failed last year and had not yet been repaired. I "borrowed" the oscillator from a donor machine and plugged it in.
The Write operations on the German 1401 worked.
Sam and I spent the rest of the afternoon discussing the implementation of the 729 Emulator's CE tools GUI interface.
Saturday November 22 - 4th SaturdayFrom Stan Paddock:
I wanted to let Robert know that our one operational drive may not bee good enough to process the NASA tapes.
However, would you please add the fact that Bob Erickson and Ron Williams sweated all day on the CT 1402?
I cannot give you a status of the machine tho.
I don't think it is working yet but that is not due to a lack of effort of Ron, Bob and Bill.
Wednesday November 26 - General
Thursday November 27 - Tape Team
Go to December 2008
go to Team Bios