Work? Logs of a Tech
Rep. on the USNS Bowditch
or
When I learned that Computers are Magic.
By John Prough, former Sperry Tech Rep.
It has been noted and commented upon by various
shipmates on the USNS Bowditch (and by extension its sister Tags ships, the
Dutton and Michelson) that the Tech Reps did not seem all that busy. It was
suggested in the short lived but interesting newsletter circulated on the
Bowditch in late 1967 into 1968, that the steward should make more noise
when checking the Tech Reps. rooms. “They could be dead and no one would
notice for a couple weeks.” This would have hurt to the quick, if I could
only stay awake enough to read the whole thing.
My personal opinion is that if a Tech Rep was not able to sleep for at least
10-12 hours a day, he better be able to find something to occupy his
time—because most of the time working on his equipment was not likely. The
General Instrument Tech Reps. seemed to spend more time working then the
rest of us. Maybe that is why there were three of them. My Sperry colleague
on the SINS also had more to do than I did on the NAVDAC. The Bunker Ramo
guy seemed to be closer to me on the work schedule.
But, like a fire extinguisher, we would be appreciated when something bad
happened. Break Glass – Wake Tech Rep. should have been clearly printed on
our doors.
I had a major concern over the meals on the Bowditch. They were too good. I
didn’t need the calories that came with those good meals. So I shifted my
waking hours to avoid breakfast and lunch. I would get up no later than 4PM,
shave/shower, and go up to dinner at 4:30PM, though I tended to think of it
as breakfast. After eating, I would take a look at the ocean, which
strangely seemed to look pretty much like it did yesterday, last week, last
month….etc. Though when the storms came, it was more interesting –
terrifying – but interesting. Particularly, when waves came crashing down on
top of the bridge – 60+ feet above the waterline. A couple of the helmsmen
were my Judo students, and I asked them if they ever saw any other ships
while standing their watches. Answer was never except when we were coming in
and out of port.
At 6PM I ran my daily Judo class in sp
ace at the bottom of hold #1. When we were in Swansea Wales in September
1967, I had asked the powers that be on the ship if they would like a Judo
class, and no one had any reason to say no. So I hopped on the train to
London, found a supplier of Judo mats (called Tatami), rented a lorry, and
brought 18 Tatami (each 1x2 meters and 2 inches thick) back to the ship.
Carried them over the side and down to the bottom of the hold. Reversed that
when the ship came into Hoboken in fall of 68.
The Judo class had a small but dedicated group, some who worked shifts would
have to wait to rotate back to having 6PM off. They all learned to take
falls very carefully since the Tatami did not hide the steel deck very much.
The class would finish at 7:30 PM. Just in time to take another shower, and
get a chair at the 8 PM movie. The movie was always a surprise. Some were
pleasant surprises and some…I never knew that Stewart Granger was such a big
star in Spain. Many were what I referred to as NATO movies. The producer was
Italian, director German, cast a bunch of unknowns from unknown countries
with the original language Greek and dubbed into English. If it was in
color, that meant no script. Couldn’t afford both color film and a
scriptwriter.
After the movie, I would hang out in the Crew’s Rec. room or visit. Around
about midnight people seemed to be sleepy. While I was wide awake. So I
retired to my room, put on my headphones, punched up some music on my tape
recorder, and read until 4 or 5 AM. Then checked into my bunk until dinner.
As we came closer to an in port, I would start getting up earlier, so that
when hit the beach I was getting up during the daylight hours.
The equipment I was there to baby sit was the NAVDAC Mark 1 Mod 0. This was
the navigation computer, which took data from the SINS (Shipboard Inertia
Navigation System) and calculated where we were every 10 seconds. It then
printed out our location on a printer next to the computer and one in the
room behind the bridge where the Oceanographers were plotting the ocean
depths. Of course where we were was Top Secret, and I certainly did not have
a need to know, so I was not allowed into the computer room except when the
NAVDAC was down. And, it hardly ever was down. So even if I had been
inclined to stand and stare at the NAVDAC’s few blinking lights in the
frigid computer room, I was not allowed. Wake me when you need me.
There were only two or three opportunities for work during the entire year I
was on the Bowditch. One was when we came out of dry dock in Malta and
during the short sea trial afterwards. When power and a/c had been returned
to the ship, the plan basically called for a through cleaning of the
computer and a reseating of all cards and test routines run to verify
everything was ok. I was staying at a hotel during the dry dock since there
was no way to stay aboard. The happy campers from Brooklyn came out about 10
days early for reasons unclear to me. I would visit the ship every 2 days to
pick up my mail and check the status of the computer room. I ran into the
leader from Brooklyn and he wanted to know where I had been. They couldn’t
find me. So I gave him the telephone number of the hotel and told him I
would be there 20 minutes after he called me. As soon as there is power and
a/c please call me. But, I couldn’t see what benefit it would be for me to
stand around and watch them sandblast the bottom of the Bowditch. This
little exchange went on a couple more times, when one day I came in and the
power and a/c was on, and so was the NAVDAC! So I immediately powered it
off, and did what needed to be done. I asked Mr Big Shot GS13 why he hadn’t
called me and why did he power up the equipment before a cleaning had been
done. He was a power tripper of the worst kind, an incompetent.
When we were on the sea trial, he jumps me about 10PM with the news that my
NAVDAC was not working, they couldn’t synch on the Loran slave station.
Wanted to know where I was that he couldn’t find me. Asked him if he had
bothered to look in the movie. But, that never crossed his mind. So I went
down and ran diagnostics and nothing seemed to be wrong. Then the thought
occurred to me as to which slave they were trying to reach. Went up and
talked to the guy testing the Loran and he said the station in Turkey since
we were too far east in the Med to get the Spanish slave. So I told him to
wait until we get back to Malta and he should be able to reach the Spanish
station since the NAVDAC isn’t programmed for Turkey. Told Mr. Big Shot and
he demanded that I immediately program in Turkey. So I told him that the
last time he was on the ship he took the secret number back to Brooklyn with
him since we didn’t need it because we never would use the Turkey station.
As predicted, when they got back near Malta, they easily tested the Loran
system using the Spanish slave station.
The one other time I worked that I can remember started on Christmas Eve
1967. One of the NAVDAC ET’s was standing watch in the computer room.
Feeling sad about spending the holidays away from family and friends, he
decided to do some PM (Preventive Maintenance). Being of the sleeping dog
school, I would have recommended against this commendable initiative. He
started by removing the covers on the NAVDAC so that he could vacuum up any
dirt or dust.
Now the NAVDAC on the Bowditch was the prototype Mark 1 Mod 0 or something
like that, I cannot quite remember. There were only five of them made. There
was one on each of the TAGS plus one at Sperry and one on the USS Compass
Island. The package design was different from the production models that
went on the Subs. Originally the Mark 2 on the early boats such as the USS
George Washington. And on the later Polaris Boats like the USS Will Rogers
it was the NAVDAC Mark 4. I spent an “interesting” sea trial on the USS Will
Rogers, but that is a different story. The Mark 2 and Mark 4 were the
typical 19 inch rack mounted equipment. The Mark 1 was much more horizontal
in its layout. See the very rough sketch. In fact I went aboard the USS
Compass Island to see what it looked like. I was shown into the room by one
of the senior Sperry reps. I asked him where this Mark 1 was, and he said
“you are leaning on it”. And sure enough, I was. I had seen the other
NAVDACs at Electric Boat. This didn’t look much like its children.
But to get back to Christmas Eve in the Med., as the ET was cleaning up the
computer, the Oceanographers called down from the bridge and asked “what’s
going on down there?” Told him to look at the printer. The Printer was
printing out gibberish rather than our location. Looking back it had been
working fine and then a little gibberish and then total gibberish.
So the ET decided again on his own while I was catching some sleep, to swap out the entire rack where the Re-circulate command logic was located. This resulted in a new problem, the half-adder was no longer adding. This meant that the computer would not run since it had no way to find the next instruction. The main memory was a Drum Memory so the next instruction address was 4 down from the current location. Could not add 4 to the current address because there was no half-adder. About this time we went off site and I got to look at the problem first hand. Didn’t help much, we couldn’t get it to run at all.
After a couple days of trying various things, I concluded that it had to be a timing problem that I couldn’t see on the scope. Drum memories are notorious for being very sensitive to the clock signal. If it was off just a little, things tended to come to a swift halt. But the NAVDAC was less sensitive than say the Packard Bell 256, a computer which was used at the Sub-base for simulators. The memory on the PB256 was a acoustic wire. This was a wire wound inside a printed circuit board case. The boards were big, around 15inches on a side as I remember. A little relay would hammer the wire at one end and a microphone at the other end would hear the hammered “1” about 50 milliseconds later. It worked like a drum memory without the moving parts. But it was very sensitive about its clock signal. If its timing was just a little off; no PB256. They had not bitten the bullet and paid Dr. Wang to use his invention; Core Memory.
BTW we were working on this all through the end of 1967 Barcelona in port. The EMO had told me not to kill myself by working day and night on this, but to get some beach time. I thank him for his consideration. Particularly, since we were not making much progress.
The prints had stated that the resistor that form the RC decay for the NAVDAC’s timing card was suppose be something like 64K ohms except for NAVDAC s/n 3 it was suppose to be 81K ohms. Well we were s/n 3 so he had rewired the timing card with 81K. I decided to wire in a resistor decade box and start stepping the resistance until we found the magic number where the half-adder worked. Just then we had sailed and were back on site, so I was kicked out. I told the ET that they know what to do, let me know when they found it. I went back to my room to contemplate the problem. After about an hour they came in and told me they had the half-adder working. The documentation lied, s/n 3 wanted 64K ohms like everyone else.
But, now we were back to the first problem of the creative Re-circulate command. It was now about 2 AM. I did not have any more idea of what to do to fix that problem. So I told the ET’s to clean every thing up, put it back together and hit the rack for some sleep since we were all tired. We would try it again at Noon. Following my own advice I went to bed. About half hour later they woke me up with the news that they had fixed it. So I immediately asked how did they do that. They put the covers back on!
Now the NAVDAC Mark 1 had two horizontal cabinets with three (as I remember it) racks of logic cards etc in each case. The bottom rack of one case had the drum memory spinning in it. It also had a slightly tilted top, which was the control panel for the computer. The control panel was a few switches and some lights representing various registers. The bottom of the other case had the power supply. This computer was an early 2nd generation design. By that I mean the first generation computers used vacuum tubes for the logic circuits. The second generation used transistors. These were great big transistors that you could see and touch and replace when necessary. The third generation started the movement into integrate circuits, and so it went. So this was an early 2nd generation in that not all the vacuum tubes had been replaced. Power transistors were a little slow in the development curve compared to their lower current cousins used for the logic cards. As a result they still used a tube to supply that clean dc voltages for the computer. To make sure that this tube did not over heat the computer, they had installed a fan in the cover for the power supply rack. This fan blew room air (a chilly 68 F) across the tube where it vented through a grill on the opposite cover.
The covers on this computer each had about 47 screws holding it to the frame of the computer. I asked someone why so many screws when maybe 12 would certainly hold the covers on the frame without any problems since there was no load on the covers except for their own weight. His reply was that they had that many screws to keep it “Navy Proof”. A dozen screws would be too tempting for some sailor to see what is in there. Whereas having 47 screws it was just too much trouble to bother with. Well, Christmas Eve had increased the design criteria. He had taken off all the covers and this started the Re-circulate command on its road to error. When they put them all back on, the problem went away.
So we had them take them off one at a time. When we removed the cover with the fan, the Re-circulate command problem. Now it was more than just taking the cover off, you also had to unplug the fan. The power cord to the fan had a quick disconnect plug a few inches from the fan. The power went from that plug straight out of the cabinet to a wall plug of 110vac. Now I know what you are thinking. It was a heat problem all along. But the cover with the fan running could be laying on the deck blowing air into the tile and the computer would be happy and do its little position calculations correctly. Unplug the fan and the Re-circulate went screwy in the other cabinet!
That is when my revelation came to me: Computers Are Magic! There was no physical or logical connection between that fan and the Re-circulate command logic. Next time the incense and holy water or whatever other hocus pocus was necessary. I have slept soundly at night knowing this for the almost 40 years of working in computers since then.
I wrote a report back to Sperry to warn the other TAGS not to disconnect the fan in the NAVDAC. Total time spent working on this about 12 days. Of course some of them counted double since I had to work during the in port.
