Friday, August 22, 2014

First power up of the 1130, in excellent shape

As of yesterday, I had determined that the main circuit breaker in the 1131 was dead - wouldn't connect the power at all. I can buy a replacement circuit breaker, used, through both Amazon and eBay, but it won't arrive for about a week. I was eager to press on with testing before that time. I first looked at the circuit breaker feeding the outlet into which I am plugging the 1131, but it was 30A not 20A.

I spent some time looking for a quick and inexpensive breaker that I could hook to the lugs in place of the bad one, even if the breaker had to hang outside the enclosure, until it came to me that I could buy a 20A breaker for my home service panel, turning the outlet into a 20A protected circuit. That would allow me to swap breakers in my panel, bypass the breaker in the 1131, and get the same protection during testing. I picked it up at a local home improvement store and was back to testing power by the afternoon.

bypassing bad circuit breaker
With the bad main breaker bypassed and the proper protective breaker in my home power panel, I started through the power testing steps again. I applied the power, verified it was delivered to the right places without any shorts, then put back the first fuse, to enable the 24VAC sequencing power. That came on but the power toggle switch on the 1131 keyboard didn't do anything when flipped.

Working with sequence box on right and power supplies on left
This, however, made sense because I read a note on one of the power ALD pages about having to jumper some EPO lines if the SAC features was installed. I do have that feature installed. SAC is used to hook up the IBM 1133 Multiplexor box, a large frame the same size as a S/360 channel or 2821 control unit, but also used for non-IBM peripheral attachments. EPO is the emergency power off circuit, attached to the infamous Emergency Pull round red switches on 1131 and S/360. The SAC puts the 1131 EPO switch in series with any EPO switches connected through the SAC.

I had to stick a jumper between specific pins of the SAC feature power receptacle (pins 23 and 28, to share some irrelevant detail), after which the Power toggle switch on the 1131 energized the contactor (a main power relay). I still had most fuses and internal circuit breakers off, but at this stage power will go to all the blowers (fans) under the logic gates and in the power supplies.
Power switch left upper position, turned off in this picture
The sequencing logic of the 1130 applies power in a specific order of voltages while a 5 second timer runs, then it verifies the presence of the key voltages (+3, -3, and +6). If those three are not all present and at proper levels, all the power is dropped and the power toggle switch locked out until reset by a CE switch. That is what occured, but once I reset the lockout with the CE switch, I turned on power for only a couple of seconds at a time while I went around checking that all the blowers were working.

Next, I inserted fuses for the 'convenience outlets' which deliver 115 VAC,  typically used to power oscilloscopes and other CE equipment. That worked fine. I put in the fuse for the 7+VAC lighting power supply and verified its proper output.

More fuses and checks - the regulated supplies for 12VDC and 48VDC were working, as were the raw voltages that will feed the +3, -3 and +6 regulators in the midpack area. Last up was to check the voltages coming out of the three regulators. I didn't want to be powering my logic cards yet, so I left the circuit breakers on the regulators off except for the regulator under test. Each of these had to be tested in less than 5 seconds since the sequencer would then drop power. All three regulated supplies were right on spec.

Power supply regulators and the lamp power supply in upper left
At this point, all the power circuits are operational. Next up was attachment of an oscilloscope to validate proper filtering by the electrolytic capacitors. Ripple was almost nonexistent on the three main supplies, with their midpack regulators, and about 50mv on the ferroresonant 'regulated' 48V, 7.8V and 12V lines.

I will flip all the breakers and try to bring up the machine itself as soon as I am satisfied that all cards are properly seated, all connectors are attached, that new air filters are installed, and finally that the compartment doors with new foam are closed. Time to carefully install filters and other foam/neoprene.

I removed some sponge rubber like strips that were placed on metal edges to protect the insulation of wiring that could rub against the edge, but with the disintegrating materials used by IBM, these had to go. I bought some adhesive backed neoprene weatherseal strips and applied those in a few places to restore the protection for the wires. I have more edges to protect, a task for another day.

My 1/4" neoprene roll arrived and it seemed perfect to line the covers that hold SLT cards into place. Although a bit stiffer than the spongy original IBM coating, these are not liquefying. The covers on the compartments on the A and B gates are a full neoprene rectangle, but the two compartments on D gate hold core storage which is a stack of core surrounded by a U area of SLT cards. The covers are 'domed' to leave space over the core, but had the foam over the SLT areas. It took a bit of cutting and trimming but I have everything sealed up nicely.
Sample of neoprene to be used inside SLT card cage compartments
I cut out a fiberglass air filter from a larger piece sold by Home Depot, placing that over the SLT card cooling inlet for the cards in the internal disk drive. There is a very high performance filter that I still have to remove and locate a suitable replacement - this keeps large particles from flowing into the drive and over the platter while the heads are on the disk - even very tiny dust grains could cause a head crash.

Fiberglass air filter in place inside duct
I began to smell something from the disk drive area that I associate with overheating electrical components. The blower motor is spinning and putting out air, but perhaps it has dried lubrication that kept it stalled initially. The spindle motor isn't turning, which it shouldn't until a cartridge is inserted, but that too might have a problem with dried lubrication.

The 1053 printer (console typewriter) can't rotate its motor due to the solidified greases throughout the typewriter, but I am lucky to have the local modification done by the prior owner. It is a switch that will turn off the typewriter motor independently of the rest of the system - very handy in this situation. Wish I had one like that for the disk drive.

Printer motor just visible behind mechanisms in upper part of picture
I still need to find a good foam or neoprene to replace the sound deadening junk inside the disk drive and in other places in the 1130 system. It should have its own adhesive backing and that should have good sticking power. Even without the sound foam and some suitable material for the blower air inlet filters, I was ready to deliver power to the SLT cards and see if this beast would come alive.

I flipped the circuit breakers for all the power supplies on, then flipped the Power toggle switch. The system came up to a normal starting state and the only error light visible was the Form Check indicator which means no paper in the console printer. The processor was sitting at T7 clock step of the I1 major cycle, which is where it will fetch an instruction as soon as Prog Start is pushed.

System up in normal state
I put the machine in SI mode (Single Instruction at a time) and pushed start. It fetched a word from core memory, with no parity error! The word was a valid format, an MDX -1 which is a one word branch to itself. Whether stepping through each T clock in SS mode or running the entire instruction in SI mode, it did what it was supposed to do, branching to itself.

This is a normal content of location zero, when running the Disk Monitor System 2, put there in case a user does a branch to location zero due to a programming flaw. I don't know how much of the core storage still retains valid contents but at least the first few locations were valid when I displayed them in DISP mode.

My lamp test button showed about seven or eight bulbs did not light in the display panel; fortunately I have spare bulbs in the likely case that the bulb itself is bad and not a wiring or logic issue. I will move on to toggle in some diagnostic routines to check out the system a bit more.

Here is a video of the first power up Video of the first powerup of 1130 system

Late last night, I spent some time digging through some of the boxes of software on punched cards, identifying what I could and moving the most interesting into the Wrightline cabinet. I found many utilities, complex variable subroutines, some application programs (e.g. Electrical Circuit Analysis Program, City Street Planner), and a few games like Moonflight and Blackjack.

I know I have the main compilers - Fortran, RPG and Cobol - even though I haven't gotten to them in the search yet, but hope that Student Language I (a PL/I subset) and APL might be in the pile too. Icing on the cake would be finding FORTH, but it does not appear that this was ever distributed by IBM or COMMON which were the sources for my decks.

No comments:

Post a Comment