Sunday, August 17, 2014

Starting restoration with cleaning and altering the voltage to match my home power

A few of the maintenance documents, particularly the log books that sat inside the machines, had developed a bit of mildew on the covers. As well the back and side covers had a bit of surface mildew to remove. I discovered through a bit of trial and error that Lysol disinfectant wipes are mild enough to cause no damage to the surfaces but remove the mildew and kill any spores left behind.

I also discovered that one of the panels in the footwell area and under the console keyboard and printer had a sticky oily film on it. I used some pure isopropyl alcohol to cut and remove the oils. Thanks to the excellent advice from David Cortesi I am doing well ridding the machine of disintegrated or crumbling air filters and seals. Lots more to do but moving forward on the cleanup.

For 60Hz operation, the 1130 can be set up for 115V, 208V or 230V. The machine was configured to run with 208V power that is readily obtained from three-phase electrical service in a commercial building, but is not an option at my house where the electrical service is single phase. Thus, I have to rewire the machine for 230V power.

This is not something that is set with a switch or in just one place. The 1130 has at least four power transformers involved (that I have discovered so far), each has multiple taps to support the various power options. Terminal blocks throughout the power supply area have spots where I have to change jumpers between terminals or move a primary winding lead from one spot to another.

Many ALD pages detailing the power have text boxes listing jumpers and terminal numbers, however IBM reuses the names over and over and over. Thus, the main wiring block for the first two transformers is TB-1, but then there are TB-1 strips on each voltage regulator and in other spots. Inside the fuse panel is a transformer and TB-7 to adjust. Mixed in with the voltage regulators is a fourth transformer that delivers voltage for the lamps, with yet another TB-1 block.

The two main transformers produce the raw voltages for +3, -3, +6, +12, +48, plus 7.8AC and 24V AC. These are connected to TB-2 which has wiring running to the various voltage regulators to make the voltage filtered and controlled.  This allows me to disconnect the regulators at TB2, so I can bring up only the raw voltage supply.

The raw supply rectifies the power to create the DC levels and has giant electrolytic capacitors to filter out spikes, hum and noise. These have to be very carefully handled when powering up electronics after years of disuse. It is pretty common to have someone apply power then hear small explosions as the capacitors blow open in their death throes.

Electrolytic capacitors are a sandwich of two metal surfaces with a chemical paste, the electrolyte, separating them. To get the high values of capacitance needed, the metal plates must be very close together yet not short circuit. The paste is slightly mushy in normal circumstances but without use for decades, it dries out.

The result is small pinholes in the paste, since it shrinks a bit as it dries. These allow the plates to form a small short circuit. If you apply full power to one of these capacitors, the tiny short circuits conduct heavily, heat up, create gas in the capacitor and burn away more paste, widening the hole. This progression can occur quickly, resulting in a high gas pressure causing the sealed can to rupture and the capacitor to become a full short circuit, blowing the fuses in the unit.

It is believed that certain procedures will help the capacitor 're-form' the electrolytic separation and avoid catastrophic failure. There are three basic methods, each of which limits the amount of current that will flow into the capacitor, so that any small short circuits won't grow in size and produce gas. With limited power, the pinholes get just hot enough to melt the paste nearby which then flows evenly, filling in the hole. The hole disappears - restoring the capacitor to its intended behavior.

One can use a Variac to adjust the input power, starting at something low like 20V and gradually increasing the voltage in steps with long periods running at each level. The larger holes are heated enough to repair themselves at the lowest level, then progressively smaller flaws create enough local heat to melt the paste as the steps reach higher voltages.

Another method places resistors in series with the capacitor, reducing the maximum flow of current to control the heat of the small short circuits inside the can. This method allows easy direct measurement of the voltage drop on the resistor, thus the current flowing into the capacitor. To obtain the benefits of progressively healing from larger to smaller flaws in the paste, the resistor should be changed to lower values in steps.

A third method sometimes used is to add an incandescent light in series with the capacitor - providing resistance - and to step up from lower to higher wattage bulbs over time as a way to manage healing of different defect sizes. There are variations in behavior of this method, since the filament of the bulb changes resistance quite a bit from its dark, cold condition to its lit state - not a steady limited current as with the other two methods. Also, the 1130 supply has capacitors hooked to a variety of voltages from 3V up to 48V, which may require more types of bulbs on hand to handle all the voltages and steps of current.

Once the major capacitors are reformed, or replaced if they don't heal adequately, I should have reliable unregulated power. The regulators could be reattached one by one and checked  out.

No comments:

Post a Comment