Email Me Daynotes Gang

Creation of Bertha - Part 1 The first steps towards creating an integrated system for housing three separate servers in one self-contained cabinet, with power, console access, and network connectivity built in.

Today, among other things, I began the project to create the self-contained server system my wife has christened "Bertha". Bertha is the cabinet of a 1984 series Triad Point-of-Sale computer system, with remote terminals and registers, line printers, receipt printers, local hard-drive storage and tape backup, plus backup power.

Keri's uncle Jon Forsgren found the cabinet, and he and I spent the afternoon examining it and stripping it down to its most useful elements.

The only part I'm using here is the central cabinet. Granted that the terminals have an interesting, retro-futuristic look to them, they're not capable of being used as anything but cash registers, and there isn't much call for those around our house. I originally had thought to make use of the tape drive; I'd been told the original drive was a Winchester system (which used disks about 18 inches across) but had been replaced with a "modern tape system". Not knowing what modern meant, I figured that it might be useful, or might not; time would tell. Since it turns out the drive uses tapes the size of paperback books, and the bus is unrecognizable (it reminds me of an old floppy drive controller - like for a 720k disk drive - but twice the width, at least 80-wide ribbon cable.) So scratch that idea.

Here's what the system looks like at the start (note: lighting was poor and pictures hurried, so they're not perfect):

The front of the cabinet

The top drawer of the cabinet

The lower drawer of the cabinet

The cabinet has, as you can see, two drawers. The upper drawer is the top third, roughly, of the front face; the section with the black rectangle. That rectangular panel is the tape drive, about 10 inches (25 cm) across; the white line is one of the daily backup tapes, still in the drive. That drawer, shown pulles open in the middle picture, houses the tape drive, the hard drive, two solid-state cooling fans, and a power regulator for the drives. Power for the cabinet comes from a 110v AC cord through the back of the drawer and down to a power bus in the base of the cabinet. That bus (originally) pulled power from a 30A 110v circuit and split it into two internal and four external outlets, after running the current through an AC power filter.

The lower drawer (shown pulled open in the rightmost picture) contains a row of 20 individual processing boards, similar to the layout of a modern Sun-based servers. Each board is powered by and communicates with the rest of the system through a unified bus at the back of the slots. The rack is enclosed in solid sheet metal, with pierced-metal vents on the front and back; three powerful (but electrically noisy) fans and a filter at the back of the drawer pull air through for cooling. This drawer is larger than the top drawer, taking up the rest of the front, including the control panel of buttons and LEDs. Masses of ribbon cables run through the cabinet, between the upper and lower drawers and between the lower drawer and the rear of the cabinet, which houses the connector ports for the terminals and printers.

The first decision was easy - that board rack in the bottom would be ideal for housing the motherboards of the systems. So the first thing we did was to pull out all the system boards. That was easy, since they're made to be replaced. I left the controller board for the front panel in place - as Jon said, those lights and buttons have to be good for something on the finished cabinet.

Next, I decided that the fans needed to go. Yes, I'll need cooling in the system, but no, I don't need to provide it with those fans - they're big, brush-based AC fans, electrically noisy and probably physically noisy as well. They're six-inch, so finding modern solid-state fans won't be difficult. I decided that since I was going to have to replace the system board power supplies anyway, there was no point in keeping the power system for the lower drawer - out it went. Besides being heavy and unsuitable for use with the systems, it had a capacitator that frankly scared me at the thought of powering it up - the capacitator itself was the size of two pop cans. (For those who don't call it pop, I'm referring to 12 oz. soda cans, like Coke or Pepsi.) Considering the thickness of the wire insulation on the leads of the capacitator, I think it was a wise decision.

Once that was completed, we moved on to the internals of the cabinet itself. The terminal hubs are mounted in a shallow rack behind the limit of the cabinets, about 6 inches (15 cm) deep and running the full width of the cabinet. After determining that they were of no possible use to me, I pulled both of the installed hubs. The racks themselves are I think about the right size to hold the AT and ATX power supplies for the PC components; we'll have to measure to be sure. While we were working back there, we pulled out the power bus.

It's a very simple system; both the internal and external outlets are standard grounded house outlets, wired through a simple terminal block to the 30A input plug. We pulled the 30A plug; I will replace it with a normal power cord, probably simply by cutting the end off a PC power cord and wiring it in directly, then using a grommet or cap to seal the hole. Other than that, there's nothing wrong with it, especially since the power source is filtered; so long as I'm running all the "delicate" power through PC power supplies, I don't expect any problems from it.

The final step in the disassembly was the top drawer. That took the most time, though not from difficulty; simply because it had interesting things to play with. <G>

Power enters the drawer through a 110v AC cord in the back, where it splits to two locations. The first is a jumpered power block to serve two solid-state fans at the back - much more delicate than the fans in the lower drawer, probably because of the more delicate nature of this drive's contents. The second is a transformer, to create 5V DC, 24V DC, and -12V DC power and supply it to the tape and disk drives. There's also one more unfiltered 110v line to another high-quality cooling fan mounted directly to the back of the disk drive. The power line to the tape drive is interrupted by a large switch on the front panel of the drawer; evidently, the tape drive is meant to be powered off at intervals while the rest of the system is running, probably while changing tapes.

Removing the tape drive required a considerable amount of careful disassembly; the front panel had to be removed, with the tape drive attached to it, then the drive itself had to be removed from the panel. Next, we were able to remove the hard drive, and our curiosity was great enough that we disassembled it. It had a smoked-plexiglass cover, which we removed, and spent some time examining the four metallic platters. Based on the age of the drive, we were guessing that it's about a 10 MB capacity drive; we could be high or low, with no way to confirm it for certain, but at any rate it's a very large drive; 8 inches (20 cm) in height and width, and about 16 inches (40 cm) in length. There's a fairly good picture of the exposed platters of the drive here.

I decided to leave the power transformer in place; as Jon put it, I need something to power the lights. <g> I don't know about that, and it may yet be removed; but the plans for the cabinet are proceeding apace. I think there will be a further two columns to this story; next week, detailed plans and preparations for installation, and the week after, I hope to finish the cabinet. Before I can do that, however, it needs to be cleaned... a bit of enamel to replace some corrosion on the front and base... and somehow, it has to make the trip from Camano Island here to the apartment.

Tune in next week... <SEG>