Ken's Alto Ethernet device
Ken discovered that his Ethernet device does not work properly when he uses long ethernet cables. He checked the signals on the scope and saw a couple of problems with the signals he produced when the arrived on chip inside the Alto.
First, the low level voltage was lifted above ground a noticeable amount. Second the transitions had some obvious distortions that had the characteristics of a improperly terminated transmission line. After some experimentation, a resistor was added on the device side that cleaned up the signals.
With that change any length of ethernet cable could be used with the device. Next up is testing over the coming week of the ability to link up Altos over the Internet, joining machines at LCM in Seattle and one in Connecticut.
HP 264x terminal character sets
In prior sessions we worked on the optional character sets such as Math symbols or Line Drawing, which for most of the terminals require ROMs on an added board, the Display Enhancement PCA. However, a few terminals come with all of these standard.
We accomplished two things. First, we located the ROMs on the main boards which support the Roman character set which is the basic standard set on all terminals. Second, we located the ROM on the newer terminals which have all the sets included.
The basic characters are implemented on two 1K x 8 ROMs, providing a full 128 characters, upper and lower case. The newer inclusive ROM is a single 8K x 8 device hosting the basic, Math, Large Character and Line Drawing sets, plus several diagnostic image sets. We identified a compatible ROM for the large one, the Harris 76641, and the other ROMs are Harris 7681 or equivalents.
Diablo disk drive head alignment
We spun up the repaired alignment cartridges, with the heads installed in the drive and the head loading solenoid reattached. The heads and solenoid were out while we spun the repaired packs, making sure that any residual tapping oil was gone.
Right after the repair, the oil would streak out from the hub under centrifugal force and migrate out towards the edges of the platter. That would cause head crashes, so we cycled the packs through long intervals of rotation and thorough cleaning. Finally they were ready.
The heads loaded properly and safely onto the surface and the oscilloscope probes were monitoring the signals from the heads, in preparation for alignment. I released the rotary actuator and moved the heads out to cylinder 105, where the special pattern is recorded.
The head loading solenoid toggled in and out a couple of times, so we immediately switched off the drive. No damage but we had to inspect and repair whatever was causing the behavior. We found that one of the contacts from the loading solenoid was loose and repaired it with epoxy.
We were out of time but are now ready to do the actual alignment next week, tightening a setscrew that minutely pushes the head outwards toward the platter hub. The initial location of the heads is about five cylinders too close to the platter rim, but when we get close will see a well defined pattern on the scope. Once we are at the correct location, that pattern will be symmetric on the scope.
HP 1000 SYSTEM RESTORATION
Working with RTE IV B
I put in a couple of hours improving my RTE system, using the simulator for convenience. I updated the startup script (a file called WELCOM) to handle some recommended tasks:
- Compacting the system cartridge (002)
- Registering overlays for large programs such as FTN4 and RT4GN
- Implementing an improved timesetting function supporting the current year
Each such task teaches me a bit more about RTE and the HP environment. Fortunately it is of manageable total size, although big enough to require months to master not just days.