SIGNAL CABLE NEEDED FROM MY PCB TO THE DISK DRIVE
The archiver I build connects to certain signal lines on the Diablo model 31 disk drive using a flat ribbon cable which has an IDC 50 pin connector on one end (commonly used for SCSI internal connections on PCs) and a Winchester MRAC-42S or MRAC-42P connector on the other end.
The drive has both male and female 42 pin connectors on the rear. A terminator is connected to one side and the signal cable attaches to the other. Depending on the polarity of the connector you use, you need the opposite gender terminator. There is also an MRAC-14S socket that brings power to the drive, connecting to the plug at the top.
Every other wire in the ribbon cable is connected to ground, as a means of minimizing crosstalk between active signals. I started with a standard 50cm cable, two IDC 50 sockets on the ends, and removed one of the end connectors. I then split out the wires and soldered the appropriate signal wire to the associated pin of the MRAC-42 connector.
AVAILABLE CABLES AND TERMINATORS
I have four terminators (three plug and one socket type) and several connectors and cables that I saved from the time when I built an archiver to use the Diablo model 31 high density drive to read and save all the 2315 cartridges from the Xerox Palo Alto Research Center (PARC) library before they disposed of them. These were all used on their Alto systems decades before.
One challenge is that the Alto used the Diablo in non-standard ways. For one thing, they hot wired the Read Gate signal so that it was always on, emitting clock and data pulses back to the Alto. That unfortunately means that the cable from the Alto days does NOT have a wire to drive Read Gate. No wire in the cable and no socket for the signal on E of the connector. The designer sometimes cut off the terminator resistors based on the exact components they used in the Alto, ignoring impedance matching.
Another challenge that is endemic to the Diablo model 31 and 33F is all the options that exist with that drive. This means lots of variations in signals used and even in what signals had terminator resistors installed on the terminator boards. It requires me to carefully check the terminators I might use against the resistors that should be installed; I might need to modify one of them if none of the terminators match what I need.
I do have two MRAC-42S connectors with the flat ribbon cable and an IDC 40 pin connector on the other side, but I needed to first determine which pin on the IDC-40 connector was hooked to each MRAC-42S socket, then determine how to shift the signals with an adapter of some type.
QUICK ADAPTER BOARD DEVELOPED
After finding the signals on the IDC-40 connector of the existing connector I had on hand, I worked out the mapping between the IDC-40 connector pins and my IDC-50 connector pins. I used KiCAD to build a quick and dirty adapter PCB that housed two sockets - an IDC-40 and an IDC-50 - then routed the signals between them according to the map.
I shipped that off to be built and made certain that I had the IDC sockets on hand to solder this together when the parts arrived. I am waiting for quite a few shipments before I can put the archiver together completely - components from Digikey, the main circuit board from PCBWay.com, the adapter PCB, the sockets, the three power supplies, braided ground wire and parts to build the MRAC-14S power connector.
POWER CABLE CONSTRUCTION
The Diablo drive asks for four power rails to be delivered via the MRAC-14P connector on the rear. It uses +15VDC and -15VDC for everything on the drive. However, to minimize noise and the voltage drop on the wiring when heavy currents are drawn by the motor and the arm movement servo, the designers ask for a second set of +15 and -15 rails on their own wires to power the other electronics.
I have bought two high current 15V supplies and a lower current dual supply that outputs both +15 and -15V, thus can feed the four power rails over wiring to an MRAC-14S socket. I used 16ga wire, as recommended, plus a braided ground wire.
I found that the MRAC-14S sockets were quite rare and commanding a steep premium on eBay. The Winchester socket pins themselves go for up to $4 each and are almost always separated from the housings to extract maximum value for the sellers. I did find an MRAC-9S socket that had the socket pins included, where the price was cheaper than just buying the socket pins. Now I have to find a female MRAC 14 body or 3D print one if the prices remain too high on eBay.
NEXT STEPS WHILE I WAIT FOR ALL THE SHIPMENTS
I can install the best set of standard density heads into the Diablo drive and set it up for the alignment process. Aligning involves using a precious CE cartridge that has a special pattern recorded on cylinder 100 that allows me to move the arm to be centered on that track, by watching signals on an oscilloscope while turning setscrews on the arm; all this while the drive is spinning and the heads are on the almost irreplaceable CE cartridge.
Therefore I need to find my least valuable 2315 cartridge, one that I can afford to lose if anything goes wrong when I first load the heads on the Diablo drive. I will clean that carefully and then verify that the heads fly without scraping, crashing or otherwise misbehaving. I will clean that cartridge's platter (as well as the CE cartridge platter) using 91% Isopropyl Alcohol and laboratory quality low lint wipes. The heads get the same treatment.
Of course, I need to cross check the Diablo drive against the options and configuration I need for the archiver. I have to plug it to respond to Select Unit 1. It must be standard density with the 720KHz clock rate - option 001 - or be convertible to that. I have to check other options to determine whether I have an independent erase gate (option 008) as that requires me to bridge some wiring in the cable assembly or modify the drive.
Since I disregard the sector counter outputs, I don't care if this was configured for 12, 16, 24 or 32 sectors other than to validate the timing of the sector marker and index marker pulses. Other options like write protect, interrupts, attention lines, and logical address interlock are irrelevant and can be ignored. Other options cover the paint scheme and mounting brackets, also not important.
This drive was originally installed in a third party manufacturer's cabinet that provided this as a second disk drive for an IBM 1130, along with providing an adapter to use a third party line printer as an IBM 1403 printer as far as the 1130 was concerned. As such, the drive should have option 001, not have option 008, and have option 026 for standard density 8 sector configuration.
The options list are not all listed on the physical drive. Many of them were just on the sales order, which I don't have, but I can look at the circuit boards and other aspects of the drive to determine which are included or omitted.
I could even complete the alignment of the heads before I have all the shipments to build the archiver, as this process is done without any signal connector. All I would need is the power supplies and a temporary way to connect them to the drive.
No comments:
Post a Comment