Can the unibone interface directly to a physical disk (ex RL02)?

[John Hudak]

As I understand it, part of the magic of the unibone is that it emulates a disk controller (at the backplane signal level), for example an RLV11, which is coupled with a physical disk file structure emulation, e.g. a RL02.

If this is correct, is it possible to have the unibone physically interface to a physical disk? (e.g. via the 50 pin Berg connector).  The unibone effectively becomes the disk controller, e.g. RLV11, or whatever.

I have the situation where I have a disk controller and a disk drive, both in an unknown operational state, and hoping that the unibone could emulate the controller, the physical disk, to help in my troubleshooting efforts. Using a logic analyzer on the controller board and/or the physical disk electronics to hunt down a dead component is a bit tedious.

Having said that, anyone know of a program to exercise a controller board to help with the troubleshooting?  When DEC QC'ed these things they had to use a test jig of some sort - guess I am thinking of recreating this somehow.

Thanks

J

[Jay Jaeger]

Sorry, but no, it doesn't, and it could not without additional hardware.  The RL drive bus uses differential rather than single ended signalling.  To make it work you would need some kind of daughter board - and that assumes that enough BBB GPIO pins are presently unused.

There is someone who has done the reverse - has a hardware drive emulator so that might help you debug your controller.  Also, you could load XXDP from your Unibone emulating a different disk (RK05, for example) and use that to do tests on your controller.

https://www.pdp11gy.com/rlfertig/Readme.pdf

JRJ

[Joerg Hoppe]

Hi,

On Wednesday, 29 January 2025 at 15:49:24 UTC+1 jjh...@gmail.com wrote:

If this is correct, is it possible to have the unibone physically interface to a physical disk? (e.g. via the 50 pin Berg connector).  The unibone effectively becomes the disk controller, e.g. RLV11, or whatever.

As said, the current Unibone is a pure software emulator, and can not interface a high-speed disk interface parallal to UNIBUS.

However UniBone has 56 (+ some more) bidirectional signal linesm, which can be operated on in real time.

With that you can  build a stand-alone drive exerciser with the existing technology (DEC made these too).

- You remove the UNIBUS interface (plugout the 8641 driver chips) and wire your own 50 pin (or so) connector.
Maybe you need some differential transmitters ... then best tinker the KiCad project and make an own PCB.

- you would discard most of the Arm-C++ and PRU-C code and use the existing software as template for own realtime/user interface stuff.

With Unibone as template there should be no development risk ... just woooooork.

kind regards,

Joerg

[John Hudak]

Hello:

Well that is what I was thinking, using the GPIO pins through a  converter box to create the differential signals needed by the physical drive. An early step would be to study the code to see where a separation could occur between the controller emulation and the file structure emulation.  One would need to create the FSM for the drive control and code it and then develop the interface board to the drive.  Alternatively, use a FPGA to execute the FSM.   Yea, either way, it would take some  work and a whole lot of understanding.

Thanks all for the clarification.

J