• Please review our updated Terms and Rules here

RK05 disk emulator

The Datapoint 9350 is also a Diablo-30 series disk, with 24 sectors. I have the controller & drive, but no documentation beyond the command description, as found in Bitsavers.
 
The photo I was referring to is two photos below that one in the Wikipedia page, "A Nova 1200, mid-right, processed the images generated by the EMI-Scanner, the world's first commercially available CT scanner.":

I'd be interested in the Data General disk controller for that configuration.
My SWAG is that's a standard 30 Series drive interfaced through the same controller as the 6045 Cartridge Disc Subsystem, and using the same software interface. As Al Kossow pointed out earlier, DGC got a lot of mileage out of their early hardware designs, using the same "DISK CARTRIDGE CONTROL" board for both FDD and removable HDD (and the fixed side of the HDD had to work the same ...). Unfortunately surviving DGC documentation of all forms is sparse.

The HP7900A (http://www.hpmuseum.net/display_item.php?hw=275) is a similar era drive/interface that would be of interest to hobbyists. You can download documentation for it from the HP Museum, and also Bitsavers:

 
My SWAG is that's a standard 30 Series drive interfaced through the same controller as the 6045 Cartridge Disc Subsystem, and using the same software interface. As Al Kossow pointed out earlier, DGC got a lot of mileage out of their early hardware designs, using the same "DISK CARTRIDGE CONTROL" board for both FDD and removable HDD (and the fixed side of the HDD had to work the same ...). Unfortunately surviving DGC documentation of all forms is sparse.

The HP7900A (http://www.hpmuseum.net/display_item.php?hw=275) is a similar era drive/interface that would be of interest to hobbyists. You can download documentation for it from the HP Museum, and also Bitsavers:

The controller in question is not the same as the controller for the 6045.
It would almost certainly be a Data General 4046 and 4047 (the other possibility is a 4046 used with a 4049; the 4049 supported four diablo 31's, the 4047 supported two. Since there is only one drive present, I think it is more likely to be a 4047) the 4046 is the card that goes in the Nova chassis, and the 4047 occupies a separate rack-mount box. A cable would run from the backplane at the slot where the 4046 was installed, to the 4047, and another cable would connect the 4047 to the diablo 30-series drive.
Like so:
unknown.png

One (arguably dubious) benefit of such a setup, is that two novae can share a hard drive (or chain of hard drives): there were two connectors on the 4047 to connect to the Novae.
You can see the 4047 in the Wikipedia image, I've indicated it here in red:

1690435128864.png

The 4046 was unfortunately not designed to work with a floppy drive like the 6045's controller (AIUI, it could not have been, as floppy drives were not sold at the time it was designed (1971) except as a part of an IBM computer)

The 6045 was somewhat similar in concept, and had a similar (but not compatible) programming model (see pages 163 and 205 of this programming reference), but was different in construction and signalling standard. DG did decide to retain the "two novae use the same drives" capability, though in this case the other Nova connected to the other end of the drive daisy-chain. In physical construction the 6045 is much more like the Diablo 44. The 6045 controller could control a 6030 or 6031 floppy drive(s). (The 6031 has two drives where the 6030 has one)

I think the 4046 and 4047 would be more relevant here, because they speak directly to a stock Diablo 3X-series drive (possibly also Diablo 44 drives, though I have not yet confirmed this) so would likely be quite happy talking to an emulated RK05.
 
Great information @Nevets01 -- thanks for "speaking up". I wonder how many 4046 & 4047/4049 subsystems survive?
At least three.
The setup in the image (at the science museum in London), one set (4046+4047) in my care, and one other set (4046+4047) that I know of.
Mine is a recent acquisition; it's dirty but in decent shape; I hope for it to one day run again.
I would be surprised if Bruce Ray didn't have at least one set, though I don't have any hard evidence to prove it.
 
The Datapoint 9350 is also a Diablo-30 series disk, with 24 sectors. I have the controller & drive, but no documentation beyond the command description, as found in Bitsavers.
Thanks! I browsed through some of the Datapoint documents on BitSavers but wasn’t able to find specifics about the disk controller and disk format.
 
The controller in question is not the same as the controller for the 6045.
It would almost certainly be a Data General 4046 and 4047 (the other possibility is a 4046 used with a 4049; the 4049 supported four diablo 31's, the 4047 supported two. Since there is only one drive present, I think it is more likely to be a 4047) the 4046 is the card that goes in the Nova chassis, and the 4047 occupies a separate rack-mount box. A cable would run from the backplane at the slot where the 4046 was installed, to the 4047, and another cable would connect the 4047 to the diablo 30-series drive.
Like so:
unknown.png

One (arguably dubious) benefit of such a setup, is that two novae can share a hard drive (or chain of hard drives): there were two connectors on the 4047 to connect to the Novae.
You can see the 4047 in the Wikipedia image, I've indicated it here in red:

View attachment 1261418

The 4046 was unfortunately not designed to work with a floppy drive like the 6045's controller (AIUI, it could not have been, as floppy drives were not sold at the time it was designed (1971) except as a part of an IBM computer)

The 6045 was somewhat similar in concept, and had a similar (but not compatible) programming model (see pages 163 and 205 of this programming reference), but was different in construction and signalling standard. DG did decide to retain the "two novae use the same drives" capability, though in this case the other Nova connected to the other end of the drive daisy-chain. In physical construction the 6045 is much more like the Diablo 44. The 6045 controller could control a 6030 or 6031 floppy drive(s). (The 6031 has two drives where the 6030 has one)

I think the 4046 and 4047 would be more relevant here, because they speak directly to a stock Diablo 3X-series drive (possibly also Diablo 44 drives, though I have not yet confirmed this) so would likely be quite happy talking to an emulated RK05.
Wow! Thanks for sharing so many details regarding the Data General controllers. It’s exactly what’s useful to me to understand where the emulator might be practical to use.
 
Best of luck with that :-}. Key documentation for @gwiley:
Thanks much, Paul. Yes, the low-level details of the sector timing are especially interesting. I need to pair up the RK05 signals with the 4047/4049 signals to confirm the similarities.

Fascinating that what appears to be the Write Gate signal overlaps the Index mark. I need to study and probably simulate that to know how the emulator will handle it.
IMG_0180.jpeg
 
George, how goes the project? I am certainly interested in one and will gladly be a beta tester. I've done quite a bit with Xilinx, though I've been looking at giving Lattice a try. Let me know if I can be of any assistance as a tester, please. I've got an RK8-E passing diskless diagnostics ready to go!
 
I have several ARAID RK05 emulators that use Magneto Optical cartridges to emulate the RK05. I have most of the documentation for it do if you need it, please let me know. Also, if you are local to the Chicago area I can lend you one for analysis of signals.

Three of them have the old Diablo model 30 connectors on them. If anyone has RK8E cables for the Diablo Model 30 or adapter/converter boards to go from the four row pin Diablo connector to the RK05 Edge connector, please let me know.

Thanks.
 
If anyone has RK8E cables for the Diablo Model 30 or adapter/converter boards to go from the four row pin Diablo connector to the RK05 Edge connector, please let me know.
Not long ago I asked a member of the cctalk list about this and sent me a few photos of the RK11-C to Diablo 31 / RK03 connector boards.
These are at the drive end, and have the Winchester MRAC 42S and 42P on them plus the standard double-height DEC edge connector socket, and a flexprint cable goes from these to the RK11-C
One of them has the corner cut out for clearance to the other connector.
He also buzzed out the pinout. I don't think he's on VCFed but I can drop him a line and ask if he's ok with them being posted.
I have already labelled them with his name and details. I think they would be not difficult to draw up to repop by someone with more "mad KiCAD skillz" than myself.
 
George, how goes the project? I am certainly interested in one and will gladly be a beta tester. I've done quite a bit with Xilinx, though I've been looking at giving Lattice a try. Let me know if I can be of any assistance as a tester, please. I've got an RK8-E passing diskless diagnostics ready to go!
Thanks for asking. It's taking a bit longer than expected, plus there's been some recent travel that's slowed things down. The emulator and tester version of the same hardware seem to have most of the basic functionality working. Wrote a debug monitor that runs on the RPi Pico, runs on both emulator and tester, to peek and poke registers in the FPGA. Can transfer data between the microSD and DRAM and can perform disk read/write operations through the RK05 bus. All of the field durations are now adjustable with registers in the FPGA so we can adjust preambles, postambles, data length, number of sectors, etc. Also three recording rates (1.44M, 1.545M and 1.6M) are select options.

Since I don't have an RK8-E, I'd like to run a rigorous test of the drive emulator using the drive tester. Have recently debugged partial functions of the drive tester to be able to run some seek looping tests and read/write looping tests. This is what I'm spending time on now mostly. Just finished the RPi Pico SW to support this and am about to debug the automatic tests with the tester and emulator connected. Tester and emulator talk to each other using the debug monitor, but I think seeing this running in real-time will give more confidence in its robustness.

Prior to vacation, did some work on the microSD carriage which moves the microSD socket out of view when the Run/Load switch is activated. There were a number of mechanical clearance problems that seem to be solved now.
Made major adjustments to the microSD carriage (left) and designed a small microSD adapter board. A batch was built at JLCPCB and is on the way to San Diego now.
microSD_carriage.png microSD adapter.png
The microSD carriage is printed upside down, so the microSD adapter card on the right mounts on the bottom of the part in the image on the left. M2 threaded inserts are heat staked into the printed part, so the adapter card is held down to the carriage with two M2 screws. RPi handler to control the servo motor to move the carriage is partially tested but needs a little fine-tuning when all of the real parts are assembled together.

Having an issue with the OLED display driver library. The OLED display shows the drive number on the front panel. Maybe I should've just used a sticker with a digit printed on it like the real RK05 :ROFLMAO:
So, right now, if I link in the display code then the UART stops working. I have a small test build that has nothing but hello world to the UART and graphics on the display, and that works, so there's something else I need to track down.
 
It's a shame, it's unfortunate how communication has developed here, I have
to say that here too. I find it strange that many other posts appear in my thread
and my thread is no longer discussed at all. Unfortunately there is no collaboration
here either. Please start your own thread or delete all of my entries. Maybe you
should have contact the forum-administrator? I will never post anything
here again. Thanks and bye.
 
I've certainly commented in this thread often. I hope that none of it was offensive or hurtful in any way. That was not intended.
Much of the discussion in this thread has been helpful to me, and that's much appreciated.
I think it's natural for discussions to evolve over time.
In any case, I can start a new thread if group members would prefer to continue the discussion there.
 
My development https://pdp11gy.com/RK05_Emulator_E.html
has been hibernated and will probably not be further developed.
I don't plan developing a PCB board either. Mainly I lack access
to a PDP-11 with RK11 controller. All further actions are only
available via my homepage or Github with all sources. Good luck.
 
My development https://pdp11gy.com/RK05_Emulator_E.html
has been hibernated and will probably not be further developed.
Sorry to hear.

I don't plan developing a PCB board either.
If you need assistance to design a PCB and have a schematic then I'm willing to help.

Mainly I lack access
to a PDP-11 with RK11 controller.
Both the RK8-E controller for PDP-8 Omnibus and RK11D for PDP-11 seem to be difficult to find. However, I have seen two RK11D controllers on ebay in the past 6 months. I think listings are sometimes regional depending on where the seller is willing to ship. So I may see something that you wouldn't and vice versa.
 
Full schematics are available for the Plessey RK05 controller as well. I can forward them to anyone interested.

I have two sets of the boards and neither one is working right now.

One set passes the drive less test but fails the read/write test and the other one fails the Drive Less test.

Please let me know if I can help.

I can also lend someone one of my ARRAID RK05 Emulators if necessary.
 
Thanks for all the information and help offers.
I don't plan to develop a PCB board and the well-intentioned
note regarding Plessey RK05 schematics is not necessary for me.
As I already reported, my RK05 emulator project is shut down.
I don't have time for it anymore and gwiley's project is almost
finished anyway. I am now mainly concerned with a MINC system
and the ingenious software ( MINC Basic and Fortran package
running from a emulated RL02 )
 
Back
Top