• Please review our updated Terms and Rules here

minimal F-11 SBC ?

saipan59

Experienced Member
Joined
Dec 27, 2010
Messages
259
Location
Colorado Springs
In the past, I've hand-built a couple of boards using a T-11 with local RAM, EEPROM, and basic I/O.
Now I'm thinking about doing the same thing with an F-11 chip.
Has anyone done this before? Is it feasible to get the basic stuff down to just a handful of support chips?
Is there enough published info to work out the details, other than just copying what's on a KDF11-A board? The KDF11 has a boat-load of stuff on it, but hopefully most of it is to support Qbus, memory management, and other stuff that I don't care about.

Any comments welcome!
Thanks,
Pete
 
There have been minimal implementations with the J11, you might look at some of those. I've never looked at the bare CPUs in enough detail to know what'd be involved with the F11.
 
There have been minimal implementations with the J11, you might look at some of those. I've never looked at the bare CPUs in enough detail to know what'd be involved with the F11.

Yes, thanks, I've seen one of those online. But I no longer have any J-11 chips to play with, but have a couple of F-11's.
There's some fairly good detail in the KDF11-AA User Guide - it seems like a LOT of the logic would be in the "don't care about that" category, but it still could require a bunch of experimenting to make it go.

Another thought I had: It might be useful to implement a bus device on a little card, and use a 40-pin DIP cable to plug into one of the empty sockets on a KDF11 board. That way, all the known-good support logic would still be there, but the Qbus would be inactive. The 'bus device' could be the console UART for starters.
Which reminds me I should look at how the 11/24 CPU board works - I think it has 2 serial ports on-board with F-11 - perhaps they are directly connected, and not an actual Unibus device??

Pete
 
I suggest that you get a DCJ11, currently on sale for 65USD. I have three sets of the F11 chips and there is quite some information but when I looked into it I found it is not enough to build your own system. I was always missing the subtle details in the documents I found. A standalone F11 CPU system would have no advantage over a T11 based system. Even more I think a T11 base system is perfect if you are going to build a PDP-11 without MMU. And why not start with a KDF11-A board? You could then start to build peripherals and memory cards.

That said, I would again start with a DCJ11. This CPU has everything integrated and the support logic required is very minimal. Here is the Original PDP-11/Hack which also inspired my projects.

Peter
 
Thanks for your comments Peter. Just last month, I sold two J-11 chips, and a KDJ11 board, on eBay. I don't think I will buy another one... :-| .
I will study the KDF11 and 11/24 boards a little more, and perhaps do some experiments.
But of course, another T-11 project would be much more "practical". It's nice that it will run with 8-bit-wide memory, and only about 3 74LS chips required at a minimum. Here are a couple of pictures of a clock I made a few years ago. It is wire-wrapped, and the displays are from some old HP test equipment (after DEC, I eventually worked for Compaq, then HP). The time base is a 3.579545 Mhz crystal from a very early color TV (it looks likes a vacuum tube, at the top of the picture). Clock osc/divider is an MM5369. The T-11 runs a hack of FigForth V1.3.

Pete
T-11_clock.jpg
t11_clk.jpg
 
I would say bad timing. The T11 however looks like a very flexible CPU and I doubt The F11 would be equally simple. So don‘t sell your T11s. I like your small T11 SBCs. Perhaps I should build a clock with one of my J11 chips.
 
I remember reading that DEC and WesternDigital worked together on the F-11 chipset - and it was the same as the MCP-1600???

If this is true (and I can't confirm this - but a very quick random sample of a couple of instructions does give some credence to this statement) then some very useful documents can be found at http://www.bitsavers.org/pdf/westernDigital/MCP-1600/.

Just remember that these documents appear to describe the macro instruction set in terms of HEXADECIMAL whilst DEC describe the macro instruction set in terms of OCTAL. That threw me for a few seconds...

EDIT1: See https://en.wikichip.org/wiki/dec/lsi-11

EDIT2: Page 17 of http://www.bitsavers.org/pdf/wester..._Microprogrammed_NMOS-LSI_Processor_Oct74.pdf confirms this...

Dave
 
Last edited:
Thanks Dave!
However, I'm thinking that those docs are describing the LSI-11, not the F-11. I just checked, and some of the LSI-11 chips actually have the 1600-series part number on them.
Am I confused? Or, is the assertion that the F-11 is also a WD design using the same architecture?
I've got some F-11's that have an "AMI" brand marking (some just say "DEC").

Pete
 
>>> I thought the LSI was the MCP-1600 but the F11 was DECs own design?

Pass... I may have been meaning the LSI as opposed to the F-11 of course...

>>> Thanks for the reference to the Soha & Pohlman paper - looks like the next step after the LD12. :>)

Not likely :)!

I have a few days off (Mondays and Fridays) so I am working on the LD12 a bit more (in between real work of course)...

Incidentally, do you know what size of jack socket the 'logic probe' connector is? I am assuming a 'mono 1/4"' from the photographs? This is the last part of the puzzle for the front panel - although I have to re-lay the FP out (yet again) as I require a bit more space.

Dave
 
A bit more investigation and I can confirm you are correct.

The F-11 chip is the "11/23" which is a DEC part.

The LSI-11 is the DEC/WD MCP-1600 parts.

Sorry for the confusion :).

Dave
 
I thought the LSI was the MCP-1600 but the F11 was DECs own design?

Correct. And also, the MCP-1600 cannot, out of the box, run PDP-11 programs. For that you need the DEC microcode chip. It's just that the LSI-11 was based on the MCP-1600, with DEC microcode.

See https://en.wikipedia.org/wiki/PDP-11#LSI-11

The LSI-11 was used in the PDP-11/03.
The F-11 was used in the 11/23 and 11/24, the Pro-325 and Pro-350, as well as various embedded stuff.
 
Correct. And also, the MCP-1600 cannot, out of the box, run PDP-11 programs. For that you need the DEC microcode chip. It's just that the LSI-11 was based on the MCP-1600, with DEC microcode.

See https://en.wikipedia.org/wiki/PDP-11#LSI-11

The LSI-11 was used in the PDP-11/03.
The F-11 was used in the 11/23 and 11/24, the Pro-325 and Pro-350, as well as various embedded stuff.

But the microcode was in standard MCP 1600 microcode ROMs, so you could actually read them and disassemble. would be interesting to have a look how they implemented a PDP-11 in MCP-1600 microinstructions. I will ask a collegue who got two LSI11/02 to borrow one so I can copy the ROM.
 
But the microcode was in standard MCP 1600 microcode ROMs, so you could actually read them and disassemble. would be interesting to have a look how they implemented a PDP-11 in MCP-1600 microinstructions. I will ask a collegue who got two LSI11/02 to borrow one so I can copy the ROM.

Eric went through that exercise with the LSI-11 microcode 4 year ago. One complication is that there are application specific translation PLAs in the control chip which at the time had not yet been dumped. Just the microcode alone is not sufficient for complete knowledge of the implementation.

Microcode, DEC LSI-11
https://github.com/brouhaha/lsi11uc

Disassembler for Western Digital CP16xx/WD21xx Microcode
https://github.com/brouhaha/cp16dis
 
Back
Top