Looking for info on memory

[Chris Zach]

It can be done, technically the 11/83 and 11/93 put the CPU in slot 3 
with the PMI memories in slots 1 and 2. Putting the CPU in slot one with 
the PMI memory in 2 and 3 would just cause all memory access to slug 
across the Q Bus.

So you can leave slots 1 and 2 open if you really want to, makes it 
easier to deal with possible PMI screwage (I always wondered if one 
could put a CPU in slot 1, then the M8013/8014 RL01 controller in slots 
2/3 but I don't remember if the 813/8014 will work in a 22 bit unit 
without catching fire or something).

Anyway, just try running the CPU in slot 3, then the memory in slot 4, 
and the DLV11J in the right side of slot 5. That would simplify things 
as much as possible.

C

* PS: Right now I'm running my 11/83 in slot 1 with a 2mb PMI memory 
board in slot 2 and the TK70 controller in slot 3. The reason is that 
the PMI memory board is from an 11/84, and it has the bug that prevents 
it from working properly in a true Q Bus system.

I *think* the bug is related to concurrent memory access from both the Q 
Bus DMA and PMI bus as just using the memory as a Q Bus memory has no 
problems. A true 11/84 works because there is nothing on the Q-Bus as 
all memory access is done through the Unibus bridge thingie which only 
does memory transfers on the PMI bus (it's not a true Quniverter for DMA 
purposes).

I also *think* the bug is related to burst mode DMA as I have run my 
ESDI controller in normal DMA mode (slower) with the PMI memory running 
as PMI without problems. However if I turn on burst/hog mode DMA it will 
crash RSX11 (but not RT11). Go figure.



On 10/19/2021 8:50 PM, Nigel Johnson Ham via cctalk wrote:
> In all the hundreds of systems I have shipped, I have never put the CPU
> in slot 3! I was VP Operations of Transduction, DEC's biggest components
> group OEM in Canada.
> 
> No, I don't know if the CPU is good (it is a dual-width KDF11-A btw),
> but it is in slot 1 AB with a DLV11J in slot 2AB.  The memory has been
> in slot 3 as well as lower.  Since there is nothing above or below it on
> the CD side there cannot be any interference in the CD connectors.
> 
> I do have an 11/73 dual width board arriving tomorrow so will see if I
> get the same results with it.
> 
> Thanks for reading,
> 
> Nigel
> 
> 
> On 2021-10-19 7:55 p.m., Chris Zach via cctalk wrote:
>> Hm, ok all my Plessy boards are dual height, 1mb boards. Are you sure
>> the 11/23+ is working, and do you have the CPU in slot 3 with the
>> memory board in slot 4 (staying away from the PMI slots)?
>>
>> C
>>
>> On 10/19/2021 6:51 PM, Nigel Johnson Ham via cctalk wrote:
>>> It's just a memory board and looks like one!
>>>
>>> Picture here:
>>>
>>> https://www.mediafire.com/file/ive7i7ax6zdibv3/Screenshot_20211019_185019.png/file
>>>
>>>
>>> cheers,
>>>
>>> Nigel
>>>
>>>
>>> On 2021-10-19 5:21 p.m., Chris Zach via cctalk wrote:
>>>> Sounds familiar, got a picture of the board somewhere?
>>>>
>>>> C
>>>>
>>>> On 10/19/2021 12:57 PM, Nigel Johnson Ham via cctalk wrote:
>>>>> I am trying to bring up an 11/23 system in a BA23 box, and the only
>>>>> memory i have is an obscure Plessey one. The only identification is
>>>>> the
>>>>> p/n 705920 with dash-100 in white ink. By counting the chips I make it
>>>>> 4MB, but it does not respond. Since it takes the full 22-bit memory
>>>>> space I can't see how any jumpers would change its accessibility. Does
>>>>> anybody have a manual?
>>>>>
>>>>> any help appreciated,
>>>>>
>>>>> Nigel Johnson
>>>>>