How should jumpers related to BG, NPG be set on Unibone to emulate RL02?

[Richard Chapman]

I managed to get my 11/34 to boot XXDP, but not RT11, which I think tells me I have an NPG problem. How should the 10 jumpers be set on the Unibone board? I have them like in this picture: https://retrocmp.com/images/stories/joerg/unibone/board-populated-big.jpg/, with each jumper on the two pins farthest from the card edge connectors. Is that right for RL02 emulation? Should I move the two related to NPG to the two pins closer to the card edge connectors? 

Thanks, 

Richard

[Joerg Hoppe]

Richard,

I managed to get my 11/34 to boot XXDP, but not RT11, which I think tells me I have an NPG problem. How should the 10 jumpers be set on the Unibone board? I have them like in this picture: https://retrocmp.com/images/stories/joerg/unibone/board-populated-big.jpg/, with each jumper on the two pins farthest from the card edge connectors. Is that right for RL02 emulation? Should I move the two related to NPG to the two pins closer to the card edge connectors?

The Jumpers look good.

Booting XXDP, but not RT11 typically indicates an IRQ problem.

To boot from disk DMA is always needed, to move data into memory.

But XXDP does not use Interrupts.

kind regards,

Joerg

[Richard Chapman]

How does one begin to debug the IRQ problem?  The only boards in the system are cpu/fpu, one memory board, front panel, DL11W,  M9312, and Unibone. 

--
You received this message because you are subscribed to a topic in the Google Groups "UniBone" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/unibone/y6haevSxeoE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to unibone+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/unibone/3325f119-1a78-4c9b-a010-f162fcca7bdb%40gmail.com.

[Wolfgang Schmitz]

Hello Richard,

that's why I pointed to testing XXDP!

I have a UNIBONE in a test bench together with a UNIPROBE.

When I start RT11 or do a DIR command, I can see lots and lots of activity on the

BR4 line (LED-wise), initiated by DMA-requests.

I do not have this activity when working with XXDP, so that's a good test to figure out

problems with BR/BG.

 

From far, I do not know where exactly the problem is.

Can you use another slot (with NPG open)?

what about SACK turnaround?

Do you have a UNIPROBE?

Wolfgang

--
You received this message because you are subscribed to the Google Groups "UniBone" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unibone+u...@googlegroups.com.

[Mikesch]

Hello Richard,

That's must be a problem with the Interrupt-Request-Chain.

XXDP needs DMA therefore the NPG chain seems working.

The DL11W is a problem - the card NEVER closes the NPG grant chain - but if XXDP is running the DL11W uses a slot with connected NPG a the backside of the backplane with a wire-wrapped NPG jumper.

You also must have a terminator at the end of the unibus! - what type do you use?

The minmal 11/34 has 9 Slots and with extension 18 slots

Normally your cards should use the following slots:

Slot#   card

1          M8266     CPU control path

2          M8265     CPU data path

3          M8267     FP11 - floating point unit

4          M9312     Bootstrap

            M7859     11/34 console

5          M7891     128K words MOS memory

6          M7762     RL11 controller for RL02 - or the UNIPROBE    (NPG grant wrapped must be open!)

7          M7856     DL11W serial board   (NPG grant wrapped must be closed)

8          G727  or  closes IRQ-grant (NPG grant wrapped must be closed)

            G7273      this card closes IRQ and NPG grant!

9          M9302    unibus terminator with SACK logic

If you have an extention up to 18 slots

the slot 9 and 10 hold a M9202 Unibus extender cable

the terminator is in slot 18.

All empty slots must have a wired NPG grant closed or a G7273 card in  the slot

If NPG is close by wirewrap G727 cards are usable instead of the big G7273 card.

Without termination or a minium of one unused slot without G7273 card or G727 then unibus and the 11/34 can not work - because of then opened IRQ grant chain!

You can short the unibus bei removing the M9202 at slot 9 and stuck in the M9302 terminator there if you have not enough G727* grant chain close cards.

For your configuration at the moment 9 slots are enough.

regards from Michael

[Richard Chapman]

Hi Michael, 

Thanks for the help — I have checked the NPG chain, and it is only open on the slot where the Unibone is (previously the RL11 was there), slot 8. I have the DL11W in slot 6, slot 7 has NPG wrapped closed and an G727. Otherwise my card layout matched what you outlined , except I did not have an M9302 in the last slot. When I put it in the A/B connectors of slot 9, I get the run light on the front panel lit up on power up, and I cannot extinguish it. When I attempt to read an address via the front panel, I get the bus err light. Either something is wrong with the M9302 (not much to go wrong, but possible I guess) or something is (more likely) wrong with one of the other cards or the backplane related to the SACK signal the M9302 turns around? 

You received this message because you are subscribed to a topic in the Google Groups "UniBone" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/unibone/y6haevSxeoE/unsubscribe.
To unsubscribe from this group and all its topics, send an email to unibone+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/unibone/eaf09446-c898-43b4-a5e5-b33ecd0e4199n%40googlegroups.com.

[Richard Chapman]

I tried replacing the M9302 with an M930, and the system does not hang, which pretty much guarantees the problem with the M9302 is an open bus grant daisy chain somewhere. Not necessarily and in fact probably not NPG since XXDP boots fine.  I will look at the pins of the 7430 chip on on the M9302 to figure out which one is the problem, and then trace it back through the backplane I guess. Hopefully just a bad contact somewhere. Of course this may not fix my problem booting RT11, but we can come back to that. I do appreciate all the help. 

[Richard Chapman]

The M9320 problem was simply that the Unibone was not running a program. the M9320 causes no problem with either a real RL11 or the Unibone if it is running a program. But RT11 still does not boot, while XXDP does . I see about 4 flickers of the RL0 light on the Unibone, then no output, using the rt11v5.5_34.rl02 image.  If it is an IRQ problem but not a BR/BG/NPR/NPG problem, how might I start figuring out what is wrong. The IRQ would  be triggered when a block of memory has been transferred,  to run a bottom-half handler of the RL11 device driver? The boot doesn't crash, just never finishes, so I am guessing it must be stuck waiting on some reads to complete that never do? 

[Richard Chapman]

I got it to boot RT11 and RSX by cleaning the contacts on the memory board with Deoxit

[Joerg Hoppe]

Am 20.03.2024 um 03:42 schrieb Richard Chapman:
> I got it to boot RT11 and RSX by cleaning the contacts on the memory
> board with Deoxit

Wow, congrats,

Joerg