Unibone will do DMA with NPG inactive

[Michael Thompson]

The RICM is resurrecting a PDP-11/40 that was assembled from parts. We used a Unibone to emulate RL02 and then RA80 disks. Both worked perfectly and allowed us to boot XXDP. We added an RL11/RL02 controller downstream of the Unibone, and it failed diagnostics because it would not do DMA. While debugging we noticed that the NPG signal was stuck inactive low. The RL11 would not do DMA with NPG inactive, but the Unibone worked OK.

Does the Unibone check the state of the NPG signal when negotiating for control of the Unibus?

[Jay Jaeger]

I had an issue with the Unibone's cousin, the QBone, if I had certain devices downstream of it, but have not seen that issue on the Unibone that I can recall.  Are you sure your jumpers are correct, such that the NPG will be passed downstream?  Did you check NPG going into the Unibone as well as coming out of it?  Did you do the Unibone tests in a backplane with just a terminator (you can't use the 11/40 CPU backplane for that, I suspect - I'd use a separate DD11.)

Also, I seem to recall that if you are using all Unibone simulated memory, then in some configurations you can have it do the DMA transfer internally, thus no NPG appears on the UNIBUS after the Unibone - and maybe none appear on the UNIBUS at all.  So, another data point for folks would be what memory you are using.

Just some ideas - I could easily be mis-remembering some of this, too.

JRJ

[Michael Thompson]

The RX211 upstream of the Unibone was holding RPG to ground. The Unibone and the RL11 are working together now.

--
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.
To view this discussion visit https://groups.google.com/d/msgid/unibone/0b4fb5ff-edd2-4985-8f32-e957f24bfef8n%40googlegroups.com.

--

Michael Thompson