Running RT11 on 11/03 with Emulex UC07 and SCSI2SD

[wa2flq]

You should be able to enable the LTC any time after RT11 Boots.

You might also try replacing the first 3 words after loading the UC07 bootstrap, but before starting


000200 106427
000202 340
000204 240

then 200g

This will replace the instructions to block interrupts while the bootstrap loads the OS.

MTPS #340
NOP

I haven't tried it. But it might work.

 

[RichCini]

With the G143R version of the UC07 firmware the PDP-11 host resident part of the FRD code is contained within bytes 0xE5F2 though 0xE6A5 of the 64KB firmware image, which is DMA copied into locations 200o though 463o into the PDP-11 host memory.

Do you know what address the code is for the hard drive bootstrap? Rather than typing the DMA code every time, I'd like to make it something I can use PDP11GUI to jam into memory.

Rich

 

[wa2flq]

Have you tried to configure the UC07 to autoboot? The LSI also needs to be set for mode 0.

No reason it shouldn't work if the LTC is disconnected.

 

[RichCini]

I haven't tried this yet...likely tomorrow.

I have a message out to the Heathkit group about something odd. On my H-11, with the LTC enabled by the panel switch I get about 60Hz on the BEVENT pin. When the LTC is disabled, BEVENT shows a 150Hz signal. Doesn't feel right which is maybe why I was having problems -- I thought the LTC was disabled but there was a signal anyway. It accomplishes this by opening or connecting a path to ground on the base of a transistor feeding BEVENT.

Kermit needs the LTC so I have to fix the above before moving on.

 

[wa2flq]

If you have the H11 backplane the schematic doesn't include a factory LTC switch. Nor do I see any Bus termination. If someone added a switch later, it could be the lines are just floating in the off position and pickup up noise.

In WH11 schematic there appears to be switch control the BEVENT signal, but always leaving the circuit connected.

 

[RichCini]

There are two versions of the H11 - one with the LTC switch and one without, but apparently there was a hack since there's a jumper on the power supply board for it. The version with the LTC uses an SPDT switch to ground the base of the transistor driver (disabled) or drive the LTC light (enabled).

No, there's no bus termination either. I wonder if I should get a terminator. BDV11 is a quad-width which won't work I don't think?

 

[glitch]

You'll typically be OK without termination unless you have a QBus jumper to another backplane...and IIRC on most QBus systems that's accomplished with one of the REV11 varieties, which often include termination anyway. One of the REV11 devices (there's a number of things called an REV11) provides 18-bit termination, DMA refresh, and bootstrap for a limited number of devices on LSI-11/2 CPUs -- you'll likely have one already if your main memory is the old 8K (4KW) MSV11-B boards as they require external refresh.

A BDV11 will work if you have a quad-width backplane.

FWIW, my MINC-23 is the only QBus machine I've got that actually has termination, and it's through a BDV11 at the end of the chassis. I suspect it kind of needs it with the integrated I/O crate.

Have you put an oscilloscope on BEVENT yet? Do you know what the waveform looks like?

 

[RichCini]

Have you put an oscilloscope on BEVENT yet? Do you know what the waveform looks like?

I just put a counter on it to see what frequency was there. Attached are jpgs for the LTC running and stopped. Looks fine to me.

 

[wa2flq]

Yes, I would expect that the termination on the CPU card should be sufficient in most cases for the H11. I cannot find a reference as to whether its 120 or 220 ohms though.
But I believe the "default" CPU position was row 2 for the run light to work, so I wonder if other non-standard wiring or board population is at work here.

Once Rich has a scope on the line, he should be able to see if the signal he sees is really a clock or just random noise picked up by the line. Even if the BEVENT is flapping, that can be solved. So I wouldn't rush out to get a separate terminator if everything else seems stable. That the UC07 works goes a long way in exercise of DMA, Interrupts, etc...