RQDX1 Not playing with the RX50

[NeXT]

Lou - N2M1Y has been helping me for almost two years get the PDP-11/23 boards that he gave me working in my BA23 enclosure.
One of the boards I received was an RQDX1 that was tested prior to mailing out. I attached it to an RX50 I bought on ebay but tested on my Rainbow 100 to confirm that it could read and write both its own floppies and other floppies and it had no issues. I had to crimp a new 34 pin and 50 pin ribbon cable because my BA23 was missing them. Lou also supplied me with a bootable floppy disk so I could get started.
My problem has been that when a disk is inserted, the drive is latched, I tell the system to boot from drive 0, the light on floppy 0 comes on alongside the motor, there is a click followed by a pause and then another click before everything stops and I get:

ERROR 17
NO DISK

There has been a fair bit of troubleshooting:
-RQDX was sent back to Lou for testing after a replacement resistor pack was built and it ran with no problems.
-RX50 was again tested on the Rainbow to ensure alignment and read/write functionality and there were no problems. The drives were also swapped and a boot was attempted from the Rainbow's drive and the problem persisted with it too.
-Both ribbon cables were quadrouple checked to ensure that there was no continuity problems and none were found.
-The Signal Distribution Board inside the BA23 was removed and checked for problems and continuity. It was fine except for pin 9 on the RQDX (DRVBUSOE) going to ground on the distribution board and it appears it's hardwired like this.
-A continuity test was made to see if the signals were making it from the controller to the drive with everything plugged in. No breaks were found.
-The power supply was replaced in the BA23 with a spare I owned. No change in symptoms.

Short of ensuring my floppy is good (and we can't rule that because the system right now says no disk is present) We have both run out ideas on what has happened. The last thing I can do is attach a logic analyzer and see if the drive is actually passing on that the disk is present but between long nights and dozens of emails I have forgotten which pin the drive and controller used to know and tell when a disk was present. Please, I need help.

 

[RSX11M+]

If I recall correctly, this cable has to be plugged in with one end intentionally inverted of normal. There are photos here in another thread.

Perhaps you already considered this.

UPDATE: I now see that is YOUR THREAD TOO! LOL

So you're still wrestling with the same system?

Next thing to check is the module containing the power switch and drive status lights. Should I re-read that other thread in it's entirety to re-familiarize myself with your situation?

 

[NeXT]

UPDATE: I now see that is YOUR THREAD TOO! LOL

Yeah I'm running into memory problems and I can't seem to remember some things but now that it's been brought up yeah, I do remember recrimping that cable so it was in the proper arrangement and the continuity checks all show it's correct.

I'm not entirely sure what I would find on the front panel. Whatever lines went from the distribution panel to it were either for:
-Fixed Write protect button
-Fixed disk write protect light
-fixed disk ready button
-fixed disk ready light
-floppy disk 1 and 2 write protect lights

That I can find there is nothing that the front panel would affect the operation of the floppy drive.

 

[RSX11M+]

There are several variations of that panel. Having a different one could potentially effect things - especially if misconfigured.

You said RQDX1 didn't you... and you don't have an RQDXE?

I'm afraid this stuff is no longer off the top of my head either. I do recall having to figure out which Drive Select was expected in various boot cases. This had to be re-arranged in some instances, particularly when installing HD / RX50 combinations.

It may be that the switches on the PCB of the Drive Select Panel determine what the controller expects... I think I have handwritten notes on this somewhere. It may also be covered in the Micro11 systems manuals.

Any of that sound familiar?

 

[NeXT]

Unfortunately I only have the RQDX1.
My front panel is the super simple one. Really all mine in particular is missing is the "MicroPDP-11" badge. It lets you write protect or put the MFM drive ready or tell you if either floppy is write protected. All other functions (including the two dip switches hidden behind it) run straight to the qbus backplane

I really doubt that it is my problem.

 

[Lou - N2MIY]

He has a BA23, which doesn't need an RQDXE. However it is indeed very possible that there is somehting screwy with the drive select panel. That is the one part of Next's system that I don't think has been scrutinized! I remember that one of the little switches is related to the LTC, and I think the other is unused. However, there are active components on that board and something could be wrong there.

I have never anywhere seen a print set for the BA23 backplane or the drive select board or the power supply. Why not?

Lou

 

[NeXT]

I remember that one of the little switches is related to the LTC, and I think the other is unused.

I read up that the other switch disabled the reset button.
I have an idea. I ran it through Lou a month ago and when the electronics shop opens tomorrow I'm going to throw a stone into the dark and crimp a ribbon cable that goes from the RQDX1 to the RX50. If previous testing was correct, it should work.

 

[RSX11M+]

Ok - just spent a couple hours trying to find my notes - no luck so far. I did review the original docs for RQDXE and it's uses with RQDX1,2,3. Basically - if you have no external drives, you don't really need an RQDXE. (my refresher - not yours)

The jumpermainia I'm probably recalling was discovering erata in the RQDXE instructions - and figuring it out myself. (very early days prior to the RQDX2)

Still, your problem is hauntingly familiar. When I upgraded to an RQDX3 (to support RX33s) I needed to figure out the way DS0-3 were resolved into LUNs. (again) All the documentation here indicates RX50s as internally configured to respond to DS0, DS1.

I happen to have an RX50 opened up here, and didn't see any obvious internal drive select to check. It is possible that this selection is ignored on some systems and could explain why a drive might test ok there, but not here where it mattered.

Oh - and there are versions of the power panel (for multiple HDs) that get much more involved in drive selects, ready and write-protect, so a fault on that board could potentially be the issue. Are you brave enough to unplug it from the BA23 drive distribution panel to see what happens when you try to boot a floppy?

Another thought is to put in two diskettes, even though you only mean to boot the one you have in. Yet another is to put in a write-protected diskette, and see that the correct indicator lights on the power panel. These would tend to confirm there's no transposition of signals in the mix.


RQDX1 - you do know it has to be last on the QBUS - right? Don't know if that would cause your remaining issue. But is does provoke me to ask if there is any other drive present on the BA23 panel?

You must have thought of a lot of this stuff already in the time you've spent. I'm sure I'm being redundant.

Sorry I'm not being of more help here... brain cells stretching.

P.S. - I've never done the direct cable experiment you referred to. (RQDX-RX50) Sounds plausible, - I'm sure you worked it out with Lou, so it's worth a try.

P.P.S. - Lou - can you provide a little more of a history on this BA23?

 

[Lou - N2MIY]

I don't know the history of Next's BA23, but I ran this card in one of my BA23s with no problem. All of my BA23s have four button panels, and the one I ran this card in had no modifications to the four button drive select panel. I ran this card on all stock hardware and I am sure I did have it as the last card in the backplane (I know that issue). Last card does not mean the bottom slot, it means last card in the bus grant chain (for Next's info...)

Lou

PS. I have modified a four button panel to allow the second RD drive to work properly.

 

[NeXT]

Oh my god this is dumb. I am embarrassed that this took so long.

So I went to the electronics shop and bought brand new ribbon cable and crimpable ends. Up until now I've always harvested from other ribbon cables because they're essentially free these days.
In a vain attempt I started going over things I already did. First up was the custom ribbon cable that would go from the RQDX1 tot he RX50 with no stops in between. Wasn't cheap. The connectors alone cost me $15.

RQDX1      RX50

8----------28
22---------2
24---------6
26---------8
28---------10
29---------12
30---------14
32---------16
34---------18
36---------20
38---------22
40---------24
42---------26
46---------30
48---------32
50---------34

Ground pins on RQDX1 are 3, 5, 9, 21, 25, 33, 35, 37, 39, 41, 45, 47, 49
Ground pins on RX50 are all odd pins from 1-33
**ALL GROUND LINES MUST BE CONNECTED OR OTHERWISE CONNECTED TOGETHER**

I immediately knew something was different from before. Like I said there are distinct clicks form the RX50 when it's in use. This time they clicked faster.
Threw in my floppy, tried a boot and....

WOAHHHHHHHHHH! SUCCESS!!! MOTHERFUC%*@G SUCCESS!

Okay, so this meant there HAD to be a flaw on the distribution board and I could completely rule out the front panel as it wasn't even attached.
The job involved was applying the ribbon cable directly to the PCB by hardwiring it, just in case there was an invisible broken trace.

This however, did not work. :|
So there was a flaw somewhere. I went back on my notes and older pinouts to see if anything SOMETHING was different.
I saw it. I had written it in a notebook a year and a half ago and while I had reviewed the same notes several times, it didn't click on me.
Pins 6 and 8 (Index and Drive Select 3) on the RX50 were going to pin 24 (Drive Select 3) on the RQDX1. I checked again and this was the case PLUS pin 8 was going to pin 26 so there was a bridge between the pins not in the PCB but in the crimped end of the ribbon cable itself. I didn't even think to see if there was pin-to-pin bridging going on in the cables. I must of mangled something when I first crimped them.

can you see the difference? Neither can I.

So I crimped a new 50 pin ribbon cable AND FINALLY IT WAS ALL WORKING.
Ugh.....that was one of the most painful things I have ever done.

 

[RSX11M+]

Happy days.

I know how you feel... this would be one of those victories I couldn't even savor they hurt so much.

Tenacity is my middle name... I suggest you consider it for yourself.

Well done, hard won.

I'm not certain I follow everything you said, but I'd guess that in your signal continuity testing you couldn't have checked for the "negative" (non-continuity) case much, or you might have found this at an earlier point. So, were you able to remove the bird's nest and revert to a clean board? (That I really missed)

Making cables is something we've all done... it's by "feel" more than anything that we get clue of a goof. I've made some mistakes myself, and it's always painful.

So... What's next? (I suggest... alcohol)

 

[NeXT]

Now I need to figure out how the address dip switches work on Qbus cards so I can add in my CMD SCSI card and DELQA, transfer RT-11 to the hard drive and start building from that. Oh yeah, I need to find a 4 port serial card so I can run the system as a "low power" alternative to my 11/84 and allow the system to pass TCP sessions on to the /84 if they are granted permission.

 

[tingo]

You fixed it - that is what matters in the end (IMHO). Congrats!

 

[RSX11M+]

NeXT, what's the "ultimate" configuration you want to run? Are you the "plan ahead" type, or the "Tinker as you go" kind?

In the ancient past we would plan a configuration to make certain the power output of the supply and the AC BUS loads of the backplane were not exceeded. There is a configuration planning guide in the manuals and the S&O catalog for this purpose.

Many of us are accustomed to just "winging it" because of either our familiarity, or knowledge based on not coming anywhere close to the load limits of a box/CPU combo. It can be very easy to make an error however, and it's never a good thing to let any of the magic blue smoke escape from these things. (it's devilishly difficult to put back)

Since you have an '84, you're probably already aware of these things - I just wanted to put it out there for posterity.

You might consider a DHV, DHQ, DZV, or DZQ serial port. I'd bet they'll be cheaper than something like a DLV11-J and more common. (always be sure to get a cabinet kit with the unit) Transfers via DLV are character at a time via interrupt. Some of these others have DMA based transfers or hardware silos that will be more tolerant of the bursty nature of IP over serial or even system to system file transfers - if that's what you're thinking. However, they're going to be supported only in later versions of RT11. They're in 5.7 for sure (I checked the SPD) but I'm not certain about earlier revs or about what's in a common sysgen or pre-gened kit.

If I recall correctly, DH's support higher baud rates too (38.4Kbps) but that may only be under RSX.

Are you going to be able to fully test your RQDX with RX50 and hard drive soon? That will answer a lot of questions about you being done with troubleshooting. Re-arranging to add other options will be simple after that.

 

[NeXT]

As I stand I really don't plan on further expanding beyond the serial ports because I can't afford (in space, power and money) any more expansions. I did the math to make sure my buses could handle the boards I have and are planned (really all that's left is a framebuffer. The rest of the systems have already been load tested).
Because electricity is so expensive now the idea was to run the /23 on whatever OS would work best in a networked multiuser environment (I think we settled on RT-11 because of this) and for the most part I really won't be needing more performance than the /23 has.
On the other hand if I do for some reason need more computing power or I need access to a 9-track drive the idea would be I power the /84 up and I would have accounts that are simply setup to push a connection from the ethernet on the /23 through a serial port to a login session on the /84. If I really didn't care about both systems being multiuser I could just run a null modem cable between the two and I'm done.

 

[RSX11M+]

Then the only additional comment I would offer is that an RSX environment might have something to offer over RT11. However, the bar for learning it is a bit higher and there are fewer users out there with experience in it.

Designating a port on a DHV as passthrough (to the 84) was a common use back in the day. RSX even has a built-in facility for it. See RSX utilites: Data Terminal Emulator (DTE) and MicroRSX File Transfer (MFT).

I would hesitate to keep a real PDP-11 powered 24/7 any more, for fear of losing it. Electricity is a factor too. If you're going to offer these for remote use 24/7 - I'd think about running Ersatz-11 (John Wilson) on a PC for the 24 hr part, and powering the actual systems only as needed.