• Please review our updated Terms and Rules here

GOTEK w FlashFloppy as RX33/RX50 floppy emulator

Man you're having a weird day. I did a full install of 3.2 onto an RD53 using the stock DECUS images on my pair of Goteks.

Speaking of which, is there a different read of the 32 disks somewhere? The DECUS ones had an error, the Pro/Communications disks were like 3 of the images but disk2 is a copy of disk 1. I'm guessing the guy who did the transfer left the wrong disk in the drive and since I have a TMS board I'd really like to install Pro/Communications 3.2.
 
Lots of problems solved in the interim! Long story short: The only images that work on my Gotek drives have track 0 blank with the boot block and directory on track 1. Does that ring any bells with anyone? The raw disk images I downloaded are missing that blank track and none boot. If I convert the TD0 images on UPDATE to raw sector images using Applesauce they show that blank track and work fine. WTF??
 
So here's a question: When I boot diskette 21 and select install it eventually asks me for the PROSYSTEM32 diskette in "slot #2". Does that mean "put in the second drive"? If so, it never seems to be aware it's there! I notice that the DEC RX-50 is signaling the system when I close the door and I can see it scan the disk. Not seeing that with the Gotek however. Can you share your configuration files and images with me? This is starting to make me nuts.
 
I solved the immediate issues. Here's what tripped me up (in no particular order):

  • RT-11 refers to the diskette drives as DZ0: and DZ1:, while the P/OS installer calls them DZ1: and DZ2:. So I was changing the wrong diskette.
  • DEC machines ignore physical track 0 on floppy formats (Rainbow excepted). This is hidden down at the driver (or controller) level. Many of the raw sector images are intended for use with, e.g. Xhome and simh which expect the image file to start with what would be track 1 on the media. Those images do not work with floppy emulators. You cannot tell by the length, but must look at the binary to see where the bootblock is.
  • A good number of the TD0 images I downloaded were taken from damaged diskettes. Applesauce will even show you them since they're maked in the metadata. Needless to say these do not work. The TD0 images from UPDATE in Sweden are fine to the extent I've checked them.
  • I'm still struggling a bit with configuration of Flashfloppy. P/OS watches for diskette change (door closed in real drive) and scans the diskette on its own before you take any action with it. Simply configuring as 'shugart' interface didn't seem sufficient. I ended up adding 'pin34 = rdy' to FF.CFG. I still need to quantify this a bit better.
  • In the same groove as the first point above, there is a three-diskette 'emergency' package that asks you to change this diskette in slot 2. That's ambiguous, since the floppy controller itself is in slot 2 of the card cage. What they really mean is "second floppy drive" - or what the installer calls DZ2:. Nothing like standards I always say.
All together I burned about 7 hours of time working through to the point where I could do a P/OS install from diskette images. I hope this saves someone else from that level of grief.

Finally, there are a number of important P/OS manuals that have not been scanned. If anyone has these would you please speak up?

"P/OS Server User's Guide"
"PROSE User's Guide"
"PRO/EDT User's Guide"
"Command Language User's Guide"
"Hard Disk System User's Guide"
"Hard Disk System for Beginners"
"Hard Disk System Error Guide"
"P/OS Hard Disk System Release Notes"
Click to expand...
 
I checked the CHM catalog and we don't have any of those. There are a couple of others like the system reference that I'll see about scanning
It doesn't seem like P/OS was liked by many people.
 
Al Kossow said:
I checked the CHM catalog and we don't have any of those. There are a couple of others like the system reference that I'll see about scanning
It doesn't seem like P/OS was liked by many people.
Click to expand...
Bummer. Johnny Billquist has them, but they're not stored at his present location. Probably also a matter of available "spare time" :)

I'll ask all my high friends in low places to see if someone can help here in the US.
 
> DEC machines ignore physical track 0 on floppy formats (Rainbow excepted)
In FILES-11 ODS-1 (P/OS and RSX) at last sector of the disk stores file "badblk.sys", and in DZ notation (for RX50 floppy drive) "last" sector 800 physically lays on track 0. (please see dzmap.txt attached)
> a number of important P/OS manuals that have not been scanned. If anyone has these would you please speak up?
It will be good to check not only common names, like "P/OS Server User's Guide", but also a DEC revision partnumber
for example, diskettes:
BL-N639C-BH (P/OS HARD DISK V1.5, DISPATCH, volume label: PRODISPATV15)
doc:
binder AV-GK48A-TH - Professional Hard Disk System Documentation (P/OS_V.3.1) included:
Ch0 - AA-EX85B-TH - Release Notes
Ch1 - AA-EX78A-TH - Hard Disk System Users Guide - Third Edition
Ch2 - AA-EX78A-TH - Hard Disk System for Beginners - Fourth Edition
Ch3 - AA-DK96A-TH - P/OS Server Users Guide - First Edition
Ch4 - AA-FC89A-TH - Hard Disk System Error Guide - First Edition

P.S. I am looking for a "CTI BUS: Technical Manual EK-OOCTI-TM-002" scan.
 

Attachments

  • dzmap.txt
    3.3 KB · Views: 5
  • DZ_dsk.png
    DZ_dsk.png
    60.6 KB · Views: 8
anasana said:
> DEC machines ignore physical track 0 on floppy formats (Rainbow excepted)
In FILES-11 ODS-1 (P/OS and RSX) at last sector of the disk stores file "badblk.sys", and in DZ notation (for RX50 floppy drive) "last" sector 800 physically lays on track 0. (please see dzmap.txt attached)
> a number of important P/OS manuals that have not been scanned. If anyone has these would you please speak up?
It will be good to check not only common names, like "P/OS Server User's Guide", but also a DEC revision partnumber
for example, diskettes:
BL-N639C-BH (P/OS HARD DISK V1.5, DISPATCH, volume label: PRODISPATV15)
doc:
binder AV-GK48A-TH - Professional Hard Disk System Documentation (P/OS_V.3.1) included:
Ch0 - AA-EX85B-TH - Release Notes
Ch1 - AA-EX78A-TH - Hard Disk System Users Guide - Third Edition
Ch2 - AA-EX78A-TH - Hard Disk System for Beginners - Fourth Edition
Ch3 - AA-DK96A-TH - P/OS Server Users Guide - First Edition
Ch4 - AA-FC89A-TH - Hard Disk System Error Guide - First Edition

P.S. I am looking for a "CTI BUS: Technical Manual EK-OOCTI-TM-002" scan.
Click to expand...
I tried searching most of those terms in Google and find only a Russian site with broken links for downloading.
 
> I tried searching most of those terms in Google and find only a Russian site with broken links for downloading.
of course, since Koshka (=== "Cat") is in beta state it sends directly for person by e-mail ask to anasana@ukr.net or anasana@gmail.com
p.s. telegram, skype, whatsapp or any other comfort messenger You use to chat - is great.
After russian aggression to Ukraine in 2022 (but, if clear, in 2014) I mainly collapse the cooperation even in a hobby layer but some friends, the few I have; so today You may see a lot of dead links occurred.
any of us uses the Linux only, please note, that the Koshka is a Windows GUI program for a Simon Owen "fdrawcmd.sys" FdInstall driver.
p.p.s the current stage - is collecting of correct GAP3 dimension size, as for official DEC media and for other counts of 3rd party formatters.
p.p.p.s. maybe +1 mess and I can write here directly to mr. Al Kossow, while to aek@bitsavers.org we speak before, last year I cant connect about of DEC PRO docs update that I scan.
 
Last edited:
To save others from the grief I've been though this week, here's a pair of little Python programs to convert between raw RX50 diskette images (sectors in logical order) and the format expected by Xhomer/simh emulators (sectors in physical order with track 0 moved to the end). If someone sends you an image for an emulator you will need to convert with simh2dsk.py in order to load that file into a Flashfloppy/Gotek. Conversely, an image read from diskette using, e.g. Applesauce or Imagedisk must be converted with dsk2simh.py for compatibility with the emulator.
 

Attachments

  • rx50_convert.zip
    1.8 KB · Views: 5
anasana said:
> I tried searching most of those terms in Google and find only a Russian site with broken links for downloading.
of course, since Koshka (=== "Cat") is in beta state it sends directly for person by e-mail ask to anasana@ukr.net or anasana@gmail.com
p.s. telegram, skype, whatsapp or any other comfort messenger You use to chat - is great.
After russian aggression to Ukraine in 2022 (but, if clear, in 2014) I mainly collapse the cooperation even in a hobby layer but some friends, the few I have; so today You may see a lot of dead links occurred.
any of us uses the Linux only, please note, that the Koshka is a Windows GUI program for a Simon Owen "fdrawcmd.sys" FdInstall driver.
p.p.s the current stage - is collecting of correct GAP3 dimension size, as for official DEC media and for other counts of 3rd party formatters.
p.p.p.s. maybe +1 mess and I can write here directly to mr. Al Kossow, while to aek@bitsavers.org we speak before, last year I cant connect about of DEC PRO docs update that I scan.
Click to expand...
I'm not quite following your statement about Al Kossow. Are you saying you have scans of DEC PRO docs for him but cannot make contact? If you have these scans I can give you a dropbox area where you can upload. Please PM me?
 
czunit said:
Man you're having a weird day. I did a full install of 3.2 onto an RD53 using the stock DECUS images on my pair of Goteks.

Speaking of which, is there a different read of the 32 disks somewhere? The DECUS ones had an error, the Pro/Communications disks were like 3 of the images but disk2 is a copy of disk 1. I'm guessing the guy who did the transfer left the wrong disk in the drive and since I have a TMS board I'd really like to install Pro/Communications 3.2.
Click to expand...
have you checked the images on ibiblio?
 
shirsch said:
the format expected by Xhomer/simh emulators (sectors in physical order with track 0 moved to the end
Click to expand...
I really hate simulation container formats with no metadata AND rearranged tracks??? unbelievable!
So now, without knowing it, we are "preserving" software with embedded knowledge that the container doesn't match reality.
 
Al Kossow said:
I really hate simulation container formats with no metadata AND rearranged tracks??? unbelievable!
So now, without knowing it, we are "preserving" software with embedded knowledge that the container doesn't match reality.
Click to expand...
Well, the good news is that one can use my Python utilities (posted above) to convert between a proper disk image and the rather odd arrangement expected by simh and friends. Very easy to tell what you have by looking at the first few sectors.

UPDATE: I can probably code a heuristic to identify the image type.
 
Last edited:
I really hate simulation container formats with no metadata AND rearranged tracks??? unbelievable!
So now, without knowing it, we are "preserving" software with embedded knowledge that the container doesn't match reality.
Click to expand...
This is not for all cases works. In "standard" formats like is DEC RX50, we always know what expect, even having from sectors dump only, and how it will be restored is a simple and clear.
(yess, about of the 81-track with one 128-byte sector and unique DEC-media-label here in FM mode exists. people mainly don't know :) )
Of course, the copy of protected diskettes for many home-oriented architectures need to save ALL flux data stream.
 
Can someone who has a pair of Gotek/FF drives connected to their Pro comment on what happens when the machine is booted with no images mounted in the Goteks? An empty RX50 drive doesn't seem to cause any issues, but with the Goteks I get a self-test fail on the floppy subsystem. Tried a bunch of things in the Gotek configuration and cannot find the magic setting to placate self-test. Finally I tried connecting a logic analzyer to watch the interaction with an RX50. Unfortunately the lines are so glitchy it's impossible to make sense of.

Based on the RX50 documentation, it does not use pin 2 and sets pin 34 true (low) when a diskette is present. That's exactly how the Gotek is setup but somehow that's insufficient. There's something going on for handshaking, but what?
 
After some A/B comparison between the RX50 and Gotek while attached to my floppy drive exerciser I learned that the RX50 holds index low when motor-on and select are asserted and no diskette is in the drive. Surely that had to be it, right? I built a small circuit to mimic this behavior and it STILL fails the Gotek during POST unless a diskette is inserted. Starting to run out of ideas. Yes, it's sort of cosmetic in that the machine will still boot eventually but I'm bothered by the fact I don't understand the behavior.

If anyone has a WAG what the firmware in the Pro 350 might be looking for to determine that an (empty) drive is attached I'd appreciate some input. I've ruled out track 0 sensor (Gotek asserts this even when empty) already.
 
Well, that's two days of my life I'll never get back. I built up various bits of external circuitry on a project board in an attempt at getting an empty Gotek/FF recognized by the power on self-test. No joy. Workaround is to ensure that the first (or only) Gotek drive has an image inserted from the start. I have no idea what the blasted firmware is looking for, but it must be some subtlety of the RX50 hardware that I'm not seeing. Compared logic analyzer traces until my eyes crossed and nothing jumps out.

If anyone smarter than I figures this out, please let us all know?
 
You must log in or register to reply here.
Back
Top