• Please review our updated Terms and Rules here

QD32 getting into ROM

User78

Experienced Member
Joined
Jul 24, 2015
Messages
69
Location
Germany
Hi All,

next topic on my MV II.

I have found a QD32 controller in the chassis of a Fujitsu M2321/M.... which I have bought last year. (Many thanks to the seller!!!!)
So the QD32 seams to be the controller of this nice big and heavy SMD drive.
I have by now finished the replacement of all capacitors in the Powersupply, voltages are all checked and OK, so I'm ready to test it.

I would like begin with entering the QD32 ROM, which ends in the following.

>>> u
>>> i
>>> D/P/W 20001F40 20
>>> D/P/L 20088000 80000000
>>> D/P/L 20088004 80000001
>>> D/P/W 20001468 1
>>> E/P/W 2000146a
P 2000146A 0B40

Here I should get a 0B00 instead of 0B40.

The System freezes when I go forward and enter the rest.

>>> D/P/W 2000146a 3003
>>> D/P/W 2000146a 4401
>>> s 80
FREEZE


QD32 base address is default, as all other configuration also. He is also configured to 22 Bit.

IMG_5169.jpg

Any hints?

Many Greetings
Ulrich
 
Hi All,

next topic on my MV II.

I have found a QD32 controller in the chassis of a Fujitsu M2321/M.... which I have bought last year. (Many thanks to the seller!!!!)
So the QD32 seams to be the controller of this nice big and heavy SMD drive.
I have by now finished the replacement of all capacitors in the Powersupply, voltages are all checked and OK, so I'm ready to test it.

I would like begin with entering the QD32 ROM, which ends in the following.

Suggest you review this thread http://www.vcfed.org/forum/showthread.php?41530-OEM-11-73-Issues

The QD32 discussed does not appear to have a bootable formatter or interactive NOVRAM editor built into the unit. There are still commands to accomplish this that can be executed by careful register deposits.


Regards,
Jerry
 
Hi Jerry,

many thanks for pointing me to this thread.
But I have not found aditional information, except, that the controller seams to have defenitly no interactive configuration menue.
So looks my asumption, that I have some issue with the QD32 is not true.

This makes it not easier for me, but now I will review the SMD drive configuration, and afterwards I attach the SMD drive, and see what happens.


Many Greetings
Ulrich
 
So I have noted and reviewd the SMD Drive configuration.

1. it is terminated
2. it's configured for Hard sectoring
3. Hard sector is configured to 34 so we have 603 Byte per sector


After turning on, and booting from cdrom, NetBSD sees ra0 on uda0, but shows an error like unknown drive.
But as I have no clue for which system the 34 Hard sector config was done, it could be, that NetBSD and any other MV2 related OS can not use this starnge config.

Possible now to change to soft or 512 Byte/sector, but than I need to touch the strange QD32 NVROM, which I would not really do at the moment.

Many Greetings
Ulrich
Unbenannt.jpg
Unbenannt1.JPG
 
A sector size of 603 bytes per sector might be somewhat normal. I think that number is the raw sector size after the track is split up into raw sectors and sector gaps. The controller then reduces those raw sectors down to the standard 512-byte host visible sectors.

If you look at Appendix D, Disk Drive Configuration Parameters of the QD32 Disk Controller Technical Manual (MSCP Compatible), QD3251001 Rev B, November 1995, some of the drives listed should have sector sizes of 600, 576, 587, 593, and 594 bytes.

http://www.bitsavers.org/pdf/emulex/QD3251001-B_QD32_Nov85.pdf

There is a 409,600-byte emulex_qd32_diag.img disk image available here:
https://computerarchive.org/files/mirror/oldcomputers.dyndns.org/mirror/os2site/sw/dec/pro/misc/

That would be an RX50 sized disk image. I don't know if this disk is standalone bootable on a MicroVAX, or if you have to boot something else on the MicroVAX first and then run the format utility from this disk.

Here are the contents of the emulex_qd32_diag.img disk image as displayed by the EXCHANGE utility on VMS. I don't know how you actually execute these utilities yet. Maybe you have to run them under MDM (MicroVAX Diagnostic Monitor)

$ exchange
EXCHANGE> dir /full /col=1 dua1:
%EXCHANGE-S-MOUNTED, the RT-11 volume _VX3900$DUA1: has been mounted

Directory of RT-11 volume _VX3900$DUA1: 25-SEP-2016 09:18
Volume ID: VMS Exchange Volume Owner: SYSTEM System ID: DECVMSEXCHNG

UEVM.EXE 262 19-Sep-1985 14
FQD01M.EXE 48 19-Sep-1985 276
FVD03M.EXE 74 19-Sep-1985 324
IVV000.EXE 8 19-Sep-1985 398
IQT12.EXE 50 21-Oct-1985 406

Total of 5 files, 442 blocks. Free space 344 blocks, largest 344.
 
Last edited:
Agree with gslick. See https://archive.org/details/bitsavers_fujitsuB03_14826297 page 4-14 for description of what is in a raw sector. The minimum size for this drive appears to be 576 bytes, so a setting of 603 for this drive is probably correct.

If the drive and controller were part of the same system, the last thing I would do would be to reload the NOVRAM. Perhaps if you get get one of the utilities loaded, perhaps there is an option to see what is loaded.

I would double check the orientation of the cables. Even if they have polarizing plugs, there is a chance they weren't the the originals. Also try to use the drive on the other B side connection. It could have been the second unit drive. See 4-16 in the QD32 manual.

What is status of any LED's on the drive itself? See 2-2 in the drive manual.
 
>>> E/P/W 2000146a
P 2000146A 0B40

Here I should get a 0B00 instead of 0B40.

The System freezes when I go forward and enter the rest.

>>> D/P/W 2000146a 3003
>>> D/P/W 2000146a 4401
>>> s 80

As far as I nderstand the controller is DMAing stuff into memory. That something is what you then start. Is the DMA operation successful? Do you have valid ,sensible looking instructions from 80 and onwards?
 
There is a 409,600-byte emulex_qd32_diag.img disk image available here:
https://computerarchive.org/files/mirror/oldcomputers.dyndns.org/mirror/os2site/sw/dec/pro/misc/

That would be an RX50 sized disk image. I don't know if this disk is standalone bootable on a MicroVAX, or if you have to boot something else on the MicroVAX first and then run the format utility from this disk.
Hi,
This would be an option, but I have no RX50 only a TK50. :-(


What is status of any LED's on the drive itself? See 2-2 in the drive manual.

Hi,

LED Status is the following:

STTG2: on
STTG1: on
STS4: on
STS2: off
STS1: off

RDY: on
FPT: off
EMG: off

So looks OK, but I think STS1 should also be on, as it shows an active Controller / Drive connection.

It looks, that there is no way, to move around the QD32 configuration and checks. I'm not really familiar with the deposits, reads registers, ... . But a good time and need, to learn about.

I have also reviewed the possibility to identify the PROM which is used. Have found in the doc, that there is a jumper "M-N" which needs to be set, to specify the PROM size 16K or 32K.
The 32K version should have the menu in.

And found out, that the jumper M-N is set, so the 32k PROM is configured and the Menu should appear.




As far as I nderstand the controller is DMAing stuff into memory. That something is what you then start. Is the DMA operation successful? Do you have valid ,sensible looking instructions from 80 and onwards?

Hi Mattis,

I would like to say yes/no, but have no clue, and need to learn first much as I can about.
Could you please point me to the right procedure to check this?


Many Greetings
Ulrich
 
Last edited:
....
So looks OK, but I think STS1 should also be on, as it shows an active Controller / Drive connection.

It looks, that there is no way, to move around the QD32 configuration and checks. I'm not really familiar with the deposits, reads registers, ... . But a good time and need, to learn about.

A simple test would be deposit a known value in long word Address 80.
Next try to load the code that starts built in diagnostic, but do not start 80. If the value is changed, then the code is probably loaded.


The loading steps are the same as used for a UC07 or the QD33. I suggest not examining the results the step as you did in your example (E/P/W 2000146a). Since the sequence is custom, perhaps an unexpected read of the CSR registers will cancel it.

I have also reviewed the possibility to identify the PROM which is used. Have found in the doc, that there is a jumper "M-N" which needs to be set, to specify the PROM size 16K or 32K.
The 32K version should have the menu in.

And found out, that the jumper M-N is set, so the 32k PROM is configured and the Menu should appear.

If the code doesn't appear to be loaded, check the arrangement of cards in your backplane. The first trees rows in a MicroVAX II BA23 are Q/CD and then Q/Q serpentine. You should not have any empty Q slots. The would prevent the card from DMA'ing the data into the memory.
 
...

LED Status is the following:

STTG2: on
STTG1: on
STS4: on
STS2: off
STS1: off

RDY: on
FPT: off
EMG: off

As I read the table on 2-4, this is normal with the drive set to hard sectoring and a good sign in general.

If this doesn't change to STS1 on and STS2/4 off during attempts to access the drive by NetBSD, then I'd go over the cable and configuration. Check conductivity of the cables end to end for each pin.
 
If this doesn't change to STS1 on and STS2/4 off during attempts to access the drive by NetBSD, then I'd go over the cable and configuration. Check conductivity of the cables end to end for each pin.

It doesn't change when NetBSD is trying to access. Cables checked and not found a connection issue on the 60p and smaller cable.

Next for me would be to go back to the controller, and check basics there. There are also some registers I can read out, to get some status information.
But here I need to learn a bit as I know how to deposit, read and so on, but are always struggeling a bit.

I have not much time this week, so maybe my next reply is a bit delayed. But feel free, to flood the thread with useful hints. ;-)

Many Greetings
Ulrich
 
It doesn't change when NetBSD is trying to access. Cables checked and not found a connection issue on the 60p and smaller cable.

Next for me would be to go back to the controller, and check basics there. There are also some registers I can read out, to get some status information.
But here I need to learn a bit as I know how to deposit, read and so on, but are always struggeling a bit.

I have not much time this week, so maybe my next reply is a bit delayed. But feel free, to flood the thread with useful hints. ;-)

The only other quick test I would suggest is to try NetBSD w/o the drive actually connected.

If the result is different, then you know you have some interaction between the controller and the drive.
If the result is the same (one available drive ?), that suggests that the controller is just presenting what is in the NOVRAM to the OS, but the controller itself is not seeing the drive.


The MSCP controllers use command packets, so you have to deposit the packet into memory, then tell the controller to execute it through the command registers. I don't have experience debugging this type of controller at that level. Perhaps someone else can suggest some basic commands packets to attempt.
 
Back
Top