Folks, I wendt to the ZuluSCSI page and rummaged around for any mention
of it being used on a Micro Channel system. No results, but maybe the
'net is full 'o satisfied users, each with a ZuluSCSI contentedly
nuzzling the SCSI bus on an IBM MCA SCSI adapter.

Please report your experiences with the ZuluSCSI. IBM or non-IBM
controller, IML or non-IML system. Version of ZuluSCSI and firmware level.

https://github.com/ZuluSCSI

I came across TWO issues with a PS/2, one in a 9577 Bermuda, the other a
Model 90, maybe solved, but I can't tell.

Synchronous transfer rate at 5MB/s on PS/2 Model 77 fails · Issue #189 ·
ZuluSCSI/ZuluSCSI-firmware · GitHub

https://github.com/ZuluSCSI/ZuluSCSI-firmware/issues/189

There seems to be an issue shifting from Stage 1 POST to Stage 2 POST.
It does not seem to be solved. MAJ Tom or the Magic Christian could
probably probulate it and see where the fumbled process occurred.

ZuluSCSI 1.1 with IBM PS/2 Model 90 · ZuluSCSI/ZuluSCSI-firmware ·
Discussion #257 · GitHub

https://github.com/ZuluSCSI/ZuluSCSI-firmware/discussions/257

Unfortunately, the OP does not identify the complex. Not enough there to
make a SWAG.

On 1/23/2024 9:49 AM, Louis Ohland wrote:
> Folks, I wendt to the ZuluSCSI page and rummaged around for any mention
> of it being used on a Micro Channel system. No results, but maybe the
> 'net is full 'o satisfied users, each with a ZuluSCSI contentedly
> nuzzling the SCSI bus on an IBM MCA SCSI adapter.
>
> Please report your experiences with the ZuluSCSI. IBM or non-IBM
> controller, IML or non-IML system. Version of ZuluSCSI and firmware level.
>
>
> https://github.com/ZuluSCSI
>
> I came across TWO issues with a PS/2, one in a 9577 Bermuda, the other a
> Model 90, maybe solved, but I can't tell.
>
> Synchronous transfer rate at 5MB/s on PS/2 Model 77 fails · Issue #189 ·
> ZuluSCSI/ZuluSCSI-firmware · GitHub
>
> https://github.com/ZuluSCSI/ZuluSCSI-firmware/issues/189
>
> There seems to be an issue shifting from Stage 1 POST to Stage 2 POST.
> It does not seem to be solved. MAJ Tom or the Magic Christian could
> probably probulate it and see where the fumbled process occurred.
>
> ZuluSCSI 1.1 with IBM PS/2 Model 90 · ZuluSCSI/ZuluSCSI-firmware ·
> Discussion #257 · GitHub
>
> https://github.com/ZuluSCSI/ZuluSCSI-firmware/discussions/257
>
> Unfortunately, the OP does not identify the complex. Not enough there to
> make a SWAG.

Louis,

This is my first reply on Eternal September using Thunderbird... hope it
goes through. I use the ZuluSCSI RP2040 with my Model 80 Reply Turbo
Processor. See Here:
https://store.rabbitholecomputing.com/ZuluSCSI-RP2040-p/zuluscsi-rp2040.htm

I specifically passed on the ZuluSCSI 1.2 because the tech specs says
"up to 4.5 megabytes/second, with a sufficiently-fast SCSI controller"
while the RP2040 supports FAST SCSI and is advertised as "up to 9.5
megabytes per second, with a suitably fast SCSI controller."

I found the RP2040 to be very reliable and never had any of the weird
issues mentioned in the Github article. I did have to fight to find
the right card to get close to top end speeds out of it. I ran it with
the Adaptec AHA-1640 and then with the three can Spock Prime and the
RP2040 would top out about 2.3MB/sec in Windows 95 on both cards, even
when using Unal Z's drivers AHA206 and SPOC206 drivers. This whole
exercise really crushed my thoughts about the Spock Prime as the Adaptec
card, being 16bit card, stood toe to toe and I could never get the Spock
to beat the Apatec card. Maybe the cache on the Spock would help with
Win95... but never saw it.

It wasn't until I got a Storage Dimensions SDC3211F (BusLogic BT-646 S)
that I saw decent speeds of mid to high 6MB/sec with rare 7MB/sec speeds
at the right block sizes. That card, with the Data Streaming ADF option,
just hauls ass with the SynchroStream controller on the Reply
TurboProcessor. The only shame with that upgrade board is that the
Pentium Overdrive and other upgrade chips like the Kingston TurboChip
and the 486 Overdrive 100mhz are just dog ass slow on that board for
some reason. The same chip in a PS/2 Valuepoint benchmarks 20% to 30%
faster. As a side note, before I got the BusLogic card, I picked up a
Corvette Turbo, but I hadn't researched a head of time to find that the
late WBST Clarke determined it wouldn't work on a Reply board many many
years ago...

Anyway, I don't use any config files or mess with any of what was
mentioned in the Github article. I just load up the images on the SD
card and go. I also ensured I had a quality SD card, UHS Class 3 or
Video Speed Class V30 or better, but that was purely about maxing out
the Bus Logic card. I never have boot issues, but it might be
system/card specific.

Below is the log from the last boot up.

[10ms] Platform: ZuluSCSI RP2040
[10ms] FW Version: 23.09.06-release Sep 7 2023 00:03:27
[11ms] DIP switch settings: debug log 0, termination 0
[11ms] NOTE: SCSI termination is disabled
[12ms] Flash chip size: 2048 kB
[14ms] SCSI target/disk mode selected by DIP switch, acting as a SCSI disk
[43ms] SD card detected, FAT64 volume size: 122208 MB
[44ms] SD MID: 0x1B, OID: 0x53 0x4D
[44ms] SD Name: 6D4Q9
[45ms] SD Date: 7/2023
[45ms] SD Serial: 0x9AA05C11
[49ms] Config file zuluscsi.ini not found, using defaults
[51ms] Active configuration:
[53ms] -- SelectionDelay = 255
[54ms] -- EnableUnitAttention = No
[56ms] -- EnableSCSI2 = Yes
[57ms] -- EnableSelLatch = No
[58ms] -- MapLunsToIDs = No
[59ms] -- EnableParity = Yes
[316ms] Finding HDD images in directory /:
[317ms] -- Opening /HD0_Buslogic _Main.hda for id:0 lun:0
[317ms] ---- Read prefetch enabled: 8192 bytes
[318ms] -- Opening /HD2_Buslogic_Games.hda for id:2 lun:0
[320ms] ---- Read prefetch enabled: 8192 bytes
[331ms] -- Platform supports ROM drive up to 1692 kB
[332ms] ---- ROM drive image not detected
[332ms] SCSI ID:0 BlockSize:512 Type:0 Quirks:0 ImageSize:8388608kB
[333ms] SCSI ID:2 BlockSize:512 Type:0 Quirks:0 ImageSize:2097152kB
[434ms] Initialization complete!

Hm, non-IBM system board. Definitely non-IML [nothing wrong with
that...]. AHA-1640 and Spock [one, two or three can] are SCSI-2 Command
Set, but firmly SCSI-1 speed [5MHz / 5MB/s]. BT-646 is SCSI-2 CCS and
SCSI-2 10MB/s.

Bummer on the Corvette, the 50 pin port is Fast/Narrow.

If only a Teutonic myth-legend would stop feasting and drinking, maybe
we could get an ADF for the SCSI-2 SE High Performance Internal/External
I/O Controller (Type 4-4)
https://www.ardent-tool.com/RS6000/RS6000_4-4.html

But only a Magical being could ever bring this about.

> Anyway, I don't use any config files or mess with any of what was
> mentioned in the Github article. I just load up the images on the SD
> card and go. I also ensured I had a quality SD card, UHS Class 3 or
> Video Speed Class V30 or better, but that was purely about maxing out
> the Bus Logic card. I never have boot issues, but it might be
> system/card specific.

How do you create the images?

Lionel Harris wrote:
> On 1/23/2024 9:49 AM, Louis Ohland wrote:
>> Folks, I wendt to the ZuluSCSI page and rummaged around for any
>> mention of it being used on a Micro Channel system. No results, but
>> maybe the 'net is full 'o satisfied users, each with a ZuluSCSI
>> contentedly nuzzling the SCSI bus on an IBM MCA SCSI adapter.
>>
>> Please report your experiences with the ZuluSCSI. IBM or non-IBM
>> controller, IML or non-IML system. Version of ZuluSCSI and firmware
>> level.
>>
>>
>> https://github.com/ZuluSCSI
>>
>> I came across TWO issues with a PS/2, one in a 9577 Bermuda, the other
>> a Model 90, maybe solved, but I can't tell.
>>
>> Synchronous transfer rate at 5MB/s on PS/2 Model 77 fails · Issue #189
>> · ZuluSCSI/ZuluSCSI-firmware · GitHub
>>
>> https://github.com/ZuluSCSI/ZuluSCSI-firmware/issues/189
>>
>> There seems to be an issue shifting from Stage 1 POST to Stage 2 POST.
>> It does not seem to be solved. MAJ Tom or the Magic Christian could
>> probably probulate it and see where the fumbled process occurred.
>>
>> ZuluSCSI 1.1 with IBM PS/2 Model 90 · ZuluSCSI/ZuluSCSI-firmware ·
>> Discussion #257 · GitHub
>>
>> https://github.com/ZuluSCSI/ZuluSCSI-firmware/discussions/257
>>
>> Unfortunately, the OP does not identify the complex. Not enough there
>> to make a SWAG.
>
> Louis,
>
> This is my first reply on Eternal September using Thunderbird... hope it
> goes through.   I use the ZuluSCSI RP2040 with my Model 80 Reply Turbo
> Processor.  See Here:
> https://store.rabbitholecomputing.com/ZuluSCSI-RP2040-p/zuluscsi-rp2040.htm
>
> I specifically passed on the ZuluSCSI 1.2 because the tech specs says
> "up to 4.5 megabytes/second, with a sufficiently-fast SCSI controller"
> while the RP2040 supports FAST SCSI and is advertised as "up to 9.5
> megabytes per second, with a suitably fast SCSI controller."
>
> I found the RP2040 to be very reliable and never had any of the weird
> issues mentioned in the Github article.   I did have to fight to find
> the right card to get close to top end speeds out of it.  I ran it with
> the Adaptec AHA-1640 and then with the three can Spock Prime and the
> RP2040 would top out about 2.3MB/sec in Windows 95 on both cards, even
> when using Unal Z's drivers AHA206 and SPOC206 drivers.  This whole
> exercise really crushed my thoughts about the Spock Prime as the Adaptec
> card, being 16bit card, stood toe to toe and I could never get the Spock
> to beat the Apatec card.  Maybe the cache on the Spock would help with
> Win95... but never saw it.
>
> It wasn't until I got a Storage Dimensions SDC3211F (BusLogic BT-646 S)
> that I saw decent speeds of mid to high 6MB/sec with rare 7MB/sec speeds
> at the right block sizes. That card, with the Data Streaming ADF option,
> just hauls ass with the SynchroStream controller on the Reply
> TurboProcessor. The only shame with that upgrade board is that the
> Pentium Overdrive and other upgrade chips like the Kingston TurboChip
> and the 486 Overdrive 100mhz are just dog ass slow on that board for
> some reason.  The same chip in a PS/2 Valuepoint benchmarks 20% to 30%
> faster.  As a side note, before I got the BusLogic card, I picked up a
> Corvette Turbo, but I hadn't researched a head of time to find that the
> late WBST Clarke determined it wouldn't work on a Reply board many many
> years ago...
>
>
> Below is the log from the last boot up.
>
> [10ms] Platform: ZuluSCSI RP2040
> [10ms] FW Version: 23.09.06-release Sep  7 2023 00:03:27
> [11ms] DIP switch settings: debug log 0, termination 0
> [11ms] NOTE: SCSI termination is disabled
> [12ms] Flash chip size: 2048 kB
> [14ms] SCSI target/disk mode selected by DIP switch, acting as a SCSI disk
> [43ms] SD card detected, FAT64 volume size: 122208 MB
> [44ms] SD MID: 0x1B, OID: 0x53 0x4D
> [44ms] SD Name: 6D4Q9
> [45ms] SD Date: 7/2023
> [45ms] SD Serial: 0x9AA05C11
> [49ms] Config file zuluscsi.ini not found, using defaults
> [51ms] Active configuration:
> [53ms] -- SelectionDelay = 255
> [54ms] -- EnableUnitAttention = No
> [56ms] -- EnableSCSI2 = Yes
> [57ms] -- EnableSelLatch = No
> [58ms] -- MapLunsToIDs = No
> [59ms] -- EnableParity = Yes
> [316ms] Finding HDD images in directory /:
> [317ms] -- Opening /HD0_Buslogic _Main.hda for id:0 lun:0
> [317ms] ---- Read prefetch enabled: 8192 bytes
> [318ms] -- Opening /HD2_Buslogic_Games.hda for id:2 lun:0
> [320ms] ---- Read prefetch enabled: 8192 bytes
> [331ms] -- Platform supports ROM drive up to 1692 kB
> [332ms] ---- ROM drive image not detected
> [332ms] SCSI ID:0 BlockSize:512 Type:0 Quirks:0 ImageSize:8388608kB
> [333ms] SCSI ID:2 BlockSize:512 Type:0 Quirks:0 ImageSize:2097152kB
> [434ms] Initialization complete!
>

On 2/27/2024 8:03 AM, Louis Ohland wrote:
> Hm, non-IBM system board. Definitely non-IML [nothing wrong with
> that...]. AHA-1640 and Spock [one, two or three can] are SCSI-2 Command
> Set, but firmly SCSI-1 speed [5MHz / 5MB/s]. BT-646 is SCSI-2 CCS and
> SCSI-2 10MB/s.
>
> Bummer on the Corvette, the 50 pin port is Fast/Narrow.
>
> If only a Teutonic myth-legend would stop feasting and drinking, maybe
> we could get an ADF for the SCSI-2 SE High Performance Internal/External
> I/O Controller (Type 4-4)
> https://www.ardent-tool.com/RS6000/RS6000_4-4.html
>
> But only a Magical being could ever bring this about.
>
> > Anyway, I don't use any config files or mess with any of what was
> > mentioned in the Github article.  I just load up the images on the SD
> > card and go.  I also ensured I had a quality SD card, UHS Class 3 or
> > Video Speed Class V30 or better, but that was purely about maxing out
> > the Bus Logic card.  I never have boot issues, but it might be
> > system/card specific.
>
> How do you create the images?
>
> Lionel Harris wrote:
>> On 1/23/2024 9:49 AM, Louis Ohland wrote:
>>> Folks, I wendt to the ZuluSCSI page and rummaged around for any
>>> mention of it being used on a Micro Channel system. No results, but
>>> maybe the 'net is full 'o satisfied users, each with a ZuluSCSI
>>> contentedly nuzzling the SCSI bus on an IBM MCA SCSI adapter.
>>>
>>> Please report your experiences with the ZuluSCSI. IBM or non-IBM
>>> controller, IML or non-IML system. Version of ZuluSCSI and firmware
>>> level.
>>>
>>>
>>> https://github.com/ZuluSCSI
>>>
>>> I came across TWO issues with a PS/2, one in a 9577 Bermuda, the
>>> other a Model 90, maybe solved, but I can't tell.
>>>
>>> Synchronous transfer rate at 5MB/s on PS/2 Model 77 fails · Issue
>>> #189 · ZuluSCSI/ZuluSCSI-firmware · GitHub
>>>
>>> https://github.com/ZuluSCSI/ZuluSCSI-firmware/issues/189
>>>
>>> There seems to be an issue shifting from Stage 1 POST to Stage 2
>>> POST. It does not seem to be solved. MAJ Tom or the Magic Christian
>>> could probably probulate it and see where the fumbled process occurred.
>>>
>>> ZuluSCSI 1.1 with IBM PS/2 Model 90 · ZuluSCSI/ZuluSCSI-firmware ·
>>> Discussion #257 · GitHub
>>>
>>> https://github.com/ZuluSCSI/ZuluSCSI-firmware/discussions/257
>>>
>>> Unfortunately, the OP does not identify the complex. Not enough there
>>> to make a SWAG.
>>
>> Louis,
>>
>> This is my first reply on Eternal September using Thunderbird... hope
>> it goes through.   I use the ZuluSCSI RP2040 with my Model 80 Reply
>> Turbo Processor.  See Here:
>> https://store.rabbitholecomputing.com/ZuluSCSI-RP2040-p/zuluscsi-rp2040.htm
>>
>> I specifically passed on the ZuluSCSI 1.2 because the tech specs says
>> "up to 4.5 megabytes/second, with a sufficiently-fast SCSI controller"
>> while the RP2040 supports FAST SCSI and is advertised as "up to 9.5
>> megabytes per second, with a suitably fast SCSI controller."
>>
>> I found the RP2040 to be very reliable and never had any of the weird
>> issues mentioned in the Github article.   I did have to fight to find
>> the right card to get close to top end speeds out of it.  I ran it
>> with the Adaptec AHA-1640 and then with the three can Spock Prime and
>> the RP2040 would top out about 2.3MB/sec in Windows 95 on both cards,
>> even when using Unal Z's drivers AHA206 and SPOC206 drivers.  This
>> whole exercise really crushed my thoughts about the Spock Prime as the
>> Adaptec card, being 16bit card, stood toe to toe and I could never get
>> the Spock to beat the Apatec card.  Maybe the cache on the Spock would
>> help with Win95... but never saw it.
>>
>> It wasn't until I got a Storage Dimensions SDC3211F (BusLogic BT-646
>> S) that I saw decent speeds of mid to high 6MB/sec with rare 7MB/sec
>> speeds at the right block sizes. That card, with the Data Streaming
>> ADF option, just hauls ass with the SynchroStream controller on the
>> Reply TurboProcessor. The only shame with that upgrade board is that
>> the Pentium Overdrive and other upgrade chips like the Kingston
>> TurboChip and the 486 Overdrive 100mhz are just dog ass slow on that
>> board for some reason.  The same chip in a PS/2 Valuepoint benchmarks
>> 20% to 30% faster.  As a side note, before I got the BusLogic card, I
>> picked up a Corvette Turbo, but I hadn't researched a head of time to
>> find that the late WBST Clarke determined it wouldn't work on a Reply
>> board many many years ago...
>>
>>
>> Below is the log from the last boot up.
>>
>> [10ms] Platform: ZuluSCSI RP2040
>> [10ms] FW Version: 23.09.06-release Sep  7 2023 00:03:27
>> [11ms] DIP switch settings: debug log 0, termination 0
>> [11ms] NOTE: SCSI termination is disabled
>> [12ms] Flash chip size: 2048 kB
>> [14ms] SCSI target/disk mode selected by DIP switch, acting as a SCSI
>> disk
>> [43ms] SD card detected, FAT64 volume size: 122208 MB
>> [44ms] SD MID: 0x1B, OID: 0x53 0x4D
>> [44ms] SD Name: 6D4Q9
>> [45ms] SD Date: 7/2023
>> [45ms] SD Serial: 0x9AA05C11
>> [49ms] Config file zuluscsi.ini not found, using defaults
>> [51ms] Active configuration:
>> [53ms] -- SelectionDelay = 255
>> [54ms] -- EnableUnitAttention = No
>> [56ms] -- EnableSCSI2 = Yes
>> [57ms] -- EnableSelLatch = No
>> [58ms] -- MapLunsToIDs = No
>> [59ms] -- EnableParity = Yes
>> [316ms] Finding HDD images in directory /:
>> [317ms] -- Opening /HD0_Buslogic _Main.hda for id:0 lun:0
>> [317ms] ---- Read prefetch enabled: 8192 bytes
>> [318ms] -- Opening /HD2_Buslogic_Games.hda for id:2 lun:0
>> [320ms] ---- Read prefetch enabled: 8192 bytes
>> [331ms] -- Platform supports ROM drive up to 1692 kB
>> [332ms] ---- ROM drive image not detected
>> [332ms] SCSI ID:0 BlockSize:512 Type:0 Quirks:0 ImageSize:8388608kB
>> [333ms] SCSI ID:2 BlockSize:512 Type:0 Quirks:0 ImageSize:2097152kB
>> [434ms] Initialization complete!
>>
Just any old empty file creator will work. I've used these two:
https://github.com/gajjartejas/EFIC-EmptyFileCreator
https://diskjockey.onegeekarmy.eu/

Once you mount them up in the Zulu SCSI and write partitions via
whatever OS you want to use, WinImage will open them if you want to drop
or grab a file.