SCSi2SD V6.0 board

[kirk_m]

Has anyone gotten one of these to successfully work on the Amiga?  I have the 5.0 board, and have had zero problems getting it working on my A3000.  The V6.0 board sports faster transfer rates, so,  I decided to order one.  I have gone back and forth with the designer, and, nothing gets it working properly.  He even sent me a replacement board of the latest revision, which fixes some of the power issues my first one's revision was known to have, but, still no dice.

I've tried multiple SD cards, tried it on both my A3000 and A500 with the GVP sidecar, but, no luck on either.

Here's the gist of the issue.  I prepare the SD card on either a PC or MAC to wipe out its partitions, then I format it to FAT32.  I then insert the SD into the SCSI2SD, and plug that, via USB, into my PC or MAC to configure it with the SCSI2SD-util.  Everything saves properly after I put the proper parameters in for what size drive I want it to work as.  I am using a 4 gig card, and have the SCSI2SD configured to only make the first 3.6 gigs visible to the computer.  I do this because on this version, 6, the settings for the SCSi2SD card are saved on the SD card, instead of in flash memory on the SCSI2SD itself, as it was done on the v5 board.  I don't want the Amiga wiping out the settings somehow, since they are saved in the very last two "cylinders" of the SDCARD.

After all this is saved, I put the card into the A3000 or A500 sidecar, and fire up HDTOOLBOX (3.9 on the A3000 and 3.1 on the A500).  Both will detect the "hard disk" with all its proper settings (SCSI ID, size, manufacturer, revision number, product name, etc.).  I can then proceed to install the RDB on the "drive."  All goes as it should.  I can then reboot and partition and reboot again, or just go right to partitioning from install of RDB, and it partitions fine.

After partitioning is complete, I reboot again so I can then format.  Instead of being detected as an empty drive (being mounted as NDOS or something similar) in need of formatting, the drive is then not detected or mounted.  I go back to HDTOOLBOX to check the settings, and, all of the information that was previously there (the SCSI ID, SIZE, MANUFACTURER, REVISION NUMBER, PRODUCT NAME) has been wiped out and the drive cannot be installed due to something like "invalid drive identifier."

IF I repeat the wiping of the SD card, redoing the board's settings via USB, etc., and reinsert the board, all these parameters are again detected  properly by HDTOOLBOX, but, once again, after installing the RDB, the drive is rendered undetectable.

My first thought is that the amiga is somehow killing off the reserved area on the SD card that hold's the board's settings after I install the RDB.  I think this because the V5.0 board had its own flash array for storing settings.  But, the card's developer swears this is impossible because the 6.0 board doesn't grant the computer access to the "reserved for settings" sections of the SD card.

Has anyone had success with getting a 6.0 board to work?

BTW, termination is properly set.

[Gulliver]

Have you tried with other partitioning programs such as rdp391 and hdisnt (both on Aminet) ?

Maybe HDToolbox is somehow incompatible in some subtle way we are not aware. It would not be the first time.

[amiadudeorwat]

What about leaving it formatted as FAT then mounting it on the amiga with Fat95?

[kolla]

I woud not use HDToolBox from 3.1 on a large "drive".
Well, I would not use 3.1 at all, actually

[Davka]

Any progress with this issue?
I'm going to order SCSI2SD for my Blizzard 1260 SCSI Kit and I would prefer the v6 instead of v5.

[nicholas]

Acard adapters are better, they just work.

I have two for sale, one is UltraSCSI and one is UltrawideSCSI.

[desantii]

Depending in the machine it's actually similar price to buy a FastAta card for an A3000 for example

[Pat the Cat]

Don't know if this is the case, but might be issue with the A3000 SCSI controller chip. A3000 loads Kickstart from HD too (mine did, anyway). That' why A3000 owners don't like changing hard drives. More of a headache to set up a new boot disk than you might think.

A3000 SCSI were pretty early, were never tested properly for SCSI ultrawide (which itself came in several variants to begin with). True Ultrawide is 32 bit data, LVD. Not single ended, if you want it all to work at LVD speeds , it's all got to be LVD (including the controller). Mix in one single ended device and it becomes pointless to use faster devices.

If the V6 adaptor assumes your controller is always true ultra wide, and doesn't drop down properly (particulary with the 3000 controller) that might be the root cause of the issue.

If it drops down and works OK with other early SCSI-2 controllers, that would seem to point at an incompatibility specific to A3000 onboard controller. And remember, a SCSI bus is always limited to the slowest device standard - any slow drives or devices limit transfer rate of every device connected on that bus. Including controllers.

[kirk_m]

No fixes.  It doesn't work for either my A3000 (which has the -08 revision scsi chip) nor in the A500's GVP HD8+ enclosure.  The author has sent me updated firmwares, repeatedly, but none of them fixes it.  I'd stay away from the 6.0 board.

[wmaciv]

Finally got mine to work, mostly.  I have a Sandisk 16MB Extreme Plus in it right now, with a small WB partition (310 MB), a large Work partition (1.9 GB), and a 1.9 GB non-partitioned area on the "backside" of the card I had to leave unformatted, as any time I try to set a partition to the end of the virtual disk, I can never format it, it always refuses to format and displays the "not a Dos disk" message.  Speed and access times are really good for what is working, 2,200,000 - 2,300,000 average.
 
 In use with A2091 with 14mHz physical hack (jumper wire; no patched software), with C= 7.0 ROM's, 2MB 60ns RAM on board.  Board is single device on the chain, native passive termination on the 2091 (resistor packs), V6 is set to device 0, SCSI termination on, 4GB volume size, and SCSI2 option enabled in options.
 
 Realize this is after wrestling with another set of cards that "should" have worked, the ones Alex sent with the cards I purchased, also SanDisk , just one speed grade below what I am using now.  
 
 Here is the weird thing.  When Alex from Inertial sent the boards (I have two), they worked initially, but as I started running into problems and reformatting, they got worse and worse about not reading drive geometry and not retaining whatever I attempted to write to the virtual disk (RDB info).  
 
 It is as though minimal fiddling = greater success?  Did I somehow overwrite a low-level formatting required to allow the cards to work?  I know SD cards have the (invisible to the average user) security partition.  Is it possible Michaels v6 board overwrites that area with the virtual volume infomation, as it is the last two blocks of the card space, and that that space is more critical to the card operation than just security protocols?  Just spit balling here...
 
 A more basic possibility that there is an issue with the Amiga side (HD Toolbox) overwriting something critical where Michael stores the volume parameters on the card, or the reverse, that the location on the SD card where the volume info is stored somehow prevents the Amiga side software from being able to write to the card.
 
 The logic seems to fit if I cannot partition the "last" segment of any given virtual volume, as though that area has to remain off limits for the v6 to use.
 
 Would love thoughts from the group.
 
 Mac

[wmaciv]

Finally! Got the v.6 working correctly. They are EXTREMELY picky about termination. Amiga 2000 with A2091, passive terminating resistor packs on the card desoldered, active terminator on the DB 25 external connector, 6.13 guru rom on board, all partitions firing on all cylinders. I have 4 SCSI devices, each set a 3 GB, and each partitioned into two equal parts (about 1.5 GB). NO errors, and that is with the SCSI 2 and cache toggle set active. I am getting consistent 2.4 MB/sec on my A2000 with stock CPU!

I have a SupraTurbo28 and GVP A2000-HC+8 on the way due to arrive next week. Curious to see if they play well or help with the speed.

This thing rocks!

Mac
 
 Oops, forgot to mention I went from 1.85 MB/sec to the 2.4 just by applying the 14 MHz hack. I actually applied the hack right after getting everything stable a 1.85 and it did not upset the setup. Woot!

Mac

[HammerD]

Finally! Got the v.6 working correctly. They are EXTREMELY picky about termination. Amiga 2000 with A2091, passive terminating resistor packs on the card desoldered, active terminator on the DB 25 external connector, 6.13 guru rom on board, all partitions firing on all cylinders. I have 4 SCSI devices, each set a 3 GB, and each partitioned into two equal parts (about 1.5 GB). NO errors, and that is with the SCSI 2 and cache toggle set active. I am getting consistent 2.4 MB/sec on my A2000 with stock CPU!

I have a SupraTurbo28 and GVP A2000-HC+8 on the way due to arrive next week. Curious to see if they play well or help with the speed.

This thing rocks!

Mac
 
 Oops, forgot to mention I went from 1.85 MB/sec to the 2.4 just by applying the 14 MHz hack. I actually applied the hack right after getting everything stable a 1.85 and it did not upset the setup. Woot!

Mac

Very cool. I should pick one up for my A2500!

[aggro_mix]

Interesting!! Might buy one to try in my A3000 then If it doesn't work I have a bunch of samplers I can use it in.

[TheMagicM]

So the word on the A2091 is, you have to desolder stuff in order for that SCSI2SD to work?   A stock rev 6 A2091 no workie with this board?  I ask because I have a stock rev 6 board with the original ROMS.

[wmaciv]

   Okay, just updated to 6.1 firmware; no issues updating, but lost significant speed on the A2091/GuruROM setup. Went from almost 2.95 MB/sec to 2.35. Not sure what happened, but probably going back to 6.0.13 for the time being.