Yamaha SCSI-IDE v769970 supported CF card list

chiark

chiark

AmiNerd
AmiBayer
Blogger
Joined
Jan 22, 2008
Posts
2,702
Country
United Kingdom
Region
Yorkshire
Seeing as there's quite a few of us with these devices now (cheers Frodegh!) it's worth saving frustration by listing what does and does not work:

Known working:
- Samsung 1GB CF (rebranded as PureMedia 1GB CF)
- Samsung 2GB CF (rebranded as Jessops 2GB)

Known NOT working:
- Transcend 2GB 133x CF
- SanDisk Extreme III 256MB CF
- SanDisk 512MB CF

Transcend and SanDisk 512MB both lock the SCSI bus completely, with inquiry light constantly on. These are the only two cards that support DMA
The SanDisk Extreme does not support DMA (!), but does not support IORDY.

The cards that work:
- Do not support DMA of any type whatsoever
- Do support IORDY signal
- are all marked as removable

DMA card? Locks the bus.
No IORDY support? returns incomplete information when polling drive.

Code: 
Name                      Works  DMA  IORDY  Rem? Notes
----------------------------------------------------------------------------------
Sandisk extreme 256         N     N     N     Y    Returns incomplete info on disk poll
Sandisk 512MB               N     Y     Y     Y    Locks SCSI bus
Transcend 1GB 133x          N     Y     Y     Y    Locks SCSI bus
Samsung 1GB (PureMedia)     Y     N     Y     Y    Works
Samsung 2GB (Jessops)       Y     N     Y     Y    Works
 

Attachments

  • YAM-IDESCSI-CONVERT.pdf
    89 KB · Views: 1
Last edited:
Great initiative! I'm curious to see if there is any logic to which CF cards work or not.

Over IDE it can make a difference if they are of fixed or removable type, but it depends on the driver.
Scsi.device and compactflash.device does not care if a card is removable or fixed, but IDEfix97 and drivers for Individual computers IDE controllers only work with fixed cards, as they mistake removable drives for ATAPI devices.

If you run windows diskpart with your CF cards on IDE or SATA (not USB!) and type 'list volume' you will see if a card is 'removable' or 'partition' (=fixed).

It would be nice to know if this plays a role when connected over SCSI.

Edit: It would also be nice to know if a non-working card works well over IDE!
 
chiark said:
Yep, the odd thing is all the cards work fine on the native IDE!
Click to expand...

Ok. Could you try them on the pc as bescribed above to see if they are removable or fixed type?
If your SanDisk and Transcend are removable and the Samsung fixed, there would be a pattern already..
But I guess that's a bit optimistic :)

(I might be up to the different drivers how they handle this for SCSI as well..)
 
They're all marked as removable... Curioser and curioser.

I suspect the way they respond to drive inquiry would be markedly different, but I can't find a tool that will allow me to do that for Windows or MacOS...

---------- Post added at 09:47 ---------- Previous post was at 09:09 ----------

Trying DSKPROBE... But what I really want is something that discloses more information about the response to the drive inquiry.

Drive type DRIVE PHYSICAL_DEVICE
name; Sec size/sec trk/trk cyl/cyl
Sandisk 256:512/63/255/31
Samsung 1GB 512/63/255/127

Are there any tools that show the raw inquiry results?

---------- Post added at 10:13 ---------- Previous post was at 09:47 ----------

...and that don't require to boot from a DOS diskette?

---------- Post added at 11:16 ---------- Previous post was at 10:13 ----------

OK, wading through the ATA specs...

I want a tool that simply shows me the results of IDENTIFY_DEVICE (Opcode EC)

Preferably with some interpretation, but that's not mandatory...

---------- Post added at 11:34 ---------- Previous post was at 11:16 ----------

Damn, found something (bus tools) but it seems that the information returned is largely concerning the USB controller rather than the disk.

Looks like I'll have to set up a PC, hook the CF to the internal IDE port and use one of the boot disks.

Not a problem, but it won't happen today!
 
Just to make sure: You connected the CF card to your PC with the IDE-CF adapter, not a USB reader? (When connected with USB all cards report as removable.)

Talking about identify_device, I never used any such tool, so can't help there.
However I did read the other day that ATA-2 expanded the identify_device command, to accept ATAPI devices etc, and I think that would explain why ATAPI compatible amiga IDE drivers fail with removable CF cards.

Edit: Ah, you just posted again. Yes, you need to use the cards over IDE to be able to study anything.
 
Cheers Frode. I have a handy mini-itx board that can serve for this purpose... In fact I might even have an old file server with PSU somewhere so it shouldn't be too difficult!
 
OK, a couple of hours of testing...

I think there might be a pattern, but could do with a little more testing.

It seems that for a card to work, it must not support DMA, but it must support the IORDY signal.

Supports DMA = locks the SCSI bus.
Does not support IORDY = returns incomplete data for inquiry

Magic combination: no dma, support for IORDY...

Anyone corroborate this?
 
Good job Chiark, that's some serious research there!

What is the name of the software you use for this?
I will try to make a bootable USB stick to try it out with my cards. (A: is long gone...)

As for IDE, I think the SCSI drivers might also play a role in this, so it might not be up to the Yamaha alone..

According to this post prepping the card on IDE for use on SCSI might make sense as well.
(Author had the 'read drive information' feature in hdtoolbox fail when connected over SCSI, but the card worked fine on SCSI once prepared on IDE)
 
Last edited:
I tried a brand new 2GB Sandisk CF (©2004), and it didn't work. I don't think it's my SCSI controller, since a 1GB Quantum IDE drive worked just fine :dry: I'm going to try an SDHC to IDE adapter soon. Speed isn't crucial, since it's for an A500.
 
Are these cards still available anywhere? I wouldn't mind one for my A1000 phoenix. I take it they are SCSI-1 compatible without doing mods for termination voltage and parity?
 
I've seen them a couple times on ebay, probably NOS. The price is usually more friendly than the Acard and R-IDSC cards', so people snap them up pretty quickly :LOL:
I don't know about SCSI-1 compatibility, my controller is SCSI-2.
 
Just a little bump: I could not get this to work with the Taifun controller reliably - would not write - and neither with the CDTV SCSI.

Works with SupraDrive XP.
 
I just did a quick test, and the v769970 recognized my SDHC-IDE adapter with a 2 gig SD card. :D There are countless versions of the adapter, but all of them are based on the KTC FC1306T chip. So if this one works, then SD cards are a safe choice for this SCSI-IDE adapter.
I'm going to format it, do a few tests, and see how it performs.
 
Well, sometimes it works, sometimes it doesn't. I believe it's my flaky SCSI controller messing with me again. I have to get a cheap SCSI card for my PC... :dry:
 
chiark said:
Supports DMA = locks the SCSI bus.
Does not support IORDY = returns incomplete data for inquiry

Magic combination: no dma, support for IORDY...
Click to expand...

I don't know about IORDY but I have read in other URL:
https://lkml.org/lkml/2006/5/23/198

just some crazy idea that probably won't work...
I guess we would have to do the contrary to disable DMA.
CF pin 44: disconnect from IDE pin 29 (and connect to +V??? it may be dangerous?)
CF pin 43: disconnect from IDE pin 21.
(IDE pin 29 is DMAACK and IDE*pin 21 is DMAREQ)

About IORDY
CF pin 42:WAIT should be connected to IDE*pin 27 (IOCHRDY) I guess.

pinout reference:*http://pinouts.ru/DiskCables/ide2cf_cable_pinout.shtml
May someone with more CF&IDE*knowledge tell me that this is a ridiculous idea?

I have read in Slashdot a 2008 article that states that most CF cards fail to implement correctly DMA modes due to a bug in the chips they use. So I guess first/older CF will work OK thanks to total lack of DMA support (Yamaha will detect it's 100% PIO so no problems related to DMA) but post 2008 cards support DMA badly and after reporting themselves as DMA ide devices they'll fail to accomplish transfers correctly. I guess cutting both DMA signals won't be enough but perhaps it worths a try.

All in all: cutting wires 21 & 29 from IDE cable should be enough to test this easily (please note I haven't tested that yet).

I still have to try out some SDHC adapter to discover why it's refusing to work with my yamaha scsi-ide adapter. It works perfectly with ACard one.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom