According to the FCCID https://fccid.io/JK9VL200IDEIO , possibly connected to a company call Sporton out of Taiwan but not much luck driver-wise 🙁

I saw one eBay seller who currently sells this particular card right now. https://www.ebay.com.au/itm/114255314650

The seller has been kind enough to manually supply me with each page of the manual, photographed one by one. I have attached it here for future reference.

The seller also happens to have the driver disk listed together with this item. Perhaps, if you ask him (Alex is his name) to send you the content of the disk, you may still have the chance. 😉

amang that is AWESOME news. Thank you so much!

Got the drivers! I'll test soon and upload them to Vogonsdrivers. So excited!

Seems like this might be a nice controller after all — it mentions support for 50MHz FSB at 0 wait state

Even with the big fake chip sitting in there. For the odd tens of bucks it's nice, for the original price of several hundreds, not so me thinks...

jheronimus wrote on 2020-06-27, 13:50:

Got the drivers! I'll test soon and upload them to Vogonsdrivers. So excited!

Seems like this might be a nice controller after all — it mentions support for 50MHz FSB at 0 wait state

Cool bananas! If you could upload them, that would be great. Does it mean a motherboard with 40Mhz base clock can utilize this card? I have an AMD486 120Mhz CPU that I am planning to use, but somehow this card just refuses to work when the CPU is set to 120Mhz (3 x 40Mhz FSB).

jheronimus wrote on 2020-06-27, 13:50:

Got the drivers! I'll test soon and upload them to Vogonsdrivers. So excited!

Seems like this might be a nice controller after all — it mentions support for 50MHz FSB at 0 wait state

How did you go with the test? Is there any major improvement in terms of the speed?

amang wrote on 2020-07-03, 00:16:

jheronimus wrote on 2020-06-27, 13:50:

Got the drivers! I'll test soon and upload them to Vogonsdrivers. So excited!

Seems like this might be a nice controller after all — it mentions support for 50MHz FSB at 0 wait state

How did you go with the test? Is there any major improvement in terms of the speed?

Well, the first attempt failed: the system could not boot properly once the HVLCACHE.SYS loaded.

I'll try again with a fresh installation — I had a lot of software (QEMM, Norton Utilities, 4DOS) that potentially could cause a conflict.

For the moment I'll share it via Dropbox. I'll post it to vogonsdrivers once I or anyone else will confirm it to work.

jheronimus wrote on 2020-07-03, 09:11:

Well, the first attempt failed: the system could not boot properly once the HVLCACHE.SYS loaded. […]
Show full quote

amang wrote on 2020-07-03, 00:16:

jheronimus wrote on 2020-06-27, 13:50:

Got the drivers! I'll test soon and upload them to Vogonsdrivers. So excited!

Seems like this might be a nice controller after all — it mentions support for 50MHz FSB at 0 wait state

How did you go with the test? Is there any major improvement in terms of the speed?

Well, the first attempt failed: the system could not boot properly once the HVLCACHE.SYS loaded.

I'll try again with a fresh installation — I had a lot of software (QEMM, Norton Utilities, 4DOS) that potentially could cause a conflict.

For the moment I'll share it via Dropbox. I'll post it to vogonsdrivers once I or anyone else will confirm it to work.

Could you attach them to a post here?

I've played around with VL cache controllers and they are pita to work on. If you tried to load an existing system with the controller that might not work, it can in some cases require full format of the drive.

EDIT: Attached to this post.

jheronimus wrote on 2020-06-27, 13:50:

Got the drivers! I'll test soon and upload them to Vogonsdrivers. So excited!

Seems like this might be a nice controller after all — it mentions support for 50MHz FSB at 0 wait state

Speaking of 50Mhz, if you notice between page 8 and 9 of the manual, there is conflicting information about how VL-Bus is normally limited to operate at 40Mhz, whereas page 9 shows specifically how this card can operate up to 50Mhz base clock.

Anyways, I managed to test the drivers myself. Three conditions have to be met in order for these drivers to work:
1/ BIOS has to be set to recognize your HDD as NORMAL drive, not LBA or LARGE. Setting your HDD other than NORMAL will cause the controller not being able to read the drive properly under DOS. A simple test is simpy to run a DIR command for one of the directories in your drive;
2/ No drive overlay is allowed in your HDD. So it's best to have your HDD in plain vanilla mode.
3/ No overclock or fancy tweaking on your motherboard / BIOS setting. I have my Am486 DX4 WB running at 100Mhz (33.3Mhz x 3)

I have a 2GB CF card for this test, but somehow my DOS system only recognized up to 504MB.

I ran two benchmarks (Speedsys and HDDspeed) under 3 scenarios:
1/ CF card installed on VL200 with those drivers loaded in DOS
2/ CF card installed on VL200 without drivers loaded
3/ CF card installed on a standard VLB controller card.

Results are as follows:

** SpeedSys

\ Scenario (1)
Random access time: 0.65ms
Buffered read speed: 1,824 KB/s
Linear verify speed: 256,146 KB/s
Linear read speed: 1,804 KB/s

\ Scenario (2)
Random access time: 0.78ms
Buffered read speed: 1,520 KB/s
Linear verify speed: 128,500 KB/s
Linear read speed: 1,507 KB/s

\ Scenario (3)
Random access time: 0.81ms
Buffered read speed: 1,453 KB/s
Linear verify speed: 128,419 KB/s
Linear read speed: 1,442 KB/s

** HDDspeed

\ Scenario (1)
Read Linear Speed at Track: 1.8 MB/s
Read Min Linear Speed: 1.8 MB/s
Read Max Linear Speed: 1.8 MB/s
Max Cache Read Speed: 1.8 MB/s (64k block)

\ Scenario (2)
Read Linear Speed at Track: 1.6 MB/s
Read Min Linear Speed: 1.6 MB/s
Read Max Linear Speed: 1.6 MB/s
Max Cache Read Speed: 1.6 MB/s (64k block)

\ Scenario (3)
Read Linear Speed at Track: 1.5 MB/s
Read Min Linear Speed: 1.5 MB/s
Read Max Linear Speed: 1.5 MB/s
Max Cache Read Speed: 1.5 MB/s (64k block)

I hope these findings would provide some insight into the usefulness of this card. 😀

Well, I've tested it on a Intel DX4-WB motherboard running at 3x33, but my hardrive is a 2.5GB Quantum Fireball 😀 Not sure if I'll have time to make a clean installation using a smaller capacity.

But since amang has proven that these are the correct drivers, I've posted them to Vogons Drivers library thread. Thanks for all the help!

Well, INIVL200.SYS is quite interesting. It is a hacked version of SMARTDrive 3.13, as delivered with MS-DOS, but it probes for the presence of the Hornet VL "Cache" Card, and does not handle any command line arguments. In XMS mode, it is set to 512KB maximum cache size, while in EMS mode it would use up to 8MB (the default value by smartdrive). Just patching the driver name from "VL200_AC" to "SMARTAAR" in the driver file at offset 10 decimal should be enough to have it controllable using the same IOCTLs as DOS 5.0 SMARTDrive. For some reason, the last IOCTL write function (set original INT13) is missing in the dispatch table, although the function number check still allows the function, and the code itself is also still present. If you insert two dummy bytes at offset 224 (hex), the similarity to the original SMARTDRV.SYS (at least the resident part) is obvious. Most different bytes are just caused by the 2-byte offset resulting from the missing table entry.

INIVL200 claims to be an ADI2 timing and mode register initialization tool, but it does neither timing nor mode register initialization.

What's the point of bundling a software disk cache with a card with RAM slots ? Are those slots just for show ?

This card seems like nothing more than an elaborate scam .

darry wrote on 2020-08-16, 17:50:

What's the point of bundling a software disk cache with a card with RAM slots ? Are those slots just for show ?

This card seems like nothing more than an elaborate scam .

As already quoted in Re: Bought these (retro) hardware today, a german computer magazine already got to the same conclusion in 1994. Furthermore, it is known that the biggest chip, that is supposed to look like a controller chip, is not connected to anything. If the card were meant as a scam from start, I don't think they would have put that much of TTL components on the cards. My guess is that they tried to create a hardware cache controller, but engineering failed to deliver. I am undecided on whether engineering cheated management and claimed to "finally got the cache working", or management pushed the engineers to deliver at least something that looks like a cache controller so that they can show it to their investors or customers.

Most likely the RAM slots are actually accessible from the host CPU, as the BIOS is known to initialize the cache memory and fails to boot if no RAM is installed. The INIVL200 also checks for some I/O ports that are definitely not provided by the ADI2 chip.

As I don't know yet what HVLCACHE.SYS does, it might do something with the onboard RAM. Also, the BIOS might do something with the onboard RAM. I doubt so, though.

HVLCACHE.SYS is an IDE Hard Drive BIOS with 32-Bit I/O support. It supports extra BIOS functions, AH=C4,C5,C6,E0-E6,E8,EC,EF that run the IDE command given in AH. If block mode is sucessfully configured using C6 (Set multiple), BIOS functions 2 and 3 (read and write) use read multiple/write multiple instead of the standard read/write commands. It has no knowledge about the ADI2 chip, it does not program any timings, and it does not interface with the RAM using I/O ports 130/230/330/300.

It might be possible that the cache RAM is completely accessible using ports 130/230/330/300 (depending on jumper settings). Both of the DOS drivers do not do that, though. If the cache RAM is accessible, SMARTDrive (or another cache) could be patched to use the cache RAM instead of XMS/EMS. Another possible use of the cache RAM would be a RAM drive.

CHKIDE.COM seems to be just a geometry dumping utility that compares the native geometry to the geometry in the drive parameter table, and HDDIAG.EXE seems to be just a standard IDE low-level format utility (with the possibility of setting the interleave factor). I wonder who needed a tool like HDDIAG in 1993 (the production date of that Hornet controller). In 1989, it might have been useful for the drives in use at that time, though. In this case, the manual correctly describes the function of the utilities.

Best find so far: INITADI2.SYS from the netware directory. Again, this driver does not do anything about the cache part of the Hornet controller, but it is a sensible initialization tool for the ADI2 VL controller chip. It has a small database of built-in timing values, and applies PIO0/1/2 according to the IDE identify data for other drives. It also enables "read-ahead" for drives that support block mode, and disables "read-ahead" for drives that don't. It measures the VL clock to convert nanoseconds into VL cycle counts, so it works independent of VL bus clock.

I have no idea why the real ADI2 initialization driver is suggested for booting NetWare (remember: NetWare uses DOS as boot loader), but is not used when booting into DOS. If the BIOS initializes the controller timings.

IDE.DSK from the netware directory looks like a slightly adapted version of Novell's own IDE.DSK from around NetWare 3.11, but with the copyright message patched to say "Hornet", so no interesting stuff in that.

It seems I analyzed all possibly interesting software as far as it is useful for understanding the card and the software. I would also look into a BIOS dump - maybe the BIOS dump shows something about how to access the cache RAM from the host CPU.

There is no Cache accessed.
The Cache is fake.
This is Fake Cache Controller.

If you place SIMM Modules on it, they are unused.
I also own this Card.

Can you extract the contents of its ROM please?