The SCSI-2 Adapter /A was OEM'd by Future Domain, based on the MCS-700. Early Future Domain branded MCS-700 SCSI controllers may use the TMC-1800 SCSI controller AND the termination resistor arrangement of the MCS-600. The MCS-700 / 600 can do Narrow-Fast of up to 10MB/s to the SCSI bus -BUT- only 3MB/s to the MCA bus. Please remember the Future Domain was designed (successfully!) to be a secondary SCSI adapter.

Future Domain Ships Trio of Single-Chip SCSI-II Adapters

"Boards based on the chip can implement MFM-compatible BIOS parameters, allowing AT or MCA computers to interact with SCSI drives as if they were standard MFM hard drives. This capability will allow use of products such as PC Kwik disk caching (Multisoft) or Speedstor (Storage Dimensions) that modify drive parameter. "

Future Domain's PowerSCSI includes Win3.11 32-bit disk support... I always wondered how, since I don't remember other SCSI controllers that also support 32-bit under WfW. Also, is this due to the FD BIOS, or is it due to the chipset???

18C50 vs 1800 Based Adapters
David Beem (HAL) pipes up with:

> What are the differences between the TMC-1800 and TMC-18C50?
    The TMC-18C50 had the improvement of a jumper-selectable termination instead of removable resistor packs.

The external SCSI connection was a high-density DB-50 instead of the 'Apple' DB-25 connection of the MCS-600. Shrouded internal SCSI connector (helpful for the right orientation every time) & un-implemented solder pads for jumper settings than could adjust some SCSI bus options on the MCS-700 as well (artifacts for floppy? The FD ISA SCSI-2 has similar jumper block). The MCS-700 power connector was more conservatively rated at 1.5 amps for the 12VDC and 5VDC pins, versus the 2 amp rating on the MCS-600. Of course "C" in the chipset number means the power-saving CMOS fabrication.

    I have to see if these different chipsets return different values for  the Future Domain BIOS call "Get SCSI Controller Information" INT 13h, Function 18h. The book I have shows only values for the older FD chipsets. There is another BIOS call that determines ANSI SCSI-1 or SCSI-2 compatibility. Of course the TMS-700 will have a much newer BIOS as well. On the versions I have the boards are remarkably similar despite the 3 year difference in production. Tiny differences that add up on the finer points of manufacture.

SCSI Chips from HERE
"The TMC-1800 chip [was] completed in 1990.  The 1800 chip, which supported the 16 bit AT and MCA bus, had several design flaws that required hardware and software to work around.  As a result it lacked the compatibility of the 8 bit chip."

U2 - NS NMC27C64Q-200 65,536-Bit (8192 x 8) CMOS EPROM datasheet 
U4 8Kx8 static RAM (FIFO)

Jim Shorney pulls a few out of the weeds and says:
U7 - REG5601U or  UC80989DWP 18-line active terminator.  PDF data sheet is here:

Alternate for U4 - CY7C185-20 - 8Kx8 static RAM.  See here:

TERMENA Enables integrated terminating-resistor. Remove if both INTERNAL and  EXTERNAL SCSI - devices are connected. 

TERMPWR Enables terminating-resistor voltage and should normally left in place.

W? is for ISA version to set I/O and Memory Base Address

Passive/Active SCSI Terminating
 Passive terminating-resistors are normally fully functioning if only internal or external devices are connected. If there are internal and external devices connected you MUST use ACTIVE -Terminating on both ends of the SCSI - cables (internal/external), especially if there are fast Hard-Disks or any other FAST SCSI-2 devices ( => 10MB/s ) connected. 

SCSI-2 on IML System
   The 71G3575 will NOT support IML. They will lay an IML track, but can't access the IML partition. No end of frustration... Do not try to use these in a 90 or 95 as an IML controller. 
   You can use the IBM SCSI-2 as a secondary controller where they work fine. Hang a CD or a scanner off them. Nice to have a standard SCSI port to use (no RS6K stuff).  

Lacuna Convenience Partition
>Err ... Peter.  All three of my 77s (9577-BTG) has a convenience partition that was laid with the OEM'ed FD-MCS-700 that came standard.  Same thing goes for my 76s (9576-BNB). 

   That's what I said. The system partition will not work with the *original* Future Domain MCS-700 without the "IBM Support BIOS" .... It works on the IDE-machines (utilizing the IBM Int4B ABIOS extension hooked to generic Bios Int13h -which is the boot / harddisk interrupt- ... attached to hardware IRQ 0Eh) and on the SCSI Models only with the reworked IBM controller BIOS.  Reason why (to my opinion): the FD-controller can utilize other hardware IRQs than only 0Eh (14). 

My 9577-BTG has the IBM-version MCS-700 with the Rom BIOS 1.01 (I think) and it has the "convenience partition" as well. I had the cached SCSI in that machine as well - it also supports the partition, but is officially not supported in the Lacunas. 

Manual install-
   Control Panel>Add New Hardware>Future Domain (left hand scroll box) > Future Domain MCS-600/700 (right hand scroll box).

   Make sure the FD/SCSI-2 settings from under IBM's system programs (refdisk or setup as you want to call it) are used. W95 knows the choices available. Make sure the I/O and IRQ are correct! If not, you won't see a CD-Rom.

a)  Future Domain V3.nn = Future Domain and supports Int 13h via Int 4Ch (SCSI-CAM). Does boot-drive scan from ID 0-6. (Peter)... but does not support IBM's ABIOS functions which use Int 4Bh, which is the one that establishes / handles a "convenience / reference partition". And which is the function that reports back the attached SCSI devices. Max drive size directly controllable with the latest Future Domain BIOS (v3.61, IIRC) it's around the 8GB mark (actually 8064MB), as limited by the Int 13h BIOS call parameters' max. values (1024-cylinders x 256-heads x 63-sectors x 512-byte sectors).

a1) Future Domain v3.4 - v3.5 (Reported by Cameron Labut)
  Cameron was experiencing an odd behavior with a 8580 20MHz system (Busmaster capable), 4MB on-planar, XGA1 w/512KB, and a Future Domain w/ V3.4 BIOS.

The system would boot with the XGA1/512KB, HDs and CDROM accessible. He swapped in an XGA2 (1MB stock) and the HDs were inaccessible (but shows up on FD boot message). CDROM is still accessible. This happened with the original 8580 refdisk and with the refdisk patched with XGAOPT.EXE . No XGA2 error messages were displayed during POST, and the XGA2 passed advanced diagnostics...

After swapping in an IBM Patriot 1.01 BIOS, the XGA2 and all SCSI devices were accessible at boot.

Ed. As we totally lack either the 18C50 or Future Domain BIOS references, we can't be sure what the incompatibility is from. The incompatibility is most likely NOT Busmaster related, since the 512KB XGA1 co-existed with the MCS-700. Since the MCS-700 is PIO, that removes another possible conflict. No XGA2 error messages suggest that the patched refdisk files are compatible with the XGA2 BIOS / hardware.

The only thing not fully explored is IF a FD MCS-700 w/3.4 and XGA1 w/1MB was unable to access HDs. Hours spendt scouring the internet only provided ONE possible significant issue, that of HD ordering. This does not mean the HD ordering was the only issue with V3.4, but suggest that V3.4 had issues.

Future Domain BIOS 3.4 and 3.5 Hard Drive ordering
Please note that the drive ordering that Future Domain implemented in BIOS versions 3.4 and 3.5 is the opposite of the order (currently) used by the rest of the SCSI industry.  If you have BIOS version 3.4 or 3.5, and have more than one drive, then the drive ordering will be the reverse of that which you see under DOS.  For example, under DOS SCSI ID 0 will be D: and SCSI ID 1 will be C: (the boot device).  Under Linux, SCSI ID 0 will be /dev/sda and SCSI ID 1 will be /dev/sdb.  The Linux ordering is consistent with that provided by all the other SCSI drivers for Linux.  If you want this changed, you will probably have to patch the higher level SCSI code.

b) IBM V1.0n = Supports Int 13h via Int 4Bh (IBM SCSI). Does boot-drive scan from Id. 6-0 and supports RefDisk Config and Diags. If using the IBM v1.01 BIOS, it has a max drive size of 3.94GB (1024-cylinders x 255-heads x 63-sectors x 512-byte sectors), again IIRC.

>     For using the MCS-700 as a secondary controller, allowing drivers to be loaded, how big can the secondary drive be?

    This is only limited by the driver's and OS's design, but has limits set by a 32-bit "Logical Block Address" (LBA) of 4 Gigablocks and the assumed 512-byte logical block size = 2 terabytes. Check your OS doc.s and any READMEs for the driver for that OS.

(Peter) 
   1.00 seems to have limit at 4GB and -probably- with ATAPI CD-ROMs. This is a "Lacuna"-specific problem when you have the harddisks attached to the SCSI controller and an additional IDE CD-ROM on the planar port.

   I'd tried that on a machine with IBM Controller BIOS Rev. 1.0 and the system refused to even recognize the CD-ROM. I switched to a 1.01 controller and -voilá- there it was. However: when I set the CD-ROM to "Slave" it failed to work properly even with 1.01 on the SCSI controller.

   There seem to be dependencies within the Boot-BIOS part of the IBM SCSI Bios on the FD-controller. 1.01 works fine with bigger HDs and CDs ... but dislikes CD-ROMs solely attached to the IDE jumpered as "Slave".

Finally, Peter sez:
   The FD MCS-600/700 can be upgraded to an IBM SCSI-2 with the IBM ROM. The two only differ by the DC-plug that the FD has and the IBM lacks. The FD was originally designed as an "upgrade controller" to add to an existing system which might not have a free DC-plug. The IBM version was intended as "additional controller" (e.g. for tapes in a MCA Server) or sole SCSI controller as in the Lacunas, which have enough DC-plugs coming from the power supply. Just in case anyone wonders why IBM saved the few pennies for the DC-plug

Future Domain MCA SCSI Adapter BIOS versions

18MC means 1800, otherwise the table wraps

BOARD     BIOS
------   ------
        1.0|1.2|1.21|350 2.0|18MC 1.0|18MC 2.0|18MC 3.0|3.2|3.4|3.5|3.6|

MCS-350  X | X |  X |   X   |        |        |        |   |   |   |   |

MCS-600    |   |    |       |    X   |    X   |   X    | X | X | X | X |

MCS-700    |   |    |       |    X   |    X   |   X    | X | X | X | X |

SCSI-2 and P70 ESDI Adventures
Jeff Hellige vents and sez:
  I've got my P70 running again under OS/2 Warp Ver. 3 and regardless of the slot it's placed in or the configuration, the MCS-700 seems to be conflicting with the onboard ESDI controller in protected mode.  POST is reports the following error on bootup:  1047000 221 (ESDI Controller Wrap Failure)
 
        If the SCSI board is left in, OS/2 will run for a while and then start locking up, which makes sense if the conflict is in Protected mode.  DOS will run without errors.  The Reference diskette diagnostics configuration report shows the wrong configuration with the board install and I've not attempted to go any further with the SCSI board installed.  With it removed the diagnostics goes all the way through without errors, showing the 
correct configuration. 

        The ROM version on the MCS-700 is 1.01.  I've tried it in both slots as well as tried changing the IRQ and such in the Reference disk setup.  All of this had no effect on the error.

   1) reinstalled the MCS-700 with the IBM BIOS 1.01 still installed. It continued to give the error and I ran diagnostics from the Refdisk.  The SCSI test gave an error of 0210000U.

   2) I removed the BIOS and reinstalled.  PowerSCSI4 would not install without a device hooked to the card so I connected an external 1gig SCSI hard disk to it.  Drivers installed fine under both DOS and OS/2.  No errors on boot and I'm able to access the drive fine under both operating systems as well.  It took 4-1/2 minutes to copy 64meg of data from the internal DBA disk to the external SCSI disk.

   I had tried to disable the BIOS in the system setup on the Refdisk, but it didn't help.  Removing the BIOS chip altogether seems to have fixed it though.  My only complaint is that it insists on formatting the external hard disk with 32k sectors!  I'll have to play with that some more.  Now the question is, what functionality have I lost by removing the BIOS chip?  Am I correct to assume that it won't be possible to boot off a SCSI disk in this configuration?

FD w/o ROM
Tony roars with:
   Setup a MCA flavor S/320 with one of the FD's (minus ROM) running an Archive Viper tape drive.  Nothing dramatic happened - it just worked. Nice to free up one of my scarcer v1.01 IBM ROMs so I can replace the brain dead v1.0 in something else. BTW, autoconfig seems to want to allocate a ROM address for the adapter by default.  I just went in after autoconfig ran and disabled the (nonexistant) ROM.

Tim Clarke
   That's because the ROM actually has only 6KB mapped-in and the controller chip has a 2KB "buffer" that is configured to be contiguous with it, to make up the 8KB total "ROM" allocation. I'm not sure if the "ROM Disabled" configuration means that the buffer is too, causing some extra I/O overhead, or not.

Adapter Memory Location
   Memory location used for the BIOS ROM 
    <Segment CA00>, Segment CE00, Segment DE00, Segment C800

Adapter I/O Location
   I/O location the adapter will use 
     0140, <0150>, 0160, 0170 

Select Interrupt Line
   Interrupt used by the SCSI controller 
     <Int 5>, 10, 11, 12 (Mouse), 14 (HDD), 15 (Rsrvd), 3 (Serial Alternate), Int Disabled

ADF File for MCS-700 /IBM MC SCSI-2 adapter (18C50 VLSI)
60E9 IBM PS/2 SCSI-2 Adapter/A or MCS-600/700 

Adapter ROM BIOS Address
    Memory address for ROM BIOS.  Generally, the BIOS is enabled to support fixed or removable SCSI disk drives.  If you use PowerSCSI software, and if the SCSI devices attached to this controller are tapes, CD-ROM drives or non-direct access devices, the BIOS may be disabled to speed system startup 
     <Segment CA00>, CE00, DE00, C800, Disabled 

Adapter I/O Port Address
      I/O port addresses the adapter will use 
     <0140h>, 0150h, 0160h, 0170h 

SCSI Adapter Address (ID)
      SCSI ID of the adapter is fixed and cannot be changed 
     <7>

Select Interrupt Line
    Interrupt used by the SCSI controller 
     <IRQ 5>, IRQ 10, IRQ 11, IRQ 15, IRQ 3, IRQ 14, Disabled 

The Magic Christian responds:
This requires a Future Domain MCS-600/700, IBM OEM'd version (Patriot),  or something else that can run the DOS utility FDDSU.EXE that comes with 'Powerscsi4'

 'The following procedure should read the firmware parameters from a SCSI drive and then write those parameters back to the media. This will normally restore a fixed disk to the factory default parameters. Not all drives will support this procedure. Future Domain will not be responsible for the results stemming from the use or misuse of this procedure.

1. Insert the Future Domain "SETUP" utility.
2. Type "SETUP" and press <RETURN>.
3. A screen will appear displaying the SCSI ID  and LUN of the drive.
4. A menu will appear as follows:
    1. Format Unit
    2. Edit Defect List
    3. Surface Analysis
5. Press the <F5> function key to invoke the Custom Utility menu.
6. Message appears "you are about to enter the Custom Utility...", answer yes to continue.
7. A menu will appear as follows:
    1. Format Unit
    2. Edit Defect List
    3. Surface Analysis
    4. Sense Byte Editor
8. Choose option 4. Sense Byte Editor.
9. A menu appears as follows:
    1. Mode Sense
    2. Write sense data to a file
    3. Read sense data from a file
    4. Print current list
    5. Edit current list
    6. Set options
10. Choose option 6. Set options.
11. A window appears and asks if you want to change Mode Select byte 1. Answer no.
12. You are asked if you want to change Mode Sense byte 2. Answer yes.
13. Enter the hex value "BF" and press <RETURN>
14. The Sense Edit menu appears. Select option 1. Mode Sense.
15. Now select option 5. Edit current list.
16. Window opens up on left side of screen. These are the Page codes of the SCSI drive.
17. Press <F7> function key. Message should say "Sense Info sent successfully".
18. Press <ESC> twice to exit back to the main menu.
19. Choose option 1. Format Unit.
20. When asked "Permission to format", answer yes.

   This should update the media with the parameters from the firmware. The drive must accept and finish the low-level format for the above procedure to work correctly. Also, some drives do not support a low-level format. When in doubt check with the drive manufacturer. "

OS/2 Switches
The following drivers support Future Domain and IBM SCSI host adapters: 
      FD8XX.ADD supports Future Domain and IBM 8-bit SCSI adapters (ISA!)
      FD16-700.ADD supports FD 16-bit SCSI adapters 

BASEDEV= ---- FD8XX.ADD --------------------------------------
       - FD16-700.ADD -| - /ET -| Search SCSI devices for Logical Units
      +- FD7000EX.ADD -+ - /FS -| Enable fast synchronous data transfers 
                                                                                           (TMC-1850 chipset ONLY)
                       +- /RD:n+ Maximum device recovery time
  -------------------------------------------|
        +- /A:n ----------------------+ Adapter number
                    |- /!DM ----------|  Disable DASD Manager
                    |    +- :unitlist +  List of unit identifiers
                    |- /!SM  ---------| Disable SCSI Manager
                    |    +- :unitlist-+ List of unit identifiers
                    +- /ET  ----------|Search for Logical Units
                        +- :unitlist + List of unit identifiers

U12 UM6116-3 Unicorn Microelectronics 2K x 8 CMOS SRAM datasheet

After looking at an MCS-350, there is no discrete SCSI controller. I >assume< the controller functionality is within one or more PALs.

AdapterID 5F77 Future Domain SCSI Adapter

Memory Location
   Memory location used for the BIOS ROM
     <"Segment CA00" (ca00-cbff)>, C800 (c800-c9ff), CC00 (cc00-cdff), CE00 (ce00-cfff), D000 (d000-d1ff), D200 (d200-d3ff), D400 (d400-d5ff), D600 (d600-d7ff), D800 (d800-d9ff), DA00 (da00-dbff), DC00 (dc00-ddff), DE00 (de00-dfff)

DMA Arbitration Level
   DMA channel used to transfer data.
     <"Level 6">, 7, 5, 0, 1, 3, 4

Select Interrupt Line
   Interrupt used by the SCSI controller
     <"Interrupt 5 (Reserved)">, 3 (Serial Alternate), 10 (Reserved), 11 (Reserved), 12 (Mouse), 14 (Fixed Disk), 5 (Reserved)" 

Use Front Panel Disk Busy Light
   Whether the front panel light is to be used by the SCSI devices to indicate that a SCSI device is busy.  The same light is also used by the internally installed hard drive.  There is no conflict if the same light is used by both devices.
      <"Use Front Panel Light">, Do Not Use Panel Light

Use MC BUS Wait (IBM Model 80)
   Extended synchronous bus cycle is used as the default fastest cycle on the transfer of DMA data to the SCSI device.  The Model 80 will not support full speed DMA writes via the uChannel bus, so this option is required for high speed devices on the Model 80.
      <"Use Wait State (Model 80)">, Do Not Use Wait State

Termpwr Fuse
   The Termpwr fuse is a PTC Resistor which goes to high resistance if too much current flows while providing Termination Power to the SCSI devices. When the overload is removed, the PTC resistor cools down and allows normal operation.