Fast/Wide RAID Controller

"Passplay" and Mylex DAC960M

@8FBB.ADF IBM Fast/Wide RAID Controller "Passplay"
@8F6C.ADF DAC960M RAID Adapter

4227110.EXE FWSR Option Disk #1, v2.31 (zipped image)
   Contains Disk Configuration Utility for "Cheetah" and "Passplay"
4227111.EXE FWSR Option Disk #2, v2.31 (zipped image)
   Readme for FWSR Option Disks
10L9849.EXE RAID Supplemental Diskette v2.0 (OS/2 command-line interface)
   Readme for Supplemental Diskette

DAC960_W95.ZIP MYLEX DAC960 Protected Mode Driver for W95 Combined DAC960 Upgrade disk Rev D1 May not apply!
   FW2.214 2.14 firmware upgrade for the 1/2 Channel DAC960
dacmcam1.exe DAC960M Drivers and Utilities Diskette, version M-1
daccf471.exe Configuration & Utilities for all DAC960 Server 95 Array RAID API Module (Non-ASPI)

Mylex files in Taiwan (list archived but all files dead?)

SCSI-2 RAID Controller
Mylex DAC960M
Passplay Configuration
BIOS Updates
Function of NVSRAM
Cable Parts
IML Limits of Passplay
HD LED Does Not Work
Cyrix/Non-SOD Type 1 Incompatibility?
Cache Size
   Generic 30 Pin SIMMs for Cache
Accessing the RAID Configuration
   Configuration Utility
Saving a Passplay Flashed with Cheetah Microcode
   BIOS Release Features
   Complex BIOS Levels Required
Slots Passplay Fits
   More than one Passplay in a System?
Low Voltage Differential (LVD) Drives
Number of Devices Supported
Running RAID Without Trays 'n Bays
Passplay and CD under NT 4 (yes, it works!)
Passplay Under W95 (it works!)
Mylex DAC960M Firmware
Specifications For FWR
ADF Sections - Passplay
ADF Sections - DAC960M

Fast Wide RAID Adapter "Passplay", "DMC960", FRU P/N 92F0335 [P] [P]

Mylex DAC960M

F1 PTC fuse (SCSI Channel 1)
F2 PTC fuse (SCSI Channel 2)
J1 Status LED header
J5 Unknown, probably factory use
J6 Unknown
P1 External SCSI port (Mylex only)
P10 SCSI Channel 1
P12 SCSI Channel 2
U5 Firmware
U9,15,21,29 UC5601DWP
U16 8Kx8 NVRAM
U17,18 NCR 53C720
U31,44,52,60 30-pin SIMMs
U55 i960CA-25
U57 93F3167 or 34G1520 or 34G1521 "Miami"
X1 40.0000 MHz osc
X2 25.0000 MHz osc

There is NO external SCSI port (P1) on the Passplay! Just the solder pads for one.

DAC960M external port is a little odd. It doesn't use a sidecard like the Cheetah, but instead has a cable attached to the front port (Channel 1) that goes to an HPDB68 port that is mounted to the slot cover.

U16 Benchmarq bq4010YMA-200 or Dallas DS1225Y-200 8Kx8 NVRAM (alt)
U9,15,21,29 UC5601DWP 9-Line Active SCSI Terminator
U15,16 NCR 53C720 Datasheet | Programming Guide
U55 Intel A80960CA25, Intel A80960CF (substitute), i960 Performance Brief
U57 93F3167 or 34G1520 or 34G1521 "Miami" MCA iface

i960CA is pin AND function compatible with the i960CF. The i960CF adds a 4KB cache (or something). People have swapped out the CA with a CF for @ 10% increase in performance.
Note: The i960Hx seems cool, but lacks a DMA controller, some pins are used for different purposes, and though 5V tolerant, needs current limiting on the Vcc line. No Holy Joy here, move along.

J1 is a 6-pin header that provides a connection for three status LEDs. Pin 1 is to the left end of the board. In each case the odd-numbered pin is the +5V source. An  external series resistor is not required for the LEDs.

LED Indicators (meaning if ON)

Pin Name Meaning
1-2 SCSI Activity One or two SCSI channels xmitting or recving data.
3-4 Write Pending DAC960 xmitting/recving data to/from host system.
5-6   DAC960 cache has data more current than on HDs.

DAC960 LED Error Flashes

  When any DAC960 controller is powered up, the 960 CPU performs a sequence of tests on its hardware. Failure in any of these tests is indicated by an error-code by flashing the Write Pending  LED on the DAC960 controller. It is also accompanied by an appropriate BIOS error message on the system. The memory parity error  (irrecoverable), during the operation of the card, is also indicated by the same LED.

DAC960 Write Pending LED post on DAC960
1 or 2 Channel Header J2 (non-populated)
3 Channel Pin 5-6 of 6 pin header J1
5 Channel Pin 5-6 of 6 pin header J1

The LED blinks will always be repeated at regular intervals.

LED Pattern Error Indicated
Constant ON/OFF at 50% duty cycle Parity error on DRAM. FATAL Error. Check DRAM.
2 blinks DRAM error. FATAL Error. Check DRAM.  DRAM support logic failure is likely.
3 blinks BMIC I/F failure. FATAL Error. 
4 blinks Firmware checksum error. Try loading Firmware using DAC960FL utility.
LED stays on at power-up Controller dead.

SCSI Channel Equivalents (non-bay values)
SCSI Ch 1 set IDs as 0-1-2-3 (CD, HD1, HD2, HD3) top 5.25 single bay, triple RAID bay.
SCSI Ch 2 set IDs as 1-2-3 (HD4, HD5, HD6) bottom triple bay
More detail BELOW.

Passplay Configuration

Both the FWR (Passplay) and FWSR (Cheetah) are only configurable through the RAID Utilities disk. You CANNOT see the SCSI Disks under "Set and View SCSI Devices" like normal SCSI drives. Boot with FWSR Option Disk, #1 ver. 2.31 in order to view or configure the array.

Both adapters use the same Utilities disk since both are based on Intel i960 / Mylex / NCR technology. The RAIDADM (manager) should work on both /A-adapters.
FWSR Option Disk, #1 ver. 2.31
FWSR Option Disk, #2 ver. 2.31
Readme for FWSR Option Disks
RAID Supplemental Diskette Version 2.0 (OS/2) RAIDSEND OS/2 ONLY command-line interface for performing various tasks on a IBM F/W RAID Adapter/A
Readme for RAIDSEND

BIOS Updates

Caution! The Passplay and the Cheetah differ in the microcode, which may not be interchanged. The Passplay (FWR) adapter uses a microcode-level 1.6x through 1.99, the Cheetah (FWSR) uses 2.xx levels. If you flash the one adapter with the code from the other you end up in non-functional adapters. Check THIS for (unverified) recovery method.

SR Flash BIOS ver. 1.61 For RAID controller WITHOUT external port! (zipped image)
SR Flash BIOS ver. 1.62 For RAID controller WITHOUT external port! (zipped image)
SR Flash BIOS ver. 1.63 For RAID controller WITHOUT external port! (zipped image)
SR Flash BIOS ver. 1.99 For RAID controller WITHOUT external port! (zipped image)
SR Flash BIOS Readme (for ver. 1.99)

NVSRAM Function
   Each NVSRAM has a lithium energy source and control circuitry which constantly monitors Vcc for an out-of-tolerance condition. When it occurs, the lithium energy source is automatically switched on and write protection is unconditionally enabled to prevent data corruption.

Cable Parts
   The mini C68 for the Channel edgecard connectors is the Molex 71660i, part# 15-92-3068, called a half pitch Centronics, or a VESA Media Connector. Suprise! AMP makes a similar part (mini-C68) AMP Part 1-557089-2 Any cable with a .025 pitch, 28 to 30 AWG will work with either connector.

A Better Cable Hack?
Allen Brandt wrote:
> A small, shoddy attempt to get something uploaded concerning the PS/2. HERE.

My Take on it:
   I am starting to have neurons fire. Actually, Allen provided the push. Al went and slit the conductors for better flexibility (in pairs).

Could you slit the flat cable up towards the controller and get the very flexible cable bundle of the IBM original? The black sheathing is available from Jameco for about $1 a foot. Well worth it, IMHO. (Start the slit with an X-Acto and use the reverse of the blade to finish parting the conductors???)

The sheathing is Techflex Cable Sleave, looks to be the 3/8" size. Sold in a 25' spool. Part #162157, Product # CCPT2X per spool $14.95 Techflex is HERE
What kind of signal degradation might occur? Each signal pair hopefully cancels it's noise out. If the Brandt maneuver can be done from the top drive connector to the adapter, it might be a close match to the real thing.

IML/Boot Limits of Passplays
   The Passplay was introduced for use with non-Flash systems. However, the Passplay does NOT support Int13 or IML. You either have to boot from a floppy or from an IBM IML capable SCSI adapter. A telling sign is the Type 5 form factor card cutouts in the case are on Slots 2-4. A Type 3 form factor SCSI Adapter has to be installed in Slot 1 to support Type 1 and Type 2 complexes that do NOT have the enhanced BIOS.

HD LED Doesn't Work
>Is it a fact that the HD LED does not work on a 9595A with a PassPlay RAID Adapter?
   The fixed disk light is non-functional with both the Server 95 A "Passplay" and Streaming-RAID "Cheetah" MCA RAID adapter. I suspect this is also the case with other OEM'ed Mylex RAID adapters.

Cyrix/Non-SOD Incompatibility?
Tim Clarke says:
   Hi gang, Just thought that I'd better warn you. After checking out the Cyrix 5x86 at 4x clocking (in Type-1 non-SOD w/cache) my PassPlay RAID adapter seems to have been "duffed up". I only get a part of the BIOS v1.05 initialization/installation message and the machine hangs (with *any* CPU) at CP:96. Looks as though the Flash ROM has been partially overwritten (just a guess).

Cache Size

Go HERE for the details.

IBM or Generic SIMMs?
   They are 30-pin standard industrial ("generic") SIMMs. Since the original concept allowed 4x1MB, 4x4MB and even 4 x 16MB cache SIMMs they *must* be generic, because IBM only coded the 256K, 512K and 1MB modules. The 4MB and 16MB are not on IBM's list.

Note: All non-SurePath (flash based) systems require a standard SCSI adapter or system board resident SCSI controller connected to an IML and/or boot hard file. The lone exception is the Type 1 486DX2-66 Upgrade complex with flash ROM.

Saving a Passplay Flashed With Cheetah Microcode

From Peter:
   During an IBM technical class after the introduction of the Server 500 the instructor told us that you *could* accidentally flash a Passplay with the Cheetah firmware - and make it non-functional with that.

He said: "You need a Cheetah adapter to flash back to Passplay level. The Passplay with the Cheetah-firmware will refuse to re-flash. Remove both - the Flash-ROM and the EEPROM - and stuff it into a Cheetah adapter. Then run the Passplay Flash diskette to reprogram it to level 1.6x (Ed.: or 1.99 actually). Then re-install the two chips on the Passplay and it will work again."

Never tried it myself however.

Passplay Code Releases

Hakan Gadler asked: By the way, why did they jump from version >1.6 something to 1.99?

From Peter:
   IBM "unified" the version numbers for the BIOS and the microcode a bit. Or - to be precise - the return codes. The later Passplay codes ran along with the same RAIDADM software that was used by the "Cheetah" and contained some better infos and other tweaks for performance / safety reasons.

1.60 Initial code worked only with early Passplay board releases.
1.61 The Passplay was widely available with this. Some trouble with drives other than the IBM 1.0GB (mostly the Maxtor MXT-540S) in the "small" array configuration).
1.62 Mandatory for use with 90MHz upgrade board
1.63 Sometimes drives were set to "DDD" (Dead) state when they failed to come on ready in a very short period of time after power on. 1.63 should fix problems with the DFHS 2GB drives (later microcodes - successor of the 0664) and the "dead drive" symptom. AFAIR IBM recommended 1.63 in all machines with more than 3 drives installed and Pentium platforms (My guess: problems with the PSU and signaling problems caused by DC-ripples). From 1.63 on you could use the same RAIDADM and later ServerGuide RAID Manager for both adapters. Previous versions seemed to have delivered slightly odd codes that caused confusion.
1.99 IBM announced the code "out of blue sky" and I wondered why, because the Passplay was already discontinued at that time. The 1.99 codes contained some fixes for "other systems" than the 95A. It was the last code announced for the Passplay.

Complex BIOS Requirements

There are however some dependencies between the RAID adapter microcode and that of the complex. You should not run the RAID with a complex BIOS below 03. The BIOS 10 has been announced to fix Y2K problems with OS/2 AFAIK. If you flash the complex to10 and keep the old RAID microcode you might run into problems. It should be 1.63 at least - especially if you have the P90 platform, which appears to the RAID microcode as Server 500 with the differing backplane layout - the return codes to the RAIDADM then might not reflect the "real" position of the drives on the backplane.

Slots Passplay will Fit

The Passplay is a Type 5 form factor card (it's significantly larger than the usual MCA adapters). There are cutouts in 95 and 95A cases that will allow the edge of the card at the bracket end to fit. Only slots 2, 3, and 4 have these cutouts.

The 95s don't have a cushion in them, 95As do.

More Than One Passplay?

>An interesting note is that the 9595 has three enlarged slots to fit in three Passplay adapters. Or, I guess, to move the one Passplay adapter to the desired slot.

From Peter:
   The machine has slotted rear wall in the positions 2, 3 and 4 that allows installation of the oversized cards like Passplay. Since there are limitations in the MCA "package density" I would say the positions are made to move cards around rather than installing 3 of them. As far as I know the MCA cooling density does not allow more than 2 cards of that size to be installed - and even then the middle slot must stay free to allow sufficient airflow between the cards.

As I wrote: it does not make sense to install more than one of the Passplay adapters anyway. For the Cheetah, which has a rear port and can use the "sidecar card" to route the Channel 1 to the rear as well for e.g. a 3518 expansion unit (and a total of 14 additional devices) things look different.

But the Cheetah isn't an oversized card. It is a standard full size 32-bit card - but it does not exceed the average card *height* - and that's the most sensible criteria in a 95A box regarding airflow. Not speaking of DC-load on a single MCA slot. That's another criteria that might forbid using too many of the "Big" cards in a machine.

Not everything that *could* be installed is supported - and not every amount of cards is a) practicable, b) supported and/or c) works reliable. Even if the 400W PSU on the 95A might appear as a bottomless pit (current-wise) the base planar and the slots have physical limits on what current you can draw among it.

> We know that we can stuff multiple Cheetahs in a 9595A, PC Server 500 or 720, but I've never seen any more than ONE Passplay in a 9595A. Does anyone know if a tech manual anywhere makes a definitive statement on this subject?

Uhm ... given that the adapter is in fact "hardware castrated" (and that this is not an affect of the adapter BIOS and the RAID management software) it were useless to stuff in a second adapter anyway. In a 9595A you simply don't have the room to install more drives ... and the adapter has not been announced for different models. The Server 500 uses the "Cheetah" instead. The "Passplay" has been designated as "Server 95A RAID" adapter in various publications and its sole usage seems to be in a Server 95A. It had a very limited purpose due to the lack of an external port (which were as well -again- of limited use if our assumptions regarding the ID-assignments / limitations are correct).

Low Voltage Differential (LVD) Drives

> What kind of drives does the RAID take? Is F/W DIFFERENTIAL SCSI the right kind? Or are LVD (low voltage differential) different and it needs them instead? I've never dealt with RAID before.

From Peter (edited):
   Remember the Cheetah's marketing name? IBM Fast/Wide Streaming RAID Adapter /A. It's an ordinary F/W intended for single-ended SCSI devices. It does however take U/W LVD drives, because these are backward-compatible with single-ended, which the old "high-voltage differential" are *not*.

If you get - for example - a set of U/W "Low Voltage Differential" (LVD) IBM DDRS 4.5 or 9.1GB drives then they will nicely run with the Cheetah. I have some of them in "Starship" - my Server 520 attached to the Fast/Wide RAID Adapter PCI. No problem. You can even mix them with "ordinary" F/W or U/W drives. Same for the Cheetah and even the older Passplay.

# of Drives Supported

From Peter:
>Is it possible to connect more than 7 drives to a PassPlay adapter? I tried to connect a short cable with an external connector on both the channels of the PassPlay adapter without any luck.

The adapter supports only 7 devices. Technically it is a Fast-SCSI adapter, but only a single channel with a crippled ID-section. It does not use the MSB of the ID signal, even if you attach a Wide device to it. The thing is a bit mixed up and screwed down. It has two ports to make cabling easier for upper and lower bay - electronically they are treated as one port.

RAID without Bays 'n Trays
   Go HERE for Peter's experiences.

Passplay under W95
   First, I am shocked. But, live and let live...

Helmut P. Einfalt wrote:
   Pete Rickard got me onto the right track:

You need to install IBMRAID.SYS (in DOS directory of RAID Diskette 1) even if and when the controller does fine under DOS. I installed it manually, although I presume that UINSTALL.EXE would do as well, however after all the time I spent with the machine I wanted to have a manual go at it...

What no Readme (by IBM) tells you is hidden in the Mylex DAC960 Readme: To work properly under Win9x, this driver must be installed *before* any memory manager such as HIMEM.SYS...

Did that, booted, and up came Win95 continuing the installation procedure it broke off at the first "real" Windows start...

Transplanted the whole system (RAID cage w/3 drives, Passplay) to the 9595-ALF (= EMEA version of -0LF).

Same setup:
(1) Spock (+ the famous 1 Gig Fujitsu brick and a CDROM)
(3) Passplay (+ 3 drives RAID5)
(5) XGA-2
(6) IBM ISDN 2000
(8) some NIC.

The Fujitsu (set to SCSI 6 !!!) contains nothing but the IML -- the rest is formatted, no OS, nothing. The Passplay array still has DOS 7 and Win95 (basic installation w/o knickknack for the moment -- the installation completed on the 95A after so many woes)...

And now comes the surprising part: Powered up the machine -- and up came Win95 without a hitch. The Win95 I had installed on the passplay RAID.

Here we are.
Contrary to what the first experiments showed, it *is* possible to install the Passplay on a 9595-xLx machine running Win95.

(1) Spock or similar with a HD of any size that contains the IML track.
(2) The appropriate driver sequence in config.sys:

DEVICE=C:\RAID\IBMRAID.SYS (or wherever this thing is)

Probably you could put the IBMRAID.SYS even before the DOS=HIGH statement, but since it ain't broke I don't want to fix it right now...

Passplay and CD Under Win NT 4.0
   For those of us wanting a kick-ass CD burning machine, the 85/95/95A systems provide the tools for success (the 9577 have the room as well).

For the 95 sized systems, the first issue is which slot to stick the Passplay in. You can use Slots 2-4, which have cutouts for the form factor 5 cards. The second issue is a cable with the correct connector and sufficient length. I used a short cable from the Server 500, with three HPDB68 drops.

For Channel 1 (connector to front of system) set the CD to ID0 (assume you are using a 50-68 pin adapter), and put the wide drive (ID1 to 3) at the end of the SCSI cable (assume in top 5.25" bay), enable termination. Make sure you have the power and SCSI cables connected...
Note: When using the short Server 500 cable, the SCSI connector may be pulled out of the SCSI HD when you push the CD all the way in. I had a Passplay in Slot 4, short SCSI cable to a 50-68 pin adapter on CD, and the HD on the last HPDB68. Make sure with that configuration that you use the first HPDB68 for the CD and the end HPDB68 for the HD.

If you use two HPDB68 that are next to each other, the length between the devices may be too long and the HPDB68 for the bottom device may be pulled partway off of the SCSI port. End result - the device is not visible to the configuration program.

Note: The CD will not show up under View Configuration, since you cannot add it to an array.

To my surprise, a CD is usable to the NT 4 setup floppies (remember to use "S" to choose Mylex DAC960 when choosing mass storage devices). I was able to use the Passplay connected CD for NT 4 setup and installation. CD continues to work without problems after install as well.

Mylex DAC960M Firmware
The Mylex Manufacturing Part ("D040") number can be located on the back of the DAC960 controller, and uniquely identifies the model and number of channels on the controller.  It does not identify the amount of memory installed, or the FW/BIOS versions, since these can be updated.

When referring to this D040 number, please use the entire number, since this will help Technical Support identify specific features.

Mfg.No.Mylex Model

Not all boards will support the upgrade to 3.xx firmware. If the controller has a revision number of D040347 or greater, the board will support the upgrade. This revision label is usually found on the back (non-component side of the board).


SCSI type SCSI-2 Fast/Wide
SCSI bus path / speed 16 bit / 20 MB/sec
I/O bus path / speed 32 bit / 40 MB/sec streaming
I/O features Streaming data transfer
Address parity and data parity
RAID levels RAID 0, 1, Hybrid 1, 5
4 ind (A, B, C, D) / 8 logical arrays
Tagged Command Queuing Yes
Processor i960 at 25 MHz
Size  Type 5 (only fits Model 85 or 95)
Channels Two (both internal)
Connectors Two internal only
Devices supported 7 devices per adapter
Cache std / max 4 MB / 64 MB (with parity)
Cache method 4 sockets for 30 pin 80 ns SIMMs
Cache configurations 4, 16, or 64 MB only
Cache write policy Write-through or write-back

AdapterID 8FBB IBM RAID Controller

 Interrupt Level
   Interrupt level for adapter.
    <"Level E>, B, A

BIOS Base Address
   BIOS base address. Each adapter needs a unique address range.
    <"C0000-0C1FFF>, C2000-0C3FFF, C4000-0C5FFF, C6000-0C7FFF, C8000-0C9FFF, CA000-0CBFFF, CC000-0CDFFF, CE000-0CFFFF, D0000-0D1FFF, D2000-0D3FFF, D4000-0D5FFF, D6000-0D7FFF, D8000-0D9FFF, DA000-0DBFFF, DC000-0DDFFF, DE000-0DFFFF

 I/O Address
   I/O address range. Each adapter needs a unique address range.
    <"1C00-1C1F">, 3C00-3C1F, 5C00-5C1F, 7C00-7C1F, 9C00-9C1F, BC00-BC1F, DC00-DC1F, FC00-FC1F

DMA Arbitration Level
   DMA channel used to transfer data.
    <"Level 8">, 9, A, B, C, D, E, 1, 3, 5, 6, 7

AdapterID 8F6C DAC960M RAID Adapter

Interrupt Level
   Interrupt level for adapter.
    <"Level E"  (IRQ15)>, A (IRQ10), B (IRQ11)

BIOS Base Address
   BIOS base address. Each adapter needs a unique address range.
   <"C0000-0C1FFF">, C2000-0C3FFF, C4000-0C5FFF, C6000-0C7FFF, C8000-0C9FFF, CA000-0CBFFF, CC000-0CDFFF, CE000-0CFFFF, D0000-0D1FFF, D2000-0D3FFF, D4000-0D5FFF, D6000-0D7FFF, D8000-0D9FFF, DA000-0DBFFF, DC000-0DDFFF, DE000-0DFFFF

I/O Address
   I/O address range. Each adapter needs a unique address range.
    <"1C00-1C1F">, 3C00-3C1F, 5C00-5C1F, 7C00-7C1F, 9C00-9C1F, BC00-BC1F, DC00-DC1F, FC00-FC1F

DMA Arbitration Level
   DMA channel used to transfer data.
    <"Level 8">, 9, A, B, C, D, E, 1, 3, 5, 6, 7

Data Parity Exception Handling Support
   Enable or disable Micro Channel data parity generation capability of this adapter (M class and up). Does not show if system doesn't support it.
   <"Enabled ">, Disabled

Micro Channel Streaming
   Enable or disable Microchannel streaming capability of adapter
   <"Enabled ">, Disabled

INT 13 Support
   This provides limited support for BIOS INT 13 function calls and is required if boot devices are connected to RAID Adapter.
  If system is Flash,  then <"Enabled ">, Disabled
     <"Disabled">, Enabled

Content created and/or collected by:
Louis F. Ohland, Peter H. Wendt, David L. Beem, William R. Walsh, Tatsuo Sunagawa, Tomáš Slavotínek, Jim Shorney, Tim N. Clarke, Kevin Bowling, and many others.

Ardent Tool of Capitalism is maintained by Tomáš Slavotínek.
Last update: 08 May 2024 - Changelog | About | Legal & Contact