Ardent Tool of Capitalism

Sound Blaster MCV & Pro MCV

@5084.ADF Sound Blaster MCV (CT5320)
@5103.ADF Sound Blaster Pro MCV (CT5330)

sbmdu.exe DOS Drivers for SB MCV v1.0 (zip)
sbmwu.exe Windows 3.1 drivers for SB MCV (zip)
sbpmdu.exe DOS Drivers for SB Pro MCV v1.0 (zip)
sbpmwu.exe Windows 3.1 drivers for SB Pro MCV (zip)

ct5320.zip Sound Blaster MCV CT5320 Installation Disks
ct5330.zip Sound Blaster Pro MCV CT5330 Installation Disk #1

NT support: Creative's site says no to NT, but see here.
OS/2 support: No drivers at Creative Labs.

sbpup.exe Sound Blaster Updates Disk SUPD-SBP-01?
sbp2up.exe Sound Blaster Updates Disk SUPD-S16-02?

Link to the original Creative Lab's Download Site (dead).

TESTPMCV Reads I/O address, Interrupt and DMA channel settings from CMOS RAM (plus DSP version from SBP) and tests those settings. If the test fails or system hangs, re-configure settings with Reference Diskette and run TESTPMCV again.

Sound Blaster MCV - User Reference Manual (by Sandy and Major Tom)
   See README.TXT for corrections
Sound Blaster MCV "READ ME FIRST" Sheet (ditto)
Sound Blaster Voice Editor - User's Manual (ditto)

Sound Blaster MCV - README.TXT

ctsbk2.exe Sound Blaster Developer Kit, 2nd Edition (for DOS only)
Developer Kit for SB Series, 2nd Ed. Hardware Programming Reference!
Sound Blaster Series Hardware Programming Guide, 1st Ed
Programming.Source.Sound
The Sound Blaster 2.0 and the C/MS Upgrade (not exactly apropos, but good background)

Sound Blaster MCV
Sound Blaster Pro MCV
Passions of the SBP MCV (Screen Corruption on 9553 w/DOS Games)
Design Considerations
Empirical Evidence
General Hardware Info
Sound Blaster Pro MCV ADF
Sound Blaster MCV ADF

Sound Blaster MCV CT5320

SB 1.x vs SB MCV

The Sound Blaster/MCV is based on the original Sound Blaster 1.x design. The main differences between the two are as follows:

Other than that, the cards are identical feature-wise.

Sound Blaster Pro MCV CT5330 (photo from pleonard)

Note: This card is called Pro MCV in most places but marked as Pro 2 MCV on the PCB itself. The "2" was likely dropped because this is the first (and only) SB Pro card for the Micro Channel bus.

CT1336A is referred to as an ISA Bus Interface.

Peter said:
RSPRG_EN should be a jumper to enable the "PC Speaker" on machines that can share the speaker between the planar and sound card. Default is "jumper set".
AUDIO PWD enables the power to the on-card audio amplifier.

SBPRO, SBPRO MCV use DSP version 3.xx

There are two versions of SBPRO. The difference is in the FM chip used. The earlier version uses a two-operator FM chip, while the later version uses a four-operator FM chip. To distinguished them, you can read the value from I/O port 388h, two-operator cards will return a value of 06h, and four-operator cards will return a value of 00h.

There are two versions of FM synthesizer chips used on Sound Blaster cards; YAMAHA OPL2 and YAMAHA OPL3. The OPL2 chip is used in earlier versions of Sound Blaster Pro, SB2.0 and Sound Blaster. The later version of Sound Blaster Pro, and Sound Blaster 16 use the YAMAHA OPL3 chip. (SB Pro MCV included as well?)

Passions of the SBP MCV (Screen Corruption on 9553 w/DOS Games)

Brandon Cobb was trying to get some 9553 and SB Pro MCV action:
   I took a photo tonite of a grand example of what often happens when a game has problems. In this case, it's the "Lasersoft" logo shown first thing when launching one of the old shareware versions of Rise of the Triad. The logo, obviously, is supposed to show a CD-ROM. But you can see what garbage is actually showing up on my screen. When the game is launched without sound options selected, the logo shows fine and so does the rest of the game; it plays fine, too. But with sound enabled, this is what I get. Then, the game either crashes or tries to start up: playing the Apogee fanfare, and then zip... darkness.

Michal Necasek Throws Down:
   Could well be DMA. The black dots are odd, given that the sound is working. The dots imply that something is corrupting memory, writing somewhere it shouldn’t be. It could in theory be DMA.

Garbage Ad Lib sounds are typically caused by “too fast” CPU/software. The OPL2 chips (and I think OPL3 as well) are very sensitive to fast writes and if the required delays aren’t observed, things go wrong. The behavior depends on the CPU, bus speed, and the chipset (“I/O recovery” cycles or whatever OEMs call it). The 9533 is significantly newer than P70 I think, so that may well be different.

The CT5330 is old and from what I have heard, a quick and somewhat ISA to MCA conversion.

On a clone I’d suggest slowing down the bus speed and adding I/O recovery cycles. If and how that can be done on a 9533 I don’t know.

Later:
   There really shouldn’t be resource conflicts on a MCA machine, unless at least one of the devices is lying.

In general, it is worth double and triple checking the sound card configuration (base I/O address, DMA, interrupt) and how the game handles it. Some games must be manually configured, others have automatic detection, some rely on the BLASTER environment variable, others ignore it. Games may be unable to handle some DMA/IRQ settings. The standard is base address 220h, DMA channel 1, IRQ 5 or 7. Anything else may cause problems with some games, while others will handle it just fine.

I have the "Sound Blaster: The Official Book”. It was written in late ’91 or early ’92, shortly before the SB16 showed up, so it couldn’t possibly have information directly related to the 9553. It does mention (page 338) problems with Windows 3.1 on PS/2 machines, especially with models 55sx and 57sx.

One suggested remedy is running Windows 3.1 in Standard mode, exiting, and restarting in 386 Enhanced mode. No idea what that’s supposed to do. Another suggestion is changing the base address from 220h to 240h (and good luck with games which assume 220h). They also suggest putting the SBPRESET utility in AUTOEXEC.BAT.

There should also be a test utility called TESTPMCV. Running that might even give some clues. Installing the SB drivers in Windows 3.1 and seeing how that behaves might also give some clue.

The CL-GD5426 is a pretty vanilla (Super)VGA chip, it doesn’t do anything outrageous. Simple, reliable chip, not a speed demon when it comes to acceleration but a solid workhorse. At least that’s the case with the ISA/VLB variants.

From Us, the Royal god-Emperor:

Prompt "Base I/O Address"

choice "220 Hex"  pos[0]=00100001b
                  pos[1]=11110000b
                     io 0220h-022fh 0230h-0237h 0200h-0207h 0380h-038fh

choice "240 Hex"  pos[0]=01000001b
                  pos[1]=11110000b
                     io 0240h-024fh 0250h-0257h 0200h-0207h 0380h-038fh

Base Address Conflict

Some of the PS/2 systems may use the address range that conflict with the base addresses of the Sound Blaster Pro MCV. Should this happens on your system, please change the base address settings on your card.

Although our base address are set at 220H & 240H, the following addresses will be decoded. Please make sure that other cards do not make use of these addresses.

    220 - 23F    240 - 25F     388 - 389
    620 - 63F    640 - 65F     788 - 789
    A20 - A37    A40 - A5F     B88 - B89
    E20 - E37    E40 - E5F     F88 - F89

Comparison of Reply 55SX TurboProcessor PE3FE and IBM 9553 Planar PE3FE

Reply uses:

   NamedItem Prompt "Cache enabled at startup"
      choice "Yes"  pos[24]=XXXXXX0Xb
      choice "No"   pos[24]=XXXXXX1Xb

After looking at both PE3FE.ADF, the sections mostly call the same pos[x] sections. But not all. There is a slight chance that the 9553 L1 can be turned off, but that depends if IBM left the registers alone. They didn't...

Can the Extra Decoded I/O be Fixed?

Michal Necasek:
   An updated ADF could prevent/flag the conflicts if it listed all the address ranges at which the card responds. This is not something a TSR can fix.

Yes, Creative did most likely do a simple ISA conversion, and not a very good one. But they didn’t decide to add extra I/O ranges, it is simpler/cheaper/faster to decode fewer address bits. It’s more like they didn’t bother making it into a proper MCA adapter.

The IBM PC only decoded the low 10 bits of I/O addresses for all onboard devices. Some adapter cards did the same, which in turn made it much harder for other adapters to use full 16 bits of the address. MCA/EISA/VLB/PCI normally don’t have these problems. The issue is that if you have one old card (like the SBP MCV?) which ignores some address bits, it has the potential to mess up the whole system because it “grabs” more I/O accesses than it ought to.

From WBST:
   Michal has this pretty much explained. I have noticed in my (ISA) SB AWE32 docs that, given a base 0220h selection, the 0620h 0A20h and 0E20h ranges are described as accessing the (optional) Advanced Signal Processor (ASP) chip. This chip and related logic and socket was never available for an MCA bus variant adapter, but they may have left that addressing capability for a future, undeveloped version. Who knows?

I note in Technical Ref. that the I/O address ranges for any one "choice" are limited to 16, so we cannot develop an ADF that caters for the upper nibble of address bits not being checked.

JP5 Jumper

>"On the Sound Blaster MCV Card the shunt across the JP5 Jumper is required only for system such as Model 70 and 80 which have higher I/O channel transfer rate. REMOVE the shunt across JP5 Jumper, if you have a PS/2 system with a slower I/O transfer rate such as PS/2 Model 50 and 55."

Darius Vaskelis replies:
   Apparently the Sound Blaster/MCV originally had the PS/2 Models 50, 50Z, 55SX, 60, and 65SX in mind. It turns out it didn't work in PS/2s that had a 32-bit bus. So, they modified the original design, but made it a jumper on the adapter itself. If you run it in a 16-bit machine, you are supposed to break the solder connection before trying to use the adapter. Otherwise you'll get unpredictable results.

U23 Variants (by Oliver Kluge)

While all audio-related parts of the card work wonderfully in my IBM PS/2 Model 80 (16 MHz 386, 1 wait state), the joystick interface chokes! I debugged the board with an oscilloscope, finding that the IO interface circuitry on the Sound Blaster are too slow to follow a 16 MHz CPU!

Creative Lab provided a jumper J5 to initiate a Card Channel Ready (CD CHREADY) signal to the bus signaling to the bus that the card will take some more time to process the operation. This really works and makes the joystick interface work flawlessly. However, the Sound Blaster fails to release this line timely to not disturb basic system operations, so unpredictable system crashes occur randomly within 0-3 minutes!

This malfunction probably stems from the fact that CL equipped the joystick interface part (not the audio part!) with Low-Power Schottky (LS) TTLs that might be too slow, especially if your machine is faster than my 16 MHz 1 wait. As any other MCA board manufacturer does, CL is best advised to use only Advanced Low-Power Schottky (ALS) or Fast (F) in the future. Perhaps the PCB layout needs better design.

This is obviously a design fault by the Creative Lab EE designers. This has to be considered a serious bug in the circuitry of the board. Obviously they have failed to recognize that a Micro Channel runs at somewhat higher clock speeds than the normal AT bus and failed to use the proper TTL IC family for this task. Low Power Schottky (LS) is just not fast enough, compared to Advanced Low Power Schottky (ALS) Advanced Schottky or Fast (AS or F). I am sending a fax to Creative Lab suggesting them to stop delivery of the board until the bug is fixed.

If a 74LS32 is soldered in there, try getting Creative support to fix it... If there are 74ALS32, 74F32 or 74AS32, the board should run OK.

And here is how to solve the problem if you already bought one, it does not run the joystick and if you are experienced in handling TTL ICs! Apply all the usual precautions against static electricity. Carefully desolder U23 (74LS32) using a vacuum solder pump. Solder in a 74F32 or 74ALS32. This should be it.

Further reading: ALS and AS Logic Families

Design Considerations

> Is the MCA Creative Sound Blaster (Pro) a usable soundcard, MIDI, W95/98 e.g.?

From Peter (edited):
   Yes, from the principle. However the card is based on the old 8-bit ISA Sound Blaster, roughly converted to MCA and originally designed for 80286 and the first 80386 machines. They tend to cause bus timeout errors on faster machines ... and since you mentioned Win98 (which won't work properly on most PS/2) these will be your preferred machines I guess.
   Best recommendation: do yourself a favor, forget the SB MCV and buy a ChipChat or SoundPiper card. They have been designed and tested to work with the later PS/2s.

>My nephew has a Sound Blaster Pro running in a 76 Lacuna with DX2/66, Win95 FAT16. Clean sound. Can't think of a solution, if timings is the problem.

Peter Wendt opines:
   I would guess that it is a problem with the relatively low DMA clock on the "older designs" PS/2s. The Lacuna supports streaming and can run DMA at much higher rates for the individual DMA requests. The SB is known for holding its IRQ and DMA for extended cycles - which ends in bus timeouts (and error 107 or blue screens) or choppy sound. The original Sound Blaster is based around the old 8-bit ISA designs, which are not exactly racers from the bus interface point of view anyway.

I had that with the Pro MCV2 as well... until I decided to remove the card. The Reply Vibra suffers the same symptom - if you run them with the original SB drivers. With the updated versions it runs much better. Have one in my 9595-S20 "aol-gate" along with a Pentium 133 just now.

Nonetheless the sound is a bit choppy when many tasks run at the same time - particularly at Windows startup. Sounds odd. So it is a basic design problem of the SB compatible cards based on ISA technology. (As is the MCV2, Pro MCV/2 and the Reply Vibra-16).

Empirical Evidence

William Walsh wrote:
   I have had no problem with my MCV Pro for the short time that I used it with Win95 and my 9595-0MT with P66 complex. It worked great there.

Ezoto said:
   I use the MCV Pro in my 9595 with Win NT 3.51. I just use the drivers that came with the NT CD and it works fine.

AOX MCMaster and SB Pro

Due to my experiences with said Model 80 and AOX (now Kingston) CPU boards I am of the opinion that the SBPROMCV is indeed borderline hardware. But after much digging I appear to have found success.

"Hint: I had to change the AOX boards "flush mode" to mode 1 or 2 which changes the way the AOX board's cache is flushed. This has resulted in much better sound from the sound blaster - including it being able to find its DMA channel!!? So, it would appear that you have to treat the SBPMCV as something like a bus mastering device due to its extreme timing sensitivities, when using it in conjunction with other devices eg Kingston boards and SCSI controllers."

Speakers Activate when Printing

>I also have an interesting "thing" happening with the Sound Blaster pro driver. When ever my printer starts to print the speakers are activated!

That's easy! The standard IRQ for the Sound Blaster (7, I believe) is the same as for LPT1. DOS doesn't use the printer interrupt, but multitasking systems (such as OS/2 and Unix) do use them. Thus, you have an IRQ problem. The simple solution is to move the Sound Blaster to another interrupt. Of course, if you're out of interrupts (which can easily happen with a fully loaded system), you're out of luck.

Bus Timeout under Windows 3.1

Some of the PS/2 system may encounter a "BUS TIME-OUT" error when starting Windows in enhance mode after bootup. We have included a driver "VDMAD.386" to counteract this problem. The program "WINSETUP.EXE" will copy this driver to your Windows \SYSTEM sub-directory and change the following line under the [386Enh] section in your SYSTEM.INI file.

[386Enh]
; change this line
DEVICE=*VDMAD
; to this line
DEVICE=VDMAD.386

8237 DMA Controller needs MCADMA=off

If your PS/2 system uses the basic 8237 DMA controller (not the extended 82037 DMA controller), you MUST manually add the line "MCADMA=off" to the [386Enh] section in your SYSTEM.INI file. E.g.:

[386Enh]
MCADMA=off

Currently, we have identified that the PS/2 system model 57 is using the basic 8237 DMA controller while systems such as model 60, 70 ,80 and 90 are using the extended 82037 DMA controller.

If you have difficulty implementing the above procedure, please change the line "DEVICE=VDMAD.386" back to "DEVICE=*VDMAD" and follows the steps to start Windows in enhance mode.

1. Start Windows in standard mode first by entering "WIN /S" from the DOS prompt.
2. Exit from Windows.
3. Start Windows in enhanced mode as usual by entering "WIN" at the DOS prompt.

SB Pro MCV will not Reset on 55sx and 57sx

Some PS/2 machines do not reset the Pro MCV card properly. Models 55sx and 57sx are known to have this problem with the Sound Blaster Pro MCV. As a result, Windows 3.1 will not start in Enhanced mode. To remedy this problem, install SBPRESET into your AUTOEXEC.BAT file as outlined in SBPRESET.TXT.

SBPRO MCV and Audiovation Coexistence

>Has anyone gotten the SB Pro MCV and Audiovation/A to coexist in a PS/2? I'm running a 9577 and I can't seem to get the little (*) to go away in the ref partition.

Yes, it works... after patching the ADF file. What's conflicting is the addresses of the game port, defined in both ADF's as fixed resources and therefore the conflict cannot be managed by the POS itself. If you choose to keep the game port on the SBPro, just as I did, your Audiovation ADF file @8FD6.ADF should look like the end of this message. Alain

Also, does anyone have a copy of FORPS2.EXE, the program that can modify other programs to work with SB/MCV?

There were some bad ADF files distributed with some SB Pro MCV cards. They were missing some 'h's in the ADF file, causing incorrect address configurations.

General Hardware Info

AdapterId 5084h "Creative Labs, Inc. - SOUND BLASTER/MCV"

I/O Address used
   <"220 Hex" (io 0220h-022Fh 0338h-0339h, arb 1)>
      "230 Hex"  (io 0230h-023Fh 0338h-0339h, arb 1)
      "240 Hex"  (io 0240h-024Fh 0338h-0339h, arb 1)
      "250 Hex"  (io 0250h-025Fh 0338h-0339h, arb 1)
      "260 Hex"  (io 0260h-026Fh 0338h-0339h, arb 1)
      "210 Hex"  (io 0210h-021Fh 0338h-0339h, arb 1)

Note: The SB MCV uses jumpers to set IRQs (3, 5, 7). There is only one DMA "choice" — DMA 1.

The Base Address includes the I/O Ports. FM music can be accessed through I/O addresses 388h and 389h. MIDI is accessed through the standard 338h and 339h I/O addresses. The joystick port uses the standard 200h-207h range.

AdapterId 5103h "Creative Labs, Inc. - Sound Blaster PRO MCV CT5330"

Base I/O Address
   <"220 Hex"  (io 0220h-022fh 0230h-0237h 0200h-0207h 0380h-038fh)>
      "240 Hex" (io 0240h-024fh 0250h-0257h 0200h-0207h 0380h-038fh)

DMA channel
   < "DMA 1">, 0, 3

Interrupt
   <"IRQ7 selected ">, 5, 3