ISA/DOS Sound cards and VIA Chipsets, what do I look for?

[hunterjwizzard]

I used to wire a cheap Casio keyboard (about the cheapest you could get with a MIDI port) up and route DOOM's music through that. I'm not going to say it was "better" than the AWE32, but it it was a lot more fun than any sound card.

...ok I am definitely doing that.

I have amongst my eclectic collection of things a Yamaha DX7, which I understand was at one point a very highly sought-after keyboard. I've also got an ISA card for controlling MIDI instruments. This is definitely more interesting.

 

[Plasma]

The original DX7 isn't multitimbral so it won't work for PC games.

 

[hunterjwizzard]

Curses. Foiled again.

 

[GiGaBiTe]

Oh see I'm finding it impossible to get an SB16 to work in there. The SB16 just doesn't play anything besides MIDI. I have an SB live in the machine right now and am having all sorts of issues, probably the hanging note bug as described.

The hanging note bug was really only a major problem with early Sound Blaster 1.x-2.x and Pro cards. A hung note would persist even after the application that was playing music closed, and even between system reboots. I've only ever seen this one time, on a customer's sound card that they sent in for recapping. It was one of the aforementioned above models.

The SB 16 and AWE 32 have a different type of hanging note problem, but it's nowhere near as bad as the early SB cards. SB Live! cards don't have this problem, but they do have another problem related to echo effects in certain DOS games like Duke 3D causing the game to crash. There were patches to some of those games to disable the echo effect, but I don't know if they're still floating around.

My goal is to be able to use this machine to play DOS games in addition to win98, and I'd like to not face any serious compatibility issues. My understanding is I can assure this by sticking with ISA. But I've had so gosh darn many issues with sound on this machine I'll take any ideas I can get.

In order to maintain DOS game compatibility on Windows 98, you have to install the VxD driver instead of the WDM driver. VxDs in general are more likely to cause system instability and conflicts because they give user applications direct access to hardware, rather than going through Windows's sound API.

Windows 98 is itself one giant incompatibility issue, there's no getting around that.

 

[hunterjwizzard]

The SB 16 and AWE 32 have a different type of hanging note problem, but it's nowhere near as bad as the early SB cards. SB Live! cards don't have this problem, but they do have another problem related to echo effects in certain DOS games like Duke 3D causing the game to crash. There were patches to some of those games to disable the echo effect, but I don't know if they're still floating around.

In that case haven't a clue what problem my SB Live is having. Its at present struggling to play MP3s normally. Game sound(even native windows sound) is right out.

 

[Eudimorphodon]

FWIW, I tried refreshing my memory about what's supposedly wrong with Soundblaster Live cards and VIA chipsets, and it seems like there's a lot of... fog, on that question. Back in the day I did actually have an SB Live! in my 1.33ghz AMD Thunderbird box with a VIA motherboard (I think it was a KT133A chipset) and I don't remember running into any problems. There's some statements out there from VIA saying they were able to replicate the problem on certain motherboards (not all), and they issued a fix to their 4-in-1 chipset drivers that supposedly worked around the issues.... so, I dunno. That's the beautiful, terrible thing about Windows 9x; there are *so* many different possible scenarios for different devices to step on each other's toes in confounding ways.

 

[hunterjwizzard]

FWIW, I tried refreshing my memory about what's supposedly wrong with Soundblaster Live cards and VIA chipsets, and it seems like there's a lot of... fog, on that question. Back in the day I did actually have an SB Live! in my 1.33ghz AMD Thunderbird box with a VIA motherboard (I think it was a KT133A chipset) and I don't remember running into any problems. There's some statements out there from VIA saying they were able to replicate the problem on certain motherboards (not all), and they issued a fix to their 4-in-1 chipset drivers that supposedly worked around the issues.... so, I dunno. That's the beautiful, terrible thing about Windows 9x; there are *so* many different possible scenarios for different devices to step on each other's toes in confounding ways.

I've now experienced issues with my ISA SB16 and my PCI SB Live. The SB Live is the same one I used with this motherboard 22 years ago and had zero problems with at the time. Now it doesn't work most of the time

Its enough of a problem that I'd rather just leave Creative Labs all together for this build.

 

[Eudimorphodon]

Its enough of a problem that I'd rather just leave Creative Labs all together for this build.

An ISA soundblaster 16 and a Soundblaster Live have essentially nothing in common, so I'm kind of left thinking here you might have something going on here that has nothing to do with these sound cards in particular. (Or, conversely, wouldn't apply to any other sound card.)

Backing up a little, what version of the Soundblaster 16 *exactly* did you try to get working in this thing? Specifically, were you using a "Plug and Play" SB16, or is it jumper configured? If it's a jumper card you may need to reserve the resources it's set up to use in the BIOS to ensure the machine doesn't automatically map some other built-in or PCI device over them, and you also need to make sure that the resource configuration that Windows sets its drivers up for matches up correctly. This symptom:

Oh see I'm finding it impossible to get an SB16 to work in there. The SB16 just doesn't play anything besides MIDI.

Sounds a lot like you have a resource mismatch, because while the SB I/O port range can be a lot of different places the FM synth is almost always fixed at port 0x388.

I think even with the PnP card you might need to muck around in the BIOS to convince the computer to assign "normal" resources for the card; my AWE32 was of the PnP variety and I have vague memories of it being two completely different kinds of a PITA depending on whether it was in an ISA 486 with a not-PnP-aware BIOS vs. a Pentium-class-or-better machine that *did* have BIOS support for configuring it.

 

[hunterjwizzard]

Backing up a little, what version of the Soundblaster 16 *exactly* did you try to get working in this thing?

Mine specifically is an SB2940. There are spaces on it for jumpers but it does not have any jumpers on it.

I can put the SB16 back in a muck with it some. I know the 16 is nothing special but I managed to hang on to this one through numerous rounds of hardware purges so it'd be neat if it worked out.

 

[Eudimorphodon]

On the Live side of the equation, did you try the drivers from Phil's computer lab? A known issue/annoyance of the Live was Creative made a million subtly different versions of it. Supposedly this driver bundle is "universal". Also, for a VIA chipset board definitely make sure you have the 4-in-1 driver set installed before you set up *anything* else. But, well, that's about where my bad memories of Windows 98 run out.

 

[Zare]

I have AWE64 standard hooked to SC-55. Sure a gold edition would be more comparable but I've noticed Roland box outputs better signal in snr terms, the audio is richer in dynamics than soundcard out regardless of sample differences. I hook everything to a soundboard, which feeds a backline 2x125W amp driving into two Infinity floorstanders. Tho SC-55 outputs a weaker signal than the soundcard, the difference is noticable and I'd rather use preamps on my soundboard to level it up than any kind of soundcard circuitry anyway.

 

[GiGaBiTe]

FWIW, I tried refreshing my memory about what's supposedly wrong with Soundblaster Live cards and VIA chipsets, and it seems like there's a lot of... fog, on that question. Back in the day I did actually have an SB Live! in my 1.33ghz AMD Thunderbird box with a VIA motherboard (I think it was a KT133A chipset) and I don't remember running into any problems. There's some statements out there from VIA saying they were able to replicate the problem on certain motherboards (not all), and they issued a fix to their 4-in-1 chipset drivers that supposedly worked around the issues.... so, I dunno. That's the beautiful, terrible thing about Windows 9x; there are *so* many different possible scenarios for different devices to step on each other's toes in confounding ways.

The problem related mostly to the 82C686B and VT8231 chipsets. The 82C686B had terrible problems because it used the PCI bus as the bus between the NB and SB chips. The chipset couldn't handle heavy bus contention that inevitably arose from bandwidth heavy cards such as video, sound and disk controllers. When the PCI bus traffic was too much, the chipset started corrupting the bus and causing all matters of problems. The SB Live! issue is just collateral damage from a much different, bigger problem.

The "fix" for the 82C686B was the VT8231, which made a special "VT Link" bus between the NB and SB chips, and unloaded the PCI bus of the burden of the additional chipset traffic. Again, as you said, there is "fog" on this issue, especially since it was over 20 years ago and not properly documented. Reports of SB Live! corruption issues fell off after that, and the later VIA chipsets seemed to resolve that problem. I had many VIA chipset boards after that time, and I never had an issue with SB Live! cards on them.

I've now experienced issues with my ISA SB16 and my PCI SB Live. The SB Live is the same one I used with this motherboard 22 years ago and had zero problems with at the time. Now it doesn't work most of the time

Its enough of a problem that I'd rather just leave Creative Labs all together for this build.

Have you tried cleaning the PCI edge connector, as well as the slot? Old cards in my horde tend to have problems with light oxide buildup causing issues, I keep Deoxit gold on hand for that.

Another could be failing capacitors. I've had to recap several sound cards with bad caps on them causing weird issues.

 

[hunterjwizzard]

I have been lucky with capacitors so far, I think because I live somewhere with low humidity.

The phils computer lab drivers look promising, if for no other reason than he's got an Audigy 2 driver for win98(I have many Audigy 2s!)

For now I first had to stop and build a shelf(mission accomplished!) and can now work much more easily on the system. I put the ISA SB16 back in, let's see what happens. I am strongly considering the option of "just leave 2 sound cards in there".