I also have the GA-5AX (with a K6-III+) but I tried the CT1330A only briefly (for works-or-not) in this motherboard. I can confirm, however, that this card IS speed sensitive even on a 486. I remember I had to set slow ISA refresh in BIOS in order to get some games work on a DX4-120 machine. I do not remember exactly which games 😵

Is that true of the CT1600 ?
Or is it true witth those games on any sound card ?

Jolaes76 wrote:

I also have the GA-5AX (with a K6-III+) but I tried the CT1330A only briefly (for works-or-not) in this motherboard. I can confirm, however, that this card IS speed sensitive even on a 486. I remember I had to set slow ISA refresh in BIOS in order to get some games work on a DX4-120 machine. I do not remember exactly which games 😵

thanks, useful to get confirmation - mine is a clone rather than ct1330a but all the same, seems it's the card then. previously i'd only used it in a 386 so would not have noticed this.

sunaiac wrote:

Is that true of the CT1600 ?
Or is it true witth those games on any sound card ?

pretty sure the ct1600 is not affected.

ratfink wrote:

sunaiac wrote:

Is that true of the CT1600 ?
Or is it true witth those games on any sound card ?

pretty sure the ct1600 is not affected.

I use a CT1600 with a Pentium Pro 200, and i have used it with my P3 rig with no problems (see specs in signature).

I don't know the game, but are you talking about digitized sound here?
Because it sounds like the game might be using a primitive method of programming PCM where you just throw bytes at the card in real time. If that's what it's doing, and if they didn't implement any timing based on the real time clock, then the playback could be speed sensitive.
Could also be that their method of timing isn't quite perfect, and the method you've used to cripple your system has somehow thrown it off.

This method of PCM wasn't recommended because it sucks, but I think all the legacy era Sound Blasters support it. DMA support wasn't in the earliest cards, but it was added pretty early. A CT1330 definitely supports DMA, but that doesn't mean the game is utilizing it.

it is the music [fm synthesis] that is the problem but only on certain screens does it do this ridiculous speeding up.

i'll have to do some more trials as i too only tend to test cards work rather than test them more fully.

The YM-3812 OPL2 chip in the Adlib, Sound Blaster 1.0-2.0 and Sound Blaster Pro 1, which produces FM music, is indeed speed sensitive. The faster the computer, the more wait states needed to be added when sending data to the chip. The Sound Blaster Pro 2 and newer Sound Blasters use a YMF-262 OPL3 chip, or a clone or derivative of that chip, and it is not nearly as speed sensitive.

This will not solve your issues if the game's music or digital audio playback routines suffer from speed sensitivity.

This does not sound like the sound card being speed sensitive. If the FM music is OK but just faster, I think it is not a sound card problem.

If the FM music would sound incorrect like wrong frequency notes, missing notes, or wrong instruments, then it would be a problem that software sends data too fast to FM chip.

Not all games test if you have OPL2 or OPL3, they just use same driver. So on a faster machine with an OPL3 it will work, but won't work with an OPL2. It is dependent mainly on ISA bus speed, not CPU speed, but faster machines can execute code in cache so ISA bus operations happen quite rapidly, without fetching code from RAM between ISA bus operations. So in this case increasing 8-bit IO wait states will help, but can of course slow down other IO port accesses like VGA.

I would rather say the program code is speed sensitive. Many games just have a delay loop that is large enough on machines it was developed with. On faster machines introduced later, the delay is shorter. But many programs have machine speed tests to make sure the timings for FM chip writes are obeyed. Monkey Island calibrates a delay loop with PIT timer interrupt, and some other game I forgot counts PIT timer ticks for delays.