Sounds very suspicious. I did not think they would make these any more, so I would have expected it to read Philips logo, not NXP. And "1099s" instead of original "1099P"- Could they be newly manufactured Chinese clones of the original maybe?

Can you try the chips in the first system again? Have you tried all the other tricks to slow things down (insert IO wait states, bring down ISA bus clock if possible, otherwise slow down the system via turbo button, jumpers, or TSR)?

To be honest, I do not expect much help from those, as the SAA1099P chips will acknowledge the bus cycle when they are ready to end it, so the logic on SB will use IOCHRDY signal to lengthen ISA bus IO cycle as much as necessary. (Maybe you can confirm the side-effect that will hang a PC if you try to write to CMS chips when they are not present).

Jepael wrote:

Sounds very suspicious. I did not think they would make these any more, so I would have expected it to read Philips logo, not NXP. And "1099s" instead of original "1099P"- Could they be newly manufactured Chinese clones of the original maybe?

Can you try the chips in the first system again? Have you tried all the other tricks to slow things down (insert IO wait states, bring down ISA bus clock if possible, otherwise slow down the system via turbo button, jumpers, or TSR)?

To be honest, I do not expect much help from those, as the SAA1099P chips will acknowledge the bus cycle when they are ready to end it, so the logic on SB will use IOCHRDY signal to lengthen ISA bus IO cycle as much as necessary. (Maybe you can confirm the side-effect that will hang a PC if you try to write to CMS chips when they are not present).

They did have the "P", which as I remember indicates a chip encapsulated in plastic, whereas a non-"P"chip is encapsulated in ceramic.

I originally tried the chips in a Tandy 1000 TX, and all seemed well. Then I tried it in an IBM PC 5150, when I instantly noted the problem. So I tried it in my generic 486 and the problem was still the same. You can't get any slower than a 5150 (except for systems like a PCjr. which doesn't have ISA slots 😀

I had a similar experience recently:

It's 286 time!

The first ones I tried were Phillips branded ones and seemed to work fine at first, but one of them failed after about 2 hours use with symptoms like you describe, which quickly deteriorated into a crackle (even when not in use).

The second batch I bought, the NXP ones, still work fine. I suspect that even when sold as "new and unused", these chips are sometimes second hand.

badmojo wrote:

I had a similar experience recently: […]
Show full quote

I had a similar experience recently:

It's 286 time!

The first ones I tried were Phillips branded ones and seemed to work fine at first, but one of them failed after about 2 hours use with symptoms like you describe, which quickly deteriorated into a crackle (even when not in use).

The second batch I bought, the NXP ones, still work fine. I suspect that even when sold as "new and unused", these chips are sometimes second hand.

Your chips and mine share the exact same markings. I guess you were luckier than I was 😉 Maybe these chips are unusually fragile.

Yes and I guess the other factor here is that these sound cards are old, and if those sockets haven't been used since the card was manufactured, then they might be dirty. Perhaps a good clean and reseating of the chips is worth a shot if you haven't already tried that?

Here is some more weirdness. I originally tried these chips in my Tandy 1000 TX, where they sounded just fine. Then they sounded crappy in my IBM PC 5150 and my custom-built 486. Yesterday I put them back in the Tandy TX, as Jepael suggested and they sounded fine again! I was not expecting that. I will try the card in my Tandy SX and TL and see whether the chips want a 286 or a Tandy or what/

The 2.0 Sound Blaster needs a PSU with -5V. Just saying in case someone forgot 😊

I got a bunch of Philips chips many years ago and the 1.5 cards have always worked. I do have a 2.0 with original Create chips (all three from an upgrade kit), but because I use ATX PSUs with ATX to AT adapters I stuck with my two 1.5 cards 😀

I'm using a 1.5, but even if I used a 2.0, all my systems could cope.

However, I think I spoke too soon, because there were missing notes in every LucasArts game I tried on the TX. Then I tried the Sound Blaster 1.0 TEST-SBC and three out of six voices on each chip were silent.