VOGONS


I de-CQMed my AWE32 PnP CT3990...!

Topic actions

First post, by Eep386

User metadata
Rank Member
Rank
Member

I've had this CT3990 lying around for the longest time in my room. As it had Creative's awful CQM b*stard child instead of a half-decent FM chip, I utterly refused to use it.

Well, today I worked up the courage and scavenged the needed parts off a CT2900 ViBRA 16S, and removed both the CT1978 CQM chip and its associated TDA1387 DAC and coupling/filtering capacitors, and I migrated the missing capacitors, resistors and resonator over to the CT3990, along with the YMF289B and YAC516-E chips. Thank goodness for SRA Fast Chip.

And it worked, first time. Lemmings played in glorious real Yamaha OPL3 sound. 😁 So now I have probably the only CT3990 I've ever seen with genuine OPL3 synthesis instead of the wretched CQM garbage.

Tomorrow I'ma take a hammer to the CT1978 chip and forever banish it to oblivion where it belongs. The CT3990 has a broken SIMM RAM slot (common problem on old AWE32s) so that has to be fixed too.

Here's a picture:

IMG_0479.JPG
Filename
IMG_0479.JPG
File size
3.45 MiB
Views
5675 views
File license
Fair use/fair dealing exception
Last edited by Eep386 on 2018-01-02, 18:00. Edited 1 time in total.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 1 of 96, by TimMer1981

User metadata
Rank Newbie
Rank
Newbie

Nice, following this thread. 😀

Got a CT3980 myself BTW, which is native OPL. 😀

"I will take your bones, still alive and in great pain, and make them into a chair. I will call it "My Screaming Chair". Every morning I will sit in it and listen to you scream. Any questions?"

Reply 2 of 96, by firage

User metadata
Rank Oldbie
Rank
Oldbie

Your verdict of the CT1978 CQM may be a touch exaggerated, but that's some commendable work right there. I think this is one of the first successful conversions of a CQM afflicted Creative card. Grats, man!

My big-red-switch 486

Reply 4 of 96, by meljor

User metadata
Rank Oldbie
Rank
Oldbie

Nice! Love such projects, wish i had a bit more skills/knowledge (but working on it).

asus tx97-e, 233mmx, voodoo1, s3 virge ,sb16
asus p5a, k6-3+ @ 550mhz, voodoo2 12mb sli, gf2 gts, awe32
asus p3b-f, p3-700, voodoo3 3500TV agp, awe64
asus tusl2-c, p3-S 1,4ghz, voodoo5 5500, live!
asus a7n8x DL, barton cpu, 6800ultra, Voodoo3 pci, audigy1

Reply 5 of 96, by appiah4

User metadata
Rank l33t++
Rank
l33t++

Nice project but I kind of disagree with OPL3 being the holy grail of FM Synthesis, to me ESFM and even Crystal FM sound better, and CQM is not worse than OPL3, just different. Also, OPL3 by itself means nothing, the kind of filtering different OPL3 cards have tend to produce completely different sounds anyway, late Vibra/OPL3 SB16s' FM synthesis sounds as far from the SB Pro 2.0 sound as ESFM does. Some of Creative's own OPL3 cards have no bass response. Some have very muffled sounds. I commend what you did here but not sure I find it necessary 😀

Retronautics: A digital gallery of my retro computers, hardware and projects.

Reply 6 of 96, by firage

User metadata
Rank Oldbie
Rank
Oldbie

There's some variation in the output EQ among OPL cards, but I don't think they ever sound like part of the instrumentation is entirely in the wrong key or something, which sort of happens with CQM in some places. It is often a subtle difference, but the CQM rarely wins at anything for me.

Last edited by firage on 2018-01-02, 13:49. Edited 1 time in total.

My big-red-switch 486

Reply 8 of 96, by keropi

User metadata
Rank l33t++
Rank
l33t++

Yes please Eep386 post a pic highlighting the changes you made , maybe I will try that as well!

🎵 🎧 PCMIDI MPU , OrpheusII , Action Rewind , Megacard and 🎶GoldLib soundcard website

Reply 9 of 96, by Eep386

User metadata
Rank Member
Rank
Member

Ask and ye shall receive. 😀

IMG_0479.JPG
Filename
IMG_0479.JPG
File size
3.45 MiB
Views
5675 views
File license
Fair use/fair dealing exception

I cleaned up the board after taking this shot, it looks much less fluxy in real life:

IMG_0480[1].JPG
Filename
IMG_0480[1].JPG
File size
4.09 MiB
Views
5674 views
File license
Fair use/fair dealing exception

Those with a hot-air station will probably find this much easier going than I did, I had to use my Elenco soldering iron with SRA Fast Chip, which necessitated me cleaning the pins with copper braid to get rid of the alloy mess. Getting the YMF289B in there, with its tiny, fine pins was sure "fun"! 😜 The YAC516-E was much easier thanks to its much wider pin pitch.

Later I'll tweak the onboard op-amps that control filtering and gain, to try to get the best sound possible out of it. Those who manage to make out my slightly blurry pics (bless my Canon SD1000...) might notice that I've changed some of the MC3403 op-amps to TL074s. I'll change the one that buffers and filters the FM to an MC33079 later.

@appiah4: I understand that there is no 'one OPL3 synthesizer chip' out there, for many it's a matter of personal preference. I have a friend who's a nut for Crystal CS4236B FM, while I grew up with the sound of the YMF289 (built-into the YMF719, which was my first FM sound chip back in the day), so for me that's 'my' OPL3 chip of choice. At any rate it's a genuine Yamaha article. 😀 YMF262 and CT1747 do not have resampling artifacts like YMF289, so FM purists will probably gravitate toward those. Though, I can fairly safely assume that almost no-one here likes screechy, chirpy CQM. 🤣

Last edited by Eep386 on 2018-01-02, 18:16. Edited 1 time in total.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 10 of 96, by keropi

User metadata
Rank l33t++
Rank
l33t++

great work!
so you only need to add the 2 Yamaha ICs? any passives needed adding/removing?

🎵 🎧 PCMIDI MPU , OrpheusII , Action Rewind , Megacard and 🎶GoldLib soundcard website

Reply 12 of 96, by Eep386

User metadata
Rank Member
Rank
Member

Well, you also need the 33.88 MHz ceramic resonator, and a bunch of passive capacitors and resistors, and some electrolytic capacitors.

Unfortunately I did not identify the values of the capacitors as I was swapping parts from the donor CT2900 into their respective locations on the CT3900. However, the resistors I did identify however are a 10 ohm resistor, a 3.9K resistor and two 47K resistors. You also need two 1uF capacitors and a 10uF capacitor. There may be more parts too, but those are the ones I can make out.

The way I did the swap, was I sat down and took a good hard look at the CT2900 and the empty pads on the CT3990 and tried to guesstimate where the appropriate parts went. The easiest way to do this, provided you've got the tools to extract the parts without damage, is to simply scrap the needed parts off a lesser sound card with YMF289B. A broken CT2900 or one of the cost-reduced later boards that omit the TEA2025 or TDA1517 makes a decent donor candidate.

For a CT3990, the 10 ohm resistor goes to R30, the 3.9K resistor goes to R50 and the two 47K resistors to R88 and R91.The 1uF caps go to C86 and C116, while the 10uF goes to C105.

Oh yeah, you also have to remove the CT1978 CQM chip at U20 and accompanying TDA1387 DAC from U19.

Maybe once I get more free time, I'll desolder those caps I didn't identify and try to identify them. I want to make public knowledge a method to evict CQM from CT3990s.

Last edited by Eep386 on 2018-01-02, 19:16. Edited 1 time in total.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 13 of 96, by keropi

User metadata
Rank l33t++
Rank
l33t++

thanks for the extra info Eep386 - documenting the process is a great idea as well 😀

🎵 🎧 PCMIDI MPU , OrpheusII , Action Rewind , Megacard and 🎶GoldLib soundcard website

Reply 14 of 96, by Eep386

User metadata
Rank Member
Rank
Member

Before I continue, let me just say this: this information is provided as-is, free of charge, with no guarantees of any kind.

OK, now that that's out of the way, let's try reverse engineering the caps I put in there...

For reference, all part references are specifically for the CT3990. Other models may have different part references, be advised.

C103, C104, C106, C117 and C119 measure between 165~177 nF. So 180nF as a relatively common and cheap to purchase in-bulk value, should work for all these caps.
C131 and C142 according to Yamaha's datasheet for the YMF289B are 12pF caps. They are used for filtering the 33.8688 MHz crystal oscillator; interestingly Creative installed them while using a ceramic resonator as well. (Guess they wanted to err on the side of caution?)

One potential roadblock is the fact that 33.8688 MHz ceramic resonators aren't common, not on Digi-Key at least. I'd imagine 33.88 MHz straight would probably work without any problems. Not sure if a 33.8688 MHz crystal oscillator would work - then again, Yamaha's datasheets indicate that a crystal oscillator of that frequency can be used there, along with the 12pF caps and 3.9K resistor as mentioned earlier. The crystal oscillator would be much better than a resonator in any case, if it works.

So, to summarize so far the parts that need to be added and where... All SMD resistor and capacitors are 0805 size except otherwise mentioned...

CT3990:
Add:
U28: Yamaha YMF289B-S
U23: Yamaha YAC516-E (the -E is important, the non-E versions have a different pinout!)
R30: 10 ohm resistor (larger than a 0805; not certain the metric size)
R50: 3.9K SMD chip resistor
R88, R91: 47K SMD chip resistors
C103, C104, C106, C117, C119: 180nF SMD ceramic capacitors
C131, C142: 12pF SMD ceramic capacitors
C86. C116: 1uF electrolytic capacitors
C105: 10uF electrolytic capacitor
C4: 33.8688 MHz ceramic resonator (MAYBE a Crystal Oscillator will work... haven't tested)

Remove:
U19: TDA1387 DAC
U20: Creative CT1978 CQM chip
C54, C55, C91 capacitors

I've doubtlessly missed something here, so expect this to change as I poke this card further.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 15 of 96, by chinny22

User metadata
Rank l33t++
Rank
l33t++

I'm with the majority here.
I don't hate the CQM enough to bother, but
Recognise true OPL chip is more desirable making it a worthwhile project, and not one many of us have the skills to do.
I can see others following in your footsteps, specially as you are kindly documenting at as well, so thankyou

Reply 16 of 96, by CkRtech

User metadata
Rank Oldbie
Rank
Oldbie

Cool that you had to swap some passive components and not worry about tying a QFP pin high vs low or something like that.

Some chin-scratching (I haven't examined the circuit that much)
1: It looks like I could do the same to my SB32 CT3600 if I so desired
2: Could the board support both the CQM and the YMF?

Hmm. Not sure if you could either switch between the two (physically) - which I imagine a physical switch would need some sort of "concierge circuit" so you don't fry the mixer if you do it live. Or - you could run both separately and externally. For instance - re-purpose the line-in and mic ports as pre-amped, line level outputs for CQM and YMF. The card would drive both, and you could just use an external mixer to select which one you want for a given game.

I don't know the signals being passed - I haven't looked at datasheets yet, etc. I am just theorizing on a high level. But you've had your hands in it a bit, so I thought I would spitball the idea.

Displaced Gamers (YouTube) - DOS Gaming Aspect Ratio - 320x200 || The History of 240p || Dithering on the Sega Genesis with Composite Video

Reply 17 of 96, by Eep386

User metadata
Rank Member
Rank
Member

Well, the YMF289B and CT1978 use the same shared data bus, and the TDA1387 and YAC516 output over shared sound left/right lines. Not sure what the software side implications of running both CQM and YMF289 over the same card would be.
I'd imagine one could just wire in a DPDT switch to select which DAC outputs over the sound left/right lines, but I've yet to identify where the TDA and YAC's output 'intersect'.
It's theoretically possible to mix them together to get on-board dual-OPL3, but you'll need to probably rework the op-amp buffering the FM side into a dual adder. Again, I need to dig into the op-amp in question.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁

Reply 19 of 96, by Eep386

User metadata
Rank Member
Rank
Member

It'd be better to toggle the chip select pins. You don't want to arbitrarily power up and down chips that aren't designed to do that - unless you mean, switching the chips while the power is off, and just using the selected chip only while the power is on.
On the YMF289B-S, the /CS pin is pin 13. Chip select on the CT1978-BAP is pin 2, going by a quick multimeter probe.

Life isn't long enough to re-enable every hidden option in every BIOS on every board... 🙁