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

[Ozzuneoj]

Would be really cool to be able to A\B test CQM and OPL3 on the same card. Since this is the first OPL3 transplant that I've ever seen done online, I would imagine there have never been any straight A\B tests like this before either since no cards have BOTH. The closest would be similar cards that come in OPL3 or CQM models (I have a couple CT2940 SB16s with CQM and one with OPL3), but even then things are bound to be somewhat different.

[gdjacobs]

Hmm, yes, a CS pin will be perfect.

[carlostex]

Great project, congratulations!!

[gerwin]

Impressive soldering there, and good to hear that it functions as intended. I had the same idea years ago for a CT3600. But whenever I see the small legs of the OPL3-L, I feel sure they will short out with my soldering. I can solder SMD resistors a little, but this OPL3-L is another matter.

[Eep386]

Thank you all for your kind feedback! You guys got me interested in a dual CQM/OPL3 switched setup, it'd be perfect for FM comparisons. No more trying to get two or three Sound Blasters running together coherently!

When I get my next paycheck, my next mission would be to acquire a (non-destroyed) CT2900 or something with YMF289 or YMF262 already on it but with empty pads for a CT1978 CQM, and try adding a CQM and TDA1387 chips to the board, and wiring up a chip select switch.

Or, if I get lucky enough to snag a YMF289B-S and YAC516-S at a reasonable price, maybe add them to a CT3600 and do my switchwork on that. I'd rather add a CQM chip to be honest as the CT1978 and TDA1387 have these relatively widely pitched pins that are no problem to work with, while the YMF289B-S's tiny pins were a royal PITA to keep clean. (SRA Fast Chip is awesome stuff, but even then there's simply no replacement for a hot-air station sometimes.)

[CkRtech]

Awesome.

I started reading datasheets out of curiousity, but am sadly dead tired at the moment. The YMF289B typically recommends a YAC513 or YAC516, but is is possible that the data and clocks could be instead run to a CS8406 rather than the YAC516? I know you can't just mount a CS8406 on the card, but it made me wonder if you could use something like that CS8406 in order to add spdif output from the YMF289B (essentially just moving FM of the card to a digital transport for use with an external DAC).

Aside from the usual LRCK, SCLK, Data, etc., I got as far as checking on the YMF289B output (Most significant bit first) and checking CS8406 flags for input and suddenly felt tired. Ha.

(I stopped short of knowing if the CS8406 can take the MSB first) Maybe the TC9231N can do it?

You suddenly get to thinking about analog mixing, Creative Labs Noise Blaster, grabbing digital audio and bypassing internal, Creative Labs mixing...

I know... I know... scope creep.

[Eep386]

Not sure if a CS8406 would work at all. I get the feeling the serial data format is a bit different somehow.

[KungfuPancake]

Hello all!

Thanks to Eep386 I was able to retrofit a CT3990 I got for cheap on eBay. Since the prices for a CT3980 and pretty much everything else with an OPL are horribly inflated in my country and I already had all the tools and the passives needed, this was a nice alternative.

I ordered the YMF289B-S, the YAC516-E and a bag of 33.8688 MHz crystals via Aliexpress. All in all this set me back about 8$.

Some helpful things that popped up during planning/assembly:

* A crystal works fine and, theoretically, better than a resonator. The 12pF caps are required for this.
* The 180nF are most likely bypass caps. The "standard" 100nF works fine and I have a whole reel of those.
* The 1µF and 10µF caps (also bypass) can be replaced with ceramics. Since the voltage is 5v max, I was able to use my 0603 ones I had at hand.
* The 10 ohm resistor appears to be 1206 and was probably chosen over 0805 because of power dissipation reasons. I'd recommend sticking to this size and if not available (I luckily had one in the scrap bin) bodge in an axial one.
* A hot air station makes the process so much less painful. I recommend getting one of those chinese knockoffs for 30$, they work very well for what they are.

Please enjoy a close up shot after flux removal while I fire up some games for a proper nostalgic OPL experience.

[Eep386]

Nice, clean job with the rework there. How does it work out for you?

[shock__]

Offtopic, but really interests me ...is U8 original from factory or something you added? Asking because ZMD is one of the very few (formerly east) German IC manufacturers.

[Eep386]

For some reason, some AWE32 and SB32 cards come with a DIP packaged ZMD-made DRAM chip, while others use a smaller surface-mount chip.
Yeah, I find the ZMD-made stuff to be quite interesting myself

[realoldguy23]

Offtopic, but really interests me ...is U8 original from factory or something you added? Asking because ZMD is one of the very few (formerly east) German IC manufacturers.

I find that interesting, too. Just checked Wikipedia: https://de.wikipedia.org/wiki/Zentrum_Mikroelektronik_Dresden?wprov=sfla1

It seems they are still in business today, but I guess they are not making chips anymore. According to Wiki they were heavily financially supported by the state at the time.

[KungfuPancake]

Nice, clean job with the rework there. How does it work out for you?

Thank you! It worked very nicely from the start. I don't have a direct comparison, but the crystal and the ceramics don't seem to have an effect on the output quality.

Next step would be to try and combine both ICs using the CS lines and some glue logic for the input lines. I'm trying to integrate that into my other side project, an all-in-one solution for ATX -> AT / USB Mouse and Keyboard -> AT & Serial and miscellaneous housekeeping.

For now I'm just really happy with the result though.

[kodi]

@all
what are your thoughts on the following, as potential way to connect (digitally) YMF289B to EMU8000 (and respectively to its SPDIF output JP3) on CT3990 card:

so, according to the information here:
viewtopic.php?f=62&t=50071&start=60#p702925

on AWE32 with real OPL3, i.e. the one build-in CT1747 chip, the signal path is:

CT1747 ---I2S---> EMU8000 ---> TDA1387 DAC

and that's why the FM Synth output is enabled via AWEUTIL. Of course, there is 2nd pure digital path:

CT1747 ---I2S---> EMU8000 ---> JP3 (5V SPDIF).

further more here:
viewtopic.php?f=46&t=38201&start=20

is list of the exact pin connections.

If you look those pin connections, I think we can be certain that "EMU8000 Pin 031 - ???" is actually GPIO pin and it's connected to OPL3 CS pin, that's why:

Code: Select all

CT1747 (EMU8000-based cards)
Pin 080 - ??? (from(?) EMU8000 pin 31)


and

Code: Select all

CT1747 (Non-EMU8000-based cards)
Pin 080 - ??? (tied to GND)


CS needs to be Low to make OPL3 active, i.e. on "CT1747 (Non-EMU8000-based cards)" it's always active and on EMU8000 probably the I2S input (EMU8000 pins 27,28,29) is always active and actually most likely AWEUTIL just toggles that EMU8000 GPIO pin 31 to enable/disable the FM Synth.

Back to CT3990: I can confirm that EMU8000 I2S Output (pins 78, 80, 81) are connected to TDA1387 DAC (U27 on the board), i.e. that "half" is the same as on AWE32 cards with real OPL3. Also, it seems to me on CT3990, EMU8000 I2S input pins (27,28,29) are not connected. That means my idea boils down to connecting :

Code: Select all

EMU8000 Pin 027 (Word Clock Input)---- YMF289B-S pin 34
EMU8000 Pin 028 (Bit Clock Input)------- YMF289B-S pin 35
EMU8000 Pin 029 (Data Input)------------ YMF289B-S pin 33


One thing that bugs me is why there is no Left/Right Select on EMU8000 I2S Input interface or maybe based on this:

viewtopic.php?f=46&t=38201&start=20#p342123

the EMU8000 Pin 027(Word Clock Input) is actually L/R clock input and then instead the above correct would be:

Code: Select all

EMU8000 Pin 027 (Word Clock Input)---- YMF289B-S pin 32
EMU8000 Pin 028 (Bit Clock Input)------- YMF289B-S pin 35
EMU8000 Pin 029 (Data Input)------------ YMF289B-S pin 33


In any way, please, give me your thoughts...

[suntac]

@KungfuPancake, you should have removed also the TDA1387 DAC (U19). Its outputs are directly connected to the YAC516's outputs.

[SirNickity]

If anyone knows off the top of their head, are there easy-to-source replacements for those SIMM slots? I've had an AWE32 with expanded memory since the 90s, and then recently removed the RAM to clean the board. Naturally, the plastic retainer on one of them broke off.... I have no idea if SIMM slots have a common footprint, or if you anyone is still making them. If this is a "common problem with AWE32s" (which I now believe!) then I'm hoping someone has already done the legwork of finding the solution.

[appiah4]

If anyone knows off the top of their head, are there easy-to-source replacements for those SIMM slots? I've had an AWE32 with expanded memory since the 90s, and then recently removed the RAM to clean the board. Naturally, the plastic retainer on one of them broke off.... I have no idea if SIMM slots have a common footprint, or if you anyone is still making them. If this is a "common problem with AWE32s" (which I now believe!) then I'm hoping someone has already done the legwork of finding the solution.

I bought and replaced slots on my AWE32 only last summer, I sourced them from an eBay seller. His dual slots fit perfectly on the card and have metal retainer clips so no more annoying damage on them.



Unfortunately he seems to have run out But eBay will likely be your only friend. The listing I bought from was called: "Amp Connector Simm/dimm Socket, 30+30 Contacts 822113-2" A simple search came up with this:

https://www.ebay.co.uk/itm/1x-AMP-82211 ... rk:10:pf:0

Only one left though.

[cyclone3d]

If anyone knows off the top of their head, are there easy-to-source replacements for those SIMM slots? I've had an AWE32 with expanded memory since the 90s, and then recently removed the RAM to clean the board. Naturally, the plastic retainer on one of them broke off.... I have no idea if SIMM slots have a common footprint, or if you anyone is still making them. If this is a "common problem with AWE32s" (which I now believe!) then I'm hoping someone has already done the legwork of finding the solution.

You can get them here for $1.36 a piece... but you have to have a $20 minimum order.

https://www.peconnectors.com/sockets-pga-cpu-and-memory/hws9229/

[SirNickity]

Roger that. Thanks guys. This one's been on my to-do list and just hasn't come off yet.

[kodi]

the part is "AMP MICRO-EDGE SIMM CONNECTOR" - attaching PDF that contains table with all part numbers for it.

so, you can search for "AMP 822058" or "AMP 822113", etc where those numbers are from the aforementioned table. Please, note some of the part numbers in the table correspond to 72-pin, not 30-pin connectors.

in other words, there are plenty of sources, depending on where you live or how much you wanna spend.