DDMA mode and enabling Sensaura..

Not a very good card... The FM implementation is lacklustre.

Perhaps not the best FM synth implementation on a PCI sound card but it has good SB Pro compatibility, and there's 3D positional audio for Windows games.
I think I found the same card on AliExpress (and your review of it, the screenshots match), the chip is an ES1930S, an Allegro (which is also an ES1989), not the ES1938 Solo-1.
Interesting that this is labelled a Maestro-2SC as that name doesn't appear in datasheets. Other ES1970 variants like the ES1970S-3D and ES1970MS-3D are labelled Canyon3D, though the latter is a 144-pin TQFP instead of 100-pin. The lack of EAX 2.0 and ZoomFX support means this really isn't a Canyon3D.
The regular Maestro-2 is an ES1968S and Maestro-2E an ES1978S.
Are you able to dump its PCI configuration? I usually use lspci -vvnn on Linux, for DOS Craig Hart's PCI is one option.

holdencars11 wrote on 2024-08-22, 13:00:

Interesting so far that I've managed to get DDMA, TDMA and whatever SiS DDMA working.

WIth what chipset?

Grzyb wrote on 2024-08-26, 00:22:

holdencars11 wrote on 2024-08-22, 13:00:

Interesting so far that I've managed to get DDMA, TDMA and whatever SiS DDMA working.

WIth what chipset?

DDMA with 440BX, TDMA with KM400 and KT266A and SiS DDMA with SiS730

I also have two Maestro 2SC sound cards bought on AliExpress, so far I have not spent much time researching the drivers, although I have to do it because for example in Quake 2 there are obvious sound problems that even affect the performance of the game, while in GLQuake, changing the AWE64 value ISA I was using for this card increased the FPS from 37fps to 40 on a PMMX 262mhz and a Rage XL PCI also from AliExpress.
As for the hardware, I found the datasheet I attached which is for the Maestro 2 and from my tests I think it is totally valid for the Maestro 2SC.

At the moment I modified these sound cards by changing the small L7805 for a more powerful one, the one that comes with these cards gets very hot, I think this is the power supply of the CODEC AC97, they didn't use de 5V directly from the PCI slot, maybe to reduce the noise coming from the PSU.
I also modified them so that they could work in any AT motherboard that does not have the 3.3v present in the PCI slots by adding an AMS1117 3.3 v with its capacitors, since the Maestro 2SC is powered with the 3.3v of the PCI slot as shown in the datasheet.

I also added the connection for the audio from the CDROM.
What I miss the most in these cards is the game port, the Maestro 2SC has the functionality, but it has not been implemented in these cards, in the future I will try to investigate the possibility of adding a game port somehow.
By the way, I think these cards are already old stock because of the date printed on them, if I'm not mistaken, 3714, could be week 37 of 2014, one of my cards also has the same date.

Interesting with the mods! I did see in one of the pictures posted on AliExpress that someone got a Canyon3d and Analog Devices codec chip instead of a 2SC...I kinda wish that I got one of those... Was keen to try the MX400 back in the day but never got around to it.

I'm yet to try the lower SPK2 jack for 4 speaker mode on the 2SC.... Is it active on the chip?... I think Canyon3D had 4 speaker, but I didn't find any mention off it on the Maestro2E data sheet... Unless it's a 2SC feature.

I like this kind of recycled cards they make in China like the Rage XL, from what it says on the datasheet of the Maestro 2, I understand that this chip was intended more to be integrated into motherboards rather than used in a standalone sound card. Maybe they were a bunch of old stock chips, or maybe recycled chips from motherboards.
I like them even more because these cards cost the same as a beer and are simply useful and functional.
In case anyone is interested, I add a couple of pictures of the mods, for powering the chip it is advisable to cut the power input from the PCI connector.

What drivers did you use to activate Sensaura? Is it Win98?

There is an archive on ALSA that has datasheets for most of ESS' PCI sound chipsets apart from the original Maestro and the Canyon3D series.

holdencars11 wrote on 2024-08-26, 16:51:

I did see in one of the pictures posted on AliExpress that someone got a Canyon3d and Analog Devices codec chip instead of a 2SC...I kinda wish that I got one of those... Was keen to try the MX400 back in the day but never got around to it.

Hmm... I see that one but I'm not sure it's a real Canyon3D. The layout of that board looks similar to yours and the 4-pins on the top left corner on the left edge (pins 76-79) are connected to the AC97 codec as SDCLK1, SDFS1, SDI1, and SDO1, according to the Maestro-2(E) datasheet. There's no datasheet for the Canyon3D but the product brief for the Canyon3D 2 mentions an integrated quad-channel AC97 codec and the pinout for those 4-pins are AFILT2, CAP1, CAP2, and LINE_OUT_L. Those 4-pins also match the pinout for the Allegro, which has an integrated stereo AC97 codec.
As I was searching for information on the Maestro-2SC, I came across this 2009 post on a Chinese forum, putting through Google Translate:

New Diamond MX400 (Made in Shenzhen)
MX400-----》The last glorious farewell work of Diamond sound card, there are fakes. The difference is quite obvious, and there are 3 features to distinguish: First: See if there is an inductor coil near the coaxial, if there is, it is genuine, if not, it is fake, second: F1 uses green fuse resistor is genuine, and the one with yellow fuse resistor is fake. Third: Look at the main chip, the real MX400 uses ESS Canyon 3D, while the fake MX400 uses ESS Maestro-2SC chip. Moreover, if the fake MX400 is installed on win98se system, it is impossible to install Diamond's original MX400 driver.

holdencars11 wrote on 2024-08-26, 16:51:

I'm yet to try the lower SPK2 jack for 4 speaker mode on the 2SC.... Is it active on the chip?... I think Canyon3D had 4 speaker, but I didn't find any mention off it on the Maestro2E data sheet... Unless it's a 2SC feature.

I have a recycled Solo-1 (pictured here) which also has a SPK2 jack. Never bothered trying it as I no longer have a 4-speaker setup but the Solo-1 is stereo-only, like the Maestro-2(E).
The PCI ID might shed some light as to what the Maestro-2SC might be.

jtchip wrote on 2024-08-26, 00:14:

Are you able to dump its PCI configuration? I usually use lspci -vvnn on Linux, for DOS Craig Hart's PCI is one option.

From Puppy Linux Wary:

1Multimedia audio controller [0401]: ESS Technology ES1978 Maestro 2E [125d:1978] (rev 10)
2	Subsystem: ESS Technology ES1978 Maestro 2E [125d:1978]
3	Control: I/O+ Mem- BusMaster+ SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
4	Status: Cap+ 66MHz- UDF- FastB2B+ ParErr- DEVSEL=medium >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
5	Latency: 32 (500ns min, 6000ns max)
6	Interrupt: pin A routed to IRQ 20
7	Region 0: I/O ports at d000 [size=256]
8	Capabilities: [c0] Power Management version 2
9		Flags: PMEClk- DSI+ D1+ D2+ AuxCurrent=0mA PME(D0-,D1+,D2+,D3hot+,D3cold-)
10		Status: D0 NoSoftRst- PME-Enable- DSel=0 DScale=0 PME-
11

I have found the AC97 1921S codec datasheet used on these cards and even with my limited knowledge, I only see one stereo output with left and right channel output and one mono output, I don't think more than two speakers will work properly.
The 4558D output amplifier also supports only two channels. Everything indicates that only one speaker output will work correctly, I doubt that the four speaker functionality would work as it should.
The CD audio connection does not work either, even though it is correctly connected to the corresponding pins of the AC97 codec, maybe it is a driver problem, since I have found several that seem to work, or my CD-ROM drive is faulty, but that would be estrange since it reads data discs correctly.

@holdencars11, thanks for the PCI dump. Looks like it's a newer revision of the Maestro-2E (rev 10h instead of 00h). Almost certainly still 2-channel output only. I wonder why they bother fitting a second output jack given that the rest of the design is so cost-reduced.

Perhaps to reuse existing brackets and not get new ones made.

Tiido wrote on 2024-08-28, 06:27:

Perhaps to reuse existing brackets and not get new ones made.

Yeah it's what I'm thinking too... Maybe when they run out of Maestro they'll switch back to 8738.

I have also seen some pictured with VIA Envy24/Tremor in pci-e form. Who knows if you get a 8738 or Envy24...

It's all a little strange that they describe it as 8738 but give people a Maestro... Why not be honest about the card and update the description? I can see why some have complained in AliExpress when they can't get the card working..

I think ESS TDMA seems to have greater success rate than CMI in DOS... I will have to do further testing.