OPL3 Copies

A while ago, a reader commented that in certain circles, it’s well known that there were “fake” OPL3 chips. This does not appear to be widespread knowledge. After a bit of digging, an interesting chapter in the history of PC hardware unfolds.

First let’s remember that the Yamaha OPL3 FM synthesizer, typically an YMF262 or an YMF289B, was an essential part of a Sound Blaster compatible sound card (let’s forget about exceptions like Ensoniq SoundScape and its bizarre OPL3 “emulation”). At the same time the Yamaha chips were somewhat expensive, and protected by a patent.

It is apparent that someone manufactured OPL3 copies, disguised them as nondescript chips, and sold them to many more or less reputable sound card OEMs. It is not clear what exactly these chips were. However, the sound they produce is normally close enough to real Yamaha chips that they were likely manufactured based on stolen “blueprints” of the originals. And the fact that companies like ESS, Crystal, OPTi, or Aureal couldn’t produce 100% accurate OPL3 clones strongly hints that there’s something fishy about these chips.

It is interesting to contrast these fake chips with the far better known instance: fake cache chips. To recap, in the mid-1990, disreputable companies sold motherboards with dummy pieces of plastic pretending to be cache chips. These chips had no actual function.

The fake OPL3 chips are the opposite. They are actual functional equivalents of Yamaha OPL3 chips (quite possibly exact copies), but pretend to be something else because they were not produced legally. This is not to be confused with the multitude of legally developed OPL3 clones. For the sake of clarity, let’s call the illicit fakes “copies” and the legal workalikes “clones”.

There is absolutely no documentation about the OPL3 copies. It’s not known how many there were, how much they sold for, or who produced them (but Made in Taiwan and/or Made in China is a near certainty). The fact that there is no hint who made them is itself highly suspicious. What’s left is evidence in the form of vintage sound cards using these chips.

The Chips

Let’s take a look at a few of the known Yamaha OPL3 (or, in one case, Yamaha DAC) equivalents.


Let’s start with the LS5122, a functional equivalent of the Yamaha YAC512-M DAC. Strictly speaking, this might not be a copy, but the chip is very shady. There’s no hint who manufactured it or when, no datasheet, no nothing. It’s only obvious what it does—compare the following two close-ups:


LS5122These are photos of two Aztech Waverider 32+ boards, exactly the same PCB and model, both manufactured within the first few weeks of 1995. Both sport a genuine Yamaha OPL3 chip, but while one has a Yamaha YAC512-M DAC next to the OPL3 chip, the other has a mystery LS5122 instead. LS512, YAC512… probably not a coincidence.


The next specimen, and perhaps the most widespread one, is a pair of two chips, LS-212 and LS-215. The LS-212 is an YMF262-M OPL3 copy, while the LS-215 is a YAC512-M DAC equivalent. 215 is 512 backwards, again probably not a coincidence.

How do we know these are really copies? The OPTi 82C924 does not include an OPL3-compatible core, yet the board supports OPL3, so the chip has to be there somewhere. The LS-212/LS-215 pair is really the only suspect.


There’s even better evidence though. A leaflet provided with “Miss Melody” sound card (ESS688-based) shows a board schematic with YAC512 and YMF262 chips very clearly marked. The actual board sports LS-212 and LS-215 instead.

The chip names seem to be deliberately chosen to resemble very common components such as the LS245 bus transceivers or LS125 quad bus buffers. It is also plausible that there’s some connection between LS-215 and LS5122.


These chips have been seen on a mid-1995 ProComp Pro-Multimedia card next to an OPTi 82C929A controller. They appear to be straightforward copies of the YMF262-M and YAC512-M, respectively.


The excellent AmoRetro has a beautiful photo of a 1995 Shuttle HOT-233 where these copies can be seen as well, again next to an OPTi 82C929A controller.


FT6116 is a Force Technologies 2Kx8 CMOS SRAM. The FT6116-100 is… certainly not. Obviously a cheap Taiwanese sound card has no need for a SRAM chip. What’s more, the Avance Logic ALS100 does not include an OPL3 code (unlike its successor, the ALS100+), so an OPL3 chip has to be on the board somewhere.


The FT6116 is clearly not what it claims to be, and it just happens to match the YMF262-M package. Probably not a coincidence.


Last on the list is the DXP44Q. This chip appears to be a pin-compatible copy of the Yamaha YMF289B, also known as OPL3-L. This was a low-power chip with power management suitable for portables (and found in the ThinkPad 701C for instance).


Once again, the OPTi 82C930A does not include an OPL3 core (unlike its successor, the 82C931). Once again, there has to be an OPL3 on the board somewhere, and the DXP44Q is more or less the only candidate. The 82C930A does include a DAC, which explains the lack of a separate DAC chip.

Brief History

These mystery chips first appeared in 1994 or so, when Yamaha was still the only supplier of OPL3 chips and there was a large market for sound cards. They quietly disappeared again around 1997 when more or less all sound controller manufacturers (Crystal, ESS, OPTi, Avance Logic) integrated an OPL3 clone into the core chip.

Creative Labs does not appear to have ever used these chips (likely due to a licensing deal with Yamaha which allowed Creative to integrate an OPL3 core into its own chips), but Aztech and many others did. The fake OPL3 chips are most commonly, but by no means exclusively, found next to an OPTi controller. They have also been seen paired with ESS and Avance Logic controllers, and there’s no reason to think they couldn’t be used with others.

The affinity with OPTi controllers could be simply a function of the fact that OPTi had well established and cost-effective controllers available in 1995. ESS and Crystal went relatively quickly from zero to integrating an OPL3 workalike (ESS1688, CS4235), obviating the need for external chips.


It’s likely that there are other OPL3 copies out there. Perhaps readers have seen some? If in doubt, photos can be used to judge what’s a likely OPL3 copy and what’s not. It’s known which audio controllers include what FM synthesis capabilities, and if a card uses a FM-less controller yet plays OPL3 music, then an OPL3 chip has to be there somewhere. It’s usually not hard to spot.

This entry was posted in PC hardware, PC history, Sound, Yamaha. Bookmark the permalink.

14 Responses to OPL3 Copies

  1. random lurker says:

    I have nothing to add but I just wanted to say that this is extremely fascinating!

  2. FlipFlop says:

    These clones of OPL3 started to appear when John Chowning’s patent of FM synthesis expired in 1995. It was exclusively available to Yamaha and no other company could use this technology back then. That’s why original OPL3 dissapeared even from Sound Blasters, cause it was replaced by Creative’s CQM (Creative Quadratic Modulation) synthesis chip.

  3. Michal Necasek says:

    This article isn’t about clones, it’s about apparent illegal copies of Yamaha’s chips. There are lots of known OPL3 clones (CQM, CrystalFM, ESFM, OPTiFM, and many others) but LS-212, DSP24S, or DXP44Q are not among those. The copies were also available in early 1995 and almost certainly late 1994, even before the Chowning patent expired.

    BTW Creative itself was highly inconsistent; I have a SB 16 (CT2940) right here that’s from late 1996 and sports a genuine YMF289B OPL3-L chip.

  4. MiaM says:

    What was the transistor count on such chips?

    Early nineties coincides with the restructuring of the former east bloc. Could theese chips have been produced by whatever were left of some of the eastern block chip manufacturers?

  5. Michal Necasek says:

    I can’t find any data on the OPL2/3 transistor count. I can’t imagine it would be more than a few hundred thousand but I really don’t know.

    I would guess it is something that East Germany or Russia could have reverse-engineered. But there was no East Germany anymore, and I really doubt these clones were sourced from early- to mid-1990s Russia. It would be very surprising if these chips weren’t made in Taiwan or possibly mainland China. They were almost certainly too cheap to be worth making anywhere else.

    It all seems like the “capacitor plague” where the purchasing departments replaced the specified chips with clones and the board designers didn’t even know about it. But in the OPL3/DAC clone case it had no apparent consequences.

  6. MiaM says:

    I was thinking that when East Germany ceased to exist there were both a chip manufacturing plant and people skilled to use it but I guess that their reputation might not have been good enough to compete the usual way, so actually manufacturing pirate copies would have made some sense.

    Looking at the board layouts it seems like some were actually designed to accept different width of the chips, i.e. rather long pads.

    The main concern for many of the people involved would probably have been to be able to lie trustworthy about not knowing what was going on. A designer could make up a lie like “someone from purchasing said something about different footprints so I designed it this way and they were happy about it” (or more likely be a bit wague about who on the design team actually made what).

  7. Michal Necasek says:

    I agree that East Germany had the people and technology. But they were basically completely taken over by West Germany. I’m sure any spare capacity they had would have been put to a better use than selling cheap copies of Japanese chips.

    The thing about longer pads also struck me, but I don’t know if YAC512-M was the only legit option. I think (but can’t remember exactly now) I saw a card which used some non-Yamaha DAC with an OPL3, probably a Crystal or Philips.

    How many people knew about it is a really good question. I haven’t come across any mentions in the contemporary trade press. Did no one really even notice at the time?

  8. CommanderK says:

    I come from eastern Germany, and believe me, there were no OPL copies made here. They cloned the VAX machines, the 8086, the Z80 (as U880), but sound chips where not part of it. I also hardly believe this was possible, as the OPL 3 originated in the 90s, way after Germany united.

  9. Ethan Queen says:

    I have in my hands an old Opti 930A based ISA sound card that has the DSP24S/DAP16S chips on it. The FCC ID is K33IFISP32.

    I’ve had it for years and was just discovered this page and found it very interesting as I am looking at a similar card.. a Shuttle HOT-237 that has a similar layout except that it has the real OPL3 chip and lacks a built in amp.

    The one I have also has a higher model opamp as well as being a 930A instead of a 930.

    Both of these cards have a built in hardware wavetable setup which was very nice back in the day.

  10. Michal Necasek says:

    Creative is the only vendor on whose cards I have not yet found fake OPL chips. Pretty much everyone else (Aztech, Shuttle, all the no-name Taiwanese vendors) used them, but typically not always, so you might get the same card with original Yamaha chips or with fakes. Clearly that didn’t result in customer complaints.

    What kind of wavetable synths are on your cards? Some AdMOS or OPTi I guess? Or possibly a Dream?

  11. Konrad says:

    KS8001+KS8002 (maybe someone wanted us to believe they are Samsung chips ?)
    74LS631+74LS74 (now this is funny, they even put them into DIP28 and DIP14, like we would never guess these are 3812+3014 in fact)


  12. GL1zdA says:

    Just found information, that the ALS100+ also doesn’t have an internal FM synthesizer. These cards have even more obfuscated OPL3 clones: https://www.vogons.org/viewtopic.php?t=74802

  13. Michal Necasek says:

    Very cool. I wonder why they bothered making the OPL3 clones look like something totally different.

  14. rasz_pl says:

    Add OPL2 clone SC801/SC802 on “Master Boomer J5QMF-002” https://www.vogons.org/viewtopic.php?f=46&t=90801
    and “Baby Boomer J5QMF-001” https://www.vogons.org/viewtopic.php?p=343769#p343769
    This card was also the source of first SoundBlaster clone DSP firmware dump for Snark Barker (Archon Electronics).

    J5QMF-001 and J5QMF-002 FCC filings from 1992 list Princeton Technology Corp (PTC) as manufacturer. PTC is a huge fabless IC design house. Are they the source of numerous OPL2/3 clones starting in 1992?

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.