keropi wrote:

^ you need to buy the GALs from suntac to get this new version , only the older CT1336-only version is free

Thanks man, I'll buy that. However, would be still interesting to know, what the difference between CT1336 and CT1336A related to CMS is. Seems funny, that a lot of people are trying to figure out the problem and one pal comes silently with a solution in his hand 😲 May be the solution can be re-reversed engineered 🤣

suntac spent time and figured this out even when the original creative pals don't work.
Stating you want to copy his work is just a dick move plain and simple.
It doesn't even cost that much, you give the man some money for his time and you get CMS working on your cards , it's only fair.

unless you mean you want to tackle the problem yourself, then go ahead the sky is the limit 😀

keropi wrote:

suntac spent time and figured this out even when the original creative pals don't work. Stating you want to copy his work is jus […]
Show full quote

suntac spent time and figured this out even when the original creative pals don't work.
Stating you want to copy his work is just a dick move plain and simple.
It doesn't even cost that much, you give the man some money for his time and you get CMS working on your cards , it's only fair.

unless you mean you want to tackle the problem yourself, then go ahead the sky is the limit 😀

You got me wrong, I really discourage piracy and didn't aim to copy his work, despite that Creative could see what we are doing here as piracy either. That's why I told, that I'll buy the chips, if you recall. However, I am really just curious, what's the difference.

You know, I guess I kind of get it, that he went to the trouble to fix something that's been an issue for years, and so he technically has the right to keep it proprietary. OTOH, everyone here benefits from standing on the shoulders of giants, and we're all in this niche community struggling against odds to keep old machinery working and available. So knowing the solution and keeping it to yourself seems counter-productive and dare I say a little selfish. Are you ever going to make enough cash from the sale of a couple flashed ICs to make it worth the trouble? Probably not. Is it just bragging rights, or wanting to be known as the guy who fixed it and now holds the keys? Ugh.. ego... how tiresome. What good is EVER going to come from that?

To me, it's a puzzle to be solved. I've already ordered an older SB 2.0, and am just waiting to find a newer SB 2.0 and then I'll buy that. I will do what I can to solve this for myself just because it makes me curious. If I am able to solve it, I will play a game or two with terrible CMS sound, and think -- cool, that was fun.
And then probably never use it again.

I'm not 100% sure whether I would be violating some code of ethics by making any such discovery public since one of the inner circle is trying to keep it under wraps, but I would be willing to be that moving dick. It would bother me to know the fix, and intentionally keep it from everyone else. I'm more interested in knowing someone else got their card working than making a buck fifty in total profit (after spending $300 for rare parts to R.E.....), or having my made-up name in lights (like anybody would remember or even give a log-jam who I am after getting their stuff working anyway.)

To each their own. Though I disagree with that tactic, and am unabashedly outspoken about it, I don't hold it against anyone. You do what you feel you gotta do I suppose. And in a tiny way, I'm kind of glad that it hasn't been publicly solved. That would spoil the fun of figuring it out.

There are 8 explicit equations in the publicly available GAL code and 5 of them are more or less flawed.
They've probably hired a programmer to do a HW-designer's job.

Holy crap! Calm down people, please, I really didn't want to give this whole topic such a bad taste direction. I was really just curious, didn't want to earn money with it or s.t. else, but just was curious.

Perfectly calm, and again, suntac is well within his rights here. I don't dispute that.

It's just disappointing to see a community working together to solve a problem (in this case from what... 2012?), and then someone does but doesn't really want to share what they figured out. Don't get me wrong, it's cool he's selling the solution. Most people are just going to want to buy the parts and be done with it anyway, and so for all intents and purposes, the problem is solved. Right up until he decides he doesn't want to deal with it anymore, or just vanishes, or whatever. And then we're back to square 1 -- except also now a couple people have a working part that they're morally obligated to not reverse engineer. It's no longer closed-source work of a public, soul-less company. Now it's a human being's intellectual property, and we can't cross THAT line. 😜

I see this stuff happen all the time and it just seems so pointless. There's nothing substantial to gain from this as a venture, so... why not document it to the best of one's ability and put it out there for anyone and everyone who needs it? I'll never understand it.

I see it following way: it is cool, that suntac found the solution, kudos man! He invested money and time to research and implement a solution and it is fine to get the money back (or at least a part of it). For the rest of the community his work is still a great contribution, since now we know, that there is a solution. First I just wanted to upgrade my card, because I never heard CMS working, just for fun. If we desperately need one working sample now, we can buy it, for the curious rest of us it is just a matter of time. Yesterday I wanted to buy a GAL already, but today I understand, that I would prefer to analyze this instead, I already connected the logic analyzer to the sound card, however, since I don't have a working sample, it will be complicated to find a proper solution. Anyway, it will be a nice time killer 😀

Anyone may reverse-engineer my GAL code, or may decide to find the solution themselves. Anyone may do whatever they wish and what makes them feel good or they think is right. 😀

When I wanted to hear the C/MS, I just designed an built this minimalistic GameBlaster-compatible device:

The PCB is completely home-made, single copper layer for simplicity. The (incomplete) filtering stage is on the bottom side.

The other side looks like this:

Very cool, I already played around with this https://www.youtube.com/watch?v=0UxQgYc3ELU

However, it's just a matter of fun to have it on the old SB 2.0 to be able to try it in some real games (there are not too many anyway)

scorp wrote:

However, it's just a matter of fun to have it on the old SB 2.0 to be able to try it in some real games (there are not too many anyway)

My device can be connected to the ISA bus with an adapter and its autodetection feature works too. SB 2.0 can't support C/MS autodetection, fortunately most games don't care. 😀

Btw, the old Prince of Persia 1.0 usually sounds weird on C/MS. The trick is to use as slow CPU as possible.

suntac wrote:

My device can be connected to the ISA bus with an adapter and its autodetection feature works too. SB 2.0 can't support C/MS autodetection, fortunately most games don't care. 😀

That is awesome, very nice work.

suntac wrote:

Btw, the old Prince of Persia 1.0 usually sounds weird on C/MS. The trick is to use as slow CPU as possible.

Thank you for the note, I'll try my 386sx-16 when I'm once so far 😀

scorp wrote:

Thank you for the note, I'll try my 386sx-16 when I'm once so far 😀

I meant something like an 8088 at 4.77MHz 😀

suntac wrote:

scorp wrote:

Thank you for the note, I'll try my 386sx-16 when I'm once so far 😀

I meant something like an 8088 at 4.77MHz 😀

hehe 🤣 ok, then I'll have to get one first. That 386sx is the slowest one, I have. However, with turbo off it's around 8088 by 8 MHz, may be it is already enough?!

scorp wrote:

hehe 🤣 ok, then I'll have to get one first. That 386sx is the slowest one, I have. However, with turbo off it's around 8088 by 8 MHz, may be it is already enough?!

The closer it is to an 8088 4.77MHz, the better the game sounds.
I am curious what causes this behavior, but there is no motivation to find out.
And my kids prefer to play PoP 1.3 with Roland sound. 😀

Hi guys, yesterday I finally got all the hardware required to analyze this. I ordered everything in China and it takes forever to get it. However, I spent today already some time to look at it and, so far, it makes at least a lot of fun 😀 Anyway, if I burn the well known equation to the GAL, I can't use Yamaha chip, but can almost use the SAA1099P chip. So far, at least I can confirm the behavior. In my case I get a lot of distortion using CMS and due to my logic analyzer, this is because Yamaha gets the CS signal and is trying to snatch the control. The problem is pin14, which controls the CS forwarding to Yamaha. If I deactivate pin14 completely, I get beautiful CMS sound, however Yamaha is then not usable. This is, what everybody here already knew anyway. I'm currently trying to understand, how to distinguish the chip selection dependent on the input from CT1366A. As far as I can see, there are only 4 lines going into the GAL, pin1, pin2, pin3 and pin5. I think, pin3 acts as CS signal and pin1 and pin2 are the selectors of SAA1099P (first, second chip). Unfortunately I'm getting a lot of noise on the (cheap) logic analyzer, so I'm not sure what pin5 is. However, I don't want to make false conclusions and need a schematics of CT1366A. At least pins description. I could swear, I've seen one here on the forum, but I can't find it anymore. Googling didn't help much, but may be somebody has it?

Ah, just looked at YM3812 pinout pin5 on GAL corresponds to WR on YM3812 and is also hardwired to pin1 of both SAA1099P. Ok, don't understand yet, why pin5 is then needed to switch CS in the equations pin15/16 at all.

Oh guys, not everybody at once please! Anyway, after 100€ for parts, more than one month of waiting for delivery and about 5 hours of tinkering, I have a working solution. I tested it on two cards with CT1366A an both work just fine. The difference is quite marginal, but it took some time to understand that. The equation for pin14 has an error, at least for CT1366A. The CS signal has to be activated with pin1 and not pin2. Could you please test following code for another revisions of the card please? I think, that I have to adapt the equation to be universal, but may be this is also working on the other revisions already.

1CHIP SB20 16V8
2
3gnd=10
4vcc=20
5
6; Input
7p1=1
8p2=2
9p3=3
10p4=4
11p5=5
12p6=6
13p11=11
14
15; Output
16p13=13
17p14=14
18p15=15
19p16=16
20p17=17
21p18=18
22p19=19
23
24EQUATIONS
25
26p13.trst = /p4
27p13 = gnd
28/p14 = p11 * p3 * /p1 ; CS of YM3812
29/p15 = /p5 * /p3 * /p2 ; CS of lower SAA1099P
30/p16 = /p5 * /p3 * /p1 ; CS of upper SAA1099P
31/p17 = /p11 + /p6 ; RST of D-FlipFlop
32/p18 = /p5 * /p3 * /p1 + /p5 * /p3 * /p2; D of D-FlipFlop
33/p19 = p5 ; Clock of D-FlipFlop