New Documentation: An authoritative reference on the YM2612

For anything related to sound (YM2612, PSG, Z80, PCM...)

Moderator: BigEvilCorporation

Shiru
Very interested
Posts: 786
Joined: Sat Apr 07, 2007 3:11 am
Location: Russia, Moscow
Contact:

Post by Shiru » Sat Nov 01, 2008 5:55 pm

HardWareMan wrote:I mean I want to look at YM2612 crystal not YM3812.
Well, then you already know how much it will cost (~$90 decap without postage). I don't think it's possible to get near quality of photos at home, if you don't have good equipment and you never did it before.

Jorge Nuno
Very interested
Posts: 374
Joined: Mon Jun 11, 2007 3:09 am
Location: Azeitão, PT

Post by Jorge Nuno » Sat Nov 01, 2008 7:34 pm

I think it is the fumic nitric acid you are looking for... (The "normal" nitric acid won't do).
It's super corrosive and it's vapours can change the the skin color to yellow :shock:

Nemesis
Very interested
Posts: 773
Joined: Wed Nov 07, 2007 1:09 am
Location: Sydney, Australia

Post by Nemesis » Mon Nov 03, 2008 11:40 pm

Shiru wrote:About decapping and FM chips emulation: http://docs.google.com/View?docid=dd8kqn9f_13cqjkf4gp (OPL2 and OPL3).
Thanks for posting that, I hadn't come across that info. Those tables they provided confirm a lot about how I believed the operator unit worked internally. Both the sin and pow tables have 256 entries, and the sin table is a perfect match for the data my algorithm generates.

The power table is a little different however. Since the entries in the power table effectively run from 0x7FA-0x400, they didn't need to include bit 10, since it's always fixed to 1, so the entries in the real physical table run from 0x3FA-0x0. It's also interesting that they listed the entries in their power table in reverse, from 0x0-0x3FA. Did they read the table backwards, or does the index get inverted before a lookup to the power table for some reason? The result will be equivalent, but once again, it's interesting to see exactly how the real YM2612 performed its operations internally.
Well, then you already know how much it will cost (~$90 decap without postage).
$90? Is that all a professional decap costs? Hmm, for $90 I might pay to get the YM2612 decapped myself. In fact, I think I will.

I'll need to pick myself up a good digitial microscope to take my own pics. Anyone got any suggestions on a particular model? What kind of magnification do you think I'd need to get a good shot of the die? 1000x? The microscope will also need to have its own lighting built in obviously. I think I could afford to spend a reasonable sum on a decent scope. It'll make a valuble addition to my collection of equipment. I'll have to do some research. As soon as I've got a microscope picked out, I'll send away a chip to be decapped. I've got a few chips sitting around here already desoldered and ready to go.

I'm not really sure how much immediately useful info a decap will give us, but it'll be fun.

Shiru
Very interested
Posts: 786
Joined: Sat Apr 07, 2007 3:11 am
Location: Russia, Moscow
Contact:

Post by Shiru » Mon Nov 03, 2008 11:51 pm

Nemesis wrote:$90? Is that all a professional decap costs?
That said in the article.
Nemesis wrote:What kind of magnification do you think I'd need to get a good shot of the die? 1000x?
This (copy-paste this link, does not work otherwise) is 100x, so probably 1000x will be overkill.

Oliver_A
Newbie
Posts: 7
Joined: Sun Nov 02, 2008 9:32 pm

Post by Oliver_A » Tue Nov 04, 2008 2:14 am

Nemesis wrote: I'm not really sure how much immediately useful info a decap will give us, but it'll be fun.
The first component I would examine is the DAC.

Lord Nightmare
Interested
Posts: 19
Joined: Sun Oct 12, 2008 10:45 pm

Post by Lord Nightmare » Thu Nov 06, 2008 1:47 am

I got the high-res opl2 pics from Olli (he goes by 'yehar' on ircnet), and unfortunately, not all the chip is imaged at high res. There is also at least one smaller rom on it (roughly 16x7 bits), which could be key-shift or other LUT tables which are not imaged at a high enough resolution yet to be fully visible and decodable.
I'll see if I can prod Olli to join the conversation here.

LN
Last edited by Lord Nightmare on Thu Nov 06, 2008 1:57 am, edited 1 time in total.
"When life gives you zombies.... *CHA-CHIK!* ...you make zombie-ade!"

Nemesis
Very interested
Posts: 773
Joined: Wed Nov 07, 2007 1:09 am
Location: Sydney, Australia

Post by Nemesis » Thu Nov 06, 2008 1:52 am

I'd be interested in taking a look at those images. Any chance of putting the archive up on a file dumping site? (if Olli is ok with that of course)

I'd also be interested in hearing what kind of equipment was used to take the pictures. Did MEFAS take the pics, or did they use their own equipment, and if so, what equipment did they use?

yehar
Newbie
Posts: 5
Joined: Thu Nov 06, 2008 3:16 pm

Post by yehar » Thu Nov 06, 2008 6:56 pm

Hi, it's Olli writing...
Nemesis wrote:I'd be interested in taking a look at those images. Any chance of putting the archive up on a file dumping site? (if Olli is ok with that of course)
Here's one "dump": http://yehar.com/adlib
Nemesis wrote:I'd also be interested in hearing what kind of equipment was used to take the pictures. Did MEFAS take the pics, or did they use their own equipment, and if so, what equipment did they use?
We only paid MEFAS to do the decapsulation. They were willing to do more, but presumably at a much higher cost. I went to the material technology lab (or something to that effect) at Oulu university here in Finland, and they were kind enough to let me use their microscope that was used there to inspect materials. It had its own built-in color digital camera, which however didn't have enough resolution for reading the bits from the memory. The microscopy system was from mid-90's probably, I cannot remember details like magnification. I got better results by using my pocket digicam through the eyepiece, but it was quite cumbersome, so I didn't shoot the whole die.

If someone has access to a better microscope with a built-in camera and is willing to take pictures of the whole die of YM3812 (OPL2) and YMF262 (OPL3), I will gladly mail the decapsulated chips to them (drop me an e-mail at o@iki.fi). You will need an optical microscope that has a lighting system available that does not try to beam light through the sample (typically on a glass slide), but illuminates it from above. Older chips are easier to photograph because the line width is broader. The circuit on YMF262 was already significantly more dense than on YM3812, and I don't think we would have succeeded in reading the ROM's from it alone.

Nemesis, I got the impression that the sine table was sign-inverted, and that the exp table was as expected, the opposite of your analysis. :-)

bobthebuilder
Interested
Posts: 22
Joined: Thu Jun 19, 2008 4:36 am

Post by bobthebuilder » Sat Nov 08, 2008 7:41 am

How many pictures would that be and at what magnification?

yehar
Newbie
Posts: 5
Joined: Thu Nov 06, 2008 3:16 pm

Post by yehar » Sun Nov 09, 2008 8:50 am

bobthebuilder wrote:How many pictures would that be and at what magnification?
Hmm maybe 300 pictures per die and 100X magnification.

Lord Nightmare
Interested
Posts: 19
Joined: Sun Oct 12, 2008 10:45 pm

Post by Lord Nightmare » Mon Nov 10, 2008 7:32 pm

opl2 rom3:
This decoding is based on http://yehar.com/adlib/ym3812_rom3_1.jpg
Legend:
0 = dark/absent gate, 1 = light/present gate
C = clear/correct for this address, C? = probably correct.
X = a pullup or perhaps pulldown between every 2nd rom line,
only present on bits which need it. This was useful for checking
to be sure if the bits present are sane or not, as such:
The X will only appear if one of the bit gates immediately above
or below it is present.
Note that all bits may be inverted, and the bit order may be
backwards as compared to the usual MSB-at-left notation.

Code: Select all

0000000 C
1000000 C
X X
1010000 C
1011010 C
XXXX X
1100000 C
1100110 C
XX XXX
1101010 C
1101110 C
XXXXXX
1110000 (unclear bits ...ooo.)
1110100 C
XXX XX
1110110 C?
1111000 (unclear bit .....o.)
XXXX X
1111010 C?
1111100 C?
XXXXXX
1111110 C?
0000001 C?
      X
Thats what I BELIEVE the contents of ROM3 on the opl2 are. I'm not sure what it is used for yet. 4 bits address in, 7 bits data out.

LN
"When life gives you zombies.... *CHA-CHIK!* ...you make zombie-ade!"

Nemesis
Very interested
Posts: 773
Joined: Wed Nov 07, 2007 1:09 am
Location: Sydney, Australia

Post by Nemesis » Tue Nov 11, 2008 4:31 am

I've purchased some gear which should do the job of capturing images of the die. I'll give it a test run once it all arrives. I'm happy to take full die snaps of the YM3812 and YMF262 once I've got a process in place.

I'm planning to take full coverage snaps of the entire die in a grid, at a suitable zoom level. Once I've got all the pics, I'll stitch them together into one large psd image or the like, so you'll be able to download a full die shot of the device, and zoom in on the parts you want to examine. If nothing else, it'll make a good poster. :D

I've also contacted MEFAS and got the ball rolling with getting the YM2612 decapped.

yehar
Newbie
Posts: 5
Joined: Thu Nov 06, 2008 3:16 pm

Post by yehar » Tue Nov 11, 2008 10:19 am

Nemesis,

Nice! :-) I also just ordered a microscope, so I will delay sending you the chips until I have tried taking the pictures myself, before end of this month I hope. I bought this Aigo GE-5:

http://cgi.ebay.de/ws/eBayISAPI.dll?Vie ... AQ:DE:1123

Manual:

http://www.analogtechnologies.com/Aidep ... ual_EN.pdf

I'll let you know if it's any good. The 1/2 " size 1280 x 1024 pixel sensor and 60X, 180X and 540X all optical magnification should give a resolution of something like 90 nm, 30 nm and 10 nm per pixel. The price of the microscope was 350 EUR including postage.

yehar
Newbie
Posts: 5
Joined: Thu Nov 06, 2008 3:16 pm

Post by yehar » Tue Nov 11, 2008 10:43 am

Lord Nightmare,

I get exactly the same decoding. If you look at a pullup, the stuff just above and below the dark dot, that will already tell you to which ROM line the pullup is connected.

yehar
Newbie
Posts: 5
Joined: Thu Nov 06, 2008 3:16 pm

Post by yehar » Thu Nov 20, 2008 6:37 pm

yehar wrote:I also just ordered a microscope, so I will delay sending you the chips until I have tried taking the pictures myself, before end of this month I hope. I bought this Aigo GE-5
OK, having played with it an hour, it is not as good as I hoped. The 540x zoom will only give a resolution of 500 nm / pixel. The magnification ratio is probably calculated for a typical computer screen, and is NOT the ratio of enlargement to the sensor as I had thought. Currently it is not possible to read bits from memory with this setup, or even to follow the traces on the die.

Now I have to use a lamp from the sides and that gives nasty highlights and not a smooth lighting. Perhaps a modification would help. The sample could be illuminated from above through a beamsplitter. Dunno.

Post Reply