DSP1 question

[Markfrizb]

What is the 3 pin oscillator??? that the dsp1 uses? And can a modern replacement be bought?

[lidnariq]

The blue blob linked to in this post in the link "SHVC-SOUND"? viewtopic.php?p=117718#p117718

It's a ceramic resonator, made of PZT. You can still get them, although ROHS compliance can make it harder.

[Markfrizb]

I should have been more specific.... The resonator I'm referring to is from a Mario kart cartridge.
I don't know what blue blob you are referring to.
Anyways, so if I order a resonator, what frequency do I order?

Thanks!!

[lidnariq]

The number on it should indicate. e.g. https://en.wikipedia.org/wiki/Ceramic_resonator

[Markfrizb]

Which number is relevant here?

Top number looks like 7600A
Lower number looks like 49_3

I assume kHz or is this in hz?

Thanks for any help. I really want to source a new part instead of robbing old parts from carts.

[lidnariq]

Best guess is 7.6MHz.

http://www.assemblergames.com/forums/ar ... 35788.html mentions 7.6MHz in the context of repairing a DSP-1 board, too.

They don't seem to make all that many 7.6MHz resonators anymore; I'm not clear if anything bad will happen if you use something too slow (7.37MHz) or too fast (8 MHz). The above linked thread also points out that you can buy a full oscillator at the right speed and put it in, too. You can probably also manage to build/buy a inductor and capacitor that will resonate at the right frequency.

edit: couldn't stand to let stupid grammarfail remain

[Markfrizb]

Thanks!

[Markfrizb]

http://www.mouser.com/Search/ProductDet ... 7M68G53-A0



This looks pretty close. Any reason why not?

[lidnariq]

Sounds good to me if the original is through-hole.

Any particular reason you think it's specifically the resonator that's broken? They don't have all that many things that can go wrong with them.

[Markfrizb]

Sounds good to me if the original is through-hole.

Any particular reason you think it's specifically the resonator that's broken? They don't have all that many things that can go wrong with them.

It's a through hole part. I was upgrading the super everdrive cart to the dsp1 and I wanted to use new parts (the HCU04 and the resonator). And for less than $1 it appears I can now.

Thanks for your help.

[nocash]

Wow, a photo showing the oscillator on DSP boards! That's one of the last and biggest SNES mysteries, I didn't believe that anybody would ever take a photo of it.
Yup, the photo looks like 7.600MHz, that's also matching up quite well with the uPD77C25 datasheet (which specifies max=8.192MHz).

But, the latest info from byuu (at least the latest that I am aware of), http://forums.bannister.org/ubbthreads. ... #Post66683 says 8.00MHz for DSP chips, where does that number come from? And which is right, or is it both true? There are various DSP1 games, plus some DSP2/DSP3/DSP4 games, plus the DSP1A/DSP1B revisions, so it's possible that some carts use 7.600MHz and some use 8.000MHz or even other clocks. Did anybody ever verify that?

PS. Duh, I just noticed that I do have a Super Mariokart PAL cartridge around here myself, that about the last unsolved mystery ;-)
Mine is using a "DSP1" chip, and a "[M]7600A, 271" oscillator (the package looks same as on Mark's photo, except that it's light blue, has a red dot on the top, and the "27" part is underlined).

[nocash]

Btw. the ST010 clock is also unclear:
http://wiki.superfamicom.org/snes/show/ST010 says 22.000MHz
http://forums.bannister.org/ubbthreads. ... #Post66683 says 10.00MHz effective (20MHz / 2)

[Near]

DSP1: http://www.snescentral.com/pcbboards.ph ... VC-1K1X-10
DSP2: http://www.snescentral.com/pcbboards.ph ... VC-1B5B-02
DSP3: http://www.snescentral.com/pcbboards.ph ... VC-1B3B-01
DSP4: (no photo, sorry)
ST010: http://www.snescentral.com/pcbboards.ph ... C-1DS0B-01
ST011: http://www.snescentral.com/pcbboards.ph ... C-1DS0B-10
ST018: http://dforce3000.de/dmp/816_1261738278.png
Cx4: http://dforce3000.de/dmp/705_1261616456.png

I own boards for every special chip except for the OBC1 right now.

I lack a logic analyzer / oscilloscope, so I cannot clock these chips. But if you want photos, I can do photos.

Most of them are just solid white boxes. I went off of the MHz values written directly on the PCBs for my numbers.

If proven otherwise, I'd love to correct the values.

As it stands, I don't even know with certainty if every instruction on all of these takes one clock cycle, or more than that, or if it's a per-instruction deal.

[nocash]

Where do you see MHz values on the PCBs??? I see them on ST018 and CX4 boards, but not on DSP1-DSP4 or ST010/ST011 boards.
Hmmm, do you mean that oscillators are really "just solid white boxes" without any part numbers? My own cartridge and Mark's cartridge (see above phote) DO have part numbers though. The snes central ST011 photo does also show "15.something" on the ST011's ceramic oscillator. If you could make photos of the DSP/SETA oscillators or scribble down their part numbers would be great! If you do so: Please also include info which game it is, and if it's using one of the DSP1A/1B revisions.

I've my SNES in storage at the moment so I can't measure the clock (using an oscilloscope won't give very precise results anyways).

The datasheet isn't too clear about per-instruction timings. As I do understand it, they do simply mean that all instructions take 1 clock cycle. With the internal memory (and with it being separated into code & data areas) it should be possible to get that 1 cycle speed. The multiply function is probably executed on every opcode (no matter if the opcode does read/write the multiply result/parameters), if so, then multiplication shouldn't take any extra time either.

[Overload]

ST010 - Exhaust Heat 2 (Japan)



ST018 - Morita Shogi 2 (Japan)