nesdev.org

The corruption pattern can change when I power cycle or reset the NES. The picture of the menu after the mode is one of the lucky time it wasn't so bad. I even was able to briefly get an almost perfect menu yesterday but when I booted a game is was as unplayable. I tough it might be the connection with the 60 to 72 pic adapter even though I tested with multimeter so to completely rule that out I soldered it directed to one of the adapter like this:

http://skinnyv.com/Cart/245in1/Direct_Connect.JPG

Test each connection with a multimeter for continuity and for short. All was perfect. Tested it ion my NES, corrupted graphic, and not the nice one like on the picture. I must say it is a very strange cart.

So I'm at a loss once again:) By the way, I was still wondering about the VCC, I've been instructed to lift the RAM's VCC pin and connect to +5v on edge connector but I don't think it was running on 3V at all, it seem to be connected directly to pin 30 of the Famicom edge connector which, as far as I know is 5v...

So I've actually got a similar failure with the horizontal lines playing around with my kazzo and a SRAM NROM cart.



It turns out dumping the cart and comparing hex files, that the very first byte in CHR memory is corrupt. It's 0xA0 and supposed to be 0x00. I still haven't figured out why exactly. But I had a similar problem where the first byte in PRG memory was programming improperly. It ended up fixing it's self when I realized the retrozone cart I was using had the /CE line on the PRG memory tied to GND. So basically anything that was being write controlled by PRG /CE was failing, but writes that were controlled by PRG /WR were working properly. Connecting the /CE pin to PRG /CE actually fixed a lot of other problems I had too with corrupt data and in ability to battery back my SRAM well also.

I don't know if all this helps much but it's one way I was able to get horizonal lines with a completely different setup. My error is in the writing process to the SRAM not reading it. Basically it makes me think that it's corrupting data for something like the first byte of several or most tiles or something. That gives a horizontal line and the follow on bytes of the tile or block of memory program properly. IDK I'm making a lot of guesses on how things are actually working but it sounds possible to me atleast.

When you checked all the lines did you compare it to the pin out of the SRAM? Like did you ensure the /CE line is actually connected to CHR A13 on the cart edge? I would recommend checking all the control signals making sure they make sense where they wire them up. That's basically why I had issues, the board is designed for EPROM or FLASH where it wouldn't be an issue grounding /CE and not letting the NES actually enable the chip. But it caused hell with my SRAM even though the pinout was the same. I know it's a different story on the CHR side with VRAM and all but it atleast gives you something to look at. Perhaps if I find my error on the CHR side it'll be related to yours as well, I haven't figured out why I'm having an issue with that first byte...

The VRAM's /CE1 line is going to the mapper. The /CE2 is connected to +5V.

Is it just me or is it strange the mapper is controlling /CE? /CE is normally A13 from what I'm used to but this isn't a standard mapper either.

What about write and output enable? I would imagine the NES has control of those...

VRAM'S /WE is conntected to famicom pin 47 CHR RAM WR. As for VRAM'S /OE, my multimeter is telling me it's connected to ground and +5v but I wil lrecheck this one as I was rushing.

SkinnyV wrote:VRAM'S /WE is conntected to famicom pin 47 CHR RAM WR. As for VRAM'S /OE, my multimeter is telling me it's connected to ground and +5v but I wil lrecheck this one as I was rushing.

I can see /OE grounded, it doesn't allow for making use of CHR /RD signal but it's not of much concern.

So what does CHR A13 line do? This is should basically one of the /CE signals for the VRAM. Double check to see if it's tied to either of the /CE pins on the SRAM.

If it's not connected I'm really surprised now that I think about it a little more. The /CE line MUST be low when writing or reading to/from VRAM, and it only makes sense for what's tring to Read and Write (the NES) to have control of /CE. For the mapper to have complete control of /CE the mapper basically doesn't give control where it belongs. If this is really the case it's a backasswards design and super difficult to time properly. I'm suprised it even works in a clone... Based on the fact you're getting different glitches every time supports the possibility of a timing issue, since the Mapper can't very well know what to do with /CE when.

I'd be curious what would happen if you connected CHR A13 to the spare /CE2 that's currently tied to Vcc. If it wasn't obvious you'd want to make sure you cut a trace or somehow removed it from Vcc, you could double check by checking the resistance of the CHR A13 line to Vcc afterwards and ensure it wasn't low (like less than 1K ohms).

I didn't check the datasheet for this part, but the 2nd CE line almost certainly a positive-enable, so it being at +5V has it enabled (both CE's must be enabled).

So if the VCC pin for the RAM is at 5V, then I guess they were running the original 3V RAM chip at that.. strange.

There's nothing stopping the mapper from taking CHR A13 (or /A13) and then having it's own output to the CHR /CE on the RAM. The only reason I can think of for doing that, is if the mapper was also controlling the "CIRAM /CE", then it could map nametables into it's the cart's CHR-RAM (which would allow 4-screen mode, and nametable banking, which are kinda rare features).

I wonder, is the voltage of the clone really 5V? Maybe NES maskROMs and mappers can run at 3.3V, I can't say I've tried it or would know what to expect.

infiniteneslives wrote:For the mapper to have complete control of /CE the mapper basically doesn't give control where it belongs.

This is true in the case of one CHR memory. But when you have a split CHR ROM (e.g. mapper 68), or both CHR ROM and CHR RAM (e.g. TQROM or its Chinese clone), or allowing nametables and pattern tables to share one memory as Memblers mentioned (also mapper 68), you need the mapper to postprocess the PA10-PA13 to come up with the enables for CHR memories.

Memblers wrote:
There's nothing stopping the mapper from taking CHR A13 (or /A13) and then having it's own output to the CHR /CE on the RAM. The only reason I can think of for doing that, is if the mapper was also controlling the "CIRAM /CE", then it could map nametables into it's the cart's CHR-RAM (which would allow 4-screen mode, and nametable banking, which are kinda rare features).

Ahh yeah, it all makes sense now. Thanks for getting back on the straight and narrow guys.

I did some test now that I got back my Yobo NOAC clone and it definitely look like the clone is not outputting as much power as a real Nintendo system. I don't have exact value as I did it in a rush this afternoon but voltage was lower on the clone.

Still no luck yet running the cart on real nes though!

Oh yea, I should have mentioned that. My Yobo outputs 4.7 V when I use the supplied wall wart. Using my own 5.5v wall wart it outputs over 5v, but not quite 5.5v. I never tested the leads on the warts, I should probably do that.

Anyway, I did notice that CV3 does not boot at all when using the wall wart supplied with the Yobo (goes to grey screen). It does boot when using mine, but fails to enter game play due to the CIRAM bug.

Sorry to necrobump, but was this ever figured out? Couldn't you just add a voltage regulator to the Vcc line on the chip, and drop it down to 3.3v that way?

It was unfortunatly never resolved, I just shelved thoses 2 cool cart and I now consider them a loss.

With a help from our friend from polish forum dic-sc7, we found a solution to make this cart working on real hardware (Famicom/NES). All you need to do is simply desolder OE pin from CHR-RAM and solder it to unused CHR/RD pin 17:

Wtf, after all the crazy shit I tried with that cart (soldering directly to a famicom adapter, replacing ram with 5V IC etc...) I can say that this simple mod seem to make the cart work perfectly on a North American NES.

Dziękuję! Go poland!