nesdev.org

I guess the easiest thing for me would be to try to locate a similar ram chip that use 5v then. I'll look around the net tomorrow and try to order one. But I seem to remember that finding 5v IC was much harder nowadays... Is it a matter of just swapping the part with matching pinout or is there anything else needed to make the 5v IC work on this specific PCB?

It's easy to find a replacement chip at 5V because it's in an SMD package.

CY62128ELLat Mouser for $3.62
CY62128ELL at DigiKey for $3.54

SkinnyV wrote:I guess the easiest thing for me would be to try to locate a similar ram chip that use 5v then. I'll look around the new tomorrow and try to order one. But I seem to remember that finding 5v IC was much harder nowadays... Is it a matter of just swapping the part with matching pinout or is there anything else needed to make the 5v IC work on this specific PCB?

You'd need to change VCC of the chip to 5V since it seems they are running it on 3V also. You can lift pin 32 and run it to 5V coming from the cartridge edge pins. Nothing else should need to be changed I don't think. It might be advantageous to check the datasheet for the RAM though to make sure that 3V inputs (if the mapper also runs at 3V) can drive it.

Good call on the Vcc supply, didn't even think about that.

As for the inputs at 3V that should be OK as long as they really are 3V. Here's what the datasheet has to say:

Vih (Input HIGH Voltage) For VCC = 4.5 V to 5.5 V : Min=2.2 V Max=VCC + 0.5 V

qbradq wrote:Good call on the Vcc supply, didn't even think about that.

As for the inputs at 3V that should be OK as long as they really are 3V. Here's what the datasheet has to say:

Vih (Input HIGH Voltage) For VCC = 4.5 V to 5.5 V : Min=2.2 V Max=VCC + 0.5 V

Yeah that's fine then. The mapper will certainly swing between 0V and 3V (if it indeed runs on 3V) since it's CMOS.

Well thanks guys, I'm ordering right away and will keep you updated!

Ok, finally got the ram after a long delay partly caused by me ordering directly from the link posted without looking at the detail. Ended up ordering the TSOP version so I had to redo the order once I got the first one.

I swapped the ram chip for the new one. Lifted pin 32 to connect it to 5v . Correct me if I'm wrong though but wasn't pin 32 already connected to 5v before doing the mod? So in any case I still lifted pin 32 and ran a wire directly to one of the 5V edge pin just in case I missed something. The result was a somewhat more readable menu and less corrupted GFX but still corrupted. The was improvement but still doesn't make the cart working properly. I tried adding the capacitor like suggested between VCC and Ground but no improvement. I even redid the solder job a second time just to be sure (it looked very well done the first time tough) and still same result.

Here is the result:

Menu before modification: http://www.skinnyv.com/Cart/245in1/Menu_Before.JPG

Menu After RAM replacement: http://www.skinnyv.com/Cart/245in1/Menu_After.JPG

Game Before modification (Sunsoft's Batman): http://www.skinnyv.com/Cart/245in1/Game_Before.JPG

Game after RAM replacement: http://www.skinnyv.com/Cart/245in1/Game_After.JPG

Replaced chip: http://www.skinnyv.com/Cart/245in1/ramswap.JPG


As you can see there was definitely an improvement! But still not usable unfortunately.

Do you have floating address pins on the RAM?

Doesn't seem like there's any.

Run one of your fingers on the pins of the RAM chip while looking at the screen. If any of them are loose it will pick the noise from your finger and the corrupted pattern on the screen will change in response to the RAM picking the 60hz (or 50 if you're in EU or AU) from your finger.

It sure looks muuuuuch better than before. The fix could be something silly like adding an pullup on the PPU data lines.

Looks like you're only one CHR line away. I'd check for continuity with a multimeter if you've got one on all the signals/pins.

If he had data lines problems he would have vertical stripes. If he had address lines problems he would have garbled graphics. The fact it's RAM doesn't help at all as the NES writes the actual graphical data on the chip...

I must confess I'm a bit confused after seeing the horizontal lines pattern...

It's almost like if the cartridge program was trying to write to the VRAM without waiting for the PPU to be ready to take the writes.

What happens when you lower the CPU/PPU power supply to say 4.5V ?

I did test for continuity and everything seem good. As for the pattern there are not really horizontal, it really depend because the corruption can vary on the menu. As for lowering voltage I'm not really sure how to achieve that safely. But one thing is that I managed to get an almost perfect menu yesterday so I'll see if the adapter is not responsible for the rest of the corruption as the edge pin connector of those pirate carts are really short. I might try to solder the cart directly to the adapter to rule out the bad contact theory even though I tested most of the pin for continuity and didn't see a problem.

Is the corruption pattern different every time you start the cart or every frame?