RGB NES / Famicom and jailbar removal

Thanks to working with skips here's probably the ultimate jailbar removal setup:



Rule #1: Have the shortest distance possible between the rgb ppu and the rgb amp (whichever amp you use)

Rule #2: Shield the ppu

Rule #3: Certain revisions of the nes pcbs are terrible, as far as I know it's revisions 10 and 11.

Skips wrote:

Rule #4: Lift pins 17 and 21 in addition to 14, 15, and 16.

Rule #5: Place a 220uf capacitor between power and ground (Pin 40 and Pin 20)

Rule #6: Ground the amp on pin 20 and power it on pin 40

Rule #7: If using an RGB to whatever encoder and not straight RGB ground your video encoder on pin 20.

Rule #8: Shield and ground PPU with copper foil tape. Use the high quality thick stuff not the thin stuff from China. Use 3-4 layers and cover the trace on top with electrical tape. Ground the shielding on the outer edge of the NES PCB.

Rule #9: Socket and shield the CPU in the same manner you shielded the PPU. This may not work for every NES revision but worked well for my revision 7.

Rule #10: Place a 1uf capacitor between power and ground on the CPU (use the outer edge and not pin 20, it doesn't like it when its on pin 20 for some reason). WARNING: This may cause slowdown on some of the NES revisions which will also result in graphics glitches as if you did not have the 68pf capacitor on pin 24 of the PPU. If it does remove it.

Also please please please do not use standard IC sockets in the toaster NES. Make sure you get a very high quality precision socket that has a nice strong death grip on the PPU/Circuit you build. Drakon has told me that he has had PPU's fall out of the toasters due to the PCB being upside down. I use the sockets from Fry's electronics. They are super stiff and unless you are tossing the NES 50 feet in the air the PPU will not fall out. Personally I solder my circuit directly to the NES PCB so it does not fall out. This way you don't run the risk of the circuit falling out. A nice stiff precision socket can easily hold the weight of just the PPU with this method. Also a good rule of thumb is if you can easily remove an IC from the socket you plan to use its not a good socket to use with the toaster NES, find something of higher quality.

Drakon wrote:

I also personally hard solder the ppu into a toaster just because it's suspended upside down when I do a commission. I can't risk having someones ppu fall out and short.

Interesting I did not know the CXA1645 amps RGB. So the ths7314 is not even needed when using this encoder then?

Skips wrote:

Interesting I did not know the CXA1645 amps RGB. So the ths7314 is not even needed when using this encoder then?

That's correct with the sony encoder you don't need a rgb amp since the sony encoders come with built in rgb amps. I highly recommend the cxa2075. The cxa2075 is a drop in replacement for the cxa1645, the 2075 will work fine on a 1645 circuit you'll only lose composite video (rgb and s-video will still look fine). The cxa2075 produces noticably sharper s-video than the 1645 and in my opinion better looking colours. Basically the cxa2075 is the newer model of the cxa1645. I've never found a video encoder that generates better looking s-video than the cxa2075.

Yeah I am definitely gonna use a cxa2075 eventually. I just happen to have a spare CXA1645 available at the moment and there are some tests I am going to run tonight. I will let you know how it goes.

*update* I wonder if only some versions of the CXA1645 amp the RGB lines? I tried mine and it still required an amp before going to the encoder.

Skips wrote:

Yeah I am definitely gonna use a cxa2075 eventually. I just happen to have a spare CXA1645 available at the moment and there are some tests I am going to run tonight. I will let you know how it goes.

*update* I wonder if only some versions of the CXA1645 amp the RGB lines? I tried mine and it still required an amp before going to the encoder.

Haha all cxa1645 chips are the same. Yours won't require an amp (maybe the circuit is wrong somehow?). The av famicom upgrade kits I use have the rgb pass through a cxa1645 and the rgb outputs go through 75 ohm resistors followed by 220 uf caps in the rgb cord and then straight into the scart tv.

I've double checked all the circuits i tested it with, all of them needed RGB amped before going to the encoder. Two of them I made and two others were bought from some third party companies for CMVS projects. All of them appear to be wired just how the data sheet says they should be. They are all the CXA1645P but that shouldn't make a difference.

Skips wrote:

I've double checked all the circuits i tested it with, all of them needed RGB amped before going to the encoder. Two of them I made and two others were bought from some third party companies for CMVS projects. All of them appear to be wired just how the data sheet says they should be. They are all the CXA1645P but that shouldn't make a difference.

I wired up a cxa1645p application circuit and installed it in a snes mini using unamped rgb, the s-video output was amped perfectly and I assume the rgb was too. Like I said the systems I built for clients don't have rgb amps just sony cxa1645 / 2075 chips and they all said the rgb looked perfect on their setups.

Hmm I wonder whats different then because even the purchased encoders are wired like the application circuit in the datasheet. Without the amp its super dark. S-video and RGB out are both going through the 75ohm resistors and 220uf capacitors and the RGB in is going through the .1uf capacitors. I even tried connecting another encoder to the RGB out lines to see if it was amplified coming out of the encoder and it was not. The RGB lines going to the CXA1645 are coming straight off the RGB pins on the PPU.

Regardless though what I wanted to test was if jail bars were still there without the ths7314. I was able to at least see the picture dimly and the jail bars were still there, about the same when using the ths7314. So that leaves distance as the big jail bar culprit (as you had stated). I think I am going to wait until I can try one of the circuits you were talking about since that would eliminate the amp and place the encoder right next to the PPU. If the jail bars are still there even un-amplified on S-Video its probably distance. I have to de-solder the current setup I have on the NES PCB so I will just wait until I can get all the needed parts (so I can at least play it until then).

Skips wrote:

Hmm I wonder whats different then because even the purchased encoders are wired like the application circuit in the datasheet. Without the amp its super dark. S-video and RGB out are both going through the 75ohm resistors and 220uf capacitors and the RGB in is going through the .1uf capacitors. I even tried connecting another encoder to the RGB out lines to see if it was amplified coming out of the encoder and it was not. The RGB lines going to the CXA1645 are coming straight off the RGB pins on the PPU.

Regardless though what I wanted to test was if jail bars were still there without the ths7314. I was able to at least see the picture dimly and the jail bars were still there, about the same when using the ths7314. So that leaves distance as the big jail bar culprit (as you had stated). I think I am going to wait until I can try one of the circuits you were talking about since that would eliminate the amp and place the encoder right next to the PPU. If the jail bars are still there even un-amplified on S-Video its probably distance. I have to de-solder the current setup I have on the NES PCB so I will just wait until I can get all the needed parts (so I can at least play it until then).

Sounds good to me.

this weekend I just completely re-wired my rc2c03 rgb nes, using that small amp board i built and showed you, drakon.

I am hoping for good results. it has less than 1/4 inch between the amp and the ppu socket, i might as well have just soldered it directly to the socket pins but the board was hitting other chips! It also got rid of that huge mess of wires since i got rid of the seperate lm1881 board, the 1881 is now included on the same board with the amp.

Im not quite ready to turn it on yet (still have a little bit of soldering to do) but i will try to get screenshots to you as soon as its operational.

Working with skips we came up with this:



Read the first post for a description.

Rule #4: Lift pins 17 and 21 in addition to 14, 15, and 16.

Rule #5: Place a 220uf capacitor between power and ground (Pin 40 and Pin 20)

Rule #6: Ground the amp on pin 20 and power it on pin 40

Rule #7: If using an RGB to whatever encoder and not straight RGB ground your video encoder on pin 20.

Rule #8: Shield and ground PPU with copper foil tape. Use the high quality thick stuff not the thin stuff from China. Use 3-4 layers and cover the trace on top with electrical tape. Ground the shielding on the outer edge of the NES PCB.

Rule #9: Socket and shield the CPU in the same manner you shielded the PPU. This may not work for every NES revision but worked well for my revision 7.

Rule #10: Place a 1uf capacitor between power and ground on the CPU (use the outer edge and not pin 20, it doesn't like it when its on pin 20 for some reason). WARNING: This may cause slowdown on some of the NES revisions which will also result in graphics glitches as if you did not have the 68pf capacitor on pin 24 of the PPU. If it does remove it.

Also please please please do not use standard IC sockets in the toaster NES. Make sure you get a very high quality precision socket that has a nice strong death grip on the PPU/Circuit you build. Drakon has told me that he has had PPU's fall out of the toasters due to the PCB being upside down. I use the sockets from Fry's electronics. They are super stiff and unless you are tossing the NES 50 feet in the air the PPU will not fall out. Personally I solder my circuit directly to the NES PCB so it does not fall out. This way you don't run the risk of the circuit falling out. A nice stiff precision socket can easily hold the weight of just the PPU with this method. Also a good rule of thumb is if you can easily remove an IC from the socket you plan to use its not a good socket to use with the toaster NES, find something of higher quality.

I am using the ths amp, and an lm1881.

I built this on a nes-cpu-11 board again because i wanted to see if i got heavy bars with it.

I have faint, almost invisible bars. sofar, rebuilding my nes project has been a success. Ive been talking about sheilding for a long time. i bet that part alone will get rid of whatever faint bars i have.

Btw, ive found that keeping pin 17 as video ground only, gives you a real nice picture.

I have it wired right into my scart socket on the nes for the 3 video grounds. All other grounds are connected to the chassis of the nes and grounded to the other scart grounds.

audio ground i have going straight form the rca on the nes to the audio ground pin on scart.

btw, the only issue im having right now is that my blue signal seems to be missing as if theres a loose solder joint somewhere on it.

the nes picture everything is green or red or orange.

But at least ive got a picture! drakon knows how much trouble i had with the old setup until i just decided to do it over with many things that i had learned since i first wired it up.

Btw, I plan on switching this to a perfectly working nes-cpu-07 that i have.

I mainly wired this up as a test to see how much i could improve the 11 just by cutting wire length to almost 0.

well that and this one already had the socket in it.



[/quote]

Btw, ive found that keeping pin 17 as video ground only, gives you a real nice picture.


Let me correct this. Drakon found this out about pin 17 it wasnt me.

when i wired the first one i didnt know about this.

Try running sync through the nes composite video circuit that amps it a little. See if that helps at all.

Drakon wrote:

Try running sync through the nes composite video circuit that amps it a little. See if that helps at all.

to get rid of the yellow, orange and green?

i read one of your old posts somewhere else that said a 220uf capacitor on sync might stop the picture from rolling.

Ever since i got home today ive been trying to find one so i can test that out.

btw, your never going to leave me alone for this, but the reason i was getting the fast rolling b/w crap before was because i had the sync input and output on the 1881 swapped.

Nice work and writeup guys.

I guess I never realized how troublesome these jailbars are. (I'd buy an RGB PPU, but I'm holding out for the Universal PPU)

i ran the sync through the composite circuit and it worked perfectly for about an hour now the picture is rolling again.

wtf?

mvsfan wrote:

i ran the sync through the composite circuit and it worked perfectly for about an hour now the picture is rolling again.

wtf?

Got it. something is loose. if i push down on the pcb the sync comes back and works fine.

Grambo wrote:

Nice work and writeup guys.

I guess I never realized how troublesome these jailbars are. (I'd buy an RGB PPU, but I'm holding out for the Universal PPU)

No need for the universal ppu, you can get digital signals right from the stock composite ppu.

mvsfan wrote:

Got it. something is loose. if i push down on the pcb the sync comes back and works fine.

I always hot glue down my hand soldered wires, they tend to break from even small amounts of wiggling.

Drakon wrote:

No need for the universal ppu, you can get digital signals right from the stock composite ppu.

I don't follow.

I want RGB from my front loader and instead of getting an RGB PPU, I'm waiting for the Universal PPU to be developed and go into production.
Is there something I'm missing here?

Grambo wrote:

Drakon wrote:

No need for the universal ppu, you can get digital signals right from the stock composite ppu.

I don't follow.

I want RGB from my front loader and instead of getting an RGB PPU, I'm waiting for the Universal PPU to be developed and go into production.
Is there something I'm missing here?

There are currently a few different groups/people that have managed to pull digital signals from the stock composite PPU through the use of FPGA. I am sure this type of mod will eventually be sold as a kit by someone in the future and does not require you to remove the composite PPU.

Skips wrote:

There are currently a few different groups/people that have managed to pull digital signals from the stock composite PPU through the use of FPGA. I am sure this type of mod will eventually be sold as a kit by someone in the future and does not require you to remove the composite PPU.

Right, I guess just from what I've seen out there, it seems to me that out of all the upcoming RGB solutions, the Universal PPU seems as though it has the most drive towards being a distributed, consumer product.