NES RGB mod

[Live_Steam_Mad]

Hi Markus, you mentioned on another topic in this forum "I received the PPU from VS. Top Gun yesterday. The Chip is similar labled like the PPU from Famicom Titler and it hold the NES Homesystem colorpalette"  

Do you mean that this PPU from VS. Top Gun is RGB output but has the Composite PPU's color palette? If so, I am very interested  

The chip in question is the RC2C05-04

Baku has put some logic gates together to enable use of the RC2C05-04 in the AV Famicom ;-

http://translate.google.co.uk/translate?hl=en&sl=auto&tl=en&u=http%3A%2F%2Fbaku.homeunix.net%2FWiKi%2Frnx (scroll down)

It LOOKS like Baku is taking +5V from pin 40, and Ground from pin 20, of RC2C05-04 PPU, to power the 2 logic chips on his PCB, and that the only other 3 things wired up are PPU pins 10, 11 and 12 (pins A2, A1, and A0, i.e. three PPU address pins). The function that has to be performed is OUTPUT = ( A 2 NOR A1 ) XOR A0 , which means that A2 gets NOR'ed with A1, and then the output of that is XOR'ed with A0, this does the "address conversion" (the "unscrambling" back to normal of the addressing of A0).

There are it seems two more functions written on his page, A 2 NOR A1 = !(!A 2 NAND !A1), !X = X NAND X, I think the guy means that you can perform A2 NOR A1 by using NAND gates as well if you want to (depends what logic chips you have to hand). Since if you want to do A2 NOR A1, you can also do it by !(!A2 NAND !A1), since ! means "inverted", and if you want to invert say A1 to make it !A1, you just do !A1 = A1 NAND A1 (that's one NAND gate on the Quad 2 input NAND gate chip used out of 4). Then you also do !A2 = A2 NAND A2, which uses 1 more NAND gate (two so far). Then you do !A2 NAND !A1 (which uses another NAND gate, which makes three NAND gates used). Then you invert the output of it (by splitting the output and using it as 2 inputs to a final NAND gate which uses the final 4th NAND gate on the chip, and performs !(!A2 NAND !A1) which is the same as A2 NOR A1. Now all we need to do is XOR the output of this with A0.

Does anyone have a schematic for this circuit? I can't find one on Baku's website as I don't speak Japanese and Google Translate didn't enable me to find it. Should be quite easy to draw up now I know the functions needed. If I ever get my hands on the RC2C05-04 and do the circuit and it works then I will post the schematic here later on.

EDIT: OK so I drew out the schematic and here is a link https://picasaweb.google.com/LiveSteamMad/RC2C0504PPUAdapterCircuitSchematicForNESRGBMod#5727339177069453170. I think when Baku writes A0' that it means the function that you perform on A0. So you take the A0 signal from the PPU socket on the PCB of the NES (PPU socket pin 12) and also pin 11 (A1) from the PPU socket on the NES and pin 10 (A2) from the PPU socket on the NES and wire them through my circuit as shown and then feed the output into the RC2C05-04 PPU pin 12 (A0), and also feed A1 to PPU pin 11 unaltered and feed A2 to PPU pin 10 unaltered. Later EDIT: this is now confirmed by looking at Baku's red and green lines schematic in his post which follows below). Of course pin 12 (A0) of the PPU will have to be lifted from the PPU socket in the same way as the RGB pins (must not contact NES PCB).

Did anyone try to email Baku about this (Drakon?) and if so what was the reply?

I also found this site that seems to be Baku and RC2C05-04 related ;-

http://translate.google.co.uk/translate?hl=en&sl=ja&u=http://bakutendo.blog87.fc2.com/blog-entry-221.html&ei=cgBDTu_ML9S5hAfV-oynCQ&sa=X&oi=translate&ct=result&resnum=9&ved=0CH0Q7gEwCA&prev=/search%3Fq%3DRC2C05%2B%2B%2522A0%2522%2B%2522A1%2522%2B%2522A2%2522%26hl%3Den%26client%3Dfirefox-a%26hs%3DrqH%26rls%3Dorg.mozilla:en-GB:official%26prmd%3Divns

Or has anyone else managed to implement the RC2C05-04 in any NES / Famicom?

Cheers,

Alistair G.

[Salamander]

Can't help with the logic gates but if your concern is palette issues you could just install potentiometers that can be adjusted externally (the actual Playchoice 10 mobo has these on the color lines anyways).  It's not a fix for the problem with emphasis bits (very few problem titles) or the gray entries but you can get a pretty close match to normal colors this way.

[Live_Steam_Mad]

Wow that's a really great idea you've had there for the sockets on the back of your NES, adjustable R,G and B, audio "stereo separation" mix, MultiAV from the SNES, I like it very much.

Cheers,

ARG

[baku]

Hi Live_Steam_Mad, I'm the author of that Website.
Excuse me. I'm not good at English.
The following is a circuit made later.


schematic


layout


parts side


solder side

[RGB32E]

Hey Baku thanks for posting!!! 

Have you tried using the Gumshoe PPU (RC2C05-03) with this circuit?

[Live_Steam_Mad]

Hi Live_Steam_Mad, I'm the author of that Website.
Excuse me. I'm not good at English.

Hello Baku, it's great to see you here, we are VERY glad to have such an expert on this forum topic  

Your English is pretty good it seems.

Your RC2C05-04 adapter circuit would mean, I assume, that we could use the PPU from ;-

RC2C05-01 Ninja Jajamaru Kun (Japan)
RC2C05-02 Mighty Bomb Jack (Japan)
RC2C05-03 Gumshoe (MDS-GM)
RC2C05-04 -- daughter board -- Top Gun

I would like to ask you please, how did you manage to use only 1 IC of (what looks to be) the HD74LS86P (Renesas brand, Low power Schottky, 74 series, logic gate chip consisting of 4 of the 2 input XOR (Exclusive OR) gates, datasheet http://pdf1.alldatasheet.com/datasheet-pdf/view/247392/RENESAS/HD74LS86P.html ) in your pictures above, because in your pictures in your web page that I linked to above, you used 2 IC's, just as you did in your different schematic with the black background that you showed above ?  

http://baku.homeunix.net/RGB/loft/DSCF3197rt.jpg

Have you got a schematic showing just the use of the 1 x HD74LS86P ?

EDIT: OK I see from your Japanese webpage that you mentioned the 74LS36 http://en.wikipedia.org/wiki/List_of_7400_series_integrated_circuits (Quad, 2 input NOR gate, just with different pinout than 74LS02) as being a better part to use, why is this? Maybe you are using this in your single chip solution above, as I can't see whether it's 74LS36 or 74LS86 written on the chip... However I cannot find the 74LS36 chip for sale anywhere on the web or even find it's datasheet.

In your single IC solution, it seems that you have connected A2 (PPU pin 10) to pin 3 of the single IC, A1 (PPU pin 11) to pin 2, initially I thought that A0 (PPU pin 12) was connected to pin 1, on the underside of the board, but it seems like instead it just passes on a black wire from A0 through the board to the top side. IC pins 4,5,6, 9,10,11,12,13 have no connection? Vcc on pin 14 and Ground on pin 7. Maybe a connection on pin 8 of the IC but it looks like no connection, just to anchor the chip in the board? Then on the top side of the board you seem to connect A0 (PPU pin 12) (black wire coming from underside) to pin 2 of the IC (which is where you have connected A1 !), A0 also connects to pin 3 of the IC (where A2 is!), so it seems like you are using a single 2 input gate on this single IC to do both the NOR of A2 and A1, together with the A0 XOR  (A2 NOR A1), that is very clever! Absolute lowest propogation delay and smallest size and lowest cost! Can you explain how you manage to combine both functions in one gate?

Please can you tell us which single chip you used in this 1 chip solution and show us the schematic?

I see that you also mentioned on your Japanese page that the total propogation delay should be under about 60ns. If I use one NOR and one XOR gate then the total delay should be under that, looking at the datasheets. Of course this delay effectively makes Video RAM memory access a tad slower for the PPU, but as long as it's lower than a certain amount then it'll work of course. The 2KB VRAM on the NES is only 150ns speed on my NES...

I just ordered a few SN74LS02N (Texas Instrument brand, SN=Silicon (I think?), N = 14 pin Dual In Line Package (DIL / DIP), propogation delay 10nS, Quad, 2 input NOR gate, datasheet http://www.ti.com/lit/ds/symlink/sn74ls02.pdf ) and some HD74LS86P XOR chips (and some Texas Instruments SN74LS86AN as a backup, I like Ti...) from Ebay and I hope to build the circuit like you have shown in your 2-chip  schematic. I will be using IC sockets for ease of soldering!!

For the 2 chip solution it will need PPU pin 12 (A0) lifting from the NES PCB on the PPU socket. Pin 11 (A1) and pin 10  (A2) on PPU socket are just left as they are and connected to pin 11 A1 and pin 10 (A2) respectively on the NES PCB. This is confirmed from Baku's red and green lines schematic above. Then the NES PCB PPU pin 12 (A0) from the solder side (not component side) is taken on a wire to a small perf. board, with the NOR chip (02) also the XOR chip on it (86), then the output result is fed directly into pin 12 (A0) of the RC2C05-04 PPU.

Can you confirm for us that you definitely got the ordinary NES / Famicom (Composite Video style) color palette from the RGB-only output of your RC2C05-04   ? Since some web pages / people say that RC2C05-04 gives the Composite (NES) palette but with RGB signal quality?

I ask because the "Spelunker" game looks a bit bright on your monitor in your picture here http://baku.homeunix.net/RGB/loft/DSCF3198rt.jpg and almost like the RGB palette   (especially when looking at the orange walls in the cavern which are normally more brown in the composite palette?). Or maybe it is darker and more like the composite NES palette in reality when you see it with your eyes and the camera is making it look bright?

You can compare the "Spelunker" game that you are using in your picture in your RGB (with RC2C05-04) AV Famicom with the normal (composite video only) AV Famicom by starting the "NEStopia" emulator, version 1.40, because there is an option in the menu ( options /video / palette = YUV or RGB) to show what it looks like!

We are all VERY impressed with your work on the Famicom and AV Famicom modifications, you seem to be the master of this science!

Best Regards from NW England,

Alistair G.

[Live_Steam_Mad]

I didn't remember to disconnect pin 17 (PPU Video Ground) and pin 21 (Composite Sync) from my NES PCB when I installed the PPU socket (for getting rid of Jail bars on some configurations / NES revisions, and for amplifying / cleaning up Sync separately from the NES to help some displays sync properly, respectively)  

However it occured to me soon after that I can use a "normal" PlayChoice 10 style, 40 pin socket in my soldered-in "precision" / "turned pin" style socket, and bend the pins up on pin 17 and 21, to separate them from the NES PCB. At least I hope so. Otherwise I had horrifying visions of having to cut out the socket and start all over again  

This is something to think about when doing the RC2C05-04 RGB mod. Pin 12 (address) and 14,15,16 (RGB) and then also pin 17 (Drakon's jail bar fix) and pin 21 (separate CSync cleaner) would all it seems need to be lifted from ground in that mod  

Pinout for the 2C05 is here;- http://baku.homeunix.net/WiKi/rnx/index.rb?cmd=search&word=RC2C05 (scroll down). Pinout for the NES / Famicom NTSC PPU (RP2C02) is here ;- http://nesdev.parodius.com/2C02%20technical%20reference.TXT. Something I don't understand is what comes out of pin 22 called " /Sync ", I thought Composite Sync came out of pin 21 on the RGB PPU's, so what the hell comes out of pin 22 then?!    It also explains there what pin 17 is actually used for     It also mentions as to WHY we need an amplifier for either our Composite or RGB signals "VOUT: the 2C02's unbuffered composite video output. This signal usually travels to a two-stage common collector transistor amplifier, in order to boost the video drive to support 75 ohm loads at 1 volt peak-to-peak."

EDIT: Ahh. pin 22 " /SYNC: this signal when zero, will force the status of colorburst control,
scanline and pixel counters/flip-flops used inside the PPU to definite
states. Generally, this is the means of which two 2C02s connected together
in a master-slave config (via the EXT bus) can syncronize together; the
master PPU's /VBL line feeds the vblank information to the slave's /SYNC
input. On Famicom consoles, this pin is always tied to logical one. On the
NES however, this pin is tied in with the 2A03's reset input, and as a
result, the picture is always disabled while the reset switch is held in on
an NES."

Also wanted to mention that benzaldehyde in this topic http://nfggames.com/forum2/index.php?topic=187.msg7595#msg7595 said "I encountered a bit of trouble with the (RGB) amp, though. At least with a toaster-style NES, application of Waltarzar's RGB amp led to notable vertical bars acting as a sort of filter over the image, and a notable interference pattern was present. After some testing, I found that the amp was insufficeintly powered. Use of an external power source (the one I use now, at 250mA, provides plenty of current) removed both problems."

And rt9342 says  here http://nesdev.parodius.com/bbs/viewtopic.php?p=16119&sid=87ab44f9a7d2257dffc05eb970340123 "I recommend trying to find an RP2C03C, as opposed to the RC2C03B that usually comes with dual-monitor Playchoice boards. The "RC" version appears to have distortion in the last column of pixels, apparently sprite-related (at least mine does). Also I had some problems with the "RC" version in the top-loader NES, which I fixed by adding a pullup resistor." However my 5 x PlayChoice 10 boards all came with only RP2C03B, so I think he means that.  If you want the RP2C03C then that is the PPU from the Sharp C1 Famicom Television. The RC2C03B is from VS. Tennis or VS. Duck Hunt.

Now looking for a VS. Top Gun daughter-board / full kit / PPU (whatever's cheapest) to get my hands on the RC2C05-04, there's none on Ebay cheap and they sell quickly on the arcade board / cabinet "for sale" forums  

Cheers,

Alistair G.

[Live_Steam_Mad]

Looking at the palette from the RC2C05-04 PPU from the VS. Top Gun game (which is I assume supposed to be the same palette as the NES so I read) in this video Vs. Top Gun Arcade Nintendo & Konami Vs. System (Dualsystem and Unisystem) and comparing it to NEStopia 1.40 running the RGB and then YUV palette, I have to say the palette from the RC2C05-04 does look to be Composite NES style as the sky is orange / beige / peach and definitely not yellow in this video (but then there's problems with color reproduction in camcorder / LCD screen combination so I can't say absolutely for sure until I see both Top Gun on unmodded NES and then compare to VS. Top Gun.

Also, you can see the palette from the RC2C05-02 PPU from the VS. Mighty Bomb Jack (Japan) game (which is I assume supposed to be the same palette as the RC2C05-04 ?) in this video Vs. Mighty Bomb Jack and Vs. Pinball for Nintendo Dualsystem Unisystem Arcade Machine and comparing it to NEStopia 1.40 running the RGB and then YUV palette.

Has anyone (Baku) ever run the RC2C05-0x and RP2C03B (PlayChoice 10) / RC2C03B (Tennis / Duck Hunt / RC2C05-99 (Famicom Titler) side by side and compared them (with 2 copies of the SAME GAME) and confirmed that the RC2C05-0x has the NES / Famicom palette?

Or if anyone owns a VS. system with Top Gun (RC2C05-04) / Gum Shoe (-03) / Mighty Bomb Jack (-02) / Ninja Jajamaru Kun (-01)and would care to compare it directly to the NES or Famicom (i.e. having the systems running next to each other) with the same game in both machines (it's possible since there was a NES / Famicom version of each of these 4 games), that could verify what the palette is like on the RC2C05-0x.

Cheers,

Alistair G.

[Drakon]

Been ages since I posted here.  I've been building these systems as commissions.  I made some interesting discoveries.

1: The jailbars you get in the toaster depends on what revision your toaster is.  The nes-cpu-10 and nes-cpu-11 boards I rgb modded had terrible jailbars....but the nes-cpu-06 board I rgb modded looks almost as perfect as the av famicom

2: Some rgb ppu chips have less jailbars than others.....all my chips were the same model (rp2c03b) from the same type of pcb (playchoice dual moniter).  The better chips have no special markings so the only way to find out is plug it in and play.  Right now in my av famicom I have a chip that shows no jailbars on my tv....with any game...also no jailbars on the powerpak

3: The chip you need to swap in the rgb nes/famicom to make the powerpak work can actually be found on the playchoice 10 arcade pcb.  74HC373n chips are right on the playchoice I desoldered one and stuck it into a rgb toaster and the powerpak worked great.

4: The only important part of the model number of the 74hc373 is the 74HC373 part.  I bought a mm74hc373n chip from a local store for 1$ and it makes the powerpak work on the rgb system

5:  The regular top loader nes 2 that comes with rf only with jailbars in the rf (nesn-cpu-01) has just as clean of a picture as the av famicom when you rgb mod the nes 2

[Hamburglar]

Been ages since I posted here.  I've been building these systems as commissions.  I made some interesting discoveries.


5:  The regular top loader nes 2 that comes with rf only with jailbars in the rf (nesn-cpu-01) has just as clean of a picture as the av famicom when you rgb mod the nes 2

I mentioned this a while back...

The funny thing is the nicest picture I got from a RGB NES was from a NES top loader, even my AV Famicom has slight vertical bars.
One thing I never got around to trying was getting rid of that RF converter/regulator box and redoing the voltage regulator section.

Also the model 2 Jailbar problem on RGB varies, since there are at least 2 revisions.

[Hamburglar]

Hi Markus, you mentioned on another topic in this forum "I received the PPU from VS. Top Gun yesterday. The Chip is similar labled like the PPU from Famicom Titler and it hold the NES Homesystem colorpalette"  

Do you mean that this PPU from VS. Top Gun is RGB output but has the Composite PPU's color palette? If so, I am very interested  

I was under the impression that the Titler PPU had the same palette problems as the Playchoice PPU?  Did he just mean that the Top Gun PPU holds the same palette as the Playchoice 10 ?

This looks pretty interesting, hope to hear from someone that has done the mod and compared colors on the same games.

[Live_Steam_Mad]

OK so I finally managed to get the 3 Moosmann amps completed, but only connected Green and Red for now (I wanted to see the effect). All 3 amp's are wired to a common ground and to a common 5V wire, 5V for amp is coming from the NES's far right pin on Cartridge slot very close to RF box, and also ground from very close to it, just like in Markus's picture on page 1 of this topic (my PCB was slightly different since it was an earlier revision). No other grounds wired yet except the audio ground return via RF box, due to Mono audio being plugged into the Sony 14" TV front input.

5V coming from 3rd amplifier into SCART pin 16 via 75 Ohm resistor. I used all Metal Film resistors of 0.6W and 1 per cent tolerance and Markus's exact same transistors for my amps, and an electrolytic cap in the amp. No other components in the path onto my SCART pins (no cap's or resistors on the Red, Green, Sync pins). Composite Sync wired to SCART pin 20 and amp'ed Green to SCART pin 11 and amp'ed Red to SCART pin 15. RGB became automatically activated on the TV because of some volts on pin 16 of SCART (thanks Markus  ) and FINALLY the picture was hell of lot brighter and in color at last! But I saw interference patterns (wavy curvy lines)   and the black was way too grey.

I have solved this though (more to come).

Here is a picture of what I'm getting now on my 14" TV ;-

Cheers,

Alistair G.

[Live_Steam_Mad]

1 more picture of what I'm getting ;-

More later on how I solved my wavy lines problem.

On the upside, I've never seen even the slightest hint of any jail bars from this revision 4 NTSC USA NES 

Cheers,

ARG

[Live_Steam_Mad]

OK I then connected the final (Blue) color channel up to the 3rd Moosmann amp and I get this picture below. Starting to look good! And finally I have the RGB palette LOL. Wow it looks really weird and vibrant compared to the NES palette. However there are still Sine waves on the screen (interference) due to not running any "proper" ground. Actually I think it's because the only ground return I get is via the RF box because of the mono audio plugged into the front of my Sony 14" Flat Trinitron CRT TV, and the RF box's ground seems to be noisy. Also blacks are grey. I solved this later on. If I plug the mono audio out of the TV, the picture vanishes (no ground).

Ignore crud on the glass, fringe patterns (aliasing between camera and screen) and horizontal shadow line (interaction between camera and scan of TV)...

Cheers,

Alistair G.

[Live_Steam_Mad]

Here's what the wavy interference looks like from using the ground on the amplified mono audio of the RF box ;-

Cheers,

ARG

[Live_Steam_Mad]

I temporarily touched the flying lead from SCART Green Ground onto NES PCB in the very corner of the PCB (3 large corner stripes of metal where the controller connectors are, on the solder side of PCB) picture went MUCH better, wavy curved interference went and blacks were black. Also happened when I touched Composite Sync Ground flying lead from SCART onto NES PCB in same place.

So I connected Green Ground from SCART and Composite Sync Ground from SCART together and then connected them onto the Moosmann amp pcb into the common ground and the picture went terrible    (wavy curved lines interference to the max, black went verrry grey). Seems as if the interference source was the Moosmann amp PCB!  

Hre's what it looked like ;-

Cheers,

ARG

[Live_Steam_Mad]

The solution I found was to solder Composite Sync Ground from SCART (only) to NES PCB in that corner place I mentioned and finally the picture has no interference  , weird how I don't need any grounds on the 3 colors   (why is this?). Now when I plug out audio from front of TV the picture remains (ground from Composite Sync to NES PCB sorts that out).

However when I did plug the mono audio in to front of TV, the picture became a little too bright and the wavy interference comes back slighty.

So I then finally to solve this also connected a ground from my stereo audio mix mod. to NES PCB (same place in corner) and plugged stereo cable from the mod. into phono stereo to mono adapter and then into front of the 14" Sony TV and bingo the interference is always gone even when I have Stereo audio plugged in, and the picture is now perfect on RGB, and the horrific buzzing noise is gone from the audio stereo mod section. Picture interference free with audio mix knob in any position. But still the picture changes brightness when you plug in the (stereo) audio to the TV at the front, same as before. Easy to compensate for using picture settings on TV. Also cannot plug back in the mono-only sound (from RF box) to front of TV as picture gets the wavy interference coming back slightly (I suspect a noisy audio ground in the RF box). To solve this I would have to take the unamplified combined mono audio from the second of the 5 pins on the NES PCB that I used in my mix mod and instead put it through a Moosmann amp or amp of some kind and then feed that into the audio RCA socket and disable the existing wiring in the RF unit for the amplified mono audio. Then the mono audio would be loud and clear and wouldn't interfere with the RGB picture.

Now only have to test this console with my 29" Sony Trinitron TV and my In76 projector. Have a feeling the TV will show a good picture but the projector might show jail bars. I haven't seen any jail bars or white borders yet on this 14" TV. Sharpness on the TV seems to be disabled (set at medium) in RGB mode.

Here's what the picture now looks like. Yaey finally I see straight vertical edges on the bricks, no more composite zig zag edges problem ;-

Cheers,

Alistair G.

[Live_Steam_Mad]

Full screen black gives no interference ;-

Cheers

ARG

[Live_Steam_Mad]

Wavey interference lines finally gone!    Any flaws in the picture you see are just camera / CRT screen aliasing and not present in real life.

ARG

[Live_Steam_Mad]

Odd, when I checked last night very late on, my memory when I later came to write this up was that the wavy lines interference on the screen were completely gone when stereo audio (down to mono via adapter) was plugged into front of TV.

However when I come to test this today I see that there are very slight wavy lines on there when the stereo audio is plugged into front of TV.

EDIT: later on in April 2012 I tested it again when I was getting together the parts to pre-amplify the stereo sound (not done yet) and I see a perfect RGB picture on the 14" TV with no wavy interference or jail bars when I turn up the brightness or contrast, when stereo audio is plugged into front of the same TV in the same way. But was this time using another RP2C03B now (batch no. 8L4 18, with heatsink present, I just sold it to Martin Larsen in Denmark, one of two I had with this same batch number), this chip showed faint jail bars on blues and oranges in SMB1 on my IN76 projector, but none on black backgrounds when contrast was 50 per cent and brightness at 64 per cent, and with lamp on high power mode. However this chip showed the usual small rendering glitch on the top right when playing SMB1. Also it had the 2 short blue vertical lines glitch in the World 1 of SMB2.

I had a listen on the TV and the background noise is quite low on the audio on the stereo, BUT it's noticeably low in volume and needs a lot of turning up on my TV (TV audio amp has not much gain).  So I went and tried the NES with my HiFi amplifier (Technics SU-V5, old and knackered with noisy switches but it works mostly) and WHY the hell does the amp keep going into protection when I turn up the audio beyond 2.8 on the dial (1/4 of the travel) ?   It does it repeatedly on test, it's totally predictable. I turn the volume down and it comes back on again. It happens with 1/4 to 1/8" jack adapter into in ear headphones or large 8 ohm speakers or when speakers are switched out and headphones not connected!!   I thought that it might be a DC offset problem on the NES audio cables but the stereo mod already has a capacitor on each of the 2 mono lines coming direct from the CPU pins 1 and 2, so DC can't be getting to the amp?  HELP    

EDIT: OK much later I have figured out the problem, it's because the output impedance of the stereo audio is only 100 Ohms! Since the stereo audio mod takes the audio directly from CPU pins 1 and 2, just before they go into the 100 Ohm resistors (R4 and R3). Then the other ends of the resistors are tied directly to ground on the NES PCB! BTW there is also 196 ohms between CPU pins 1 and 2 (internally inside CPU) LOL. Anyway, so the RCA outers on the stereo mod are connected together, and are connected by a long wire to ground on the NES in the corner of the PCB, and then go (if you think about it) back through the 100 Ohm resistor to the CPU pins and therefore there is only 100 Ohm impedance on the output, which is very low, whereas from memory the output impedance of a typical cassette deck is over 2K Ohms, so the amp. is somehow objecting to it and cutting out! Therefore I will need an additional amplifier circuit (can't use the one in the RF box, it's ground is a source of noise, I know from trying the mono audio into the front of the TV it made the picture noisy!) to increase the output level (useful since the CD recorder that I am going into to get the stereo audio from the NES at present needs to be on maximum input level) and increase the output impedance, then I could drive my Technics audio amp. directly without it cutting off when I turn the volume up.

Fortunately the RGB picture is perfect when going into my HiFi with the stereo audio instead of going into the TV, headphones / speakers makes no difference thank goodness. I can cure the amp protection problem by plugging in the stereo from the NES to my HHB CDR800  (Pioneer PDR-05) CD Recorder and then going into the HiFi amp!  The background hiss on the sound isn't perfect but it's really pretty good and there's no buzzing.

As far as picture glitches are concerned, I am getting a flashing horizontal line of red-ish pixels about 1/3 inch wide on the far right of the screen 3/4 of the way up on Super Mario Bros 1, but it appears only in one place (when I stand on the 1st pipe and go right slightly) and only comes back every few seconds so it's hard to take a picture. I get glitches at the top right when the screen is rendered every time a new level appears.

On Super Mario Bros 2, the glitch was worse, I got 2 vertical flashing blue lines about 1/3 inch long one above the other separated by 1/3 inch. Is this what they mean by "glitches that need the 68pF capacitor being put between pin 20 and pin24 of the RP2C03B"    Do I need to lift pin 20 or pin 24 from the NES PCB at all for this?

Finally BE WARNED that after my NES was left on for a while ages ago when I was testing, it reset, and tonight after about 1 1/2 hours of tech / playtesting the damn thing halted when playing the music in SMB1. When I got concerned and checked the PPU by putting my fingers on it, IT WAS MAD HOT (maybe 60 -65 degrees C, a bit more than hand hot), especially the die in the centre of the chip!   And that was like 20 seconds or more after it got turned off as it took me a little while to turn the PCB over to get to it. I know that at 70C CPU's can start to bomb out (my AMD Athlon XP 2800+ does at about 70C on it's internal thermal sensor but my Core 2 in my laptop auto switches the laptop off at 88 C approx).

So another question please, why does my RP2C03B get very hot on this revision 4 PCB, and have I done it any permanent damage? My 2A03 CPU was a little warm but that's about all. The heatsink on the NES 7805 regulator was very hot and my PSU was warm. Why does Markus's PPU stay cool?!

If I need a heatsink (which I have convinced myself I do) for my PPU, which I could make out of Aluminium, what do I use to bond the heat sink to the PPU? And how do I bond it?

HELP!

Alistair G.

[Salamander]

@Live_Steam_Mad:  First part sounds like a ground issue, are pins 4 and 21(chasis) of your SCART socket also grounded?  Second part does sound like graphical glitching because of the missing cap.  The playchoice 10 hardware has a 68pf cap in place between pin 24 of the PPU and ground, you'll want to add that (do not lift pins).  It's normal for the chip to heat up during use but as you can see most of these PPU no longer have the heat sinks and some, like the Titler, don't even come with one.  Unless it gets so hot that you can't keep your finger on it for more than a couple seconds, don't worry about it.

[Live_Steam_Mad]

@Live_Steam_Mad:  First part sounds like a ground issue, are pins 4 and 21(chasis) of your SCART socket also grounded?  Second part does sound like graphical glitching because of the missing cap.  The playchoice 10 hardware has a 68pf cap in place between pin 24 of the PPU and ground, you'll want to add that (do not lift pins).  It's normal for the chip to heat up during use but as you can see most of these PPU no longer have the heat sinks and some, like the Titler, don't even come with one.  Unless it gets so hot that you can't keep your finger on it for more than a couple seconds, don't worry about it.

I don't understand why I would need pin 4 of SCART grounded yet (audio ground) as I am not using ANY audio on the SCART socket at all. I am (for now) feeding the audio direct into the front of the TV or direct into my HiFi amp. On the 14" TV this makes the picture on input ->1 and the sound in input ->2, this was enough to test it out with. I shall however hook up the mono and / or stereo audio to the SCART later on and report if it affects the picture.

I see that on my SCART connector that I am using (off a SNES cable that I pulled apart), the metal shell (just a sheet of metal surrounding the SCART pins) has a tag on it for soldering (a ground?) onto, and isn't connected to anything else at all (apart from maybe the ground on the PCB of the TV itself when the SCART is plugged in to the TV?). This I think is the pin 21 of SCART that you are referring to. In which case, if I solder a wire onto this tag, where do I solder the other end of the wire to? The NES PCB ground?

    In which case again I don't understand how that helps with my audio HiFi amplifier going into protection every time when I turn it up more than 1/4 volume, as I haven't got the audio going anywhere near the SCART socket at all, just direct off the 2 CPU pins, via 2 caps, into dual audio 10K log (whoops should have got the linear one) potentiometer, combines with mono audio from pin 2 of the 5 pins on the NES PCB, and comes out on 2 RCA's as unamplified mix audio, then going to my HiFi. My audio amp works great with anything else, CD recorder etc, and the latter item can take my NES same audio output with no problems, loud and clear sound when put back through the HiFi amp / speakers / headphones.

>The playchoice 10 hardware has a 68pf cap in place between pin 24 of the PPU and ground, you'll want to add that (do not lift pins).

Yeah I just checked my PlayChoice 10 PCB and there is a connection (tested continuity on multimeter) from PPU pin 24 vertically down to a solder pad, then across on the underside to another pad, then downwards to a 68pF cap, through the cap and then across on the underside to a large ground that runs vertically upwards to the ground going all round the edges of the PCB. See picture below. Looks like I'm doing the 68pF cap mod then.

Sorry but I really think the PPU needs a heatsink. Later EDIT: Maybe RP2C03B was developed before RC2C05-99 (Titler PPU) and emits more heat (maybe why Nintendo added a heatsink on the PlayChoice 10)? My RP2C03B is hitting about 70+ Celcius in my NES by my reckoning (was very hot indeed on my finger and I had to switch it off and turn the board over which took maybe 25 seconds or more before I got my finger on it) so I'm going to be using some Ebay thermal paste and a heatsink off Ebay or made by myself, because the chip locked up on me twice already after being left on for a while

Cheers,

Alistair G.

[Live_Steam_Mad]

Here's what the graphics glitch looks like on SMB2 ;-

Cheers,

ARG

[Live_Steam_Mad]

Here's my RGB mod. I don't know why my PPU runs so very hot. Maybe it's my revision of my NES? Has the 4MHz Ceramic Resonator got anything to do with it (that other NES's don't have) ?

ARG.

[Live_Steam_Mad]

OK so I tested it with my 29" Sony CRT and it's terrible... but that's just due to the TV   TV is a KV-29K5U. It's blurry even on maximum sharpness (whereas it was very sharp indeed with my 14" Sony FD Trinitron on default sharpness) since this Super Trinitron has convergence problems on red in particular and has always had a crap picture with colored screens (but tested at 550 lines horizontal resolution when I tested the pure luminance performance!), AND the colors look off (a bit pale and puke inducing, and the sky is purple-blue which really bugs me). Not good. The picture from the NES itself was perfect though LOL.

The wavy interference came back by the very slightest amount when I plugged the stereo audio in on the front of the TV. You can't see any lines but you can just about see the slight noise in the picture, it's very hard to see though. Sound was great.

However I got what looks like faint Jail Bars    All I had to do is turn the brightness down slightly and they disappreared completely. Doing so also helps increase the apparent color saturation and make it look nice and vibrant on the undergound section of SMB1. Can't see any problem with the picture at all when I do this, so I'm happy (just need a better TV!). Next up is testing on the In76 projector. I see from the markings on this RP2C03B that it is a later production chip compared with my white and other ones. Maybe they reduced the jail bars on these later RP2C03B's? Or maybe it's the blurriness of my TV masking the problem and I'll get them big time on my pj?

Here's what the jail bars look like on the 29" Sony (I overexposed the picture to show it, it's only a small effect in reality, and one thick vertical colored line that you can see is simply a reflection in the glass) ;-

Cheers,

Alistair G.

[Live_Steam_Mad]

Here's what I am getting on the 29" Sony CRT. Ignore the aliasing patterns, they don't appear in real life.

Note the purple sky which was blue on the 14" Sony!! Why is this   Help!!  

I see that the PlayChoice 10 PCB has 3 off 490 Ohm pot's, one on each of the RGB lines. Maybe I could use similar ones on my NES and adjust these to make the sky blue? Maybe there was wide variation between arcade monitors and this is what the pot's were for?

ARG

[Live_Steam_Mad]

See how you can make the jail bars vanish by simply turning the brightness down on this 29" TV ;-

ARG

[Salamander]

@Live_Steam_Mad:  About the mixed audio fed into 2 RCA connectors, have you grounded the chasis/ring on both of these?  Additional grounds inside the SCART socket can be made common to the NES PCB (usually pins 4, 18 and 21 but you could also include 5, 9, 13, 17).  If you are set on heatsinking the PPU then you need to remove the existing epoxy to get a clean working surface.  From there you'd need to use an adhesive rather than a thermal paste to get a strong bond.  The epoxy can be removed using a hair dryer and the edge of a sharp knife or razor blade, just work slowly and don't overheat the chip (unsocket it from the PCB and away from the NES when you do this).

[Live_Steam_Mad]

@Live_Steam_Mad:  About the mixed audio fed into 2 RCA connectors, have you grounded the chasis/ring on both of these?  Additional grounds inside the SCART socket can be made common to the NES PCB (usually pins 4, 18 and 21 but you could also include 5, 9, 13, 17).  If you are set on heatsinking the PPU then you need to remove the existing epoxy to get a clean working surface.  From there you'd need to use an adhesive rather than a thermal paste to get a strong bond.  The epoxy can be removed using a hair dryer and the edge of a sharp knife or razor blade, just work slowly and don't overheat the chip (unsocket it from the PCB and away from the NES when you do this).

Yes I grounded the audio on the 2 RCA connectors hanging from my NES! Otherwise I got the most horrendous loud buzzing imaginable from  the audio being both unamplified and ungrounded. They share a common ground wire and go to the corner of the NES PCB where the controllers are which seems to be the quiestest part of the PCB in terms of ground. I will try adding mono and / or stereo audio to the SCART from the NES and then try grounding the audio (pin 4) on the SCART and also pin 16's ground (RGB selection pin) i.e. pin 18, and also the shell (pin21) by wiring them all to that same place on the NES PCB in the corner. I will also try grounding the RGB pins (5, 9, 13). Pin 17 (Composite Ground) is already going to the NES PCB, that's what got my picture to stay on the screen in the first place (otherwise it disappeared when I plugged out the mono audio from the front of the TV).

I have no idea what adhesive was used on the PPU originally but there are a lot of remnants of it on there. I don't like the idea of removing it from it's socket or applying heat so I'll try taking the stuff off another PPU that I have with no heat sink on it, as practice first. I'll try just removing it with a very sharp modelling knife. When there are only traces of the stuff left I will try some White Spirit (Mineral Spirits) to remove what's left.

Why do you say to use an adhesive instead of a thermal paste (Thermal Interface Material, TIM) or self adhesive thermal pad? what's wrong with using either of these latter two?  I had the idea to use JB Weld as I have some and it's electrically conductive I think since it contains metal powder, so I think it will also be thermally conductive. But JB Weld is extremely strong and you would destroy things trying to remove it if ever the need arose in the future inless you could heat it by putting the PPU in the oven. It mentions on Wikipedia that Epoxy's strength is degraded by heating it to over 177C, but I don't like the idea of subjecting my PPU to that!

I am finding it hard to believe that others who have done this RGB mod all have their PPU's running cool after a couple of hours of use   Has no-one else's NES locked up because of overheating the PPU like mine does? My PPU is very hot to the touch and I can see why Nintendo gave it a heatsink on the PlayChoice 10 (which had to be left on for many many hours as it was in video arcades).

 I should imagine that just a short height heat sink for a standard 40 pin DIL chip would suffice. I checked and the PPU heatsink on my other RP2C03B's is 15mm high and the EMF shield on the bottom of the NES is 2mm below the top of the expansion socket and would foul the standard heatsink, and the plastic case would also interfere with the heatsink and maybe melt the case a little. So I had the idea of cutting a hole in the shield if I can't get a "low profile" heatsink for the PPU. The standard heatsink hasn't got much surface area (which suprises me now as I know how hot my PPU gets) anyway so isn't that great for cooling.

I also find it very odd how my Composite PPU that was in there before never gave me any problems with lockups even when I soak tested it for a few days of operation (don't know how hot it got but I imagine not very since it I don't remember it being a ceramic package on the composite one whereas this RGB one is always in a ceramic package, presumably since it runs way hotter). 

Why then does the Composite PPU stay cool and the RGB one run very hot?

Cheers,

Alistair G.

[Live_Steam_Mad]

OK I finally got this thing plugged back into my IN76 projector after it's RGB mod and... I'm a little disappointed actually. The Composite NES was vibrant enough for me as it was, on my pj, and the new palette didn't blow me away like it did on the 14" TV. Nice to finally have vertical edges though. I see that the picture is still a little soft, the same as it was before in NTSC versus my PAL NES which was quite noticeably sharper, the 525 line softness  annoys me. The pixels had harder edges in PAL.

Also the sky is a definite purple in SMB1 in RGB on my pj just like on my 29" TV whereas it was a definite blue on my 14" TV, that also VERY much annoys me (but it looked about exactly the same color as it was when I had this NTSC NES with the composite PPU, that also gave a purple sky). The PAL NES gives a blue sky on my 29" TV and pj. I did find that I could tweak the red control downwards on my pj and turn down the brightness, which made the sky a sort of compromise purple blue but it still annoys me. The picture is getting a little dark now since the bulb is at 5500 hours so it's long past it's rated life, running with the lamp power set to high put some much needed punch back into the picture but I can't stand the extra noise or shortening of lamp life...

I see that previewing SMB1 in the NEStopia emulator v1.4, the NTSC Composite PPU palette has the sky as purple, and it just about doesn't change at all the sky colour when you switch to RGB. LOL damnit, that means that the RGB value used in the game itself is purple and so would the real PlayChoice 10 arcade machine be on the sky when playing SMB1. You can see this when looking at Youtube videos of the PlayChoice 10 running SMB1   No way around this one it seems.

BTW I used an RGB SCART to S-Video converter from JS Technologies for the NES to pj. Oddly enough it was giving me PAL (!) color from this converter (according to the pj OSD, changing it to NTSC color on the pj resulted in no change to the picture!).  I got very slight jail bars, not enough to bother me, and very slight extra image noise when I added the stereo audio into my Pioneer DVD/HDD recorder and on into my Sony audio amp. to my headphones. This time I got a little buzzing on the sound unlike before but the buzz was at quite a low level.

EDIT: Later on I have tested the RGB SCART from my NES into my Pioneer LX60D HDD / DVD recorder which is connected via Component to my projector, and the display is near perfect, only very mild jail bars visible and only on plain blues / orange and not on black backgrounds, and the bars are quite faint.

So far my favourite is the 14" TV. Maybe I should try getting the largest 4:3 one Sony did in the FD Trinitron line so it's as good but bigger LOL. I would like to next try hooking up to a VGA monitor, my PC monitor is a 20" viewable flat screen Mitsubishi Diamondtron (4:3) with 0.24mm Aperture Grille pitch (Iiyama Vision Master Pro 514)., but I will need a CHEAP and SHARP (!) RGB SCART to VGA converter. Any good ones out there?  Or a homemade one I can build from parts?  Of course, the screen would I imagine still be purple on the sky on SMB1...

Here's what the jail bars look like on my pj (image over-exposed to show them, they do not look like this in real life). Ignore the larger vertical lines that are due to me using sheets of A3 paper as a screen (don't laugh!). Later I will show a picture with the brightness turned down where I get no jail bars.

Cheers,

ARG

[Live_Steam_Mad]

Here's what the colors look like on the projector. Jail bars there but really rather faint and don't bother me hardly at all. All I need is a bit more sharpness and a new pj lamp and I'm happy LOL.

Cheers,

ARG

[Live_Steam_Mad]

With the brightness turned down (like with the last picture above), the jail bars are not visible at all in the dark sections ;-

Notice how the pj deinterlaces / line doubles the picture so no scan lines visible.

Cheers,

ARG

[Live_Steam_Mad]

For those of you who want to attempt this mod but were "perf-board challenged" like I was when I started this mod (never done any perf. board soldering in my life) here is the info needed to translate the Moosmann amp onto a perf. board. Note that you are going to be soldering the reverse side and as such you have to reverse this top down diagram below (black and white) if you want to get the connections correct on the solder side of the perf. board. That's why I have also included a picture - diagram showing the solder side later on.

Moosmann amp ;- http://translate.google.co.uk/translate?sl=de&tl=en&js=n&prev=_t&hl=en&ie=UTF-8&layout=2&eotf=1&u=http%3A%2F%2Fplayoffline.wordpress.com%2Fmod%2Fnesrgb%2F

I used an Antex brand TCS 50W iron on medium temperature and the 1mm single eliptical face, angled (chamfered) faced bit, and Radio Shack 60/40 Tin/Lead Rosin Core Solder, 0.032" diameter, 2.5 oz (70g), code 64-005, that I had to import from USA (as I wanted what Ben Heck' had used in his book - I was still fairly new to soldering when I got it a while back).
   I used a Tamiya brand standard Craft Knife with the black handle (replaceable blades, 60 degrees point, medium duty) to cut the perf. board when I wanted to make a break in the board's vertical rails. The perf' board came from Maplin ages ago. I used an 8x magnifying loupe as I cut the board. Be careful (usual disclaimer). You score the copper 5 times on each of the two break lines and then cut with the knife carefully at an angle into the board looking with the loupe and the copper comes off. When you solder, the plastic (Paxolin) will melt a little around the copper and run clear.
  When making breaks where there is only a single hole needed and 2 breaks either side vertically, try to make the single hole have as much material as possible left around it, this aids mechanical strength and stops the copper from detaching from that small section. I did have one failure of the board due to this issue, but the copper came off without me doing anything to it, just a bizarre thing. Only happened on one hole though.
 
If you haven't soldered before, see ;- http://www.tamiyaclub.com/forum/index.php?showtopic=53807&view=findpost&p=361706, and practice soldering say 12 AWG mulistrand copper Silicone covered wires onto e.g. the tabs on the endbells of RC car motors or something, then try doing the same on smaller gauge PVC covered copper wires. To be honest I can't recommend this RGB mod to beginners as soldering is an art that is leared over a few months on quite a lot of wires of practicing.

Before you do ANY soldering, you MUST cut the board first in the correct places. See my picture - diagram later on for the solder side of the board to see where to cut (pink lines).

Next after the traces are cut in the correct places, we clean each pair of vertical rails where we want to add the first component (say, the capacitor) by rubbing with a cloth and a little Isopropyl Alcohol on it  a couple of times to remove the Copper Oxide / crud / finger grease (don't rub too hard, just fair pressure) otherwise the solder won't take. Now tin the rails with solder and have it run on them nicely (bright and shiny).

Then use a Swivel Head Pin Vice with 0.7mm drill bit and drill the 2 holes for the 1st component legs, drill from the top side of the board. Do NOT drill with hardly any pressure at all, ESPECIALLY where there are places where the vertical rail has been cut on both sides of a single hole or the copper might detach from the paxolin. Use a cutting mat, don't mark your table!

Before I put in a component I use a "resistor color code" chart off the web (first link in Google if you type that in) to check the resistance value is correct, and a multimeter to check it as well. For capacitors I test them also using a multimeter ("charge" them up by setting multimeter to resistance measurement, then "discharge" it by looking at the smallest DC voltage on the meter, you can see it working). I test the transistors by putting the meter on hFE (forward gain test I think) in the NPN or PNP, you should get 300 or so on the NPN, and 260 or so on the PNP.

Now put in the component, turn the board over to it's underneath, and bend the legs on the solder side of the board, and then hold the board down with say a pair of medium sized pliers (just to weight it down) but with something else underneath (I used a pen) to prop up the board at an angle so as to not damage the component underneath, and solder the legs into place. Don't leave the iron on a component leg for more than about 3 seconds (counting from 0 of course) if you can help it otherwise after more than about 6 seconds the component may well be damaged (according to the datasheets that I have seen where it mentiones "solderability" at 260C), and especially be aware of this for the transistors and capacitor. The transistor may be Silicon but it's encased in plastic and has connections to the silicon that can desolder themselves internally I believe if soldering for too long. Capacitors can explode when too hot (use eye protection, I am a spectacle wearer). The key is heat transfer using MOLTEN solder on your tip, solid solder won't conduct anywhere near as fast, and a yellow cruddy tip won't hardly conduct at all. Also, clean your tip before soldering a new joint as this puts new FLUX (from the fresh solder) on the tip and flux brings down the melting point requirement and helps make strong, conductive joints, and flux gets exhausted after about 10 seconds...

   I leave just a little bit of leg showing on the top side of the board so I can if needed get a multimeter probe in there and test things. Don't have the legs go across any other holes or else you will find the drilling about 5 x more difficult as you will be drilling through Tin legs not soft solder in that case. Later on you can use the component legs to bridge from one rail to another horizontally where needed (orange in the picture). When happy with the connection (should be shiny, I like to redo them at least once until I'm happy with them) then cut off the remainder of the leg, don't have the leg overhang the rail that you are soldering to AT ALL. Also you can use bits of leg left over to bridge connections where needed. I use Tamiya brand Modellers Side Cutters (straight blades) to cut off the excess leg.

When I have soldered everything in place I check first the vertical connections and breaks (between vertical rails) and then the horizontal ones, with the loupe, in case I have bridged anything by accident. Solder has "surface tension" like any other liquid that you can exploit if you accidentally bridge two connections by running your iron along the break needed and the solder will split, as it doesn't like to flow onto the plastic between the connections. After checking with the loupe I also check for continuity with the multimeter and check that I haven't got continuity where there shouldn't be any (across breaks). Then I brace myself and test with the NES and a TV.

Once you have done one amp, the other 2 are just a repeat, and this time should be faster to do as you merely copy the components positions from the 1st amp.

I have to say though that building this amp took up 10 x more effort and time than it took to do the PPU swap itself   Still, I didn't fancy doing those other IC based amps with their Surface Mount chips...

First here's the top down (i.e. component side) diagram. I drew the ground as brown to match Markus's amp but then realised I didn't have any brown wire spare so I used white wire instead. The ground wire runs into the very top right of the board on this diagram ;-

Cheers,

Alistair G.

[Live_Steam_Mad]

Here is the solder side of the perf. board. Orange are horizontal bridge connections. Pink are BREAKS in the board. Green are connections due to components on the top side (component side) of the board. This picture is showing the underside of the board (solder side). Green E, B, C are  the NPN transistor, black C,B,E are the PNP transistor. When doing the 2nd amp's ground I decided to use an extra connection just above the Red out, simply to add mechanical strength there where it is weak since there are only single hole connections going across the board at that point. The actual ground wire went in to the left of R2 as with the 1st amp. I did the same with the 3rd amp. I used 1 more vertical rail on the very left on the 1st amp than I really needed to, my mistake. You daisy chain the ground from 1st to 2nd to 3rd amp. Same daisy chain idea with the +5V (black) wire. Don't hook up anything from the SCART to this PCB, the ground noise is incredible!

I didn't show the rest of the connections for the 2nd and 3rd amp's as they are simply the same as the 1st amp.

ARG

[Live_Steam_Mad]

Here's what the 3 amp's look like when finished ;-

I will later on cut up the perf board to make the size the minimum as is practical. For now I connected the color wires onto the bottom of the board for convenience and speed. No blue wires connected in this shot below as I wanted to see the effect of just red and green connected first! The final +5V black wire that you can see on the end of the daisy chain of my 3rd amp' is going off to my SCART pin 16 (through a 75 Ohm resistor) to force the TV into RGB mode. The white Ground wire ends it's daisy chain at the 3rd amp and doesn't go any further.

Cheers,

ARG

[Live_Steam_Mad]

I tried to remove the adhesive off an RP2C03 with no heatsink and it was impossible to scrape off, it doesn't look like epoxy to me, either that or it's gone ultrahard. The knife blade couldn't remove the slightest bit. However then I realised that the raised central hump that contains the die is NOT heatsinked at all   . It's only the 2 far left and right sides of the chip that had adhesive on that are heatsinked. How stupid. So the heatsink has very little surface area and is really touching in the wrong place. Doah.

Thus, a low profile heat sink for 40 pin DIL, mounted on the flat main die area in the middle with some thermal paste or a thermal pad should work I think, and actually should be more effective than the original, and I won't have to clean off any of that ultra hard adhesive remains. Except where some of it sticks up a bit too much and I'll sand it back down to clear any new heatsink.

Suitable heatinks I think would be e.g. http://www.farnell.com/datasheets/94188.pdf  DIP4197, or http://uk.digikey.com/1/1/4211968-heatsink-40-pin-dip-glue-blk-508700b00000g.html etc.

Pictures below to illustrate the point LOL.

Cheers,

Alistair G.

[Salamander]

It's an epoxy resin.  Nintendo used something nearly identical to cover SMD chips in the Wii which is how I know it can be removed using heat and a sharp knife.  I cleaned mine off just to make it pretty.

[Drakon]

I've safely removed plenty of heatsinks by just grabbing the heatsink with vice grips, sticking a small flatblade screwdriver in that hole in the side and prying the chip and heatsink apart.

[Live_Steam_Mad]

It's an epoxy resin.  Nintendo used something nearly identical to cover SMD chips in the Wii which is how I know it can be removed using heat and a sharp knife.  I cleaned mine off just to make it pretty.

Yep you're right about it being removable with heat and a sharp knife. I used a 1.6KW hair dryer and it has a 2" x 1" diffuser on it and I set it to medium air flow and maximum heat and put a pair of old cotton pajamas on the kitchen lino floor and put the PPU onto the cloth and then blasted one end with the hair dryer for 30 seconds, no more. The cloth of the pj's grips the legs of the PPU automagically and stops it moving. Then I pulled some cloth over the rest of the PPU to stop me getting burned (WOW it gets HOT maybe over 120C or so I imagine, it's nearly burning me through the cloth LOL, don't wanna risk any more than 30 seconds on the hairdryer unless I damage the PPU) and then scraped the black stuff with a Tamiya brand Fine craft knife (30 degrees point) which had a sharp blade but blunt point after my Brother had been at it. Probably for the best as I don't want to dig into the ceramic too much. On one side it came off without too much effort, after only a couple of tries of the hairdryer, but the stuff on the other end of the chip was layed on thick and gave me quite a lot of grief and it took much scraping and goes with the hairdryer, quite annoying. But I got there in the end, and got all of it off so the chip looks way better now. I bent a few PPU pins but they are easily straightened again. Moral is don't press too hard or move forwards / backwards too forcefully when scraping LOL.

Later EDIT: I used a thick bed blanket the second time around on my next PPU  so that the legs didn't get bent hardly at all.

This RP2C03B that I scraped clean (later EDIT: which was batch no. 8L4 18, with no heatsink present, I just sold one with the same batch number to Martin Larsen of Denmark but that had a heatsink on it) is a different one to the one I previously used (which was the 9F3 27 batch one with the straight legs, not the other one from the same batch number which had horridly folded up bent legs that I straightened), I decided to use it this time in my NES so I could fit a heatsink later on and I wanted to see if this second one worked as well. It worked in my precision socket first time, so that makes 2 working ones so far from the lot I bought. It's got a number on it which is lower than the one I was using before. It performs much the same and has the same glitches in SMB2 although this time no blue lines on the World 1 itself, only in the green hills first section, I also saw a white border on the opening screen of SMB2 and on the later intro screens just before the game starts, which I don't remember seeing before. No white border at all in the game though.

SMB1 looked great no jail bars but the sky was a lot more purple blue this time than I remember it being last time on this 14" Sony CRT   Wierd.

 Also when I plug the stereo audio in to my CD recorder and then on to my Technics amp I now don't get any visible picture flaws at all whereas with the previous PPU I got slight noise when I plugged the audio in. Didn't try plugging the mono audio into the front of the TV this time. The music and graphics from SMB2 on the 14" TV in RGB are absolutely mesmerizing   Everything ultra sharp on the default sharpness (medium).

Once again after playtesting for maybe 30 min's, the PPU was rather hot, maybe a little under 60C, not burning my finger though but was very hot when I put my upper lip on it as a test (!). I'm still going to try attaching a heatsink to it. Can't hurt can it?!

Picture below shows the PPU after it's scrape clean showing the results. Also a pic' showing the jail bars I get in SMB2 if I turn the 14" TV brightness wayy up and I overexposed it to show them. When the brightness is just over half way with contrast on 80 per cent the image is dazzling bright and no jail bars visible whatsoever.

Cheers,

Alistair G.

[Salamander]

GPU RAM heatsinks (like these: http://tinyurl.com/3fbcx5l) would probably do the trick and usually have a thermal adhesive pad backing.  If you want something more substantial try Arctic Silver thermal adhesive.  I'd personally like to see how the C1 PPU was prepared since the heatsink there is different from what is in the Playchoice and covers the entire chip.