Help fix a NES Clone (Phantom System)

[Fisher]

Hi.
I'm trying to fix an old NES clone.
It's almost fully functional, but I still could not fix the power button.
I tried to draw the schematic and end with this:

Phantom System power switch

4013 Toggle.png (9.38 KiB) Viewed 10258 times

Probably has some mistakes...
I've already exchanged both transistors and the 4013. No deal.
I measured the resistors and they look fine, also the zener diode.
Maybe I should adapt another power switch...

Does someone have the full schematics for this clone?
Looks like I also have an bad 74'368 since I'm having trouble with the second controller, only A button works and it is read like A+B+Start+Left.

Thanks in advance.

[lidnariq]

Probably has some mistakes...

Eh, the schematic is pretty close to plausible as is.
* Is the 680Ω resistor in series with the plug the only way to get power into the device? It's too big for the typical power consumption of the 2A03/2C02 (which is about 400mA), so I assume it's too big for the clone ICs used here.
* T2's collector and emitter are backwards.

Where does VCC come from? (VCC1?)

The bottom half of the 4013 prevents the soft button from working for the first 22ms after power is plugged in. The top half is what actually controls the output.

I've already exchanged both transistors and the 4013. No deal.

What test equipment do you have access to? (BJT tester? 'scope? voltmeter?)

I'd just want to make sure simple things like:
* is pin 4 actually 0V like it should be after the first 22ms?
* is pin 3 high (8.2V) while the momentary button is pressed, and 0V later?
* do pins 1, 2, and 5 toggle when the momentary button are pressed?

Everything here should be low enough bandwidth you could even get away with using a soundcard as a 'scope — just make sure that you include some overvoltage protection & current limiting before plugging anything into your computer.

Looks like I also have an bad 74'368 since I'm having trouble with the second controller, only A button works and it is read like A+B+Start+Left.

That sounds more like the controller is damaged, than that the 74'368 is.

[Fisher]

Corrected the schematics.
Not sure if there are more mistakes...
Does it makes better sense?

New toggle

4013 Toggle.png (10.86 KiB) Viewed 10227 times

Could this work without the 2nd flip-flop?
It's a little difficult to trace, since the board has some damage and a little oxidation...
The schematics would make things clear if I could get it.

I only have a simple multmeter that measures resistense, AC and DC volts and some amperage.
I'm trying to find a proper socket to use on the IC.
As soon as I put it back I'll check these voltages.
Thanks for the tips

I also tought it was the controller, but I swapped them and got the same results.
Maybe it's some damaged trace or connector?
I tried to draw the joystick 2's output circuit.
Here is player 2's 9 pin connector as from under the board, with the 10k resistors omitted:

Code: Select all

IC-8, 12<------|  |-----> IC-8, 6     
               |  |
IC-8, 10 <--O  O  O  O--> CPU, 38
       
GND <-----O  O  O  O  O----> VCC
IC-8, 15 <---|  |  |-------> IC-8, 2
               \/
            CPU 39

Code: Select all

              IC-8 74HC368
                 .--\/--. 
      CPU Pin 35 |01  16| VCC
Controller Pin 4 |02  15| Controller Pin 2
      CPU Pin 28 |03  14| CPU Pin 30
               ? |04  13| ?
      CPU Pin 27 |05  12| Controller Pin 7
Controller Pin 8 |06  11| CPU Pin 24
      CPU Pin 26 |07  10| Controller Pin 6
             GND |08  09| CPU Pin 25
                 `------´

Maybe if I redo the traces to CPU pins 35, 28, 30 and 39 fix the issue.
I had a huge problem with some VRAM traces that when touched made continuity...
Maybe this can be the same case?

[lidnariq]

Corrected the schematics.

You swapped the base and collector on T2 instead of the collector and emitter.

I suspect it should look more like

Code: Select all

               Vcc          
                |           
                v           
              |/            
         -220-|             
       |/     |\            
Q ]-1k-|        ---[ 7805 IN
       |\
         v
        gnd

The 10k resistor shouldn't be necessary, even though I believe it's present.

(When pin1 Q is high, current flows through the NPN BJT from the base to the emitter, causing current to flow from Vcc via the PNP from the emitter to the base to the NPN's collector; allowing current to flow through the PNP from the emitter to the collector)

Could this work without the 2nd flip-flop?

The bottom half? Probably. As I said, it's there to make sure that the famiclone starts powered off, and that you can't power it up before the capacitors have had some time to charge.

I also tought it was the controller, but I swapped them and got the same results.

Maybe I misunderstood what you meant?

I thought what you meant was "if you press any of A, B, Start, or Left, the game thinks the A button is pressed" ?

But maybe you meant the opposite, that "if you press the A button, the game thinks all buttons are pressed". If so, that indicates that either the LATCH signal ("OUT") or the SHIFT signal ("CLK") aren't making it through.

[Fisher]

Hopefully now it should be correct:

Toggle rev2

4013 Toggle.png (10.55 KiB) Viewed 10177 times

I soldered the socket on the board.
It was very weird, I checked the tensions and all were no bigger than 1.3v, even at the zener.
I removed the IC from the socket and the zener tension went normal.
Fortunatelly, I had found another 4013 on the junk pile, and it worked!
Looks like that was it!! Thanks!

Now I'm having trouble with the push button.
The ones I got became around 150 Ohms when pressed.
Can this circuit be easily modified to accept this kind of push button?

But maybe you meant the opposite, that "if you press the A button, the game thinks all buttons are pressed".

O yeah! That was what I meant to tell. Damm language barrier!!
I also socketed the player 2's IC, switched it and no deal
It keeps the same problem...
I tested the outputs and all passed the continuty tests...
The Zapper also don't works.

[lidnariq]

Now I'm having trouble with the push button.
The ones I got became around 150 Ohms when pressed.
Can this circuit be easily modified to accept this kind of push button?

Sure. Increase the 560Ω resistor until it works reliably, decreasing the 1µF capacitor proportionately. (i.e. preserve the RC time constant somewhere vaguely near 560µs)

I tested the outputs and all passed the continuty tests...

Check for capacitive load, possibly parasitic. The CLK signal is only low for for ≈560ns, so it's easy for that to get too blurred.

The Zapper also doesn't work.

I assume you already checked that 4017D3 and 4017D4 go through cleanly?

[Fisher]

Great! Thanks for the info!
I tried some R/C combinations and the one that worked fine was 1M/47pf isn't that too much?
On the joystick I noticed that the buffer's pin 1 should be connected to pin 15, wich wasn't happening.
A bridge solved the problem.
Shouldn't M2 be present on the joystick port through the buffers?
I couldn't find any connection to it.

[lidnariq]

1MΩ+47pF seems ridiculous, but if it's reliable, then sure, whatever: the 4013 is a CMOS part so has approximately no input current.

In the Famicom, M2 and /RD4016 are combined to produce a signal that should be
* Driven high during not-M2 of a cycle when reading from $4016
* Driven low during M2 of a cycle when reading from $4016
* External weak pullup the rest of the time.

In the NES, it tentatively looks like /RD4016 just goes directly to the controller jack instead.

Tangent: this should mean that the DPCM controller glitch should cause only one bit deletion on NES decks, but multiple bit deletions on original HVC-001 Famicoms.

[Fisher]

Yes. It's working just fine.
Thanks again for all the help!!
Interesting, /OE2 goes to controller port and the buffer.
If I understood right, it's what enables the buffer.

There's still an umpopulated part on the board.
Since it's plugged on the PPU it may be someting related to the composite video transcodification.
In this board I have an UM6528 PPU, the colors seem a little differen from the UM6548.
Or maybe it's the output circuit or board layout that gives diferent colors.

It still have very noticeable jailbars... on the other clone I removed the video buffer form the board and soldered directly on the PPU, then used a coax I got from an old notebook wifi antenna. It's almost unoticiable now!
I found very strange that it has an audio and video output on the cartridge slot!!

Many thanks man!!
You rock!!

[zxbdragon]

[Fisher]

Oops...
Looks like I totally miscalculated the capacitor!
It should be 560 pf for the 1M resistor. Correct?
Should I exchange It?
I mean, won't this make the circuit work out of spec and lead to problems in the future?
I also have some doubts about transcodification, but I think I 'll ask it in other thread.

[lidnariq]

It should be 560 pf for the 1M resistor. Correct?

Yes, but...

Should I exchange It?

If it works, it works. Unless you're noticing problems with button bounce (e.g. sometimes you press the button and the clone stays on or stays off), I wouldn't worry about it.

I mean, won't this make the circuit work out of spec and lead to problems in the future?

I really doubt it would cause problems to arise at a later date but not right now.

[l_oliveira]

UM6528 PPU is for NTSC-M video standard and should be driven by a crystal of 21.47727Mhz.

UM6548 PPU is for PAL-M video standard (as used in Brazil) and should be driven by a crystal of 21.45666Mhz.

Both color standards use the same B&W basis TV broadcast system "M" (60Hz).

If you use a wrong color crystal on the PPU you will have wrong colors due to difference on signal phase generation at the color information.

[Fisher]

If it works, it works.

Well, I have been testing it on the last 2 days and everything seems fine, so I'm leaving as-is.
Thanks

@l_oliveira
Thanks for the info.
Do you know someway to "transcode" the thing internally?
I don't think it's possible, since Playtronic, when launched the official NES console in Brazil, has added an daughterboard just to do this and don't use "alternative" parts.

The Phantom System also has some unpopulated part on the motherboard. Could this be related to the transcodification?
Maybe the clone PPU had come on NTSC-M color system only at first and the project was originally made to use them.

I have some doubts about the color systems... maybe I should open another thread?
I'm not sure about these system's differences. Is it only the color subcarrier frequency and the alternating phase generation?
I also have a transcoded MegaDrive that gets black and white, with messed colors in some games. I just can't find a way to fix it

I have scanned a "naked" board (wich was really bad) of this clone.
It is here and here.
If someone wish to try a educated guess of how the unpopulated part (the circuit around CI-10) should be I'll be very happy!
Thanks in advance!

Edit: Looks like the main crystal's frequency is divided by 6 to get the color subcarrier. Is it?
If it's right, I only need a circuit to make the alternating phase generation.
How complicated would this be?

[l_oliveira]

The unpopulated circuitry is meant to make a NTSC PPU work on PAL-M TVs by replacing just the color crystal with the 21.45666 one.

That's what Playtronic NES units do. Legit Ricoh NTSC PPU and a "correction" circuit which changes the video to PAL. Unfortunately the colors will be a little off with that transcodification method...

Since your unit doesn't have the circuit you want to fit it with either a UM6548 PPU or keep your UM6528 PPU and install a NTSC clock crystal.

I honestly prefer NTSC NES than PAL-M NES but that's me. All of my TVs can accept NTSC signals.