Amplitude of the SNES Mini / SNES2 / SNES Jr.'s RGB

Started by micro, May 12, 2013, 04:54:43 AM

Previous topic - Next topic

micro

Every now and then the question arises what kind of resistors to use when making the RGB mod on the SNES Mini.

With an oscilloscope and a homebrew reproduction of the SNES Test Program Cartridge I've tried to measure the amplitude of the RGB lines. Well, actually I've only measured the R-line ;)



The resulting waveform on the red line during the red test screen:


The amplitude of the red signal is 1.7 Vpp. You can also see there's a fixed DC offset of +2 V which get's filtered out by the capacitors inside the SCART cable.


Provided that there's a 75 Ohm resistor to ground inside the TV, installing a 75 Ohm resistor will give us an amplitude at the TV's input of 0.85 Vpp. That's more than the standard 0.7 Vpp.
So it would practical to use a higher resistance value to reduce the amplitude. The calculated value to get an 0.70 Vpp amplitude would be 107.14 Ohm. A 100 or 105 Ohm resistor would be suited.


While installing those 100 Ohm resistors inside the SNES would give the right amplitude, one questions remains: Would it have negative aspects regarding impedance matching?





RGB32E

#1
I've experimented with this myself and even tried bypassing the built in enconder with a THS7374.  I tapped RGBS directly from the PPU.  When preparing I noticed that the only components between the PPU RGBS output and the encoder were a 75 ohm resistor to ground, followed by a 0.1uF capacitor.  From your posted information that would lead me to believe that the RGB signals are 0.85 Vpp at the encoders input.  Therefore the signal level needs to be reduced to 82% of it's existing level.  Adding a series resistor between the PPU and 75 ohm termination resistor could "fix" this so that you get the proper levels when using a 75 ohm series resistor on the encoders output.  Using the simple voltage divider equation/circut it looks like a 15 ohm resistor could work to get 1.4Vpp output from the RGB encoder.


         75
0.7 = ------- * 0.85
      15 + 75


Other threads:
http://shmups.system11.org/viewtopic.php?p=865727#p865727
http://shmups.system11.org/viewtopic.php?f=6&t=45397

ApolloBoy

Isn't 0.85 Vpp usually within tolerance though? I've always used 75 ohm resistors on my SNES mini mods and never had any issues.

RGB32E

75 ohm resistors alone do give a stronger signal than normal.  The 0.85Vpp output gives a brighter picture.  Not a problem maker per say, its just not ideal.  I don't have a scope to measure the signals but the 0.85Vpp seems to explain the brighter than normal picture.

micro

Yes, I've also got 75 Ohm resistors between the encoder output and the multi av port at the moment. I can tell that the picture is little bit brighter than it should be.

@RGB32E: Interesting! But altering the RGB signal levels between PPU and encoder would also effect the brightness of the composite video signal, I could imagine. Or is the SNES Mini's composite video signal also known for being too bright?

RGB32E

Quote from: micro on May 14, 2013, 05:52:22 AM
Yes, I've also got 75 Ohm resistors between the encoder output and the multi av port at the moment. I can tell that the picture is little bit brighter than it should be.

@RGB32E: Interesting! But altering the RGB signal levels between PPU and encoder would also effect the brightness of the composite video signal, I could imagine. Or is the SNES Mini's composite video signal also known for being too bright?

I don't think anyone would really notice if composite video is brighter than it should be.  Composite video runs on the hot side from many different consoles, so I'd think this would go unnoticed.  Higher Vpp RGB input would result in brighter composite video output.

niall

Bumping this as I found it really useful.

I'm wanting to calibrate all of my displays, so feeding them all a standardised 0.7vpp RGB signal is my long term goal.  Rather than individually calibrating each of RGB and getting this impedance mismatch, is there a type of potentiometer that can do 3 analog signals at once - assuming they are each consistent with each other to begin with?