Re: [stella] Re: [Fpga2600] Where am I going wrong?

Subject: Re: [stella] Re: [Fpga2600] Where am I going wrong?
From: Chris Wilkson <ecwilkso@xxxxxxx>
Date: Fri, 3 Jun 2005 00:20:08 -0400
You're using a DSO (digital sampling oscilloscope) and it looks like a
software frob so I don't know if it's real, but you have glitches during
your sync pulses.  That's a murder/death/kill right there.  A big one.

Regarding the voltage levels, there's a *lot* of give in acceptable NTSC
signals, i.e. what will work with most TVs.  This is why people were able
to get away with using all those numbers over the years.  But the newer
the TV, the less capable it will be when trying to deal with an out of
spec signal.  (odd that, and sad...but true)

Here are the official NTSC monochrome levels (this goes into a 75 Ohm load, BTW):
sync  0V
blank = 40IRE = 40V/140 (~286mV)
black = 47.5IRE = 47.5V/140 (~339mV)
brightest white = 140IRE = 1V

The peak-peak amplitude of the colorburst is 40IRE (= 40V/100).  If you are trying
to duplicate the 2600 color palette, you'll want the same saturation level for all
15 colors.  The earliest machines had a saturation ratio (active:colorburst) of 1:1.
Later machines increased this ratio for more vivid colors.  The ratio varied from
rev to rev, but the number 2.2 comes to mind as one of the variations.  But you
shouldn't trust that number.

If you are trying to keep things within spec, you'll want to reduce the white level
so that the chroma peaks are less than a volt.  (this will hurt the contrast, but
many video systems will severely distort the signal if it's more than 1Vpp...
oh well)

Keeping with the default 1:1 saturation ratio (active color amplitude = colorburst
amplitude = 40IRE = 40V/140) you need the following:

highest point of brightest color = 140IRE = 1V
lowest point of brightest color = 100IRE = 100V/140 (~714mV)
=> white level (halfway) = 120IRE = 120V/140 (~857mV)

Hope this helps.  Let me know if you have more questions.  And get rid of those
sync glitches!!  :)


On Thu, 2 Jun 2005, Adam Wozniak wrote:

> Hash: SHA1
> On Wed, 1 Jun 2005, Adam Wozniak wrote:
> > Here's a scanline...
> >
> And slightly better:
> I'm running "Space Invaders" with the b/w switch on.  When I turn on
> colors I can't sync at all.  Maybe I'm modulating too much.
> I had some bad playfield logic which prevented pixels from being drawn
> on the rightmost edge of the screen.  When I fix that, I lose video
> entirely.  This leads me to believe something in my horizontal sync
> still isn't right.
> I've found different references for various voltage levels.  One source
> says black at .3, another .4, and another says "blank" at .40 and black
> at .48, while some don't mention "blank" at all!
> Here's what I'm using:
> sync  0.000V
> black 0.297V
> lowest point of colorburst 0.155V
> highest point of colorburst 0.44V
> brightest white 0.918V
> highest point of brightest color 1.061V
> lowest point of brightest color 0.776V
> Any help, advice, or references greatly appreciated.
> - --
> adam@xxxxxxxxxxxxxxxx
> KG6GZR             
> Version: GnuPG v1.2.1 (GNU/Linux)
> iD8DBQFCnxsnyvXf5Z0z5zERAuXlAKDHKOGrRRh2LoySQgf1kRynZZUO1ACglqQ0
> z+ffu61yq+dHXi6x+vVsyOA=
> =R707
> Archives (includes files) at
> Unsub & more at

Archives (includes files) at
Unsub & more at

Current Thread