The UAV Blog - Let's try this!

[RSS Bot]

I know it's been hard to find specific UAV specifics in a hurry so I'm making a blog to put the most pertinent in one place. Over the next few weeks I'll be putting up install documentation and links to the best info in the forums.

Now, a little bit about UAV itself. For a long time I've wanted a video upgrade for the A8 computers that would straighten out Atari's missteps with the factory circuits and also address an issue within GTIA itself. I also wanted an upgrade that wouldn't require me to do extensive butchering on the board. I planned to make it conform to the appropriate video specifications since "good enough" video was beginning to cause compatibility problems on modern TV's.

When I began looking at the output of various Atari systems on an oscilloscope (always remember to load the output with 75 ohms first!), I discovered that the levels were all over the place and color carrier rarely resembled a sine wave. Some of the existing attempts to fix the video were resulting in a cleaner picture, but were causing the levels to be even more out of spec. I decided that my new upgrade needed to produce a perfect color carrier, needed to find the best overall balance between brightness and color saturation, and needed its own regulator to reduce the amount of system noise present in output. This last one has proven to be quite a challenge since ultimately, everything in that little box shares a common supply.

There was one other issue that I designed UAV to address, and that's luminance skew. GTIA's 4 luminance pins are not skew matched or directly clocked and don't always change in tandem. The result is noise on the pixel boundaries with the worst case being the shade 7 to 8 transition in GTIA mode 9. To address this, UAV has a circuit to monitor the luminance pins to detect a change in shade, then a small delay before the new state is latched and displayed. This delay is controlled by the blue potentiometer on Rev C and D boards (Rev A and B were a little different in operation and never left the lab). This results in a razor sharp pixel edge and the adjustment also has the side benefit of changing the artifact colors. Adjusting for something close to the XE's Red/Blue scheme is probably the best aligned setting, though.

The next step was design a board that would work in any Atari computer or video game system having a 4050 video buffer chip and fit into as many places as possible. This resulted in a tiny SMD board that I originally assembled completely by hand. Now I order them mostly assembled after staying up night after night trying to fill Rev C orders. For systems without a 4050, it's still fairly easy to solder the board in with wires to the appropriate points on the board. It's always been a design goal to not disrupt the original video circuits if possible. This is primarily so that the upgrade is reversible, but also so that the RF output can be retained if desired.

I'm going to be posting install information created by myself and others here soon, and thanks to everyone so far who has purchased the board and waded through page after page of forum threads to find the relevant instructions. And, even more thanks to those who posted their own instructions and helped others with their questions. It has not gone unnoticed.

http://atariage.com/forums/blog/695/entry-14173-lets-try-this/