Repairing Atari's 1976 classic Stunt Cycle video game|
By, William Stephens
(clicking any picture in this article will bring up an enlarged version)
When it first came in, Stunt Cycle was D.O.A.
The first check was the power supply/fuses which were all in good shape, power was flowing and even the CRT was glowing nicely. A few quick twiddles of the knobs on the Black and White Motorola XM501/701 monitor and the image started to appear, but it was wavy fuzzy and blurry... time to recap the monitor!
The re-cap of the monitor was performed and the monitor was re-aligned and adjusted for nominal performance. That's one thing down and an undefined number of things to go.
Now when operating Stunt Cycle, only the busses, ramps background and score counters appeared. The game had no action beyond that... no little motorcycle running around, and no sound. The screen image on the other hand was crisp and strong thanks to the re-cap.
This tells me that the game is almost working. Seeing that this game uses no microprocessor, a complete crash is impossible because there is no central commander of things. Each thing like ramps, buses or the score counters are all self generating and relatively independent of one another. Everything seems to be working except the motorcycle, I guess I should focus on circuitry for the motorcycle, duh!
A visual check showed nothing obvious so I started off checking the horizontal generator controlled by the handgrip control which is in reality a 5K potentiometer. A scope probe was put on pin 3 of the 566 at L11. The scope showed a typical square wave. When the throttle was advanced or retarded, the frequency of the square wave changed appropriately. The board has the driving signal for action, just none is currently apparent.
Next I checked the ROM and asociated circuity on page 1 of the schematic using a logic probe. This is where the data to draw the motorcycle is located. Everything seemed to be operating well in this section... it looked like addresses were sequencing and data was flowing, at least somewhere besides the CRT.
At this point I was out of hunches and started checking states on any and every named signal. In probing around, one stood out. It was the "cycle reset" and "not cycle reset" lines, they showed an open circuit... could I be this lucky?
Those signals are driven by a dual retriggerable one shot multivibrator, a DM9602N at board location N6. The duration for the "shot" is determined by an electrolytic capacitor. I checked the values of these capacitors and found them to be out of tolerance by nearly 100%, I checked the input lines and found proper activity but found open circuits for the outputs. BINGO!!!.
I checked stock and found I had a DM9602N and by luck the axial 100uf and 1uf capacitors. I removed the multivibrator and the caps and installed new ones.
Once the game was plugged back into the cabinet and plugged in, correct operation resluted. The little motorcycle was jumping busses and making the full compliment of annoying noises again.
During these tests, I noticed that waveforms were a little out of shape and the sounds were not exactly right. Old tired electrolytic capacitors are responsible for this. Now that it works, some attention will be paid to re-capping the circuit board, afterall those caps are 30 years old!