I finally got around to testing the Sega Star Trek board set.
This machine is based upon a G-80 arcade machine which is a
vector based graphics engine. Instead of the typical Raster process. This means the graphics of the machine are crisp and clean.
I had to take a different approach to debugging this machine as I couldn't just hook it up to any old monitor. It requires a vector monitor to operate; which I currently don't have.
It turns out that a good electronics work bench would have an Oscilloscope. Mine happens to have two. Many ?all? Oscopes have an X-Y mode. So I decided to hook the video output of the machine to the oscope. The Yellow "Y" goes into the "Y" for the Oscope (aka Ch2). The "X" (green) goes into Ch1 (X).
Unfortunatly; when I first powered up the board - I got nothing but a dot in the center of the display.
I did some reading and came across the
G-08 FAQ which didn't really provide much troubleshooting hints save for the comment that the soundboards can be removed to minimize the boards being debugged.
More reading. Read the Manual. looked at schematics. First thing I noticed is that the Red (D3) Test LED was not easily viewed when in the cage. Second; the test switch was broken. Fixed that with a spare switch and some rework wire.
The Test button now worked. but gave inconsitent results. Some times the LED would come on and stay on. Sometimes it flicker rapidly. some times steady. Umm. More looking / investigating. I figured at this point; I needed to actually bench test the board. Check for obvious problems... IE Reset stuck low. Check the Interrupt lines. Etc of the Z80A cpu. It was then I noticed someone had cleanly clipped away Y1, R10, and R11. Not desoldered. but cut off near the board. <sigh> That'll do it ... no clock gen means mr cpu an't getting what it neads to fetch instructions from memory.
Searched high and low for a 8MHz crystal. I just didn't have one. Radio Shack - didn't even bother. Frys - hoped but they don't carry crystals at all.
On a parts board - I did find a 4MHz crystal. I remembered from my reading that the board acutally runs at 4MHz. I also remembered that the schematics indicated the Clock was divided down by 2 thru a D flipflop (7474)... and then again to 1/2 clock frequency. I looked at the schematic again and formulated a plan.
1) Re-Populate the crystal with a 4MHz crystal.
2) Verify that pin9 of U10 (7474) was operating at 4MHz.
3) Verify pin 11 of U10 was 1/2 input (4MHz / 2 ) = 2MHz.
4) put U10 into a socket.
5) Lift pin 9 on the U10 package.
6) bridge pin 9 to pin 11.
Steps 2&3 confirmed my understanding of the clock circuit. The designers generate a 8MHz clock from the crystal using the inverter and nand gates. Then use the "D" flipflop to half that clock to 4MHz which goes to the clock input of the Z80A.
Steps 4-6 uses the 4MHz crystal to generate a 4MHz clock which is fed directly into the Z80A.
I bench powered the board and verified that the Z80A was getting a 4MHz clock.
I put the board back into the cage and power it up. The board came up and I was treated to imagery on the Oscope. WOOT!
Since the Oscope doesn't have a "Z" input the electron beam of the oscope can't be turned off went necessary; so the image shows up as just a continous beam being drawn on the screen.
This made me really happy - a little troubleshooting and the boards were all right.
Next; I need to put the sound boards back in and verify functionality.
I also need to secure a vector monitor.
The G08 Electrohome monitor used in these games is a real PIA - It's known to catch fire. As such; many are toast... proably in a landfill somewhere. I have found a person selling me one untested - which means I'm going to have to figure out how to test that puppy. I'm proably going to have to recap the monitor and do
some design changes to keep mine from catching fire.
The PSU in the sega is turing black from heat; I'm likely going to need to rebuild it too.
More later - stay tuned.