pauldon
New member
Root causing clock signal problem on Williams Stargate
I'm seeing RAM errors on boot – almost always a 1-2-5, sometimes a 1-3-1 and occasionally an entirely different chip is implicated. RAM replacement and swapping changes nothing. The board is running 4164's with the harness modified to allow for this.
I've gone through the power and ground just to make sure that's not the issue. Voltage at the main board is 4.96VDC. I've run a ground jumper directly from the power supply to the main board during troubleshooting just to help ensure grounding is not an issue.
Looking first at the reset on the CPU, at pin 37 I see about 1 second of Low, then it pulses high in sync with the RAM Error LED flash (high while each of 1, 2, 5 are displayed) then goes low when the RAM LED goes dark.
Moving to the clock generator I think I have found the issue but so far have struck out finding the root cause. Simply put the clock is not a square wave. The output on pin 6 of the 74LS04 (7J on the board) that should have a 5V 12Mhz square wave on it has a 2V near sine wave (no clipping to square at all). I thought that chip was probably bad so replaced it – no change. Tracing back I see the same signal at pins 2 and 4 which likely rules out the 74LS107 downstream being bad and causing me to see strangeness on the output of that one inverter. Going downstream from the oscillator, the 7474 (7G) that should create the 4Mhz signal produces a 3V Malformed square wave at pins 8 and 9. I replaced it anyway. I don't have a 74LS107 (7K) handy yet but just for good measure I'll replace that soon but don't think it's going to help.
Vcc at all of these chips is 4.9V or more.
Today despite things measuring fine, I replaced the capacitors and resistors in the oscillator circuit to no avail – no change at all.
Ideas? I'm extremely puzzled by the 2V wave on that 74LS04 (7J) while Vcc is fine. Anyone happen to be sitting next to their game and scope? You are seeing a nice square wave on pin 6 of that chip, yes?
A couple pictures of the scope traces I mention are attached.
Thanks
Paul
I'm seeing RAM errors on boot – almost always a 1-2-5, sometimes a 1-3-1 and occasionally an entirely different chip is implicated. RAM replacement and swapping changes nothing. The board is running 4164's with the harness modified to allow for this.
I've gone through the power and ground just to make sure that's not the issue. Voltage at the main board is 4.96VDC. I've run a ground jumper directly from the power supply to the main board during troubleshooting just to help ensure grounding is not an issue.
Looking first at the reset on the CPU, at pin 37 I see about 1 second of Low, then it pulses high in sync with the RAM Error LED flash (high while each of 1, 2, 5 are displayed) then goes low when the RAM LED goes dark.
Moving to the clock generator I think I have found the issue but so far have struck out finding the root cause. Simply put the clock is not a square wave. The output on pin 6 of the 74LS04 (7J on the board) that should have a 5V 12Mhz square wave on it has a 2V near sine wave (no clipping to square at all). I thought that chip was probably bad so replaced it – no change. Tracing back I see the same signal at pins 2 and 4 which likely rules out the 74LS107 downstream being bad and causing me to see strangeness on the output of that one inverter. Going downstream from the oscillator, the 7474 (7G) that should create the 4Mhz signal produces a 3V Malformed square wave at pins 8 and 9. I replaced it anyway. I don't have a 74LS107 (7K) handy yet but just for good measure I'll replace that soon but don't think it's going to help.
Vcc at all of these chips is 4.9V or more.
Today despite things measuring fine, I replaced the capacitors and resistors in the oscillator circuit to no avail – no change at all.
Ideas? I'm extremely puzzled by the 2V wave on that 74LS04 (7J) while Vcc is fine. Anyone happen to be sitting next to their game and scope? You are seeing a nice square wave on pin 6 of that chip, yes?
A couple pictures of the scope traces I mention are attached.
Thanks
Paul