Board history: fully recapped, some minor corrosion found and confirmed cleared (traces/joints check out with continuity). PSU voltages all nominal. All chips power up and get warm.
Symptom: power on, HDD spins up, floppy light flashes briefly (normal self-check), no chime, no chime-of-death, no video output (tried both a VGA adapter, previously proven working on my Centris 660AV, and a genuine Mac monitor, no sync/picture on either).
What I've checked and ruled out so far:
- RAM: symptom persists identically with all RAM removed (still no chime/chime-of-death), so RAM isn't the root blocker. Currently testing with a known-good matched pair in Bank A.
- Cuda (341S0788): VDD stable and clean through power cycle, no ripple. Clock oscillating correctly at 32.768kHz. Manually grounding Cuda's RST pin causes a clean shutdown, confirming Cuda is alive and responding correctly. Cuda's own RST pin sits high at idle and never pulses automatically, traced this back through a 330K pull-up to the S2 (manual reset switch) footprint, which matches the schematic (sheet 1/12), this appears to be normal behavior, not a fault, since Cuda generates its own internal power-on reset from its own clean VDD.
- Main CPU clock: found and confirmed via schematic (sheet w/ U44 MC88916-80 clock driver, fed by a 20MHz can oscillator near L30). Oscillator confirmed solid 20MHz, with clean 40MHz output from the driver chip feeding the CPU.
- CPU reset (RSTI, from schematic near U44/U52): probed with proper triggering - shows a clean, brief low pulse at power-on before releasing high and staying high. Looks like correct, healthy reset behavior.
- TA/TEA at the CPU: initially looked like a stuck repeating pattern when probed with the board inverted (likely bad probe contact), but re-probed properly with RAM installed and got irregular, varied pulsing on both; consistent with real bus activity rather than a stuck retry loop.
- Bus transceivers (4x 74ABT245D between CPU and RAM): OE shows normal-looking activity, DIR sits low (not yet confirmed if that's correct per schematic). Two-channel scope comparison of A-side vs B-side on the same data lines shows them mirroring correctly, so data appears to be passing through cleanly.
Where I'm stuck: everything checked so far looks individually healthy (RAM ruled out, Cuda good, clock good, reset good, TA/TEA active, transceivers passing data correctly), but there's still zero chime, zero chime-of-death, and zero video sync activity even after 10+ minutes. Given the chime is generated by the CPU in software very early in boot (under a second in, from what I understand, or is the DSP responsible for the chime?), and video init also happens early, I'm having trouble reconciling "CPU appears to be executing" with "neither of the two earliest visible boot milestones ever happens."
Someone mentioned bad bus transceivers as a known failure mode on these boards, checked mine and they look OK per the above, but wondering if there's something board-specific about the 840AV I'm missing (an AV-specific ASIC or DSP3210 dependency that could hang boot before either chime or video is reached?).
Thanks in advance.
Symptom: power on, HDD spins up, floppy light flashes briefly (normal self-check), no chime, no chime-of-death, no video output (tried both a VGA adapter, previously proven working on my Centris 660AV, and a genuine Mac monitor, no sync/picture on either).
What I've checked and ruled out so far:
- RAM: symptom persists identically with all RAM removed (still no chime/chime-of-death), so RAM isn't the root blocker. Currently testing with a known-good matched pair in Bank A.
- Cuda (341S0788): VDD stable and clean through power cycle, no ripple. Clock oscillating correctly at 32.768kHz. Manually grounding Cuda's RST pin causes a clean shutdown, confirming Cuda is alive and responding correctly. Cuda's own RST pin sits high at idle and never pulses automatically, traced this back through a 330K pull-up to the S2 (manual reset switch) footprint, which matches the schematic (sheet 1/12), this appears to be normal behavior, not a fault, since Cuda generates its own internal power-on reset from its own clean VDD.
- Main CPU clock: found and confirmed via schematic (sheet w/ U44 MC88916-80 clock driver, fed by a 20MHz can oscillator near L30). Oscillator confirmed solid 20MHz, with clean 40MHz output from the driver chip feeding the CPU.
- CPU reset (RSTI, from schematic near U44/U52): probed with proper triggering - shows a clean, brief low pulse at power-on before releasing high and staying high. Looks like correct, healthy reset behavior.
- TA/TEA at the CPU: initially looked like a stuck repeating pattern when probed with the board inverted (likely bad probe contact), but re-probed properly with RAM installed and got irregular, varied pulsing on both; consistent with real bus activity rather than a stuck retry loop.
- Bus transceivers (4x 74ABT245D between CPU and RAM): OE shows normal-looking activity, DIR sits low (not yet confirmed if that's correct per schematic). Two-channel scope comparison of A-side vs B-side on the same data lines shows them mirroring correctly, so data appears to be passing through cleanly.
Where I'm stuck: everything checked so far looks individually healthy (RAM ruled out, Cuda good, clock good, reset good, TA/TEA active, transceivers passing data correctly), but there's still zero chime, zero chime-of-death, and zero video sync activity even after 10+ minutes. Given the chime is generated by the CPU in software very early in boot (under a second in, from what I understand, or is the DSP responsible for the chime?), and video init also happens early, I'm having trouble reconciling "CPU appears to be executing" with "neither of the two earliest visible boot milestones ever happens."
Someone mentioned bad bus transceivers as a known failure mode on these boards, checked mine and they look OK per the above, but wondering if there's something board-specific about the 840AV I'm missing (an AV-specific ASIC or DSP3210 dependency that could hang boot before either chime or video is reached?).
Thanks in advance.
