Originally posted by: Wildstar
Originally posted by: AaronE
^Wildstar, have you ever met Guntz? You two might get along
I have intricate knowledge of how this was made and how it operates I've been sworn to secrecy though :X
I will tell you wildstar that your previous estimations of the output voltage are off by at least a factor of -4x.
Perhaps the output voltage isn't nearly that high. It only needs to be a little bit above the max voltage to kill the CPU and PPU on the console. I'm just giving an example of how you can instantly kill the console with electricity. The chips are 5v tolerant chips. If you hit chips with over 7.0v, it's likely you'll be causing damage if not destroying the chips. 7.0v is part of the maximum voltage specifications of the 65c02s chip. The PPU and CPU uses a 65c02 core with additional custom on-die components. Again, 10v would kill the chips.
I seriously doubt that 10V is anywhere near the breakdown voltage of the chips. Running them a little bit overvoltage won't kill them immediately, but a 2x increase in voltage will quadruple the heat dissipation of the chip. Actually it's greater than that. Assuming the diode junctions have a Vdrop of .8V, and there are two of them between the GND and VCC, this gives the chip around 3.4V internally to work with. At 6.4V [1.6V + sqrt2 * 3.4V], the heat dissipation would be twice what it is at 5V. At 8.4V [1.6V + 3.4V*2], the heat dissipation would be four times what it is at 5V. The chip could definitely handle momentary overvoltage by a volt or so or across one or more of it's pins, but the black resin the chip packages are constructed from can only dissipate so much heat production. As a result the internal junction would temperature would rise even moreso above ambient temperature than the increase in wattage. Pushing the chip far out of tolerance can in some cases permanently alter the silicone even if it still apparently functions, causing intermittent problems with other I/O logic chips in the circuit. A circuit that does not outright fail but occasionally misbehaves can be difficult to diagnose.
As an example, the lockout "zapper" circuit used in unlicensed game carts provided -5V to the input pin on the lockout chip. This caused the chip to malfunction in such a manner that it did not trip the reset switch, allowing the console to function normally. However despite being operated repeatedly out of spec such that the chip could not function properly as long as the inverted signal was present, this caused no long term damage to the lockout chips, which would resume proper operation as soon as the out of spec signal was removed. Nintendo later tharted this workaround by adding a diode and shunt resistor to the lockout schematic. This safely dumped the inverse voltage source to ground without damage to the console or game cart, but the game cart refused to operate on such systems.
A catastrophic failure of the IC chip can be caused in one of two ways. Frying it by running the chip out of tolerance for extended periods by exceeding the absolute ratings of the device could get the silicon hot enough internally so that it breaks down. Generally the chip will develop burn marks or welts in the casing indicating visible damage. So I think placing -5 or +10 volts on a single I/O, likely may not be enough to destroy the chip but will likely impede proper operation.
The other method is to create a voltage potential inside the chip that exceeds the breakdown voltage of the silicon junction. A quick static charge will do this. Low current CMOS generally operates by pairs of MOSFET transistors on each output. These transistors conduct when a charge at their input causing the N-type or P-type silicon wafer inside to change types, allowing current to pass across the transistor. Unlike Bipolar transistor, each MOSFET have a thin insulator isolating the collector from the Base and Emitter. If a charge of sufficient strength to "zap" the collector is present, this insulating layer is destroyed causing current from the input to leak into the body of the transistor. Current flows from the Collector into the Base or Emitter and the MOSFET ceases to function properly.
In an IC circuit, leakage current caused by a "zapped" collector, internally or externally, will cause the chip to draw excessive current and get hot when connected to a rated supply. Sometimes the faulty chip will blow the regulator or fuse, and other times, as in the case of overvoltage supplied to the VCC of the PCB, one dead chip will sink so much current that the supply droops, preventing cascading failures of other healthy chips on the board.