@betajet I once fixed a nasty ground bounce by replacing a 74ACT245 with a 74LS245
Would you have any information on the now-defunct 74ACQ Fairchild logic family? (Q for 'Quiet'). Not a lot of online technical information available even in the data sheets other than a one-page primer (google the 74ACQ245 data sheet page 7), but from that I suspect they might have done something to limit the 'shoot-thru' current which can also contribute to ground bounce. I wonder how successful that effort was, and why they are no longer manufactured.
A recent linked-in discussion revealed to me why a 50% duty cycle digital square wave can have strong evenharmonic radiated emissions. I had always wondered why that happens in the EMI measurements. The culprit is shoot-thru current which occurs on every transition and is certainly not 50% duty, therefore both even and odd harmonics are generated.
roshan wrote: Bypass capacitors are used to provide in-rush currents to the part. It may be helpful in VDD bounce, but it will definitely not solve the ground bounce issue altogether.
This is true. The ground bounce is due to inductance between chip ground and the PC board. Here's a helpful diagram from Wikipedia's Ground Bounce article:
Assume the bottom line driver is high, so the capacitor is charged. When the driver tries to pull its load low, it needs to draw a large amount of charge from the capacitor. This requires a large current to flow through L. Inductors don't like to change current flow, so instead the chip GND blips high. This causes a positive blip on all the other line drivers in the package. If any of those are clocks or control strobes, have fun debugging!
One way to fix ground bounce is to reduce L, for example by having lots of ground pins. Another way is to have slew rate limited outputs so you don't get abrupt changes in ground current. I once fixed a nasty ground bounce by replacing a 74ACT245 with a 74LS245 I had in my parts bin (Pack rat? Moi?) Those nice, slow 74LS transitions did the trick.
Seems like you are trying to model the lead inductance ( on VCC and GND pin ) of the leaded part in Fig 2 . In which case, you will have to also show the lead inductance on the output (Vout) pin .
It would make more sense, if you showed a capacitive load on the output pin to demonstrate the discharge path of the current when Q1 is turned OFF and Q2 is turned ON
Bypass capacitors are used to provide in-rush currents to the part. It may be helpful in VDD bounce, but it will definitely not solve the ground bounce issue altogether. This is because the part is referenced to its internal ground, where as bypassing happens externally....so there is a difference in internal ground ( what the part acually is referenced to )and external ground ( what the external circuitry is referenced to ) potential.
Thanks for your explanation--I think I already understood this and your explanation was still helpful. But I think figure 3 should show the capacitor connected directly to Ref A and Ref B rather than as shown.