Select Page

#### Below are the questions asked during the live event, along with their respective answers.

Q: In asymmetric stripline configuration, if the spacings between two planes from the signal layer is 3:1 ratio, can we assume almost all of the energy flows between signal layer and the closer plane?
A: It is easy to see the energy flow by considering the first wave action that brings energy to the entire line. The wave has deposited static energy in the two capacitances. Since energy equals 1/2 CV x V, The ratio of energy is simply the ratio of capacitances or about 3:1.

Q: Can I make the location of a decap unimportant by decreasing the spreading inductance of the capacitor by bringing power ground planes close to each other upto around 2-4 mils?
A: A decoupling capacitor located on a circuit board is connected to two transmission lines, one to the switch on the die and one to the transmission line carrying the logic signal forward. In the die there are other connected logic lines. The dominant impedance is the decoupling capacitor itself which looks like a few ohms. Bringing the power plane closer to the decoupling capacitor changes the characteristic impedance of this transmission path. Ideally if this impedance is changed to 50 ohms there will not be a reflection at this interface. As I pointed out in my talk, the best you can do is match the impedances along the entire energy path. The capacitor itself can be considered a 3-ohm source. To say that the location is unimportant is not valid because even if the characteristic impedance is controlled there is a delay and the energy that is moved by wave action is limited by 50 ohms in series with 3 ohms. That takes many reflections.

Q: What happens physically with the current along the line when there is a standing wave?
A: The standing waves indicate that some of the energy in the line is involved in a resonance. This reduces the energy that could be radiated by the antenna. If the sine waves are considered a series of step functions the resonances I showed in my talk take place. If the line is properly terminated there are no reflections at any frequency and thus no standing waves.

Q: Do fields penetrate thin conductors?
A: Yes. For logic the spectrum is wide and penetration is very complex. For logic, the penetration in the first nanosecond is important as this defines the amount of copper that is involved in losses for the first wave. In logic it is important to understand what the first voltage will be. In logic it is all transient behavior and in carrier systems it is all steady state behavior.

Q: What is the spacing of return current vias?
A: The spacing in theory should provide a controlled characteristic impedance for the energy path in either direction. If trace separation is 5 mils I would expect via spaces would be about 5 mils. Remember, the energy path is a controlled cylinder of space.

Q: Does skin effect affect characteristic impedance?
A: I am sure it does. So does trace proximities, variation in dielectric constant, humidity, etc. What is important is the control of gross anomalies, not all the trivia. Characteristic impedance control is always in the right direction. In practice I doubt there is tight control of characteristic impedance.

Q: Do fields use the holes in vias?
A: No. A field requires an E field and this field must be perpendicular to the walls. There can be no voltage across the opening therefore there is no field.

Q: Are gaps permitted in ground planes?
A: Yes, as long as the gap does not influence the shape of any field carrying energy (signal or power). If the gap forces return current to take a strange path, it is not permitted.

Q: Do ground planes shield against radiation?
A: In general, no. Conducting surfaces reflect fields, they do not absorb energy. They can modify the shape of an interfering field. If a ground plane is a controlled path for return current there will not be radiation.

This means that ground planes can limit radiation but this is not shielding. Shielding requires a conducting closure around a signal. Conductors that penetrate an enclosure can violate the shielding.

Q: How does the natural frequency of a decoupling capacitor relate to its performance?
A: The character of a capacitor in circuit theory involves the low frequency capacitance and inductance. Here all the inductance and capacitance are involved. For logic signals the length of time it takes to propagate through the capacitor is important. The time it takes a wave to move through a capacitor depends on the dielectric constant. If the relative dielectric constant is 10,000 the wave velocity is c/100. In other words the initial supplied energy depends on parameters that are different than those that control natural frequency. In practice the energy that can be supplied from a capacitor involves the characteristic impedance of the connecting traces. This limitation dominates in most applications.

Q: Why are gaps required when using ground-to ground vias?
A: The path taken by energy flow is the shape of the space between conductors. A path between ground planes is itself not enough. Gaps are needed so that the field carrying the energy does not spread out. This spreading raises the characteristic impedance and causes reflections. This in turn causes radiation and adds delays. The key to good design is controlling the shape of the energy path so there are no reflections.