Beyond Design: When Do Traces Become Transmission Lines?
At low frequencies, traces and components on a PCB behave simply as lossless lumped elements—as taught in Circuit Theory 101. But as the frequency increases, the copper trace and adjacent dielectric(s) become a transmission line, the skin effect forces current into the outer regions of the conductor and frequency dependant losses impact on the quality of the signal. The PCB trace now behaves as a distributed system with parasitic inductance and capacitance characterized by delay and scattered reflections. The behavior we are now concerned about occurs in the frequency domain rather than the familiar time domain. This is the real world of high-speed design.
Beyond Design: Plane Cavity Resonance
Plane pairs in multilayer PCBs are essentially unterminated transmission lines—just not the usual traces or cables we may be accustomed to. They also provide a very low-impedance path, which means that they can present logic devices with a stable reference voltage at high frequencies. But as with signal traces, if the transmission line is mismatched or unterminated, there will be standing waves: ringing. The bigger the mismatch, the bigger the standing waves and the more the impedance will be location dependent.
Beyond Design: When Legacy Products No Longer Perform
As IC die sizes continue to compact due to demand for smaller and faster technology, and as switching speeds continue to improve, rise and fall times are creeping down into the sub-nanosecond realm, a territory previously reserved for microwave engineers. It is a common quandary that established products that have worked flawlessly for years suddenly stop performing reliably, due to a new batch of ICs that is used in the latest production run.
Beyond Design: Transmission Line Losses
In an ideal world, the entire signal waveform would uniformly decrease in amplitude, over distance, and the rise time would remain constant. This reduction in amplitude could easily be compensated for by applying gain (cranking up the volume) at the receiver. However, as signals propagate along a lossy transmission line, the amplitude of the high-frequency components is reduced, in magnitude, whereas the low-frequency components are unaffected. This selective attenuation of high-frequency components is the root cause of ISI and collapse of the signal eye.
Beyond Design: FPGA PCB Design Challenges
The primary issue is generating optimal FPGA pin assignments that do not add vias and signal layers to a PCB stackup or increase the time required to integrate the FPGA with the PCB. Engineers generally do not consider FPGA pin assignments that expedite the PCB layout. Hundreds of logical signals need to be mapped to the physical pin-out of the device, and they must also harmonize with the routing requirements whilst maintaining the electrical integrity of the design.
Beyond Design: The Dark Side–Return of the Signal
I guess we all think of a copper plane as a thick, solid plate of copper that can basically handle any amount of current we sink into it. It also serves to make the circuit layout easier, allowing the PCB designer to ground anything, anywhere without having to run multiple tracks. That may well be the case with DC or very low-frequency analog circuits, but certainly not in the case of high-speed design.
Beyond Design: Return Path Discontinuities
PCB designers generally take great care to ensure that critical signals are routed exactly to length from the driver to the receiving device pins, but take little care of the return current path of the signal. Current flow is a “round trip” and the critical issue is delay, not length. If it takes one signal longer for the return current to get back to the driver—around a gap in the plane for instance—then there will be skew between the critical timing signals. Return path discontinuities (RPDs) can create large loop areas that increase series inductance, degrade signal integrity and increase crosstalk and electromagnetic radiation.
Beyond Design: Microstrip Coplanar Waveguides
The classic coplanar waveguide (CPW) is formed by a microstrip conductor strip separated from a pair of ground planes pours, all on the same layer, affixed to a dielectric medium. In the ideal case, the thickness of the dielectric is infinite. But in practice, it is thick enough so that electromagnetic fields die out before they get out of the substrate. CPWs have been used for many years in RF and microwave design as they reduce radiation loss, at extremely high frequencies, compared to traditional microstrip. And now, as edge rates continue to rise, they are coming back into vogue. This month, I will look at how conformal field theory can be used to model the electromagnetic effects of microstrip coplanar waveguides.
New Functionality Improves Designer’s Productivity
I originally came up with the concept of an online impedance calculator way back in 1994 when I was working on the PCB layout and design for a new generation of SPARC 20 servers. We basically reformatted a Sun SPARC 20 pizza box motherboard to fit into a 5.25-inch drive slot.
Beyond Design: PDN–Decoupling Capacitor Placement
The impact of lower core voltages and faster edge rates has pushed the frequency content of typical digital signals into the gigahertz range. Consequently, the performance of decoupling capacitors, that are required to complement the power distribution network (PDN) and curb signal induced fluctuations, must also be extended up into this range. However, rudimentary design rules, adequate for frequencies below 100MHz, may not be suitable for today's high-speed digital circuits. The symptoms of an inadequate PDN design are increased power supply noise, crosstalk and electromagnetic radiation leading to poor performance and possibly intermittent operation.