 |
| Figure 1. "Pollution" occurring on PDN traces. |
Board pollution is noise occurring on the packages and interconnects (traces and planes) that carry current from the VRMs to the consumer devices.
One place it can originate is from the VRM itself, for example, with the switching noise the VRM generates (Figure 1). That can be a real concern if the board capacitance means you have a resonance around the switching frequency that would act as an amplifier for the switching noise and cause all kinds of problems with other devices on the board.
 |
Figure 2. On-die Vcc (blue) can
affect the board power rail (green). |
You can also see board pollution from the IOs. To return to the example of multiple IOs switching over time affecting the Vcc rail (Figure 2), the low-pass filtering effect caused by the package lead inductance and the decoupling capacitors is filtering out some of the higher frequency effect, but you still see unexpected movement of several tens of millivolts on the board power rail (green trace). You might potentially see something similar on the ground of the board, as well, because you'll have a ground bounce type effect.
 |
Figure 3. Core logic activity (blue) can
affect the board power rail (green). |
Signals from the processor core can actually affect the board interconnects (Figure 3). Logic activity in the core can couple out through the connections onto the PCB, whether it be through the core voltage rails or by some other means. That noise is going to be a smaller amplitude than what the core sees, because there's a big impedance peak present on the board due to the PDN itself, which filters and attenuates noise transfer both from the die to the board and from the board to the die. One reason we tend to look at board PDN power integrity separately from on-die PDN power integrity is that they're somewhat insulated from each other by this impedance peak, which essentially acts as a filter.
When it comes to board pollution, one might well ask, "how do I figure out where the noise on my power rail is coming from?” Because the PDN is affecting everything in the system, that noise could be coming from anywhere. We have found that the most effective way to track down board PDN noise is by using a spectral approach to identify problematic aggressors.
 |
Figure 4. Spectrum of power rail shows
noticeable aggressor peak at 2.4 GHz. |
In Figure 4, we're using our Spectrum Analyzer software to look at a noisy power rail. The software is identifying the five highest peak frequencies in the spectrum. The biggest peak occurs at about 1 GHz. A lot of the peaks look like they may be harmonics of each other. There's a very noticeable peak at 2.4 GHz that's probably some sort of RF component, whether a component on the board or something picked up from the environment. Once you know from looking at a spectrum what the frequencies of the aggressors are, it's easy to correlate them with the known frequencies of components in the system to figure out what's bleeding out onto the PDN and causing your noise.
No comments:
Post a Comment