See how easily you can do a PDN measurement using a network analyzer or VNA in the video here. Building solid power distribution systems (PDN’s) is maybe the most important learning from the signal integrity training overall. Because it matters. Both for the logic functions, the signal quality (integrity) and for the radiated emissions (EMI). It is all about getting the right bypass capacitors on the board and making sure you get the nice low impedance across the frequency range that you are after.
When you build something, you have to test it as well. And think about how you want to test it preferably before you build it. With PDN, the simplest way to test is by using either a
- Network Analyzer (VNA), or a
- Spectrum Analyzer (SA) with Tracking Generator (TG)
This short video shows how easy these measurements can be made using a network analyzer. The same method can be used with a spectrum analyzer (with TG option).
(also available directly on YouTube)
The two boards used in the video are actual designs, one mounted with capacitors only and one bare. For the purpose of demonstrations, both boards have been equipped with SMA connectors, so the measurements can be done easily with both hands-free.
Simpler methods for probing exists, and that may be a topic for another post. Usually probing it not a big issue, but obviously it’s easier when the boards have been prepared for probing.
In reality, the measurements work just as well on a PCB with all parts mounted, where you have the benefit of also measuring the capacitance on-chip and package.
The measurement is often referred to as an S21 measurement, as we measure the amount of energy that can pass through from port 1 to port 2 with this PDN placed as a big ugly load in the middle of the transmission line.
Really we do not care about the phase in this PDN measurement, only the amplitude. So you could argue that using a network analyzer for PDN measurement is overkill. To put it very short: the difference between a network analyzer and a spectrum analyzer equipped with a tracking generator is that the network analyzer provides both phase and amplitude information – the SA/TG does amplitude only – this does not matter for PDN measurements. Use what you got. This little network analyzer is quite handy for these PDN measurements.
UPDATED April 2020: The VNA/network analyzer used in the video is the Rohde & Schwarz ZNL3 (5kHz – 3GHz). This is a really nice instrument in the lower end of the price range, yet quite capable and more than adequate for this type of measurement.
Notice how the two different test point locations give pretty much the same result up to the frequency where board resonances start to dominate. This is another way of saying that you can move the bypass capacitors around relative to the IC’s with no significant impact. The location of bypass capacitors is not that important.
Why is it 25Ohms in the calculation of the impedance magnitude? Shouldn’t it be 50 Ohms, since this is the reference impedance of the VNA-ports?
Try doing the math. Remember, 0dB means the receiver sees half the generator amplitude.
Do we need to modify the setup when we have active components on the board while measuring?
Good question. Turns out that is not an issue. See also some of the other posts about this in the PDN category.
Hm, I don’t understand why that is. Could you please point me in the right direction? I was on the understanding that if your insertion loss is 0dB, then all of your signal gets through (no return loss), and therefore the impedance is zero. Any help would be greatly appreciated.
The impedance “Z” of the PDN is a shunt impedance in the S21 measurement. So at 0 ohms, nothing gets through, and at infinite impedance, everything gets through (0dB). Hope this helps.