PDN Impedance Measurements using the Bode 500 and Picotest PDN Probes

Ғылым және технология

by Florian Hämmerle - OMICRON Lab
The power delivery network (PDN) describes the complete system delivering the power from the power source (i.e. a battery or a AC/DC power supply) to the power sink (i.e. a processor or FPGA). The PDN must be able to provide a sufficiently low source impedance such that the voltage drop does not exceed the limits when the maximum current is flowing. The faster the current transients, the more challenging this gets. Adding decoupling capacitors to the PDN helps to provide a low impedance over a wide frequency range but also brings the risk of resonance peaks in the PDN. A simple PDN impedance measurement can help in verifying the PDN design. It will reveal critical resonances that could lead to system failures. This presentation will cover the basics of dynamic output impedance and PDN impedance, focusing on measuring low impedance using the Bode 500 in conjunction with Picotest PDN probes.

Пікірлер: 2

  • @CHkorkskruv
    @CHkorkskruvАй бұрын

    Thank you, Florian, for a very interesting video! I have a couple of questions. - At 9.20 you show two impedance plots. Is the blue one only a dream, or would it be realisable? How does it decrease the impedance seen by the load at the high frequency peak? - An interesting demo would be to use the Bode/Picotest combination on some real, third (or fourth) party PCBs to demonstrate the difference between a good and a poor PDN. - The formula introduced at 13.30 might need more explanation. - Would it be possible to make a (very delicate and expensive :) ) Kelvin-type probe with four needles, two for the excitation (current), and two for differential measurement of the voltage difference across the point of interest, e.g. a capacitor? The differential measurement would use both the input channels of the Bode. All three coax shields could by connected at the probe, so ground currents would still flow in the measurement cables, but as the difference is measured, this would not matter. Well, I am sure you tried this already :) - At 43:30 the impressive noise floor of the Bode 500 is shown. It is not clear to me how the voltage across the short is measured, with channel 1? - How important is it at higher frequencies, > 1 MHz, to have the PCB under test powered on? Looking forward to the next video from you, C

  • @OMICRONLabTutorials

    @OMICRONLabTutorials

    29 күн бұрын

    Hi! The blue curve at 9:20 is a simulation done by Picotest. It follows the "flat impedance approach". Steve Sandler is doing trainings on how to achieve this. I think the blue resonance peak is lowered by removing the low-impedance valleys. Thanks for the hint to the explanation of the 2-port shunt-thru formula. Regarding the noise-floor, the CH1 is not connected, the receiver 1 is routed internally to the signal source as a reference signal like in the S21 measurement. Powering on the pcb is important to see the dc-bias loss of ceramic capacitors when they are charged. For a more detailed discussion, please contact us via support@omicron-lab.com.

Келесі