The images are so blurry! I'd really like to see the numbers on the frequency axis of that graph of the coax response, because I have a similar length of cable that I'm using for some things on a scope
Thanks for sharing your comments. You are correct and made some good points. While I have made homebrew divider probes, I can't usually get good pulse performance above 1GHz. I do use the Lecroy P65 and P66 transmission line probes. These are much better matched than I could do myself, so I can achieve 7.5GHz and 47pS rise time with 0.75pF loading. If you can tolerate the low impedance, and in most power applications you can, these are a great way to go. I also like that they aren't keyed, so they work with any 50 Ohm equipment.
Tektronix offers a great 1GHz passive 3.9pF probe and their TPP0850 800V/800MHz passive probe, with 1.4pF. I measured the rise time of the TPP0850 at 970pS.
I also have the Lecroy ZS2500 and ZS4000 active probes, which are excellent when you need the speed and these are very low loading at <1pF.
I published an article recently in Power Electronics on measuring ripple that covers more on this subject of probing and cable matching. There is also a chapter dedicated to it in my upcoming measurement book, which should be available in late 2014 with the tentative title Power Integrity: Measuring, Optimizing and Troubleshooting Power-Related Parameters in Electronics Systems
You gave some fair advice but failed to mention why certain probe types work best in some situations. Source impedance often dictates what type of probe you can and should use. I often find myself creating my own resistor divider probes with simple coax and a couple SMT resistors. Very low capacitance and minimal loading on high impedance outputs. Active probes are great but sometimes you don't need them or want them. Been fooled lots of times relying on an active probe. Particularly nasty if you think you can use them for power supply ripple and droop analysis.