On Thu, May 28, 2020 at 08:36 PM, Jim Ford wrote:

I was able to adjust the probe to get a nice flat response when connected to
the calibrator on the 54504A. Fortunately it also looked flat with the
calibrator on my Tek 7904.

Hi Jim,
The artifacts that I was referring to don't show up on standard 'scope calibrators. Those only allow you to adjust "LF" behavior with a 1 kHz square wave (flat top).
I meant fast-edge behavior, like that visible with <=1 ns steps in the first few ns or in the edge itself. That's where the differences are. You won't just see a flat or single-curved top but all kinds of ripples, resonance-like behavior etc. I know about only one "classic" Tek 'scope that offers that kind of fast step: the 485 in addition to a nice 1 kHz square wave. It allows LF adjustment *and* HF adjustment. The latter is done with separate adjustment regulators in the compensation box of the probe, underneath the plastic cover. *And* you need a well-matched input to the probe, like a probe-shaft-to-BNC adapter
The calibrator's edges in the 7904 are no faster than 250ns so they are useless for showing the differences between e.g. the P6139A and P6137, much less perform the HF adjustment I was referring to.

And I'd check the input capacitance of the scope (7 pF in the case of the 54504A) versus the output capacitance of the probe (8 pF in the case of the P6139A), and
if they are close, you will probably get good results when you connect the probe to the calibrator and adjust it for flatness.

That's not correct: 8 pF is the nominal *input* capacitance of the P6139A and they don't have to be nor will be equal. Otherwise, you wouldn't be able to adjust an 8pF or 1.5 pF FET input capacitance probe to several different 'scopes, like the 2465B with its 15pF nominal input capacitance or a 7A26 with 22pF nominal. It's all about RC times, probe input, cable, probe output, 'scope input. The publications I referred to show it all.

Nothing to worry about for general 'scoping in the sub-100MHz domain though. Who said that 100 MHz actually is DC?

Raymond