Here are some more Shure data points - the cartridge body used is a 1000e - one of the p-mount variants of the M97x - in this case tested at 130pf, with a Jico SAS stylus, at a range of differing R loads.
The "simple" lines are the electrically modelled response, that one should get in a "perfect" world (ie: where the mechanical/magnetic mechanism was perfectly linear...) - which provides insight into the level of "imperfection" in the transducer itself!
View attachment 263929
For the M97xE you can get closer to a neutral response, by using a high-ish C loading, combined with a high-ish R loading... so with the SAS here - the 65k or 97k loading provided the flattest midrange... and up to 10kHz... but then you get a substantial peak, which needs to be controlled with a fairly substantial C load to roll it back down.
Later I did another series of tests at 505pf C load:
View attachment 263930
Note: Me75p is another M97x p-mount variant ... I have a small collection of bodies!
Here you can see that with the increased C load and light increase in R load 55k - I achieve a very good neutral response through to about 15kHz - after which it completely drops off - it is within +/- 1db up to around 14kHz - and within +/- 3db up to 19kHz.
You have to trade off flatness in the 4k to 10kHz frequency range, for extension and rise around 14 / 15 khz
The SAS needle, as good as it is, has too much effective mass to achieve the flat frequency response of the beryllium tube originals....
The resulting resonance at around 15kHz is what needs to be tamed (along with the slight droop in the lower highs/upper midrange... which is probably magnetic related rather than cantilever.. and inherent to the body/generator design I think).
I have a feeling I have some measurements where I got closer to an ideal loading with the M97 family of bodies and the SAS stylus....but there are multiple "optimal" loadings as you need to choose the aspect of performance you are most willing to sacrifice, and the one you want to optimise for most.