ntp and impact on jitter?

The clock in the top corner of Debian was off by one hour (day light savings time) and I found this annoying. So after a search I found a discussion suggesting installing NTP. So I installed NTP and it resolved the off by one hour problem. The NTP installation appears to have installed three things:
ntp:all, ntp sec, and python3-ntp

The basetime jitter on this machine now appears to be higher bordering on 25us, when my previous recorded value was much lower 10-17us

Is it possible that NTP could be impacting jitter to this extent?

Looking at the installer, there are a total of nine NTP related programs /files installed, so I'm reluctant to uninstall all of them in fear of breaking something, or perhaps this has nothing to do with the change in jitter?

FWIW I'm using a Mesa7i95t to control my servos.

Words of wisdom are welcome, Thanks

NTP can affect jitter as it adjusts LinuxCNCs servo thread rate

Note that 10 usec of added jitter is unimportant unless you are
using a parallel port or GPIO to generate high speed signals.

PCW wrote:

NTP can affect jitter as it adjusts LinuxCNCs servo thread rate

Note that 10 usec of added jitter is unimportant unless you are
using a parallel port or GPIO to generate high speed signals.
 

That makes sense, my assumption is that he Mesa board is generating the step pulses asynchronously to the possibly jittery data it might be getting from the Linux real time thread? does that make sense?

When I finally close the outside loop using the magnetic scales (not there yet), then jitter might be more important? Not sure how that outside (position) loop is controlled, via Linux or within the Mesa board logic (more like true real time)?

Good question for Mesa.

Yes, the steps are generated at a constant rate between updates from LinuxCNC.
Critical timing like samples of the current stepgen or encoder position are retimed
by a DPLL in the firmware, reducing the actual sampling jitter to  less than 250 ns
even with 100s of usec of servo thread jitter.