RPI4 latency - is this normal

Does the attached latency plot look even half way normal for a rpi4 ? I get jitter result of around 75000 after a 10 seconds or so.  Pinging 10.10.10.10 gives reasonable consistent times of 0.105, but every now and again it spikes up to 0.150 or 0.190. This plot is from a second
pi booting from an SSD (to allow me to test in the house where i have more space), but the actual pi in the workshop is running from an SD card. 

I'm running and rpi4 with mesa 7i96, 7i85 and 7i73 and steppers. I've got a config that works to a point, but i keep getting following errors.  I can mill manually commanding the axis for a couple of inches and it then throws a following error.  It will also throw a following error whilst homing sometimes.  Is latency on the servo thread (1,000,000) going to be the issue, or is it as likely to be settings in my hal/ini (which are both still work in progress)? 

Here's the main settings for X and Y.  Gearing on Z means that axis is set slightly differently.  Stepper drivers are generic leadshine (i think, from stepperonline).  whilst there are some entries in here from linear encoders they aren't currently being used for feedback.

Apologies, i guess it is. I'd been searching the forum before starting a new thread, found a couple of older posts that I thought may be related and posted there also. I then found out how to do the latency histogram and then wondered it if was related to the other posts at all, hence starting this one, suspecting half my problem may just be my settings above.

What settings are you using for your rpi
Any over locks wifi by audio and video?

Aaroncnc wrote:

What settings are you using for your rpi
Any overclocks, wifi, bt, audio and video?

FORCE_TURBO = 1 as per PWC's recommendation in another thread.
I also set ARM_FREQ = 1800 (as per pi400, I have a massive heat sink so it is running cool).
I am pretty sure BT is turned off, WiFi would also have been turned off when I plotted the histogram.

Try turning off audio as well.

You have the minimum ferror limit very small:
MIN_FERROR = 0.0025
(2.5 microns)

I would increase that to at least 0.01

As far as determining the source of following errors, the best way
is to first widen the following error limits (so you don't get a fault)
and then plot the following error (with HALScope) along with joint
velocity.

Also if the RPI is connected to the internet. make sure you have
ntp installed. The default RPI time demon (timesyncd) makes step
changes is the clock which causes changes in the servo thread
period too fast for the 7I96's DPLL to follow.

thanks PCW. Turns out i had the parameter specified twice, with one being redundant. Things are now much better with regard the following error.

I'm still wondering if the latency histogram looks normal? i haven't seen anything else on here that looks like that.