Mesa 7I52 with 3x20/7I68

Hmm sounds like maybe a Hal file or external cabling error

The reason it does not sound like a I/O pin error is that these are multiplexed encoders
so channel 0 shares all the same FPGA / flat-cable signals as channel 1

Thanks Peter, appears to be another Delta drive.

I am starting to think about the PC base period and jitter (59us and 39us respectively). Read through a number of your posts dating back however never anything damaging to the drive is mentioned. The sum of those two periods mentioned gives a max frequency of 10.2khz. Guessing the system must meet something similar to the nyquist critera so that 10.2k may require a little shave (0.7).

With a 320PPR encoder the servo/drive is spun to 30rpsec, running unloaded. If the drive is pushed faster momentarily could you see issues with the above, missing encoder pulses or something unexpected with the step + dir pulses?

The multiplexed encodersused by the 7I52 will count up to a few MHz with the default settings
(and jitter << a servo period cannot cause lost counts)

Assuming a 2500 line encoder signal (10,000 counts/turn) this few MHz count rate
allows up to 12,000 RPM before starting to lose counts

The host jitter has little effect on the stepgen
(and the effect of jitter spikes on motion can be minimized by judicious
use of the max-error PID pin)

Certainly nothing that's done on the step/dir pins should harm a drive

Maybe is some noise problem.
Signals are differential ? The servo is on isolation transformer or have filter in power input ?
Have you tried shielded properly grounded cat6/5 ftp cables ?
Termination resistor sometimes patch the problem of noise
and if attenuation problem exists possible solutions are use of shorter/thicker cables
more voltage/current in path of quadrature signals etc.
Aliasing happens if Fsampling < 2*Fmax (nyquist) so its not possible
also i think that when you have aliasing the results are more severe from missing some pulses.

Peter, indeed reading thru the specifications on the 7I52S last week there was no question of the step pulse generation capability. The hardware is solid, confirmed pulse shape and timing. I did recall some older threads referring to base periods and reading encoders hence the query. Will look at the PID max error now.

@VRE - Differential step + dir output to the Delta drive controller so should counter most noise induced on both lines, encoder output is ABZ which will verify next.

Some quadrature encoders have inside 3 trimpots that configured properly from factory
but with motor vibrations sometimes the wiper of trimpot moves a little and encoder looses the calibration.
You must verify with oscilloscope what happening at different rpm.
I have found one old encoder at garbage that it was normal at low rpm but at high rpm loosed pulses
i verified that with oscilloscope tuned these trimpots and worked again fine.
Be careful these trimpots are tricky to config they want very very tiny movements and you must mark with a permanent marker
the initial position before touching these trimpots because the tuning is not easy.
( check this quantumdevices.wordpress.com/2012/01/26/...ntal-encoder-design/ )