additive offset error

so it seems that when it warms up, all is fine. this problem was happening in december last year, and now it's cooling down a bit, so it's returned. i geuss it is either the encoder or the mesa pwm/encoder board having a hard time counting.
im using a HENGSTLER 5000 line encoder (rs components 260-3910)

and a mesa 7i33ta daughter board.

ill edit my config files and change to a spare encoder input soon to see if it helps.

by the way, with a count of 444444.444444 counts per second (1mm = 3333.33333 counts) at 8000 velocity, this is well within the cards 10mhz read speed. and i limit speed to 6600velocity.

Sounds like something is marginal I would check
1. That the encoder 5V is OK
2. That the 7I33 5V is OK
3. That the 7I33 is jumpered for RS-422 inputs
4. That you are not exceeding the 300KHz max square wave rate of the encoder

No that's the encoder spec, not sure if its a count rate or square wave rate, count rate would be bad, square wave rate would be fine...

Hostmot2 encoders with filter-on take 15 clocks to respond so maximum count rate on a 5I20 (33 MHz clock) is ~4.4 MHz and about 5.3 MHz with the other FPGA cards. This is configurable (downward) with a global register that sets the quadrature counters filter clock rate. but HostMot2s driver does not have code to access this register.

YUP! 300KHZ seems to be the ceiling thanks for the help on this. -this has only been an issue lately, new encoders can obviously exceed there speed limit, leaving you tearing out hair later on..