Read error and following error Mesa 7i96

Re:- setp hm2_7i96.0.dpll.01.timer-us -50 Would it help if I changed it - to - setp hm2_7i96.0.dpll.01.timer-us -100

I am getting a following error as well this is on a small mill.

.
I have commented out all the setp pid.x.maxerror .0005 lines

tommylight wrote:

The joint following error is caused by the loos of communication with the 7i96, so that is not the error that needs fixing, the lost connection is.
What are the ping times ?
Try raising the servo period to 2,000,000 and also raising the watchdog time to 10,000,000 for testing.

If I change the servo period to 2,000,000 I believe I have to change the P to 500 !! is that correct

A little explanation as to why you can get a "error finishing read!" error with
poor Ethernet latency:

The hm2_eth driver times the read response of the Ethernet card. If this exceeds
80% of the servo period (default value), the read will be aborted. If this happens
5 times in a row (also a default value) , the driver disables communication and
issues a "error finishing read!" message.

You can plot the timeout error on halscope to see if its happening by monitoring
hm2_7i96.0.io_error. Also plotting hm2_7i96.0.read.time will give you an
idea of how close to the limit you are. Time is in CPU cycles on X86 so you need to
divide by CPU GHz to get ns.

A little explanation as to why you can get a "error finishing read!" error with poor Ethernet latency:

Thanks Peter.

Unfortunately this mill is about a 60 mile round trip. It is a small mill that does not need to be faster that 2000mm/min on rapids

I have commented out all the setp pid.x.maxerror .0005 lines

Re:- setp hm2_7i96.0.dpll.01.timer-us -50 Would it help if I changed it - to - setp hm2_7i96.0.dpll.01.timer-us -100

So I would like to try As Tom suggested " change the servo period to 2,000,000 " and change the P from 1000 to 500

I can do the above by phone.