I have been checking the backlash on my router and found that the feed rate affects the amount of backlash. At a slow speed, 100-500 mm/m, I get around .1 to .18 mm backlash but at a feed rate of 1000 to 10000 mm/m I get 0.0 to 0.02 mm. I get this result whatever distance I move, 1,10,100 or 500 mm. What reading should I use? I normally cut wood at 2000 to 3000 mm/m so I’m inclined to have no backlash compensation. What would you recommend?
I assume you measured the backlash by measuring dimensions of the part. If you run at different feed rates, the cutting forces are different. This results in different flex (machine, router bit, etc). The flex adds up to the backlash. The two combined results in parts having wrong dimensions.
It is easy to compensate for backlash because backlash is "always" the same. Compensate for the flex is quit difficult because flex depends on the cutting forces.
Sorry I should have added that the backlash was measured using a 1 metre long optical scale that I use to tune my stepper motors, there is minimal load. I go past the reference point the move forward, zero the gage then move forward a set amount the go back to the zero point and observe the backlash on the gage.
It is possible, the deceleration causes the backlash to "disappear" at higher feed rates. This is possible if the drag/friction in one direction is higher than in the opposite direction.
You could measure the backlash from Y- to Y+ and from Y+ to Y-.