[Flyingparts]

Currently running a horizontal multipallet cnc with a fanuc 31i model b controller. It is equipped with x,y,z servo and spindle load monitoring. As I understand spindle load monitoring is hp/torque generated during the cut. So values can range from 0 to 150% safely determined by time and tool length bending moment. What i want to understand is x,y,z loads... I can't find any information online for safe percentages, and general information. Onto my issue: the servo loads (x,y,z) bounce violently during a cut. Example : circular milling (with an iscar ffx4 xnmu high feedmill) when moving in the x or y the load meter is bouncing from 0-45% with fluctuations in the +- 30% load. My thoughts are the cutter is deflecting and becoming unstable. I'm conflicted though because the spindle load stays pretty consistent with +-5% variations. This issue is not isolated to just circular milling and high feed mills. Straight line milling with climb/conventional tool paths same issue. The most consistent toolpath is helical ramping with minor load variations.

[joeavaerage]

Hi,
when I first built my mill, it being equipped with three 750W Delta servos, and also given that the servos have an analog output that reflects servo load that I would monitor the loads.
What a waste of time!

I found that 99.9% of the time the reading was so low, maybe 50W or so that it was hard to read accurately, and then in the briefest of moments the servo load would jump up to many hundreds of W.

The physics is that a servo draws large currents (and indicates an increased load) only while it is accelerating or being asked to produce a constant thrust, while when it is not accelerating the current drawn is just enough
to overcome friction and whatever cutting load is being presented to the axis, often vanishingly small. Once I came to understand this I realised that monitoring servo load is of little value.
If you had a large production machine that was constantly or at least very heavily loaded with cutting forces for protracted periods of time, then maybe it would be interesting or informative, but for normal
machines under normal loads I don't see it as particularly useful. Spindle load is a different kettle of fish, spindles are often heavily loaded for protracted periods.

Craig