Analyses to optimize milling spindles

Photo of Spindle

Milling spindles are used to manufacture a variety of shapes and molds (e.g. for mobile phones) out of various materials. The spindle is assembled on a milling machine axis which operates on the work piece – in some applications with up to 100,000 rpm. Such spindles need to fulfill extremely high accuracy requirements within the range of some microns (μm)!


To save costs respectively increase the production volume of molds, the manufacturing time needed to be reduced. This could be achieved for example by milling at a higher speed as well as by increasing the feed motion of all milling machine axes in the x-y-z directions. Accordingly, the unbalance force (due to the higher speed of the spindle) respectively the inertial forces (movement of axes) increased due to increased acceleration and deceleration of all mechanical parts. This led to higher vibration levels and a decreased accuracy as the work piece surface deteriorated due to the vibrations.

One important parameter to increase accuracy is the mechanical construction of the milling machine in total and of the individual spindle in particular. Stiffness and damping have a remarkable impact on the vibrational behaviour during the milling process and to the accuracy of the manufactured work piece.
The second parameter that influences the manufacturing time is the high speed of up to 100,000 rpm to increase material removal. At this speed, unbalance forces have a remarkable impact.

Considering both parameters, the overall vibration levels MUST be relatively low to achieve the required accuracy.


The Overalls and Tracking Modules of the VIBROPORT 80, an acceleration sensor and an optical speed sensor were used to perform several vibration measurements at variable speed; the Report & EXaminer Software was used to post-process the data in order to learn more about the system behaviour.

In view of the high speed, a robust speed measurement was important, hence several parameters needed to be considered, such as the operating speeds, the surface and diameter of the shaft in regards to the reflective tape or trigger mark, external influences like noise or electromagnetic fields etc.

First a measurement of a coast down of the spindle was performed in the Overalls Module (f(n) mm/s rms vs. speed). Next the spindle was started again and the coast down was repeated using the Tracking Module. The raw time data (velocity vibration by single integration of the sensor signal and the trigger signal over time) was acquired and stored. Then, using the post-processing capability of the Tracking Module, a 1st order Bodé diagram was calculated from the stored raw vibration data. All these files were transferred to the Report & EXaminer Software to calculate a waterfall diagram and a spectrogram, which can be considered to be a “fingerprint” of the milling spindle.

Finally, the correctness and consistency of all measured data and the post-processed (calculated) data was verified by using special cursors.


The calculation, analysis and evaluation of the plots were the starting point for the machine experts on site to optimize their milling spindles, e.g. by changing the stiffness or damping and thereby reducing the overall vibration level.

By tracking such diagrams of a spindle over its life time, deviations to the reference measurement (taken at production of the spindle) can help identify further areas for optimization, detect developing deterioration and help increase maintenance intervals.