Speed (and feeds) Racing

I don’t think it’s a secret that the number one factor a lot of people care about when purchasing a machine is rigidity, yet comparisons between machines tend to be personal impressions by those who own the machines and not quantified in any meaningful way. I’m aware that there are industrial methods for measuring rigidity, but I don’t think these are easy for anyone with a machine to conduct. As a result I’ve never seen tables comparing true data between machines.

I’m working on doing a linear rail upgrade on the X-axis of my Shapoko 4 (maybe the first?) and I’d like to design a cutting benchmark or some other test to see how much this upgrade changes the stiffness. Could we come up with some kind of speed test cut similar to how people print boats or blocks on their 3d printers that can give us measurable results to compare different machines? Just throwing this out there to see if anyone has ideas, and this seems like a Winston video idea in the making.

There’s more than just the V Wheels vs linear rails at play (pun intended) here.

The drive belts, frame and extrusion deflections all contribute too. The machine deflection also varies substantially for the same stiffness as frequency increases as the mass tends to damp higher frequency vibration.

My measurements on an SO3 indicated that the major contributor was the V Wheels, now that I have linear rail on the X axis my deflections are much smaller but the machine still deflects and vibrates when cutting, just quite a lot less.

Cutting speed measurements are quite sensitive to cutter, stickout, workpiece material (this specific sample when we’re talking wood), feed rate, RPM etc. It’s a worthy goal, but somewhat more complex than one might hope.


A test that works with different machines is hard, like Liam said, everything matters.

However, making a test that you run on your stock machine and then again after its modded will tell you how much better its performing. IMO, desktop machine performance can (or should) be boiled down to the MRR it can achieve in certain materials with minimal chatter in its weakest direction, tested both at high axial/low radial and low axial/high radial cuts.


To expand a bit on this point, I upgraded my machine because there were cuts it couldn’t do. When I upgraded the machine, my main question was “can it do this cut now”. I’d recommend doing something similar for any upgrade.

To get at an actual test, mine was:

  • Setup
    • ~3mm single-flute endmill
    • Fix feed per tooth per manufacturer recommendation
    • Max RPM
    • ~75% radial depth of cut
  • Test
    1. Set axial depth of cut to 0.1mm
    2. Do a cut and see if the machine sounds like a dying cat
      • If yes, stop test
      • If no, increment axial depth of cut by 0.1mm and go to step 2

On a Shapeoko you might want to substitute the 3mm endmill for a 6mm endmill

I’d make it a bit more flexible than that and leave it as the highest MRR the machine can reliably achieve on a 2D toolpath. If you have to use high-radial+low-axial cuts to get that MRR (which is the case for the Nomad for example), I don’t think it’s that a big deal, so long as you know.

Don’t forget workholding! Also chip evacuation and clearance.


This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.