Web Design
Initial Observations

Purpose of a CNC Router

The router spindle has to be guided along three axis reliably and with precision. The position of the cutter head should not be affected by the cutting force nor should there any play as the direction along any axis is reversed.

The motors are stepper types. That means they turn by an exact angle for each drive pulse. The basic stepping angle is 1.8 degrees but the driver can further divide this angle to smaller fractions. As long as power is applied to the motors, they either turn or stay put. There is no free-wheeling with a powered stepper motor

The software and driver electronics is inherently reliable, provided that the pulse structures are matched to the motor. In other words, after the router has been moved back and forth 10 inches a thousand times, the final cutter position will be the same as at the start of the run.

The positioning accuracy should not be affected by the back pressure from the cutter head either.

This will be true if the motors are powerful enough and if there is no ”give” in the machine’s structure and drive mechanisms. Let’s look at the drive gears first.

Chain drives are often maligned because chains are not exactly precision devices. However, their pitch accuracy is actually quit good. Chain drive, by their nature, have no or very little backlash. Here I have named the bad boy, backlash. Backlash is defined as the dead zone of the motion as the driving motor reverses direction. Care still has to be taken to tighten the chain hard. How hard? As much as the overall structure can bear. Chains place a lot of radial pressure on all bearings along the way. This also true for the motor bearings for the Y axis in the design we are discussing. The X axis motor is only subjected to torque since it drives shafts with bearings elsewhere.

Long chains such as the one used for the X axis can also vibrate which introduces “jitter” into the motion.

The Z axis used a lead screw. This a special screw with tight tolerances for the pitch and diameter of the thread. That screw penetrates a matching nut to translate the rotation to a linear motion. When that screw and the nut wear over time, there will be an increasing play between the two causing backlash. There is help, special screws are made which compensate for wear by applying a constant pressure against the screw.

OK, so we can conclude that the drive mechanisms are tight and predictable. Next, we inspect the major structural elements such as the gantry box and the side panels that carry the gantry. The gantry box is a very stiff assembly using 3/4-inch MDF faced plywood. It is bolted together with 1/4-inch machine screws and cross dowels, a very secure fastening method. The side panels have widely spaced V-groove bearings giving the gantry a wide and stable stance.

Test Runs

The Mach3 software came with a test pattern which produced a very shallow but highly curved engraving. This means a lot of complex X-Y motion but no strain on the cutter head to speak of. The process which lasted about 20 minutes ended successfully with no apparent problems.

The next project required the removal of large amounts of maple and cherry involving cuts almost two inches deep. The cutting of several pieces in a row used up some 20 hours, filling two trash cans with wood chips.

This, time, several problems showed. The repeatability of the cuts was no longer accurate and the walls of the cutout pockets showed chatter marks that got worse with time. About half way through the job, I inspected the fit of the Y and Z parts within their rails and found noticeable play. The V-groove bearings for both axis were then re-adjusted for a very tight fit. This helped for a while but the “slop” developed again. I also observed that the Y carriage had more play in the center of travel compared to the ends.

A closer inspection revealed the real culprit. The V-groove bearings contact the edge of the L-shaped aluminum rails. The rails have what is called a structural profile, a sharp corner and a rounded corner. The hardened bearings cold-formed the edge into a new shape. The picture below shows the detail. the side with the overhand used to be a 90 degree sharp corner.

This is a very serious problem without a long-term cure. Other users of this type of linear bearing have suggested using shims under the rail or even filing the edges. Others replaced the aluminum rails with steel. The latter approach has its problems too. Straight steel angles are hard to come by and they will wear too albeit more slowly.




It was time for some serious re-engineering. See next page.

[Home] [Initial Observations] [Linear Bearings] [New Construction] [Other Details] [CNC Process] [Project Anatomy] [Sample CNC Cut] [Aztec Calendar] [References]