One of the leading causes of bit breakage on CNC routers is excessive runout or eccentricity at the cutting tool (called TIR - Total Indicated Runout) As recently reported in the trade press, excessive runout can have a significant effect on the life of carbide tools. In a test of of solid carbide drills, improving TIR from 0.0006 " to 0.00008 " tripled tool life. Ironically, in the survey, the average tool user considered 0.0005" (0.0125mm) TIR to be acceptable.
While researching the source of the runout, a surprising fact emerged. Most of the low-cost (<US$150.00) routers and laminate trimmers that we tested were actually quite good (TIR < 0.0003 in. measured inside the hollow collet taper). However, most of the the stock 1/8 in. collet adapters were just plain junk. Our measurements of stock adapters ranged from 0.0030 in. TIR to 0.0100 in. TIR. We did not find a single adapter with less than 0.0030 in. TIR, enough to snap off a 0.0313 in. (1/32 in.) carbide cutter, even under a low to moderate chip load.
The aluminum collets found in many hand grinders were the worst of all with TIR measuring as bad as 0.0150 in. The situation improved substantially with more expensive routers, although some models did not offer collets for tools smaller than 1/4 in.
To test our carbide cutters using a couple of these low cost routers we designed our own high-precision collet with run-out less than 0.0002".
Before measuring the TIR, make sure that your spindle bore and collets are clean and free of any corrosion.
Measuring spindle TIR is quite straight forward if you have access to a "Dial Test Indicator" (NOT a Dial Indicator) with a resolution of at least 0.0005 in. (0.0001 in. is preferred). If you don't have one, go buy one. If you are on a budget and can wait for a couple of weeks, go to eBay and search under "test indicator". Starret, Fowler and Minutoyo are brands to look for.
To start, load a calibration blank (or a broken bit with at least 3/4 inch of smooth shank) into the collet You need a straight, smooth shank to perform this measurement. The more the shank sticks out of the collet, the more any eccentricity will be magnified, but you need at least 1/2 in. inside the collet to insure that the bit is properly seated.
Clean the exposed shank by dabbing it with tacky putty. Silly Putty also works pretty well.
Position the ball tip of a dial test indicator as close as possible to the center line of the the shank and as far down on the shank as you can get.
Adjust the position of the indicator stand until there is about 0.010 in. (0.25mm) of pre-load indicated on the dial. We do not recommend using the jog on your CNC since an accidental rapid movement that exceeds the range of the indicator can permanently damage its internal mechanism.
Rotate the collet using your index finger to determine the orientation of the spindle that corresponds to minimum indicated deflection. The pressure of your finger against the collet nut can give a false indicated deflection so remove it when reading the dial.
When you have identified the point of MINIMUM deflection, rotate the "0" on the dial as close to the indicator needle as possible. With a sensitive instrument, exact positioning can be quite difficult so don't obsess about it. Record the position of the needle. (0.0005 in. in the picture on the right)
Place a white dot on the forward facing facet of the collet nut to mark your starting position.
Rotate the collet nut one facet using your index finger. Remove the finger and read the dial. Continue in this manner until you determine the spindle orientation that gives MAXIMUM deflection. Record the amount indicated on the dial. (0.0055 in. in the picture to the left).
Subtract the initial minimum deflection recorded above from the MAXIMUM indicated deflection to get the TIR for this spindle / collet combination.
The measurements shown in these pictures reveal that this spindle / collet combination has 0.0050 in. TIR making this spindle / collet setup useless for CNC machining with micro-tools.
Assuming that you have a long enough calibration blank, move the probe of the test indicator to a point at least 1 inch (2.5mm) from the face of the collet and repeat the above steps. In many respects, this is the most important measurement because most micro-tools stick out of the collet about 1", so this is where all of the cutting is going to take place. Excessive TIR here can show up in the finish of your final product and / or with a broken bit.
Record the TIR measured at both locations and the approximate angle (phase angle) between them (think of TIR as the hands of a clock that stick out from the true center, or axis of rotation). If the max TIR at both positions on the calibration blank are on the same side of the bit, most of your eccentricity is caused by the collet bore being slightly off-center. If there is a significant angle between them, most of the runout is caused by the collet bore being at an angle relative to the tapered exterior. This is the worst type to have because it makes the TIR grow larger as you move out from the collet face.
The moral of this story is, "Before you start cutting, measure the runout of your spindle!". Even though you will be itching with impatience to get out in the shop and cut some wood with your new machine, breaking expensive bits is no fun at all (except, maybe for your tool vendor). Adding TIR determination to your standard set up procedure will save you a lot of headaches, and will help you come up to speed much faster.
For a more detailed discussion on the effect of TIR on system performance, please read the section on Selecting the Best Grade of Collet for Your Application.