What follows is a description of a clamping fixture, a grinding tool and a software program that provides a limited drill sharpening capability for a CNC Taig mill running TurboCnc.

I, like many reading this, learned to hand sharpen common twist drills either by trial and error or were fortunate enough to have an experienced machinist teach them the technique.

With practice, one can get quite proficient at mastering the necessary rolling hand movement and eyeballing the tip features for symmetry and necessary angles and clearances. With skill and some luck, sharpening bits smaller than 1/16" in diameter is possible. Quite probably hand sharpening is more than sufficient for the home machinist but I found the idea of building a machine to sharpen drill bits an interesting challenge.

I own and enjoy playing with a Taig CNC mill running TurboCnc. Every now and then I had a passing thought about the possibility of using it to sharpen drills.

Armed with a search engine and the web, I went on a quest for the geometric specifications for a "conventional" twist drill tip. Seemed these specs would be necessary information given that I wanted to build a drill sharpening setup. What I found was:

• There are many types of drill bit geometries (conical, four facet, six facet, split point, and on and on…)
• There is precious little (as far as I could tell) in terms of real manufacturing specifications on drill bit tip geometries readily accessible on the web.

For example, take what may be the most common twist drill, the 118-degree included angle two-flute twist drill. The first order dimensions for this tip are readily available. But what is not readily available, for example, is the precise geometry of the cutting lips relative to their corresponding heels. After much searching, I finally made the connection that this type of tip is what is referred to as "conical". In fact, the lip is a section of a cone. What I have not found as yet, are the geometric parameters that describe this cone.

After some head scratching, I concluded that even given I could determine the parametric values for the conical geometry, using the CNC mill as a machining platform to sharpen conical tips would require either some sort of rotary table/apparatus or a conical shaped grinding wheel. Neither of these options was appealing to me.

It seemed like a good plan to consider alternative tip geometries and to limit the range of sharpening capability to make things simpler.

The first simplification was to limit the design to sharpen bits using a four-facet tip geometry. One big plus for a four-facet tip is that it simplified the CNC motions, eliminating the need for a rotational axis. Also on the plus side, various sources (see links at end of article) indicate that a four facet tip offers the significant benefits of a point tip (to some degree self-centering), more accurate holes and lower required thrust.

After some experience using four facet drills, the only reason I can imagine why it is not used more commonly is that it requires more machining time and therefore is more expensive.

The second simplification that seemed reasonable was to limit the capability of the design to only two fluted, 118-degree twist drills with a diameter range of 0.5 inches down to maybe 0.05". This certainly covers the vast bulk of the bits I use.

After thinking about the problem and doing some investigation into the designs of commercial drill sharpening machines, a few key design issues came into focus.

The first relates to how to hold the bit during sharpening. One important observation is that the two flutes of the bit are axially symmetrical. It is attractive to consider grinding one lip, rotating the bit around its axis by 180 degrees and repeating the grinding for the other lip.

Two fundamental approaches to implementing this rotation came to mind. One is a fixture that would allow the bit to be unclamped from its initial grinding position and then rotated 180 degrees around its axis and then reclamped and the other lip ground. The other was to clamp the bit in a holder and then the holder (rather than the bit itself) would be unclamped, rotated and then reclamped.

I decided to go with the holder approach; it was not apparent to me how to accurately rotate the bit itself 180 degrees.

Another design challenge was how to clamp to the bit in the holder. The design should hold the bit securely in a repeatable way, providing support along most of the body of the bit so that the tip is kept from vibrating during grinding. One approach would be to use a set of collets that span the range of drill diameters. This clearly is a very sound approach but it seemed to be a lot of work to build the collets with sufficient precision.

I thought about using a conventional drill chuck but was concerned that it would not support enough of the drill.

I finally settled on a V block arrangement; basically the holder would consist of V slot running through a block with a way of applying pressure along the top of the bit to seat it into the slot. To accommodate the full range of diameters requires three or four brass "rams" of different thicknesses. For the smallest bits, the ram has to have a tapered profile to fit within the V slot. Implementing rams seemed like a much more tractable implementation problem than the collets, and fewer are needed to span the diameter range.

The V slot approach is pretty simple and relatively straightforward to machine with sufficient precision. It presents one major challenge, however. The location of the drill axis relative to the rotational axis of the holder block is a function of the diameter of the bit. Stated otherwise, for an arbitrary drill diameter, the rotational center axis of the holder is not that of the bit.

This means that after cutting the facets on one flute, when the holder is flipped (rotated 180 degrees) to cut the other flute, the tip of the drill translates left/right and up/down depending on the diameter of the drill. The collet approach, where the bit is inherently concentric with the body of the collet, would not have this problem. As far as front/back movement when flipped, that was dealt with in my design by machining the V slot so that it is accurately centered between the front and back faces of the holder; because of the centering in this dimension, flipping produces no front/back offset.

The solution I adopted for the up/down and left/right translation is based on the CNC capability of the machine. The amount of the translation is readily calculated based on the geometry, given the diameter of the drill to be sharpened and therefore can be compensated for by programming the machine.

Here is where it would have been great to have a G-code driver capable of parametric input i.e. the capability to run a G-code program supporting run-time input of data. The currently released version of TurboCnc does not support parametric input (I believe it is planned for the next release/ V4.0. Note: subsequent to writing this article, V4.0 is available).

To get around this limitation, I wrote a C language program (a DOS .exe) that is "hard wired" to generate the G-code program for sharpening. When executed, the program asks for the drill bit diameter; it then calculates the offsets, generates G-code commands containing diameter specific values and writes a G-code text file specific to sharpening the given diameter bit.

Front View Holder with Brass Ram on Right

Holder with Drill Mounted; Cap Screw Presses Ram Against Bit

Back/Side View of Holder

Mill Mounting Fixture on left; Note Insulating Phenolic Base

While conceptually sound, the above-described process turns out to be a tedious, largely because of the need to initially align the fixture on the mill relative to the grinding wheel.

The mitigation I adopted was to make use of the touch probe input capability of TurboCnc. The holder mounted base (that sets the included angle of the tip) has a rib that fixes its front/back alignment relative to the mill table by seating into one of the table slots; that alignment taken care of mechanically. The remaining alignment needs are to determine the left/right tip location relative to the grinding wheel and the center axis of the drill bit.

As you can see in the accompanying photos, the bottom brass "cap" that captures the grinding wheel against the arbor is concentric with the grinding wheel and somewhat larger in diameter. It is designed to serve as one contact in the touch probe circuit.

The idea is that the brass cap can be used to touch sense the drill tip and any required fiducials on the holder. The other contact of the touch circuit is the drill/holder/fixture that is insulated off of the mill table via a phenolic standoff (to prevent shorting of the contacts via the body of the mill).

After mounting the fixture in a slot on the table at an arbitrary left /right position, alignment consists of jogging the grinding wheel/brass cap such that it is about 0.1" to the right of the drill tip with the brass cap centered on the first drill tip to be ground in the up/down dimension. The up/down alignment needs only to be accurate enough to assure that the brass cap contact picks up the top tip facet accurately.

The front/back position is jogged such that the grinding wheel is roughly on line with the drill bit axis (within +/- 0.1" or so).

At this point, the diameter specific G-code file that was generated off-line is run. The first action is the mill moves the brass cap into position to sense the front side of the bit holder. Once the touch of the side is registered, the machine can "infer" the front/back location of the drill axis (the holder thickness is a known, fixed quantity and the V slot is in the center of the holder).

The brass cap is then moved into position in front of the drill bit tip in line with the drill axis that was just determined. The cap is advanced toward the tip; when touch is sensed, the tip position is now known to the machine. At this point, all three required positional relationships (tip to grinding wheel) have been established.

Next, the program requests that the spindle is turned on (this could be done automatically with a spindle controller) and the first facet profile is cut.

The grinding wheel then moves down and the second facet is cut. The reason for moving the wheel down is to distribute wear across the face of the wheel. For consistency, both first facets of the two lips are cut on the same location on the wheel; equivalently, both second facets are cut on the same (lower) location on the wheel.

Once the two facets on the first lip are cut, the program asks for a "tool change"; the tool change in this case is the request to manually flip the holder in the fixture and reclamp in preparation for grinding the facets on the second lip.

The program is then continued and sharpening is completed. At the end of the program, the brass cap/grinding wheel is repositioned in the position set prior to running the program. If an additional grinding cycle is required to remove move material, it is a simple matter to flip the holder to its starting orientation rerun the program.

Fixture and Holder Mounted on Mill. Note Diamond Grinding Drum with Brass Cap on Arbor. Wires are Touch Probe Inputs.

Shows Holder in "Flipped" Position to Grind Second Lip.

The initial positioning of the bit in the holder is critical in that the angle of the lip should be set at 15 degrees as illustrated. I determined this empirically by trial and error, lacking specs. How much the bit extends from the holder is not very critical since the program adjusts; extending about 1/8" is fine.

Initial Bit Mounting Showing 15-Degree Offset.

I was never able to get hard specs on the angles of the facets but did find a reference that stated that 10 degrees for the lip facet and 30 degrees for the heel facet were good numbers. That is what I currently have programmed in and I see no need to tweak them.

The program is currently set to remove 0.005" per pass (relative to the sensed left/right position of the tip). The program could be generalized to include inputting the amount of material to remove per pass, different angle selection, total amount of material to be removed, etc..

The grinding wheel used is a diamond coated drum sold as a replacement for the one used in the Drill Doctor bit sharpeners (part number SAO1326GA). It can be found on the web for about $20 from a number of sources (search on the part number). I purposely chose a diamond drum to minimize the problem of wear affecting the accuracy. The drum seems of good quality and value and should be considered for other tool sharpening setups where diamond is needed.

As with any grinding operation on a mill or lathe, I take pains to capture the grinding dust, some of which is abrasive and cause serious wear on the sliding surfaces. This is probably less of a problem with the diamond wheel in the sense that it shreds less abrasive.

While I indicated that the offsets that result from flipping the holder can be calculated geometrically, I took a more empirical approach. I clamped in a number of various diameter pieces of drill rod, measured the diameters and offsets and then did a linear regression fit to the data to generate the offset calculation function used in the G-code generation program.

What does seem to me unequivocal based on the results of this project is that four facet drills, at least for working brass and aluminum are superior to conical tip drills in terms of less "walking", significantly less thrust required and to my eye, roundness and accuracy of finished hole. I am particularly impressed with how little the bit wanders when drilling deep holes.

TurboCnc site:

Darex is a manufacturer of precision drill bit sharpeners:

An interesting article on drill point geometry:

A very clever affordable drill bit sharpener that has received some very good reviews (replacement wheel used in my setup):