KaVo Spindle Mount

5 Bears Home Homebrew CNC bench mill

The Sherline spindle has been mounted and tested as of this posting. While it is a fine little spindle, it is not capable of keeping up with the potential feeds and speeds of this system. Some initial trial cuts with small carbide tools (from 0.125" to 0.250") showed a typical feed of no better than 4 ipm with a modest cut of .062" deep, give or take. Beyond that, the spindle begins to bog and chatter a bit, and it also, sadly, begins to make a mysterious noise sounding suspiciously like bearing rattle. Don't let this description prevent you from buying a Sherline spindle, especially the ER16 variety, as it is perfect for a small stepper mill. Steppers need to move slower than servos anyways.

Now properly motivated, I wanted to get my KaVo 4041 spindle mounted, and this mounting block will help me do it. A very similar block from the factory is over $400 U.S., but it includes a coolant channel internally. If needed, I can later add this to my own block. With a bit of work, then, and $5 worth of 7075, I can create a precision mount for my KaVo spindle.

This entire machining sequence was pretty straightforward. The body of the KaVo spindle is ground stainless steel and measures 44.98mm diameter. It is designed to be gripped in this area with a split-ring style of clamping block. Thus, the plan was to mill a largish block of aluminum, bored through exactly 45mm, +0.03, -0.00, split in front, and mountable on the same block which currently carries the Sherline.

Any time you create a clamp in this manner, it is very important not to bore oversize. You want a close sliding fit of shaft to bore even before you split the clamp. If the bore is oversize, you may be able to clamp it up, but it will never be very secure compared to a proper boring of the correct diameter, and accuracy will also suffer.

Plus, it's poor engineering. We can do it properly!

After an initial 1/2" drilling, the tedious boring process was begun. This was a deep and large boring, and I used the longest boring cutter which would fit in my small boring head. Depth = 2.772" / 70.4mm. Pass after pass, massive swarf generation. As I got close to final diameter, the feeds were diminished and a telescoping bore guage was used to track the diameter of the cut.

You must remember that in a boring operation like this, the cutter will spring away from the wall of the workpiece to some degree, and it will be in direct proportion to the size of the outfeed. In other words, with aggressive cuts the boring head will cut undersized relative to the true mechanics of the boring head.

A common beginner mistake is to take a series of heavy passes right up to, say, 44.9mm, and then dial in another 0.1mm, hoping for a 45mm bore. When the final pass is taken, you will get the 0.1mm plus whatever spring in the bar is removed due to the light final cut. With a steel-shanked cutter of this length, you could expect to see cutter spring of up to 0.005" or more with a 0.100" outfeed. A carbide shank would flex far less, but I can't afford them!

To summarize, then, as you approach the target bore diameter, you should begin to progressively reduce the outfeed of each pass, and do a "practice" final, finishing pass perhaps 0.2mm / 0.008" shy of target. This checks for the true diameter of the head and bar, and you can fine-tune RPM for best finish. From this practice pass, execute subsequent cuts with an outfeed of just a few thou, and you will nail the target.

Successfully bored, I flipped the block and drilled and tapped the three pinch-bolt holes. The holes were first drilled for tap size (#8 for 1/4 X 20) nearly through the width; then, the holes were opened up to clear the bolt body in the upper 1/2 of the drilled hole. Only the bottom 1/2 of the hole is tapped. This allows the bolts to ultimately pinch the clamp around the spindle body once it is split.
For a neat appearance, I counterbored the three pinch-bolt holes. Counterbores in general are pretty expensive, and for sizes smaller than a #10 cap screw, the commercial varieties are pretty much needed. But for every cap screw counterbore larger than #10 machine, I simply use an appropriate endmill plunged to the correct depth. For 1/4 X 20, a 0.375" dia. mill plunged for 0.300" works great. Increase plunge depth as needed for lock washers or other spacers.

Here, the far hole has been counterbored with the 0.375" end mill. Slow RPM and strong downward feed (don't hesitate, keep it cutting) will kill any chatter and create a clean counterbore.

After all three pinch-bolt holes were machined, the block was reoriented once more to machine the back of the block. This machining replicates exactly that found on the Sherline spindle block. This block was engineered to get the KaVo spindle centerline very close to the Sherline spindle centerline, so both spindles are roughly in the same location when swapped.

Two 1/4" X 20 holes were drilled and tapped, and a 3/16" keyway was cut in the same pattern as the Sherline. I am a bit concerned that only two bolts are used, but two good bolts, torqued hard, should do the job.

The KaVo spindle slides perfectly into the new block. The boring was nice but not quite the near "hydraulic" fit I wanted, where you could feel the air literally being forced out the other end as the spindle body is inserted. The mounting block currently looks chunky and unattractive, but don't worry, it'll get dressed up later! I still need to split the block to allow clamping action.

Since this photo is tall, I'll use this space to discuss the philosophy of this spindle and HF spindles in general.

HF (High Frequency) spindles such as this, made by a number of manufacturers, use an integral AC 3-phase motor, usually of a high rotational precision and a high energy density, i.e. lots of power in a small package. This motor, measuring 72mm / 2.8" diameter and ~ 82mm / 3.23" long, generates 500 watts or 0.67 hp.

Powering an HF spindle is not a trivial process. A fairly bulky power supply is required, c/w large toroidal transformers, big capacitors, and the necessary control electronics to vary the frequency output. My KaVo power supply was purchased surplus in the same sale as the spindle, which had just been rebuilt with new bearings. The good news is that, if rated appropriately, a KaVo supply can easily power another brand of HF spindle like a Hofer.

The speed of the motor and hence the spindle is proportional to the frequency of the 3-phase AC delivered. Often, they have tachometry feedback for display and control. Typical speeds are 4,000 to 50,000 RPM, meaning it needs a small cutter for best efficiency.

The spindle shaft is integral with the motor. There are no gears or belts, and when correctly balanced, they are scary smooth! At 50,000 RPM, I can easily hold this in my hand and feel very little besides a faint "rumbling" inside, for lack of a better term. Quiet, too.

This spindle uses shop air to pneumatically actuate the collet closer mechanism, very cool, and fast too. It also uses filtered air to pressurize the body and bearing assembly, keeping dust and moisture from entering (and destroying) the expensive bearings. The pneumatic collet closer will possibly allow automated tool change via a solenoid valve to control the air, and an appropriate milling cutter carrier.

Collets available for this spindle range up to 6mm. There are only two imperial collet sizes available, 1/8" and 1/4". This limits me a bit, but both 1/8" and 1/4" collets hold a huge variety of tooling, and I can always pick up a few metric-shanked cutters. Industrially, these spindles are used for engraving, printed circuit board drilling, light milling, die work, and many other tasks. The correct power supplies are both heavy and expensive.

My hope is to execute high speed, lighter feed milling with carbide cutters, both ball-nosed and flat. Coolant will be required (or air to clear the chips), and the spindle itself generates significant heat which will hopefully be carried away by the 20 lbs of aluminum which makes up the bulk of the spindle assembly.

In the next installment, we will finish this mounting block with the first real work of the CNC mill!