Build - Electronic Indexing Head

[Feb. 3, 2017] - started placeholder page

[Dec. 26, 2017] - added actual build notes

[Jan. 7, 2019] - added notes on final version

[Jan. 24, 2019] - final tests

The drawing to the left is a temporary placeholder, pending completion of the project; it will be replaced by a photo of the completed indexer. The picture shows the mechanical components ("fixture") of the indexer, which will be used to hold a part for milling. From left to right, the major components are as follows:

• stepper motor

• motor support

• flexible coupler

• adapter fitting (connects coupler to collet holder)

• ER32 collet

• collet support

• baseplate (at bottom)

Adding a Beeper

Adding a simple beeper (I used a "piezo buzzer"like the one on the left - actual size about 22mm diameter) is very easy; these cost around $1 or less (2019).  Simply connect the two wires on the beeper as follows:

• black: pin 13

•    red: ground (GND)

Actually the beeper will work even if the two wires are reversed, but I found that mine worked better as above than when reversed. Sometimes it is suggested to add a 100 ohm resister, but this is really not necessary.

The indexing program I wrote includes the option to use a beeper. I also wrote a simple program ("beep-test" - available for download in the Coding section below) which can be used to test beeper function and find the best frequency to use with your beeper.

The flexible coupling has a 5mm opening on one end (for the stepper motor shaft), and an 8mm opening on the other end. The collet holder has an internal hole in the straight shaft  - about 0.46" I.D. The problem is how to make an adapter that will fit into the shaft of the collet holder without slipping. The shaft is hardened, so there is no easy way to thread it - making a simple threaded fitting is not an option. My solution was to make a "stem bolt" (see Ref. #7 below). A stem bolt is a hollow cylinder cut at an angle, such that a when a bolt is used to pull the two ends together it wedges itself in a bore.

Adapter Fitting for Collet Holder

Here is the adapter I made:

Making the adapter was fairly straightforward:

First I turned a piece of steel in the lathe to fit the internal bore in collet adapter (note, the internal bore in the collet adapter was fairly rough, and there was an internal burr at the entrance to the hole which I was able to remove with a carbide countersink). I first turned the steel to a slightly oversize diamter, and then used abrasive paper to reduce the diameter until I had a very close fit. I then drilled holes for internal threading and a clearance hole. Next I revered the piece in the lathe and turned it down to fir the flexible coupling. With the turning complete, I milled flats on the large diameter (so I could hold it while I tightened the stem bolt), and also milled two flats for the flexible coupling set screws.

Note: In the plans for this build, I show an alternate simpler design (untested) for the adapter.

The final step was to cut the piece at a 45 degree angle. I did not really have a good way to hold the piece for this cut, so I ended up clamping it on the flats as shown in the picture to the left (I left the angle gauge in the vise to equalize the pressure across the jaws). Since the angle and position of this cut are not critical, I lined up the angle by eyeballing it against the angle gauge. Similarly, the slitting saw was positioned by eye against a line marked on the diameter with a blue sharpie and a caliper. I started the saw cut very slowly so that the blade would not be distorted by the pressure of the cut.

The cut went very well, and the finish was very good. Some minor de-burring was required, and there was one other small issue - the through-hole end of the piece had a few partial threads in it which I removed with a small Dremel burr.

When the assembly was inserted into the collet holder, and the socket head cap screw was tightened, it locked into place easily.

Stepper Motor Support Plate

The stepper motor support plate (made from 1/4" aluminum plate) was drilled in the mill for the motor opening and the bracket screws. The motor opening was finished to final diameter with a boring head.

Collet Holder Support Plate

The collet support (made from hot rolled steel plate) was similarly drilled and bored in the mill. The bottom tapped mounting holes created a slightly raised area around the holes, so the bottom was lapped on 220 grit sandpaper. I use a thick glass plate as a lapping plate for this purpose (hidden under the sandpaper in the picture, and held in a wooden frame).

The tapped side hole (to be used for the rotation lock) raised a burr on the inner surface; this burr was removed with a Dremel tool using a small spherical grinding attachment. This was probably a bit too aggressive (even though it worked); if I had it to do over again I would use a hand-held diamond burr (see the write-up below on the bushing).

Collet Holder Refinements

The face of the collet holder which contacts the bushing on the collet holder was fairly rough as purchased. In order to provide a smooth mating face, I used a carbide tool to clean up the face. You can see the tool in the photo to the left ( a very sharp and very good quality brazed carbide tool which I picked up at an auction). I held the collet holder in a 4-jaw chuck, using copper "soft jaws" to protect the part. I took a series of very light cuts (less than .001" per cut) until I had a smooth face.

You can also see a slight undercut on the collet shaft where it meets the face - this was already on the part as purchased. If it hadn;t already been there, I would have added it.

Bushing for Collet Holder

The bushing was made from bearing bronze (I happened to have some on hand; brass would probably also work). The raw stock was mounted in a four jaw chuck, and supported on the right end with a dead center. The photo to the left show the raw stock before being center drilled for the dead center.

The end was turned down for a close fit in the collet holder support plate. The photo shows the support plate being used for a test fit.

Before cutting off the bushing, the center was drilled out with a succession of drills of increasing size, ending with an 11/16" drill (I could have gone to 3/4", but that was a bit to close to required final I.D.).

In this photo the bushing has been parted off. Before parting off, a small undercut was made on the left hand end, and the internal face was squared off to ensure a tight fit in the collet support plate.

The partially finished piece was then mounted in copper "soft jaws" and bored out to a smooth fit for the collet holder. The final lathe operation was facing the end to make sure it is square, followed by slight chamfering of the I.D. and O.D.

Here is the finished bushing. I de-burred the tab, both inside and out, using a spherical diamond burr held in a small collet (this worked well, and gave me better control than using a Dremel tool).

Here is a closer look at the rotation locking tab (see Reference 5 below). In the finished fixture, a screw is rotated to push the tab against the spindle of the collet holder, locking it in place. A bit more than a quarter-turn of the screw is sufficient to lock the spindle in place.

After assembling all of the parts for a test fitting, I found that the collet holder was sticking a bit in the bushing. I put together a quick-and-dirty "flapper" using 240 grit abrasive cloth to polish the bore. It only took about 20 - 30 seconds of polishing to get the collet holder turning smoothly. I also use a bit of sewing machine oil to lubricate the bushing.

Initial Version : I found this version did not have sufficient torque.

Here is the fully assembled fixture (not yet including the rotation locking screw).

When I connected the motor to the electronics, however, there was one "small" problem - the motor was not putting out enough torque to turn the collet holder. As a first step to resolving this problem, I tried adjusting the motor controller to provide more torque.

Checking the stepper controller, I found it was set to the lowest running current setting (0.3A) which was of course also the lowest torque. I raised the running current setting in steps (taking care to unplug the equipment first), and found that I could get fairly good collet rotation at a setting of 0.8A. Since I wanted to make sure that the collet rotated reliably, I set the running current to the rated current for my motor (1.4A).

The motor is now running with these controller settings:

• Running current : 1.4A (max rated current for the motor)

• Stop : 20%

• Excitation : 16 (i.e., 16 microsteps)

• Decay : 0%

I found that if I change the stop setting the motor does not rotate, and as I have no idea what decay setting is optimal, I am leaving it at 0%.

With the above settings the motor moves 0.1125 degrees/microstep, which means that in order to obtain the optimal resolution I have to make the coding a bit more complicated. For example, suppose I wanted to make a 48 tooth gear. To move in 48 equal segments, each segment would have to be 3022/48=66.667 microsteps. Since the motor cannot move a fractional step, each segment has to be rounded up or down, and this needs to be done in a way that equalizes the small error across all segments, as well as making sure the motor returns to the exact starting point with each 360 degrees of rotation. Note that in the example provided, if we simply rounded to 66 microsteps, we would accumulate an error of 0.667 with every increment, or a cumulative error of 32 microsteps after a "full" rotation. Not only would this maximize the error (just 0.01%, but still . . . ) for every increment, it would also not return to the exact starting point after a "full" rotation(which would cause an obvious problem if we wanted to do multiple rotations for some reason).

Final version: After first trying another stepper motor with higher torque (but still not enough) I switched the motor to a stepper motor with planetary gear reduction. The motor is run with micro-stepping set to 1 (that is, no micro-stepping).

This motor is running with these controller settings:

• Running current : 0.8A (max rated current for the motor)

• Stop : 20%

• Excitation : 1 (i.e., no micro-steps)

• Decay : 0%

Here is the datasheet for the above motor (click on the image for a larger version). Note that the gear reduction is a nominal 27:1 but an actual 26-103/121:1 (or approximately 26.851239669:1); this makes little difference in actual operation, but is important to get right in the operating code.

Test cuts : Stepper Motor with Planetary Gear:

For this test, I mounted an aluminum disk on a mandrel in the collet, and used a keyway cutter to make 17 equally spaced (I hope!) cuts.

Prior to making the first cut, I set the stepper to rotate through a full 360° rotation in order to remove any backlash from the system. Before making each cut, I used the rotation lock (brass knob on the left) to lock down the collet.

After rotating back to the initial position, I repeated the cut. Ideally, this double cut will be identical in width to all the other cuts; however, if the position is off a bit, this will widen the cut.

Aluminum blank after cutting.

As an aside, I noticed a tendency for the mandrel to slip in the collet when I attempted to unscrew the nut on the mandrel. For future use, I will mill two flats on the mandrel so I can hold it with a wrench.

Final Test - Success!

As a final test, I made some gear blanks and set up the indexing head with an involute gear cutter. The picture to the left shows the gear blank before cutting.

The gear to be made:

• diametral pitch = 26

• pressure angle = 14.5°

• number of teeth = 30

Usage Notes

As with any mechanical system, the indexer has a small amount of backlash. It is therefore advisable when using the indexer to be aware that moving forward and reverse the same number of steps will not return the indexer to it's original position. Further, it is recommended that the indexer be rotated a full rotation in the direction it will be used, in order to take up any slack in the system. All further movements should then continue in the same direction with no reversals.

Further, the indexer rotation lock should be locked after every rotation before beginning and machining. Fortunately most of the cutting forces are at right angles to the direction of rotation, but even small movements due to cutting can accumulate over many small rotations.

Refinements

Thumb switch

To make triggering rotation of the stepper motor a bit easier, while operating the mill and working with the fixure, I added a "thumb switch" to the circuit. This was done by simply splicing a switch into the keypad connections (see next picture). I made the thumb switch using a momentary on-off switch I had in my parts box, along with some PVC tubing and two-stranded speaker wire.

The thumb switch lets me push the "A button" very easily while I am at the mill, while keeping the electronics out of the way of swarf. It also makes sure I rotate the motor in only one direction during milling operations, eliminating another potential source of error.

Splicing in the switch: I wanted to be able to, in effect, push the A button on the keypad using a handheld switch. To do this, I spliced in a momentary on-off button switch into the keypad connections. I determined which two pins to connect to by just experimentally shorting out pairs of pins until I found th two which triggered "A" (for me, these were the first and last pins).

I then spliced the two wires for the switch into the pins. It's a bit difficult to see in this picture, but if you look closely you can see two wires coming out of the black plastic sleeve on the pin. I did this by carefully removing the plastic sleeve (there is a tiny little plastic latch tab you have to lift up; you can then push the pin out). With the plastic sleeve removed, I soldered in a second wire (one of two wires from a two-wire pair) on each of the two selected pins, and then replaced the plastic sleeve.