Creating a Harmonic Drive Mount.

The harmonic drive is an AC Servo motor connected to a harmonic reduction gear set in this case to be used as a 4^th axis for CNC milling. A face plate and chuck will be mounted onto this. Work held can be turned and milled at the same time.

As received the ability to turn the motor doesn’t exist and the adaptor board for the encoder doesn’t have a standard configuration.

An open source project from Germany, called a STMBL, can run an AC servo equipped with an encoder.

Interfacing to this can be done in a number of ways including step/dir so I created a custom board to take single ended step/dir/enable signals and translate them to the differential format required by the drive. The drive fault signal was passed back turned into a single ended open collector signal to parallel with other drive fault signals.

As I had 3 STMBL drives I built 3 adaptor boards. The encoder signals were also not properly configured on the motor so I first rewired them to a DB-25 connector in Bergerda AC Servo format. I also created a small interface board that changed the DB-25 signals into the two RJ-45 Ethernet form factor connectors that the STMBL uses.

Clearly there is no standard for motor encoder signals between manufacturers so it was best to use one standard and format everything to match that. Hence the Bergerda AC Servo drives I purchased for other axis.

Next I needed to mount the motor for both testing and configuration and ultimately also create a pattern for sand casting a permanent mount. I used a 3D printer to make the mounts and since the mounts were larger than my printer I quartered the CAD drawing and printed 4 pieces.

Here’s 1 quarter and a first try at how I’d mount things.

Things look pretty good. Side view. Back view

I ran out of red so I ended up printing it in brown PLA and the mounting screws hold it together. Here’s what it will look like on the mill.

Next the pattern. Using the CAD software I expanded the basic dimensions by 3% to handle aluminium shrinkage while the casting cooled. I also removed the sharp edges and added draft to make it easy to pull out of the green sand.

Once again the pattern was too large for my 3D printer so I printed quarters.

The shoulders for the drive mounting as shown in the ¼ real sample will be milled from the casting.

Once the pattern parts are glued together the bodywork starts. First some sanding and then standard auto-body filler.

There was only 20% infill used so parts of the pattern are rather flexible. But still after some sanding and a coat of primer they came out really well.

It really does look this good in real life and once painted is ready for a trial into the green sand. It was painted with standard spray can paint and the intention was not for a super glossy finish. It came out perfectly.

Now to try it in the sand.

The pattern was laid on the board, the drag set over it and then sand sifted and then packed around it. Here flipped over it shows no voids in the sand.

Pulled from the sand very cleanly.

Next step. Actually cast it. But first how to gate and vent it?

As shown in the trial adding gates and a sprue and risers allows feeding from the bottom.

Alternatively…

Still fed from the bottom but with a runner and one riser on the far end. Not as easy or probably not as good a method.