Jun
24
Taig Milling Machine Stepper Motor and Controller Upgrade
By karlw
A friend was selling some 200 ounce/inch, 3.5 Amp stepper motors for a great price so I thought it was time to upgrade the taig cnc milling machine. With the old steppers installed, the machine was accurate but slow. Since the old controller couldn’t handle 3.5 amps I decided to upgrade that too. I found a 5 axis controller on ebay for a great price. This controller uses the Toshiba TB6560 stepper driver that can handle up to 36 VDC at 3.5 Amps. I wanted a board with at least 4 axis’ so that I could add a rotary axis to the mill but more about that in another post. I went to the local electronics supplier (Sayal elelctronics) and Parm set me up with a Hammond electronics 24 VAC, 10 Amp transformer (a local manufaturer). I picked up a bridge rectifier, large capacitor, switch, fuse, computer power cord socket and indicator light to complete the power supply. The entire conroller and power suppy were mounted in a computer case.
Transformer, bridge rectifier and capacitor.
Switch, fuse and power socket.
Mounted indicator light and switch.
TB6560 stepper controller board and 200 oz/inch stepper motor.
Power supply and controller mounted in a computer case.
Close-up of computer case.
After getting all of the stepper motors wired to the controller board it was time to start testing. I noticed that the motors started losing steps and stalling out at speeds greater than 20 inches per minutes. I hooked the oscilloscope up to the step line on one of the axis to see what was going on. Instead of a nice square wave it looked more like a sawtooth because of the slow rise time. This is what was causing all the missed steps at higher speeds. The culprit was slow rise times of the optocouplers situated in the circuit between the parallel port and the drivers. An easy fix was to take the optocouplers out and jumper them with a wire. The parallel port is no longer isolated from the driver chips but that has never been an issue with my other drivers on different machines. The machine can now run easily at 40 inches per minute.
In this photograph, on the left, you can see the waveform of the slow rise time of the optocoupler. The waveform on the right shows nice square waves generated after the optoisolators were jumpered. This makes a big difference considering that there are 8000 steps in one inch.
Karl P. Williams
Subscribe to RSS Feed
Follow us on Twitter
Our Mailing Lists
Visit Kwartzlab TV
Karl, that’s so pro! Best project pics, as always.
Sounds like the driver board designer didn’t spec those opto-isolators correctly. You really need fast ones to run at 5kHz (based on your 40 inches/minute at 8000 steps/inch). The debugging you did really illustrates the usefulness of the oscilloscope. I imagine the board is running much more efficiently with a crisp waveform going to the gates on the FETs.
Great job.
DW
Thanks Darin! I actually had to make a couple of other changes to get the board running at full speed. The wrong value of capacitor that determines the internal clock speed of each driver chip were used by the manufacturer and had to be replaced with 100pF caps. Another flaw with the board was a current limiting section of the circuit that needed to be removed for each driver chip. These were all minor changes and it was a lot of fun to debug.