|TB6560 3 Axis Board from Amazon|
This board cost less than 40.00 USD. It seemed to good to be true so I tested it.
I got two boards to see how they worked and needed one to replace the motor driver for a 3020 CNC Mini Engraver that had blown out one of it's TB6560's. Failure was induced by the failure of a motor connector on the mechanism and raw wires touching each other while enabled.
One of the two boards did not work right away but with a soldering iron to touch up some poor solder joints and seating the integrated circuits the bad one started to work.
After muddling my way through the signal pin assignments and setting up SuperCam with an ArmPod to run it I found the motors made a very anoying high pitch sound when enabled. By adjusting the current setting to the lowest the sound became tolerable.
Spinning up the motors was disappointing. The motors stopped after about 12,000 steps per second and they ran ratty from 10,000 up. Using the oscilloscope, it was obvious that the optical isolation was unreliable above 8,000 steps per second.
When I took a moment and analyzed the ground plain of the circuit board it became obvious that the optical isolation circuit was pointless because the grounds are not separated. The circuit provides no benefit so I removed the IC's for it and wired around it. This presented step pulses to the TB6560 in excess of 32,000 steps per second.
This makes the board useful in microstepping mode with 1600 steps per revelotion making the it possible to get 1200 RPM out of stepper motor.
It was looking good running motors on the bench. It was a little noisy but for the price it looked good. Before I went threw the steps of getting running a Taig Micro Mill I thought it best to test it for loosing steps. When I held on to the shafts of the motors at rest I noticed a little quiver.
I set the stepper motors for X,Y and Z with the flats up at coordinate position 0,0,0. Then I ran a gcode program that was a swirl pattern for X and Y on an imaginary machine in this case setup as a 3020 mechanism. When the toolpath was completed I expect it to return the motors to 0,0,0 and the motor flats to up as they where when it started. The results were bad. It either lost or gained steps on all three axis.
So, I took out all my fancy modes and returned it to the way that I received it and throttled back the maximum steps per second of 8,000 and retested. The results were the same and the motor flats were ending up in what appeared to be random offsets from where they should be.
Using the same parameters with a board of my design I ran the test to see if I was missing something. The results were perfection good consistanct flats up.
I made three little g-code programs to test with. Again I set the stepper motor flats up at zero. Ran a toolpath that returned at end to 0,0,0.
With the first toolpath it moved the Z axis back and forth a quarter turn. It was set up for 1600 steps per revolution and 4mm per revolution so this in code was from 0 to 1 and back. I did a copy and paste about a hundred times. Looks like it runs consistent and good. Stops at zero with flats up.
Then I made a test where all three axis move 0 to 1 and then back. Made it run about 150 times and it seems to work perfectly.
Then I made a test where Z moves to 1 then X and Y move to 1 then X and Y moves to 0 then Z moves to 0 and repeat a couple hundred times. It fails with a random aspect to the results each time the program the flats end up someplace different. You can see the motor flats drift off the return to zero during the execution of the toolpath.
Next step is to hardwire the Select line to enabled on the TB6560's and test again. But I hadn't thought of that then.
My conclusion was that the board could not be made to work well as CNC motor driver for a micro mill. It would introduce drift and be a source of inaccuracy. I suspect the TB6560 is not properly engineered to be used as a driver for repeatablity needed in machine control.
Bottom of the board shows the TB6560's screwed to the heatsink visible from the top side. There is a wire soldered to connect the three individual motor enables together. Then on the ArmPod I had to disconnect two output pins on the 74LS244's, outputs were fighting each other.
This is the a schematic found from a google search and I can't remember where it came from. It missed one of the limit switch signals but that was easy to find with a beeping meter.
Parallel Port Signal Pins, Connector is Male
Configuration of the USB ArmPod, U2-3 and U3-18 pins lifted out to function.
Home Page at http://www.super-tech.com
Published 2/10/2017 8:11:17 AM
Copyright 2016 Dennis Bohlke