Question Sabertooth 2x32, and a Kangaroo x2 controller Auto tune

Would it be possible anybody could help me with my recently constructed motion simulator, as I am having problems with the control of the sim?

I am a sabertooth 2x32, and a Kangaroo x2 controller.

For the position sensing input to the Kangaroo, I am using two TT Electronics 6127V1A180L.5FS hall Sensors.

And I have implemented a limit switch non-contact control at 75 ° on the top stroke, and 75 ° on the bottom stroke, this is controlled with hall switches via a Arduino with a bank of relays set to digital high when activated by the sensors two per motor.

I tried to tune the rig on Saturday, I used DEScribe software via a TTL Serial Converter to do a limit switch tune, after a 10 degrees upward motion and a 10degrees downward motion it began to shake violently the motors were oscillating badly and then the sim came to an abrupt stop.

On checking why, the motors had stopped I had two 25A fuses in line from the power supply that had fused, the motors are rated at 22Amps so maybe I need a 30amp fuse ?

The motors are 24V 400watt motors from simkit

I implemented the tune via the Kangaroo, is this normal behaviour when tuning with the kangaroo.

The dip switch settings I used are:

Sabretooth’s dipswitches

Off

DIP switch 1 to the Off position

Off

DIP switch 2 to the Off position

Power Supply mode

DIP switch 3 to the ON position

Packet / plain Text Serial

DIP switch 4 to the ON position

Address 128

DIP switch 5 to the ON position

No Emergency stop

DIP switch 6 to the ON position


Kangaroo Setup Dip switch settings.

Digital input

DIP switch 1 to the ON position

potentiometer position, (Analog feedback)

DIP switch 2 to the Off position

position control

DIP switch 3 to the ON position

independent drive mode

DIP switch 4 to the ON position


Any help from the forum members would be gratefully received.

As I am about to pull my hair out, I thought this was a walk in the park.

Have you tried working with the DEScribe software to set the parameters?
https://www.dimensionengineering.com/info/describe

Ive used the kangaroo myself with encoder feedback and 1 limit for homing (but needed to use crash limit for tuning). It was working fine for me but Ive since changed over to AC servos so I've forgotten many things about it.

I'll brush up on the manual and look into it for you. But it might take me until the weekend to get a chance to look at it.

In the meantime, it will be worth for you to get familiar with using DEScribe to do the settings.
I used beta ver 3.9 with no problem.

If your still struggling, maybe post screenshots of all the settings in each tab of DEScribe and I'll check it out on the weekend. And perhaps pictures or wiring diagram to the kanagaroo just to double check everything.

Hi JAD
Thank you for your reply, as you suggested I have downloaded the DEScribe manual and will read the data.
I think the severe oscillations on the motors are a normal aspect of a Kangaroo tune process to test the motors parameters, will need to find a way of resetting the motors as you cant turn these by hand.

I will post more data and pictures tomorrow

Hi JAD
Here is the circuit diagram / wiring diagram.

The tune does go through severe oscillations.
For interest, this is what the tune looked like for my setup.

In this example, I used the on board button of the kangaroo.
But later in the rig, I used DEScribe to start the tune process.

Notice in my video that the first steps of the tune is to seek the limits before it goes through the ocsillation routine. From your description of what happened in your tune, it seems like it never touched the limitis before going through the routine.

The wiring diagram looks fine.

Have you checked with DEScribe, that all the movements, feedback, and limits register correctly?
You can do this before tuning in the DEScribe interface.

Check the tune tab (on ver 3.9) where you can manually drive the motor with the slider, and observe the feedback and limit switches.
You want to see that all of this is registered properly and in the correct directions before performing the tune.
I dont have a setup to look at myself so Im going by memory that this can be done without an initial tune.

I talk about my sabertooth/kangaroo setup here using DEScribe.

I waffle a bit in this one, hope you find the what you need in it.

If this part all checks out ok but still having problems,
then Im a bit unsure what to check next.
Need to think about it.

Send lots of pictures, it might prompt some thoughts on what to check next.
Others in the community might spot something too

Hi JAD

Thank you for your reply, since our last conversation, I made contact with Dimension Engineering regarding the problems with the sim.

I was told by Charleen the engineer at Dimension Engineering the reasons my sim failed first time round.

The first was the fusible rating for my motors was to low at 25Amp, as the motors can draw more than this under load, although the rated current is 22Amps, so now I am using 40Amp fuses, as the Sabertooth can supply up to 64Amps of current.

And I had a second problem that might have caused problems, and this was the Power clamping resistors, I was told that my two 5.Ohm 25 Watt were underrated and that I needed 2 x 3.3 Ohm 200 Watt resistors to dissipate the regenerative current, I have now implemented this on my power supply.

And I need to redesign my magnet placement on the motor arms as the arms are 8mm in diameter, the magnets are 5mm in diameter and are pressed in to the arms, this causes the magnetic field to be disturbed by the metal surrounding them.

I have included some pictures for reference.

And I will down load and save you video reference .

Again Thanks for your help

I found your build thread. Looks good.
Thats a nicely laid out case.

The discussion you had with Dimension Eng sounds about right.
Just curious if the fuse and resistor change meant you could get a tune to complete all OK? Was this the whole solution.

And was the original magnetic setup causing the limits to trigger early? like you descibed where it only traveled 10degs in the tune.

Hi JAD

I think the original magnetic setup did cause problems by giving false position information to the Kangaroo, and I haven’t done a second tune as yet, I need to redesign the magnet trigger as the ones I was using 5x3mm have a very large pull strength of 0.76Kg, so there is a possibility the magnet triggering the hall sensor before the sensor got to its position.

I will be replacing the magnets with a smaller magnet 3mm x 0.5 with a pull strength of 0.08 kg, I also found that you can trigger the sensor from left to right with different results, depending on the magnet polarity being north or south, it can be head on in one direction, and be one centimetre in the other then it will trigger and it can latch if the sensor goes beyond the magnet.

So today I will be 3D printing two magnet holders to place the new magnets in and set the polarities on the magnets and do a bench test.