Society of Robots - Robot Forum

Electronics => Electronics => Topic started by: TaoBaiBai on August 25, 2012, 08:57:03 AM

Title: Motors lose control
Post by: TaoBaiBai on August 25, 2012, 08:57:03 AM
I m using sabertooth 2x25 motor controller to drive 2 24v Maxon RE40 motor. When I use RC to output a perfect RC waveform to S1 and S2 ports in sabertooth,it always lose control and  when I connect 3 0.1uF capacitors to the motor terminals and case,it is sometime lose control and When I connect 2 0.1uF Cap to S1 and S2.1 out of 10 times it  still loses control even I off my RC.But when I off the RC it will stop 2 seconds then run  again.I really cannot figure out the reason,EMF and noise should be filtered out by Caps.I doubt that is there anything wrong with the motor controller since the Input is correct and output is wrong?
Title: Re: Motors lose control
Post by: Soeren on August 25, 2012, 12:43:51 PM
Hi,

When I use RC to output a perfect RC waveform to S1 and S2 ports in sabertooth,it always lose control
How do you make sure it's a "perfect RC waveform"?
When and how does it loose control andin what way?


and  when I connect 3 0.1uF capacitors to the motor terminals and case,it is sometime lose control
How often?
Does the behaviour change during the day?


and When I connect 2 0.1uF Cap to S1 and S2.1 out of 10 times it  still loses control even I off my RC.
When you turn off your transmitter, what exactly happens?

Try pull down resistors (say 10k for starters) on the inputs instead and tell us what happens.


But when I off the RC it will stop 2 seconds then run  again.
Definitly add pull downs.


I really cannot figure out the reason,EMF and noise should be filtered out by Caps.
A few caps may not be enough! Did you twist the wiring pairs?


I doubt that is there anything wrong with the motor controller since the Input is correct and output is wrong?
How do you check the correctness of the inputs?
Do you have access to an oscilloscope?
Title: Re: Motors lose control
Post by: TaoBaiBai on August 25, 2012, 07:33:18 PM
Hi,

How do you make sure it's a "perfect RC waveform"?
When and how does it loose control andin what way?
I used an oscilloscope the check and see the frequency and duty cycle are right.

Every time I on the power the motors stop be4 I operate the transmitter, after I use transmitter to make the motor run I cannot stop it by stopping the transmitter,most likely the motor will run forward and i have to output a backward signal to stop the motor.This seems the motors are still under control but the be movement dosen't match the signal.



How often?
Does the behaviour change during the day?
I have tried many ways to improve the  behavior.Without Caps 90% it will free run without control.Adding Cap to motors terminals  and case,it will be 50% and Adding Caps to S1 and S2(transmitter signal) ,it will be 10%.

Definitly add pull downs.
Since I have off the transmitter the S1 and S2 cannot pick up the eletcrostatic  I think and pull down resistor will not work for this case.
I only see the waveform when I on the transmitter,I will try  and see the waveform when I off it and try pull down first.
I will let you know.


A few caps may not be enough! Did you twist the wiring pairs?
How to connect more than 3 Caps?Parallel the case and terminals? Twist the 5V and ground?

Thanks a lot.
Title: Re: Motors lose control
Post by: Soeren on August 26, 2012, 12:49:19 PM
Hi,

I used an oscilloscope the check and see the frequency and duty cycle are right.
Dioes the input signals still look OK when it misbehave?


[...] after I use transmitter to make the motor run I cannot stop it by stopping the transmitter,most likely the motor will run forward and i have to output a backward signal to stop the motor.This seems the motors are still under control but the be movement dosen't match the signal.
Not knowing the internals of the driver, my best guess here is, besides the lack of pull downs, that motor noise swamps out the control signal in some way.


I have tried many ways to improve the  behavior.Without Caps 90% it will free run without control.Adding Cap to motors terminals  and case,it will be 50% and Adding Caps to S1 and S2(transmitter signal) ,it will be 10%.
Did you by any chance scope out the supply wires to the motor (and the controller)?
Which kind of caps did you use? (Ceramic caps are the best, due to their low ESR)
It sure sounds like you got rid of some of the noise, but a fast 'scope will tell you exactly how the caps affect things and immediately reveal the impact of larger caps (or parallel mounted extras).


Since I have off the transmitter the S1 and S2 cannot pick up the eletcrostatic  I think and pull down resistor will not work for this case.
It's easy to test, so please humour me - you'll probably be surprised :)


I only see the waveform when I on the transmitter,
Yes and when you turn it off, the inputs are more suceptible to picking up noise.


I will try  and see the waveform when I off it and try pull down first.
I will let you know.
Good, try adding parallel caps to the ones you already have as well.


A few caps may not be enough! Did you twist the wiring pairs?
How to connect more than 3 Caps?Parallel the case and terminals? Twist the 5V and ground?
I wasn't referring to those, but do double them up. I was talking about the entire circuit needing more than just caps (and more of those as well). Scope around where wires start and end and deal with noise where it is generated as much as possible. Keep wires twisted and impedances low - what wire gauge do you use for your controller and motors?
Title: Re: Motors lose control
Post by: TaoBaiBai on August 27, 2012, 02:54:22 PM
Hi,

I scope the sabertooth 2x25 motor controller, it can recreate the sigal.when I off thetransmiiter, it will continue the last singal that I sent to it and pull down resistor dosent work because the circuit not pick up the signal,but produce the signal,this explain why I off the transmitter,  the motor still run.When I on the transmitter and make the motor run then stop, no matter how many Caps I connect to motor it still continue turn a little bit.I scope the controller output and input,the output has a spike and the motor will turn a bit, and every time i make a change over the controller and motor ,the output will mismatch the input.The catalog say the controller can find the midpoint and range of the transmitter.I suspect it is because the circuit inside the controller make a wrong detection and output a perfect forward waveform with the stop input signal.I have add the .1uF and 0.01uF to the output 5V and GND in the controller,it got improve but it still has a  low chance got spike and I have to on and off the motor controller to let it find the range and mid point of the transmitter.Do you know how to improve the situation or disable the signal recreation system?
By the way, what do you mean by twisting the wire,what wire and which part of wire should I twist.The wire I m using is a  single core normal jumper wire,it's diameter looks like 0.6-0.8mm.I have also tried  2-3 mm multicore wire and  beside the increase in curent, the hehavior is still the same.

Thaks
Title: Re: Motors lose control
Post by: Soeren on August 27, 2012, 05:16:03 PM
Hi,

[...] Do you know how to improve the situation or disable the signal recreation system?
Yes, it seems to be an RTFM issue ;)

Quote from: Sabretooth User Manual p13 first para
If switch 6 is in the UP position, then the Sabertooth is in standard R/C mode. This mode is designed to be used with a hobby-style transmitter and receiver. It automatically
calibrates the control center and endpoints to maximize stick usage. It also enables a Timeout Failsafe, which will shut down the motors if the Sabertooth stops receiving correct signals from the receiver.
I 'd guess that your switch 6 is in the DOWN position

Quote from: Sabretooth User Manual p13 second para
If switch 6 is set in the DOWN position, then Microcontroller mode is enabled. This disables the Timeout Failsafe and auto-calibration. This means that the Sabertooth will continue to drive the motor according to the last command until another command is given. If the control link is possible unreliable – like a radio - then this can be dangerous due to the robot not stopping.

Keep the caps on the motors anyway, or you may get other problems along the way.


By the way, what do you mean by twisting the wire,what wire and which part of wire should I twist.
All wire pairs, like the 2 wires to the motor, or the wires from the battery to the controller. Twisting them (3..4 turns per inch) together will dampen common mode signals and should be done at all times, whether you have problems or not as it protects fairly well from external signals as well as from radiating.


The wire I m using is a  single core normal jumper wire,it's diameter looks like 0.6-0.8mm.I have also tried  2-3 mm multicore wire and  beside the increase in curent, the hehavior is still the same.
Jumper wire won't do, so keep to the thicker wire - and now you know what the problem is and with a little luck, you read the entire manual, to prepare for the next time it does something different from what you want ;D
Title: Re: Motors lose control
Post by: TaoBaiBai on August 27, 2012, 07:18:05 PM
Hi,

Thanks,but the problem is no matter the switch 6 is up or down it still cannot off signal recreation...I think i should forget about it and just off the robot be4 off the transmitter.Dose the two wires to motor mean power and GND twist together and do I need to twist the signal wire?

Regards