So my company got two and one works and one only does circles due to one wheel is not getting any power
the wiring looks fine anyone else have this problem.
Sorry you are experiencing this. All the robots had a live driving test before leaving the factory. Please check the following.
- With the robot powered up, for the wheel that seems to have a problem if you try to move it manually does it resist or does it free-wheel?
- If it resists what is the nature of the resistance - is it smooth or is it chunky?
- If it free wheels with the robot on can you look at the odometery topic to see if as you move the wheel the odometery also changes.
- Thank you for checking the wiring, I would double check all the wiring starting from the axle of the wheel outwards, you should see an inline socket and plug make sure its pushed firmly together (its very tight) , then check both wire to board connectors (there are 2 one for odometery sensors and one for power). Make sure the pigtails of the cables are in them and that they are tightly screwed down.
OK let me come and see it in person if possible - reach out to me on dc@ubiquityrobotics.com to arrange. We must get to root cause for this issue.
oops type should be dc@ubiquityrobotics.com
I am having the same issue. One wheel smoothly resists and has power and the other free-wheels but gets no power. I have checked all of the connections to the best of my ability. All the pigtails look good and I disconnected and reconnected the inline socket and am confident it is connected properly. I’m new to this, so I don’t yet know how to check the odometery.
This reply is to @rprinc50 Please of course first verify your pc board did not burn up like the customer who started this post. We have done many reliability fixes since that customer had a burned up board. So first of all verify you did not have such a horrible event happen to your board.
Assuming it was not a horrible destructive event, please refer to this page just to check a few things. Let us check for if the motor cable is not fully inserted (EASY to have happen if you have been removing them).
See 2/3 of the way down on this page: https://learn.ubiquityrobotics.com/unboxing
Now on to your specific question which will verify correct wheel encoder operations.
Start with the red switch pushed in so that the bright RED led on the switch board is off.
Reboot the robot using ‘sudo shutdown -r now’
Run this command on an SSH window: rosrun tf tf_echo odom base_link
Raise the robot up with a small block of wood or something so the 2 large drive wheels are not touching the floor.
You will see a set of numbers for ‘Translation’ and we are going to focus on the next set of numbers under Rotation. Specifically look at Rotation in RPY (degree). The 3rd number on that line is rotation about the Z axis (up).
Try this FIRST on the wheel that seems to work
Now rotate manually one wheel 1/2 revolution and the 3rd number on a working wheel encoder unit will show in the area of 50 degrees of rotation (very roughly).
Rotate that wheel back to where it was using reverse direction of rotation and the 3rd number is supposed to go back near 0.
Next do this same test on the wheel that does not move and see if it’s encoders are operating properly in the same way with 1/2 wheel rotation like above.
Let me know the results and we will take it from there.
The wheel encoder wires are the 6 thin wires you see at the end of the motor cable going to a 6-way terminal block.
Thanks for the quick help! Nope, no fire problems, fortunately. The right wheel is registering rotation. The left wheel is not. Here’s a photo verifying the motor cable is properly connected
Ah that is a good clue. So you do not get the rotation on the test I went over for the left wheel. Now on to next step of debug. when we say Left wheel that is like it is described in a car so if you sat in the magni and drove it forward we have a front wheel drive situation so the left wheel is opposite side of where the switch board is. Let’s agree on that first.
Do ALL the things below with large fat red power to battery totally disconnected from battery.
We want to verify that each thin wires go into the 6-pin terminal strip and we will discuss all this from the back of the board. White wire is the only one that really does not matter, any other wire being not connected leads to faulty encoder readback and could explain you not seeing odom change.
From the top of this jack the wires are white, red, black, yellow, green and blue at the bottom. If that is true we want to be sure that each wire is stripped and going into the jack with the screw tight. Try to see if you can see if each wire is stripped. If a wire was not stripped then the screw tightens but only on platic insulation.
Worse case is one at a time loosen from the top the red location and ease it out just a little to see if wire is stripped. If you pull the wire out all the way it can be slightly tricky to get it back in so ease it out just 2mm max to see if wire is stripped then put back in and tighten it. If any wire simply pulls right out without even loosening the screw, that is the problem.
So left thumb and index finger grasp the red wire and anchor your palm against chassis, loosen the 2nd screw down with screwdriver in right hand and pull wire out just 2mm or less to see if it is stripped. Push it back in and tighten. Red and black are 5V and ground so those are super important. Then yellow, green and blue are the 3 encoder phases as these are 3-phase brushless motor encoders.
After we verify all encoder wires are ok on left side another test is to plug the right wheel into the cable to the left side. if the odom can change in this situation there is a faulty wheel.
We want to decide if the MCB has problem, the cable wires are the problem, or the wheel itself is the problem.
I am attaching a picture showing the cable wires as they should be and as I have described. We would be talking about the left cable that goes to the left wheel.
Sorry this is dragging you into this but we might be able to fix this with minimal delay if we go through these things.
Thanks! So to get on the same page with you, it’s the left wheel that works and the right one that freewheels. I checked all the wires for the right wheel and they all seem secure and are stripped. So I swapped cables and the left wheel continues to work and reads on the odom but the right wheel still freewheels and does not read on the odom, so it would appear the wiring is not the problem. I presume then it is an issue with the wheel itself? I can send additional pictures/video if that would help.
You are then saying the problem follows the physical wheel itself. Let me re-phrase how I interprit your report.
You have a wheel that lets say had a piece of tape on it that said ‘L’ (it did not but let call that wheel ‘L’. The other wheel had tape on it saying ‘R’.
Also you had a cable from the MCB to wheel ‘L’ that lets say had ‘L’ on it and the cable that went from right side of mcb to wheel ‘R’ had tape with ‘R’ on it.
I am not sure on your ‘I swapped cables’.
The first test would be to plug wheel ‘R’ into cable ‘L’ and see if rotating wheel ‘R’ shows odom changes. I am not sure if you did this or not.
What you ‘may’ have said is you swapped cable ‘R’ and re-screwed it all so that wheel ‘L’ goes through cable ‘R’ and cable ‘R’ goes into the left side of the board.
I think you said that this worked but this does not tell is if wheel ‘R’ was the real problem, it only tells us that cable ‘R’ is ok.
So clarify, sorry but we have MCB, cables and wheels.
I would have liked to know if wheel ‘R’ in cable ‘L’ going to MCB board side ‘L’ was ok or not.
Likewise if wheel ‘L’ into cable ‘R’ into the ‘R’ side of the mcb was now bad or good
The
Sorry for the confusion. Wheel ‘R’ in cable ‘L’ going to MCB board side ‘L’ was not okay. It continued to freewheel and did not show odom changes. Wheel ‘L’ in cable ‘R’ going into MCB board side ‘R’ continued to work and show odom changes.
Ok, that is likely the first case of a wheel going bad unless it was bad from the time you had it.
Please send an email to support@ubiquityrobotics.com.
In the subject please say Wheel failure. ATTN Mark
State what your location is in the world and in the USA if that is the case.
in the email, please tell me if it EVER worked? We have a 100% fully assembled test of each Magni so if you got it bad that means we have a quality control issue. If you got it and it worked THEN failed that means we have to mark this as a reliability issue.
Can you also state in the email when you received the Magni and so I can track it can you take a picture of the label on the front of the MCB board and note the rev of the MCB board which should be in bright white printing along the left side of the board.
Thanks and sorry you have this failure. We will get this sorted out.
I await the email,
Mark
Thanks Mark! I will send that email. For the record here, I was never able to get it to work.