I bought a mini RC car off Amazon this week and I can't wait to start putting it together tomorrow. Your videos help give me ideas on how to play around with the motors in the kit before I assemble the entire thing.
Advantage vs mini cheetah: The Moteus supports a motor temperature sensor (fw changes need though). I recall you mentioned you were disappointed in the mini cheetah's lack of temp sensor.
Nice video! As an idea for a future video, you could make some robot using ROS or maybe some kind of tutorial for beginners! It would be really interesting!
Nice controller. It would be great if they add support for optical encoder feedback (magnetic encoder can be left as an option). I made similar controller about 10 years ago using Microchip dsp/microcontroller dsPIC33FJ128MC802. I also used Field Oriented Control, but in outer control loop I used MRAC (model reference adaptive control) to cancel the noise (highest harmonics caused by low-cost brushless motor construction; AC servo motors don't have this noise). Optical encoder interface would increase positional accuracy. As example, I have DC brushless servo motor with 1024 (up to 4096) resolution and it can fit this controller, but what if we need to drive servo motors with higher resolution encoder (65536 pulses or 1,000,000 pulses).
Hi there, would the sound that the motor makes be the readout noise of the angle sensor? Would it be able to filter them out with a second order low pass filter?
It's an optical illusion as the arms cross the windings. Watch a point on the edge, it moves at constant speed, watch the arms and you see surging depending on the number of gaps visible.
BLDC controllers market is ramping up :) I found another project that might be interesting : github.com/byDagor/Janus-Controller . Meanwhile onmy end the odrive arrived so i'll continue with integration of bldc's into inmoov arm.
@@Skyentific for now i'm a bit stuck with the adaptation until i find a good fit - bldc and controller. testing different variants and my odrive just arrived . unfortunatelly from the moment i find something interesting until i order / receive it ... takes time. So ... slow progress for now until i gain the more knowledge . but i'll try to make short videos each week with what i did that week
Is it possible to precisely synchronize all actuators via CAN bus? For example, if robot arm draws a picture with a pen, there would be hundreds of motion commands per second for each actuator. And when command to one actuator is sent, controller starts sending command to the second actuator, but first actuator has already started moving, and trajectory of the robot arm becomes wrong. With stepper motors you could use single controller and STEP/DIR interface, which sends pulses for all motors and they are precisely synchronized.
Great question. I think you cannot really synchronise them. But you can control them with 1kHz speed. This means the the difference in the start time between different motors will be only 1ms. And I think this is almost simultaneously.
@@Skyentific Yes, but you also must check if all actuators finished their movements before sending new command, and, also check for errors to do emergency stop and quickly engage motor brakes if something goes wrong, which gives additional lag. ODrive is the only controller that supports STEP/DIR? Seems you don't like ODrive and prefer Mini Cheetah and some other controllers. Why so? What are ODrive's disadvantages in your opinion? Maybe you would make brushless controller comparison video?
@@TheNamelessOne12357 moteus controllers don't have a concept of a command being "done". You actually command a fixed velocity trajectory (that can optionally stop at a fixed position) for it to follow until the next command. Robots like the quad A1 use this to update many servos in synchrony with at least decent results: th-cam.com/video/aNhv88M6U0Q/w-d-xo.html
@Skyentific is right. 1kHz speed (1ms period) for servo motor position control is just fine. I had a chance to work on PA-10 Mitsubishi robotic arm (high class, 7-dof, 10kg payload, 0.1mm repeatability) and I found it's motion controller work with 10ms (100Hz) robot move command input. So, I'm close to believe that this motion controller use around 1kHz (1ms) speed for servo motor position control loop.
If only this was true... Indeed you can use FD capable CAN controllers with 2.0, but not at the same time, because Classical nodes will generate an error frame every time FD tries to send anything.
Мне конечно очень далеко да такого уровня. Сам построил несколько CNC router, 3д принтеров и лазер. Вот смотрю замечательные видео и думаю построить sentry paintball gun turret. Думаю воспользоваться вот такими моторами
Hi! Thanks for the mention and kind words!
You really deserve it!
I bought a mini RC car off Amazon this week and I can't wait to start putting it together tomorrow. Your videos help give me ideas on how to play around with the motors in the kit before I assemble the entire thing.
I saw this one before. Nice to see a review about it, thanks.
Very nice review! I just bought my second moteus, totally agree on the overall quality and support from mjbots.
Finally, mjbots and skyentific in one video.
Very good.
I've got mine last week and confirm it's awesome!
Excellent review, thanks from Colorado.
Excellent video! Thanks for making it, love your channel.
Thank you for your overview! I bought one Motheus kit after it. Good luck in your projects!
eagerly waiting for the affordable robot arm project.
Me too!
mee too. this cost 200$. i need 10. so to make a robot i need 2000 just to stat it. that is crazy. 2000 costs a whole car!
@@bambumbambu Look at Robotis dynamixel 12a. Depending on the application maybe they can work. I think a box of 6 is under 250$.
@@TheRealFrankWizza thank you. but dynamixel is 13.5W and this one is 500w
@@bambumbambu Well, then 200 dollars each is an affordable price.
Advantage vs mini cheetah: The Moteus supports a motor temperature sensor (fw changes need though). I recall you mentioned you were disappointed in the mini cheetah's lack of temp sensor.
Yes, exactly! You are right!
Nice video! As an idea for a future video, you could make some robot using ROS or maybe some kind of tutorial for beginners! It would be really interesting!
Thank you 👍 always interesting 😊 wishing you a Happy Fun Creative 2021 🎁
Thank you a lot!
awesome, was waiting for this one. go mjbots!
Thank you for sharing. 👍👍
Thank you for commenting!
@@Skyentific thank u for commenting on a comment now comment on my comment to your comment
Nice overview...might get one fr myself soon! :)
Great ingenuity !
Nice controller. It would be great if they add support for optical encoder feedback (magnetic encoder can be left as an option). I made similar controller about 10 years ago using Microchip dsp/microcontroller
dsPIC33FJ128MC802. I also used Field Oriented Control, but in outer control loop I used MRAC (model reference adaptive control) to cancel the noise (highest harmonics caused by low-cost brushless motor construction; AC servo motors don't have this noise). Optical encoder interface would increase positional accuracy. As example, I have DC brushless servo motor with 1024 (up to 4096) resolution and it can fit this controller, but what if we need to drive servo motors with higher resolution encoder (65536 pulses or 1,000,000 pulses).
Hi, Where did you get the MIT cheetah controller board from? Thanks.
watching this on a p!
Cool! :)
Hello sir! Is it possible to make whole Lego mindstorm kit with help of 3d printer?
Hi there, would the sound that the motor makes be the readout noise of the angle sensor? Would it be able to filter them out with a second order low pass filter?
I think all performance orientated STM32 has FD CAN. Question is if transceiver IC supports it, and you can modify software on STM32.
Could that problem of the dampening be related to the cogging of the motor, would a very coggy motor be a problem?
👍
Are you still focusing only on the content of the channel?
Is it possible to find a smaller Controller? I'm building a smaller Quadruped Robot with smaller brushless motors. Thank You!
At this point in your research, would you recommend the brushless motor and controller over stepper type motors for robotic applications?
06:51 Are my eyes fooling me or does the motor not turn at a constant speed? Is there a linearity issue with the magnetic sensor?
It's an optical illusion as the arms cross the windings. Watch a point on the edge, it moves at constant speed, watch the arms and you see surging depending on the number of gaps visible.
I will try yo investigate this. Just to check.
BLDC controllers market is ramping up :) I found another project that might be interesting : github.com/byDagor/Janus-Controller . Meanwhile onmy end the odrive arrived so i'll continue with integration of bldc's into inmoov arm.
Nice project. Thank you for link. Looking forward for inmoov progress!
@@Skyentific for now i'm a bit stuck with the adaptation until i find a good fit - bldc and controller. testing different variants and my odrive just arrived . unfortunatelly from the moment i find something interesting until i order / receive it ... takes time. So ... slow progress for now until i gain the more knowledge . but i'll try to make short videos each week with what i did that week
Hey, I'd be interested to hear your thoughts on the holotron vr exoskeleton videos that they've put out recently
How does this compare against an ODrive?
Is it possible to precisely synchronize all actuators via CAN bus? For example, if robot arm draws a picture with a pen, there would be hundreds of motion commands per second for each actuator. And when command to one actuator is sent, controller starts sending command to the second actuator, but first actuator has already started moving, and trajectory of the robot arm becomes wrong.
With stepper motors you could use single controller and STEP/DIR interface, which sends pulses for all motors and they are precisely synchronized.
Great question. I think you cannot really synchronise them. But you can control them with 1kHz speed. This means the the difference in the start time between different motors will be only 1ms. And I think this is almost simultaneously.
@@Skyentific Yes, but you also must check if all actuators finished their movements before sending new command, and, also check for errors to do emergency stop and quickly engage motor brakes if something goes wrong, which gives additional lag.
ODrive is the only controller that supports STEP/DIR? Seems you don't like ODrive and prefer Mini Cheetah and some other controllers. Why so? What are ODrive's disadvantages in your opinion? Maybe you would make brushless controller comparison video?
@@TheNamelessOne12357 moteus controllers don't have a concept of a command being "done". You actually command a fixed velocity trajectory (that can optionally stop at a fixed position) for it to follow until the next command. Robots like the quad A1 use this to update many servos in synchrony with at least decent results: th-cam.com/video/aNhv88M6U0Q/w-d-xo.html
@Skyentific is right. 1kHz speed (1ms period) for servo motor position control is just fine. I had a chance to work on PA-10 Mitsubishi robotic arm (high class, 7-dof, 10kg payload, 0.1mm repeatability) and I found it's motion controller work with 10ms (100Hz) robot move command input. So, I'm close to believe that this motion controller use around 1kHz (1ms) speed for servo motor position control loop.
Do you like these more than ODrive?
You should check out simpleFOC
Doing this right now! :) One big disadvantage of simpleFOC: absence of the torque control. They plan to implement it.
спасибо за обзор !
Nice accent ;) are you native russian speaking?)
Can you please compare it with Odrive
I will try to do this. Also, I want to compare the basic performance with MIT controller.
@@Skyentific Thank you, there is also a similar type of motor controller called Tinymovr that have similar functionality.
For me it seems like ODrive has best performance.
What is the precision of the position control? Is it comparable to a stepper motor?
What is the motor called?
CAN FD compatible with CAN 2.0
If only this was true... Indeed you can use FD capable CAN controllers with 2.0, but not at the same time, because Classical nodes will generate an error frame every time FD tries to send anything.
a comment! ;)
Una impresora 3D con motores sin escobillas para varear
Спасибо за видео!
Spasibo za comentarij!
Looks like those bots are taking the first comment in a channel...check below!
Офигеть, очень интересно.
Мне конечно очень далеко да такого уровня. Сам построил несколько CNC router, 3д принтеров и лазер. Вот смотрю замечательные видео и думаю построить sentry paintball gun turret. Думаю воспользоваться вот такими моторами
Мля... А чего по русски не говоришь? Лучше получится же! Да и акцента не будет ;-)
You're is russian?
Русский что ли ))? Странный акцент.