Reivax Posted July 27, 2016 Share Posted July 27, 2016 20 minutes ago, EricGhost said: Sorry if I do not believe it, when you switch the wheel on the gyro is stabilized and no acceleration is involved Huh, no. You cannot separate stabilization and acceleration. Stabilization is the goal, and acceleration is how this goal is achieved. There is no stabilization if there is no acceleration. Link to comment Share on other sites More sharing options...
EricGhost Posted July 27, 2016 Share Posted July 27, 2016 31 minutes ago, Reivax said: Huh, no. You cannot separate stabilization and acceleration. Stabilization is the goal, and acceleration is how this goal is achieved. There is no stabilization if there is no acceleration. Strange because I switch on my wheel on static conditions and it does not fold, I leave it parked against the wall switched on and it does not fold, it does not even accelerate and move around the room to be stabilized it just stay still and stabilized !! Link to comment Share on other sites More sharing options...
HEC Posted July 27, 2016 Share Posted July 27, 2016 15 minutes ago, EricGhost said: Strange because I switch on my wheel on static conditions and it does not fold, I leave it parked against the wall switched on and it does not fold, it does not even accelerate and move around the room to be stabilized it just stay still and stabilized !! Of course - because it's not moving but motor is keeping it upright. It's balancing (only) the weight of the empty wheel but as even empty wheel is still "top heavy" it uses some power to do so (for example on wheels with mobile app showing the current / W consumption like KS you can see that). Try to switch on the wheel while it's slightly tilted to front or rear to see what happens EDIT: On my KS-16 if I switch the wheel while standing in the stand (or leaning against the wall or lamppost) I can see in battery information screen (or Data logger) current jumping around 0.1 - 0.2A and consumption between 0.5 up to 10 - 11W as wheel is balancing itself while stationary / not moving. Link to comment Share on other sites More sharing options...
Reivax Posted July 27, 2016 Share Posted July 27, 2016 14 minutes ago, EricGhost said: Strange because I switch on my wheel on static conditions and it does not fold, I leave it parked against the wall switched on and it does not fold, it does not even accelerate and move around the room to be stabilized it just stay still and stabilized !! It can be still for 2 reasons: it is perfectly balanced just by itself, and then it would also be still with no power at all but mostly : it is still because of the friction against the wall. But the motherboard sends power to the engine to balance the wheel exactly the same way it does when you lean forward, except much less power is needed. Link to comment Share on other sites More sharing options...
esaj Posted July 27, 2016 Share Posted July 27, 2016 Everybody's making very good points here. I'm not personally a fan of tilt-back (well, if you can set it to 30km/h, then it's fine by me ), but shutting down the motor entirely after certain speed seems pretty stupid in my opinion. Sure, it probably cannot run any faster, as the back-emf has reached high enough value that the battery cannot push the voltage higher, but it could at least keep on running at max speed, which should give the rider a (small) chance to correct the situation when noticing that the pedals start to tilt forward (because it cannot accelerate any more). Then again, maybe it's a safety issue, and the manufacturers want to make sure the wheel does not keep running by itself in case the rider falls off? I made this self-balancing robot earlier this year, and it uses the basic "inverted pendulum" -principle, it keeps balance by riding two motors back and forth, the closed loop is formed by the MPU-6050 gyro/accelerometer, and a PID-control in an Arduino Uno-board: Here the PID-values are near to perfect, but you can clearly see how it still accelerates faster back and forth to keep balance. After getting the PID-values "correct": (the high pitch noise is the PWM-frequency resonating in motors or something, I used lower frequency here). It is still doing very, very small acceleration back and forth to keep balance. So I'm fairly sure that wheels even at standstill do miniscule accelerations (you can actually usually hear a slight "whirr" or such when the wheel is standstill), but they're so small that they're almost unnoticeable. Link to comment Share on other sites More sharing options...
Reivax Posted July 27, 2016 Share Posted July 27, 2016 4 minutes ago, esaj said: So I'm fairly sure that wheels even at standstill do miniscule accelerations (you can actually usually hear a slight "whirr" or such when the wheel is standstill), but they're so small that they're almost unnoticeable. Yes, that's exactly what I meant by not differentiating stabilization and acceleration, thanks for the illustration And it's a cool hack too Link to comment Share on other sites More sharing options...
HEC Posted July 27, 2016 Share Posted July 27, 2016 5 minutes ago, esaj said: So I'm fairly sure that wheels even at standstill do miniscule accelerations (you can actually usually hear a slight "whirr" or such when the wheel is standstill), but they're so small that they're almost unnoticeable. Yep - you can hear the KS clearly "whinge" while standing "still" leaned against the wall or in stand. You can also see it clearly on LED lights when your battery level is just at the edge threshold between two display states as it nicely switches for example from four green LEDs to 3 yellow ones and back ... Poor dog - did he not liked the PWM noise? Link to comment Share on other sites More sharing options...
Keith Posted July 27, 2016 Share Posted July 27, 2016 32 minutes ago, Reivax said: Huh, no. You cannot separate stabilization and acceleration. Stabilization is the goal, and acceleration is how this goal is achieved. There is no stabilization if there is no acceleration. While I would agree that @EricGhost clearly does not understand either what is happening in the wheel or even the basic physics of moment arms and action/reaction, it might be the term "acceleration" that is confusing the issue. it is not acceleration that keeps the rider balanced it is torque. When a rider leans forward he exerts an unbalancing force on the pivot point (axle) of the wheel, that causes the body of the EUC to start to tilt forward, the sensors see that tilt and the control board compensate for that by increasing the torque applied to the wheel, thus keeping the rider upright. The byproduct of that torque is acceleration. For example, if the rider pushes against the peddle with (say) a force of one foot pound (Newton meter or whatever) then the wheel has to produce a reaction torque of one foot pound to counter it and keep the EUC balanced. That torque is can only be generated by putting power into the wheels motor which will then start to accelerate the wheel and that acceleration will (absolutely must ) remain constant all the time the rider exerts that same force by leaning forward. As motor speed increases maximum torque decreases. At the absolute maximum speed (more accurately RPM) of the wheel torque is absolutely and completely zero- that is basic electric motor physics. So some point before that the EUC has to try and discourage the user from continuing to lean forward. It has only two possible methods of doing that: 1) beep like mad to tell you you are about to embark on a suicide mission. 2) tilt back to make you feel uncomfortable enough you stop trying to lean forward. The problem with 2) is that the only way to tilt back is to apply even more torque so it lifts the front of the peddle even though the rider is pushing it down, that means the act of tilting back makes the wheel accelerate even harder and reach the not enough torque to hold up the rider point even quicker. As for the idea there is no torque when the wheel is standing still, that is just nonsense. Ifyou prevent the wheel moving and push down on the front of the peddles the peddles feel rigid but only because the wheel is applying an equal and opposite torque to counter that push. If you prevent the wheel moving it will not accelerate because the torque that is opposing you is pushing against the wheel which is being prevented from moving. Push down hard enough and you will be unable to prevent the wheel moving either it will wheelspin or move and start to accelerate. Link to comment Share on other sites More sharing options...
EricGhost Posted July 27, 2016 Share Posted July 27, 2016 Yes of course there is a torque involved to balance the cog offset from the wheel center of rotation which in the end create a torque=Weight*COG offset. And torque is related to angular acceleration like force to linear acceleration F=MA, but those accelerations are related just to the stabilization loop for the standstill!! But there is for sure another loop that manage acceleration to move forward or backward !!! I do not say that the wheel will not move I say it's response/performance goes down and you'll feel that you need to push more fwd to get the same acceleration and then you can decide to increase your fwd and smash your face, stay still and get less acceleration and speed or slow down by moving a bit backward, which is much better than feeling speed going up constantely or progressively and then : beep, beep-beep ,... crash Link to comment Share on other sites More sharing options...
esaj Posted July 27, 2016 Share Posted July 27, 2016 5 minutes ago, Keith said: As for the idea there is no torque when the wheel is standing still, that is just nonsense. Ifyou prevent the wheel moving and push down on the front of the peddles the peddles feel rigid but only because the wheel is applying an equal and opposite torque to counter that push. If you prevent the wheel moving it will not accelerate because the torque that is opposing you is pushing against the wheel which is being prevented from moving. Push down hard enough and you will be unable to prevent the wheel moving either it will wheelspin or move and start to accelerate. Or lean it front (or back) against a wall so it cannot move or get back up, the wheel will try to apply torque (accelerate) to get back to "straight up"-position, but cannot, and if the firmware logics aren't smart enough to give up, usually it'll fry a mosfet or two, or the motor A non-turning motor is kinda like a short-circuit... Link to comment Share on other sites More sharing options...
Keith Posted July 27, 2016 Share Posted July 27, 2016 15 minutes ago, EricGhost said: Yes of course there is a torque involved to balance the cog offset from the wheel center of rotation which in the end create a torque=Weight*COG offset. And torque is related to angular acceleration like force to linear acceleration F=MA, but those accelerations are related just to the stabilization loop for the standstill!! But there is for sure another loop that manage acceleration to move forward or backward !!! No that is nonsense, what is your wheel fitted with magic beans. You only have one motive part in an EUC - the actual wheel motor. That is the only thing you have to stabilise the wheel and the only thing it can do is exert torque by applying power in a clockwise or anticlockwise direction. Go on you tube and take a look at rider balancing on a peddled unicycle, he jigs the peddles backwards and forwards to stand still. To move forward he leans forward and then peddles like mad to keep the wheel under him. There is not/ cannot be a secondary loop controlling acceleration, leaning forward means the wheel has to apply the torque to stay underneath you, acceleration is nothing more than a convenient byproduct of that stabilisation. The only way to have acceleration and stabilisation seperate in a single axle device like an EUC would be to have a bloody great weight below the CofG that moved back and forward to balance the rider. If you think that some other "magic" force is at work why on earth doesn't it also balance the wheel to the left and right when you are standing still as well? Link to comment Share on other sites More sharing options...
esaj Posted July 27, 2016 Share Posted July 27, 2016 I wonder if @EricGhost meant that there could be multiple different calculations done to come to a single value that is used for the motor control. Like one PID-loop (for example, it could be some other control scheme too) to calculate the wanted motor power for stabilization purposes (pitch-feedback), another using a current-value feedback to control torque (current-feedback), tilt-back control (ie. overshooting the correction, probably current/speed/voltage-feedback) and yet another for something like allowing the pedals to tilt forward/backward a little when making a slow-speed turn (roll- and speed-feedback), and then all this is smashed together to come to a one final value that is used to control the PWM for the motor? 43 minutes ago, HEC said: Poor dog - did he not liked the PWM noise? Yeah, but what she really hates is anything that makes a beeping noise... Link to comment Share on other sites More sharing options...
EricGhost Posted July 27, 2016 Share Posted July 27, 2016 12 minutes ago, Keith said: No that is nonsense, what is your wheel fitted with magic beans. You only have one motive part in an EUC - the actual wheel motor. That is the only thing you have to stabilise the wheel and the only thing it can do is exert torque by applying power in a clockwise or anticlockwise direction. Go on you tube and take a look at rider balancing on a peddled unicycle, he jigs the peddles backwards and forwards to stand still. To move forward he leans forward and then peddles like mad to keep the wheel under him. There is not/ cannot be a secondary loop controlling acceleration, leaning forward means the wheel has to apply the torque to stay underneath you, acceleration is nothing more than a convenient byproduct of that stabilisation. The only way to have acceleration and stabilisation seperate in a single axle device like an EUC would be to have a bloody great weight below the CofG that moved back and forward to balance the rider. If you think that some other "magic" force is at work why on earth doesn't it also balance the wheel to the left and right when you are standing still as well? "Another loop" was a wrong choice of words, What I was meaning is that there is a set of parameters controlling how much accelerate related to the COG offset, how fast to respond,.. and so on ... those parameters can change according to the wheel capability to provide a required performance and the user can feel the downgrade and manage the situation Link to comment Share on other sites More sharing options...
EricGhost Posted July 27, 2016 Share Posted July 27, 2016 1 hour ago, esaj said: I wonder if @EricGhost meant that there could be multiple different calculations done to come to a single value that is used for the motor control. Like one PID-loop (for example, it could be some other control scheme too) to calculate the wanted motor power for stabilization purposes (pitch-feedback), another using a current-value feedback to control torque (current-feedback), tilt-back control (ie. overshooting the correction, probably current/speed/voltage-feedback) and yet another for something like allowing the pedals to tilt forward/backward a little when making a slow-speed turn (roll- and speed-feedback), and then all this is smashed together to come to a one final value that is used to control the PWM for the motor? Sorry but English is not my mother tongue so maybe I did not send the proper message, I'll try again. While the physics for all the wheels is the same I do not think all the wheels respond to the pitch fwd o bwd in the same way, for example the KS16 itself has the 3 bicycle modes : learning,cycling,play; which means three different behaviours. So if the system has not enough energy to substain its most brilliant behaviour it shall downgrade its performances runtime, then you'll feel that you're not accelerating like before so what do you do? you pitch more forward? then you get the proper face plant you deserve. Link to comment Share on other sites More sharing options...
EUCMania Posted July 27, 2016 Share Posted July 27, 2016 EricGhost suggested to give less torque or acceleration than the rider pushes for, such as if the rider wants 2g of forward force, the machine only gives 1g. The problem with that is the rider tells the machine the force he wants by leaning body forward. If the machine gives less than 2g, then the rider's body and pedal will automatically leaning forward more, creating more demand of forward force from the machine. This is called "pedal softening". Once the rider sense this softening, he may want to reduce his body's forward leaning. However, to do this, he has to push more force on the toe, the front part of the feet. This further pushes the pedal forward, asking the machine to provide more forward force. Once you get into this vicious cycle, you have to jump off or faceplant or both. So, the machine absolutely must provide the force you ask for to balance. Ninebot used to try to give riders a feed back by vibrating the wheel when close to the limit. But it causes more faceplants. Ninebot abandoned the vibrating in later firmware upgrades. Link to comment Share on other sites More sharing options...
US69 Posted July 27, 2016 Share Posted July 27, 2016 2 hours ago, EricGhost said: Sorry if I do not believe it, when you switch the wheel on the gyro is stabilized and no acceleration is involved But let's say it MUST accelerate but how much? this is the issue here ! Moving the COG fwd or bwd is the command for acceleration and by consequence speed variation but stabilization is always on, for sure with the same quick fwd cog on different wheels you'll get different accelerations, or not, but the same stabilization and if a wheel gives you half an acceleration of another wheel you'll not fall down because it just half accelerate. What I'm saying is that is possible to correlate system status/available energy with system performance, in almost real time (of course if you jump from the wheel front first nothing to do) so that you can perceive the limits you're pushing, which is much better than going after my crazy command till the last moment and then shut down or throw me away A simple temptative is the today KS16 of reducing automatically the max speed based on Voltage , this is a simple proportional math, but today it is possible to do much more. I accept more that there could be a costing issue because industrial can mean something like 10000$ of card Controller Puuuh... OK...the only Problem in your thinking is that all self balancing unicycles work from leaning forward or backward.... Lets say you are at 30kmh max....and now you want to lean Forward and you want that the wheel just stays on 30km??? The hull and so the pedals are "leaning" against the wheel! What you want is to put energy into the System without consequences..... on a System that is working on putting energy into it ...... Would be fairly difficult to Programm such a board as it is only in the mind of the Driver what he wants....will he Speed up? will he break? will he stay at his Speed? But congratulations: Since nearly 15 Years of self balancing vehicles you found the easy solution....i wonder why this Multi.million Dollar Company segway, ninebot or inmotion didn't found it and they all work with bloody stupid tiltback on max Speed :-) Link to comment Share on other sites More sharing options...
US69 Posted July 27, 2016 Share Posted July 27, 2016 1 hour ago, Keith said: While I would agree that @EricGhost clearly does not understand either what is happening in the wheel or even the basic physics of moment arms and action/reaction, it might be the term "acceleration" that is confusing the issue. it is not acceleration that keeps the rider balanced it is torque. When a rider leans forward he exerts an unbalancing force on the pivot point (axle) of the wheel, that causes the body of the EUC to start to tilt forward, the sensors see that tilt and the control board compensate for that by increasing the torque applied to the wheel, thus keeping the rider upright. The byproduct of that torque is acceleration. For example, if the rider pushes against the peddle with (say) a force of one foot pound (Newton meter or whatever) then the wheel has to produce a reaction torque of one foot pound to counter it and keep the EUC balanced. That torque is can only be generated by putting power into the wheels motor which will then start to accelerate the wheel and that acceleration will (absolutely must ) remain constant all the time the rider exerts that same force by leaning forward. As motor speed increases maximum torque decreases. At the absolute maximum speed (more accurately RPM) of the wheel torque is absolutely and completely zero- that is basic electric motor physics. So some point before that the EUC has to try and discourage the user from continuing to lean forward. It has only two possible methods of doing that: 1) beep like mad to tell you you are about to embark on a suicide mission. 2) tilt back to make you feel uncomfortable enough you stop trying to lean forward. The problem with 2) is that the only way to tilt back is to apply even more torque so it lifts the front of the peddle even though the rider is pushing it down, that means the act of tilting back makes the wheel accelerate even harder and reach the not enough torque to hold up the rider point even quicker. As for the idea there is no torque when the wheel is standing still, that is just nonsense. Ifyou prevent the wheel moving and push down on the front of the peddles the peddles feel rigid but only because the wheel is applying an equal and opposite torque to counter that push. If you prevent the wheel moving it will not accelerate because the torque that is opposing you is pushing against the wheel which is being prevented from moving. Push down hard enough and you will be unable to prevent the wheel moving either it will wheelspin or move and start to accelerate. Nice to have somebody with decent english to explain it better than i can!!!! Thanx! Link to comment Share on other sites More sharing options...
EricGhost Posted July 27, 2016 Share Posted July 27, 2016 1 hour ago, KingSong69 said: But congratulations: Since nearly 15 Years of self balancing vehicles you found the easy solution....i wonder why this Multi.million Dollar Company segway, ninebot or inmotion didn't found it and they all work with bloody stupid tiltback on max Speed :-) Never give up, try harder and you'll get it, and if people safety is involved the more the reason to try harder :-) history is plenty of "wrong ideas" that changed the world, it's always worth a try :-) Focus on the solution not the problem ..... Still I think I did not exposed my point properly, see my latest posts ..... Wow 2500 views, we have an issue here, something interesting has been arised, isn't it Link to comment Share on other sites More sharing options...
Reivax Posted July 28, 2016 Share Posted July 28, 2016 12 hours ago, EricGhost said: Never give up, try harder and you'll get it, and if people safety is involved the more the reason to try harder :-) history is plenty of "wrong ideas" that changed the world, it's always worth a try :-) Focus on the solution not the problem ..... Still I think I did not exposed my point properly, see my latest posts ..... Wow 2500 views, we have an issue here, something interesting has been arised, isn't it When I was talking about blind faith... You say it's science, but you do not understand how the balancing works. So, yes, it is science, and science also explains why some things can't be done. Yet, even when people explain to you how it's working, you keep knowing better, so this is going nowhere. Probably some day you will come up with an engine made only with magnets that will require no other source of power to run forever. Link to comment Share on other sites More sharing options...
Frode Posted July 28, 2016 Share Posted July 28, 2016 On 27.7.2016 at 4:33 PM, esaj said: usually it'll fry a mosfet or two, or the motor Or a fuse! That nice little selfballancing model you have in your video, is it made from an available kit that I can buy somewhere? Link to comment Share on other sites More sharing options...
esaj Posted July 28, 2016 Share Posted July 28, 2016 40 minutes ago, Frode said: That nice little selfballancing model you have in your video, is it made from an available kit that I can buy somewhere? Not as a kit per se, but it's built from readily available parts (from Aliexpress, of course ) in that video (excluding the frame I made from scrap wood): Arduino Uno -clone (around 2.5€), you could also use Arduino Nano (something like 1.75€/piece) MPU6050 -gyro/accelerometer on a breakout board (something like 1.5-2€ / piece) L298N dual H-bridge motor driver on a breakout board (something like 1.40€/piece) Two motors with gearbox and tire (about 2€/piece?) Some wires and miscellanous parts Some battery for the motors (something like 6-12V, depends on the motors) Some battery for the Arduino (unless powered from the same battery as the motors) I must warn you though, the L298N is very inefficient (internal voltage drop of several volts, don't remember the specifics), and overheats very easily (it enters and exits thermal shutdown once it starts to heat, and the robot starts to wobble, then falls over). I've later on designed and built two H-bridges with power-BJTs (that were poorly cooled, as I designed the board layout "too tight" and still overheated ). After that, I made one H-bridge with power MOSFETs, that seemed to work a lot better, but haven't (still, it's been several months ) built a second one to actually use on the robot. Also got some HIP4082's mosfet-drivers later on for driving the H-bridges, but also haven't had the time to work on new bridges with those The batteries I first used were 9V and AA -cell holders, but I switched it to 7.2V (nominal) 6S NiMH-pack, which is seen in the video. That cost maybe something like under 20€. Currently that pack (and a 3S LiPo-pack) is going to power my linear power supply (basically just a box & circuitry for undervoltage-protection, fuses, and some fixed and adjustable outputs), as I have some projects that need very steady voltage (my current power supply is based on a PC-switching supply, which adds ripple in the output). The rest of the robot is "just" building the frame & the software for running the robot... Basically, you read the tilt-value from the MPU6050, I've used ready-made libraries for it, and drive the motors according to the input, in my case, I used a PID-loop where the P-, I- and D-factors can be controlled with the potentiometers on the breadboard in the robot. I looked at the kits for building one, but they cost something like 90€, and I thought I'd learn a lot more just by doing it from more "basic" pieces. More details are buried somewhere in my ramblings in the off-topic section: Link to comment Share on other sites More sharing options...
EUCMania Posted July 28, 2016 Share Posted July 28, 2016 Once the EUC cannot balance the rider, what should it do to protect the rider? 1. release an airbag attached to your body before sending you to fly. Trouble if it is a false alarm. 2. extending another wheel to the front to help balance and a pole in front of you to help you to upright your body. This is hard to be done right. The pole may penetrate your body. Looks none of the above can be done , if possible, without a big price tag. So, the best solution so far is to wear enough protection gears, stay within the safe speed, acceleration zone of the EUC. Currently, many EUCs give warnings such "over powering" when they sense the power demand is close to danger zone. The EUC manufacturer can and should provide graphs in the manual describing the max force you can push the euc at various speed, and max acceleration at various speed for various load. These will help riders to understand that at higher and higher speed, the acceleration the EUC can handle is smaller and smaller. A height of bumps vs. safe speed graph is necessary, which so far we just learn it by our mistakes. Link to comment Share on other sites More sharing options...
Keith Posted July 28, 2016 Share Posted July 28, 2016 Or 3. Perhaps a forward thinking EUC manufacturer could go into partnership with Martin Baker http://www.martin-baker.com Link to comment Share on other sites More sharing options...
EricGhost Posted July 29, 2016 Share Posted July 29, 2016 On 28/7/2016 at 6:56 AM, Reivax said: When I was talking about blind faith... You say it's science, but you do not understand how the balancing works. So, yes, it is science, and science also explains why some things can't be done. Yet, even when people explain to you how it's working, you keep knowing better, so this is going nowhere. Probably some day you will come up with an engine made only with magnets that will require no other source of power to run forever. 1- science said speed of light is a limit, yet today science says that by deforming space in front of a spaceschip is possible to travel at higher than speed light, science just tell you what you cannot do but just with your actual knowledge, there is always more 2 - I was speaking about the controller performances not about the basic physics, even if I used improper terms my understanding I think is clear see link the simple models http://arxiv.org/pdf/1511.04171.pdf the complex ones http://www.araa.asn.au/acra/acra2010/papers/pap149s1-file1.pdf https://www.researchgate.net/publication/275370743_Balancing_and_Velocity_Control_of_a_Unicycle_Robot_Based_on_the_Dynamic_Model Link to comment Share on other sites More sharing options...
EUCMania Posted July 31, 2016 Share Posted July 31, 2016 One possible safety device to prevent over speeding and overlean: The pedal it self can move backward/forward, driven screw powered by a small motor. As the EUC's speed increases, the pedal moves more backwards, so that the the max possible forward torque the driver can apply to the EUC, when he puts all his weight on the forward edge of the pedal, become smaller to stay within the safe range. Link to comment Share on other sites More sharing options...
Recommended Posts
Archived
This topic is now archived and is closed to further replies.