May 14, 20187 yr Wow, I'm intrigued, what's that? All the best, David https://bmsbattery.com/home/857-all-in-one-hub.html They have an app that talks to the controller over Bluetooth 4.0 I am sure d8veh is going to put me right on this. I think that the controller uses the 4 dollar MPU-6050 gyro/inclinometer to work out how much assist to give, replacing the expensive torque sensor and LCD. Edited May 14, 20187 yr by Woosh
May 14, 20187 yr https://bmsbattery.com/home/857-all-in-one-hub.html They have an app that talks to the controller over Bluetooth 4.0 I am sure d8veh is going to put me right on this. I think that the controller uses the 4 dollar MPU-6050 gyro/inclinometer to work out how much assist to give, replacing the expensive torque sensor and LCD. Yes, it's just a regular sensor like you have in your phone (unless you're a Luddite, like Flecc), but it's only a sensor. It's what the controller does with the information that matters, the same as a torque or cadence sensor. Nevertheless, it's still a brilliant idea to automatically adjust the power as you go up and down inclines, but I don't think many people would want that.
May 14, 20187 yr have you tried out their all in one hub? I must say I am tempted to buy a sample. I guess they use the inclinometer to replace the assist level. You only need to tell the controller how much you want to pedal (eg 70W), the controller works out how much current to supply. Edited May 14, 20187 yr by Woosh
May 14, 20187 yr have you tried out their all in one hub? I must say I am tempted to buy a sample. I guess they use the inclinometer to replace the assist level. You only need to tell the controller how much you want to pedal (eg 70W), the controller works out how much current to supply. No, I haven't tried one. Yes, the gyroscope thing makes a very clean install and solves the problem of buttons on the handlebars. With a torque and/or rotation sensor and a battery in the hub, you don't need any wires at all. I guess I'm a bit old-fashioned with these things, but I prefer to have control myself rather than hand over to automatic systems. I guess that eventually they'll get these sort of systems perfect, then I'll have to let go of my old ways.
May 14, 20187 yr With a torque and/or rotation sensor and a battery in the hub, you don't need any wires at all. I have a bottom bracket controller on a new bike. Unfortunately, Hatti does not think it will sell and she knows our customers better than I do. She also said it'll take in water...
May 14, 20187 yr No, I haven't tried one. Yes, the gyroscope thing makes a very clean install and solves the problem of buttons on the handlebars. With a torque and/or rotation sensor and a battery in the hub, you don't need any wires at all. I guess I'm a bit old-fashioned with these things, but I prefer to have control myself rather than hand over to automatic systems. I guess that eventually they'll get these sort of systems perfect, then I'll have to let go of my old ways. Guess I am a bit of a Luddite too. To me, Add complications, Add unreliability.
May 14, 20187 yr I think that the controller uses the 4 dollar MPU-6050 gyro/inclinometer to work out how much assist to give, replacing the expensive torque sensor ... Ok, I'm scratching my head, but I can't work out how that could replace a torque sensor. How would it do that ??
May 14, 20187 yr Ok, I'm scratching my head, but I can't work out how that could replace a torque sensor. How would it do that ?? The torque sensing is not needed, what is needed is an algorithm to calculate how much current is required. This is just me theorising, if you need to know the torque, you may need the MPU-9150 instead of MPU-6050 because the 9150 has a magnetic sensor built in on top of the gyro/inclinometer. A magnet ring on the BB can be used to derive the RPM while the gyro can be used to calculate the speed, the inclinometer gives the gradient. You compute the required power from speed and gradient, subtract the motor output to get user input, then deduce the torque from user input power/rotational speed in rads.
May 14, 20187 yr The torque sensing is not needed, what is needed is an algorithm to calculate how much current is required. This is just me theorising, if you need to know the torque, you may need the MPU-9150 instead of MPU-6050 because the 9150 has a magnetic sensor built in on top of the gyro/inclinometer. A magnet ring on the BB can be used to derive the RPM while the gyro can be used to calculate the speed, the inclinometer gives the gradient. You compute the required power from speed and gradient, subtract the motor output to get user input, then deduce the torque from user input power/rotational speed in rads. OR You use a twist throttle.
May 14, 20187 yr The whole concept is to have everything in the hub - no wires at all. You do some basic setup with a smart-phone to set the power according to your weight or fitness. The hub detects when you pedal via its own torque or cadence sensor inside the hub. You start riding and the system knows whether you're going uphill or downhill via the gyro sensor. It adjusts the power accordingly to keep the pedal effort within the parameters that you set with your phone. Like everything, there are advantages and disadvantages.
May 15, 20187 yr It adjusts the power accordingly to keep the pedal effort within the parameters that you set with your phone. Ok, but that is all well and good if all you want to do is cycle at a constant input effort all the time. In real life and in complex situations, a cyclist would be varying and controlling the power delivered to the bike, often in a dynamic and constantly varying way. To do this on a normal bike, your brain is constantly scanning with your eyes and altering the power via your leg muscles. If you want to augment this with an ebike motor, you have to somehow tap into the info that the brain is producing if you want to maintain this fluid, dynamic cycling experience. You could try drilling electrodes into your head (!?!), however a far less painful method seems to be to measure the force exerted onto the pedals! i.e. use a (good quality) torque sensor. I don't see how anything else can produce the amplified 'natural cycling' experience that many people find very satisfying and prefer. I completely understand that some people don't want to/can't put the effort in to adequately operate a torque only based system, hence why a switchable system does sound ideal. However you can only switch off something that is already there. You can't switch in something that is missing in the first place, so I personally am all for pedelecs (as distinct from electric mopeds) that have torque sensors! Anything else just isn't cycling
May 15, 20187 yr I don't see how anything else can produce the amplified 'natural cycling' experience that many people find very satisfying and prefer. it is not practical to respond immediately to variations in the torque sensor output which itself varies cyclically with the crank angle. The torque sensor helps to maintain an average ratio of cyclist pedal power / motor power. That job can equally be done with the data collected from the gyro.
May 15, 20187 yr it is not practical to respond immediately to variations in the torque sensor output which itself varies cyclically with the crank angle. The torque sensor helps to maintain an average ratio of cyclist pedal power / motor power. That job can equally be done with the data collected from the gyro. Yes, I can imagine that you might not want a direct, instantaneous 1 to 1 relationship (although pondering this as I write - wouldn't that be exactly what happens with a powerful cyclist just cycling normally on a non powered bike?) But you could (and they probably do) start playing around with peak detection algorithms with smoothing etc to try and counter that if you didn't want it. However I'm still puzzled as to how data from a gyro sensor can detect how much power you are/want to input ?? (but I'm still pondering. There is a power meter out there that works out rider input without actually using a torque sensor! I forget its name, but it sits on the front bars and measures inclination, speed, and also wind).
May 15, 20187 yr Ok, but that is all well and good if all you want to do is cycle at a constant input effort all the time. In real life and in complex situations, a cyclist would be varying and controlling the power delivered to the bike, often in a dynamic and constantly varying way. To do this on a normal bike, your brain is constantly scanning with your eyes and altering the power via your leg muscles. If you want to augment this with an ebike motor, you have to somehow tap into the info that the brain is producing if you want to maintain this fluid, dynamic cycling experience. You could try drilling electrodes into your head (!?!), however a far less painful method seems to be to measure the force exerted onto the pedals! i.e. use a (good quality) torque sensor. I don't see how anything else can produce the amplified 'natural cycling' experience that many people find very satisfying and prefer. I completely understand that some people don't want to/can't put the effort in to adequately operate a torque only based system, hence why a switchable system does sound ideal. However you can only switch off something that is already there. You can't switch in something that is missing in the first place, so I personally am all for pedelecs (as distinct from electric mopeds) that have torque sensors! Anything else just isn't cycling The motor does have a torque sensor in it. The gyro sensor adjusts the eco/tour/sport or level 1/2/3/4/5 setting automatically.
May 15, 20187 yr However I'm still puzzled as to how data from a gyro sensor can detect how much power you are/want to input ?? it's just a programming task. This paper shows how to calculate speed and distance using gyro data: https://www.nxp.com/docs/en/application-note/AN3397.pdf From the speed and gradient (you get it from the same I2C bus), you calculate the required power.
May 15, 20187 yr it's just a programming task. This paper shows how to calculate speed and distance using gyro data: https://www.nxp.com/docs/en/application-note/AN3397.pdf From the speed and gradient (you get it from the same I2C bus), you calculate the required power. In the aircraft world and other applications such as submarines, space flight, and artillery positioning it is called "inertial navigation" and has been in use in one form or other since the 1960's. In those days the hardware was just a little bigger - the 1970's gyro for the Jaguar aircraft was a 2 man lift into the airframe!
May 15, 20187 yr the 1970's gyro for the Jaguar aircraft was a 2 man lift into the airframe! how time has changed! The new chip measures 4mm x 4mm x 1mm.
May 15, 20187 yr .....or we could could use a tried and tested torque sensor that sits in the BB and is operated by pressure applied to the cranks through pedals. It's not only in hills that we need more power it's for acceleration too. A torque sensor is simple and effective. I understand it's not perfect for everyone hence the reason the twist throttle overrides the TS and we have a cadence sensor only mode on our Torque models.
May 15, 20187 yr how time has changed! and the operational computer memory was "core store" with a capacity of 6Kb - it was wired by specially skilled ladies who could thread the copper wire through the ferrite cores by hand. It also weighed about 40 lbs when fully cased and ready for installation. To do diagnostics we had to unload the flight program and reload the diagnostic program with a 1" mylar tape that had been perforated for the code and optically read. The reader/loader was in its own servicing trolley with a 28v aircraft battery to provide for use on the flight line without mains electricity - we used to tow it with a landrover!
May 15, 20187 yr The motor does have a torque sensor in it. The gyro sensor adjusts the eco/tour/sport or level 1/2/3/4/5 setting automatically. Are we talking about this one? https://bmsbattery.com/home/857-all-in-one-hub.html Are you saying that it can directly sense the force that the user is applying to the pedals? If so, how does it do that?
May 15, 20187 yr Are you saying that it can directly sense the force that the user is applying to the pedals? If so, how does it do that? No, they don't need to. With the gyro, they can do one step better than cadence sensing. They can work out the average ratio user input power / motor power without needing a torque sensor.
May 15, 20187 yr No, they don't need to. With the gyro, they can do one step better than cadence sensing. They can work out the average ratio user input power / motor power without needing a torque sensor. I don't think the gyro can predict when you need a little more power to lift the front wheel over an obstacle in quite the same way that a torque sensor is going to respond to your input.
Join the conversation
You are posting as a guest. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.