Interesting to see regenerative braking used on a bike. For railroads, regenerative braking is much more powerful. In the US it is called "Dynamic Braking." Electric and diesel-electric locomotives use it to control heavy trains (10,000 + tons) while descending long grades. The dynamics can take the entire load of the train to maintain a constant speed without need to use its air brakes or drag its brakes all the way down the hill. Since locomotives do not use regenerative braking to recharge batteries, the power from their traction motors is fed into large, air-cooled, resistor banks where the energy is transformed into heat. Standard dynamic brakes can not completely stop a train. Their effect fades out between 3 or 4 mph and air brakes are used for stopping. Just thought you might like to know.
Steel wheels on rails are a lot more efficient than rubber tyres due to a lower rolling resistance. I've been told a railed vehicle of the equivalent weight over the same distance would consume 1/7th the energy of that a rubber wheeled vehicle. So a lot of that energy could be captured back in comparison.
You would not believe me, but I'm a locomotive engineer and everywhere in the world electirc trains, electric locomotives and diesel electric locomotives have dynamic braking with resistors lol, even diesel hydraulic locomotives and trains have hydro dinamic braking. But apart from that, newer electric trains and locomotives like siemens vectron are putting their ED braking back to grid and it's recuperative braking ;) More on the same subject, all diesel - electric locomotives are hybrids and they have been around for 100 yrs, but now I guess is just recognised as technology evolves, so it can be used better then just a heavy load transmision that can sustain grater forces :)
@@fauzirahman3285 that explains why it was possible for a strongman to pull 1000 ton train, yet pulling 700-1000 cars even on perfectly flat ground would be impossible. Hysteresis losses in the rubber and also the slight rubbing from expanding and contracting contact patch area (strongly depends on tire pressure) would make the force requirements too much. In fact because there are so many cases where the brake pads in cars slightly rub against the wheels, it might even be impossible to turn 4000 car wheels at once. Even if each pad rubbed against the wheel and created 250 grams of resistive force at the edge of the wheel, it would still take a tonne of force to simply turn all the wheels!
Mining haul trucks do the same thing. They will have a diesel generator and electric wheel motors. Most of the time they go down hill they'll use the regen to slow them, and like the trains, dump it across a big resistor bank. Sometimes you can hear a high power fan turn on to help cool the resistors.
Where Regen really access is at saving your brake shoes. Ebikes tend to wear brake pads down faster because you generally travel at higher speeds and they are heavier. Regen breaking mitigates that problem nicely.
On a bicycle I’d assume most of the “wasted energy” is overcoming wind resistance. With a fiberglass shell or even a polycarb windshield I bet that your efficiency would go way up. This was a great we’ll put together video, but I’m waiting for a wiring guide to this setup!!! How does the recharge thumb throttle fit into the wiring? Can you ask the motor to spin up while it’s spinning in reverse while the e-brake is active? Would you get more recharge by slowly easing into the thumb throttle, like in electric motor vehicles? I am so interested in your unique approach to ebike engineering, I hope to have my video up soon detailing some of the work I’ve done on my bike, because, although we have similar performance, our approaches are vastly different! Thanks for the great videos man. Keep it up
The regen is all done within the speed controller (VESC6), I simply wire in two inputs (input ADC1: twist throttle & input ADC2: thumb throttle), then configure in software that I want the ADC1 to activate throttle and ADC2 to activate regen. When the brake is activated, the throttle from the ADC1 is cut off. In terms of how much intensity of regen braking to use, I'm not entirely sure and will need to do some testing on it. Thanks!!
I would say the used energy for bikes really depends on where/how you are driving and the specifics of your bike. But in general i would assume that at the speeds most people drive every day it would be the friction of the tires.
For typical bikes the crossover between air restance and rolling resistance is at about 15km/h. Since drag increases with the square of the speed (rolling resistance only linearily), drag is quite quickly the relevant type of resistance. For example going 30km/h instead of 20 (so 50% faster), you need almost 3x (+200%) the power. All "non-refundable" unfortunately.
didn't know about those figures specifically, that whats one of the things I was getting at. Since most of us aren't willing to compromise the design of a upright bike to make it more aerodynamic, I wouldn't worry too much about that % of "wasted energy". With that being said though, EV's like the Nissan leaf have some pretty undesirable industrial design quirks due to the engineers trying to maximize aero. The battery in the first leaf was only 24 kWh, less energy than found in a single tank of gasoline. (81,891 BTU vs. 115,000 BTU) it's actually pretty remarkable what we've been able to do with EV's in the last decade alone, especially when considering that gasoline cars have so much more stored potential in almost every circumstance. Yet we still are getting comparable ranges and performance specs because of gains in regen efficiencies and aero.
The trick is just ride downhill. Thing is that regenerated energy doesn't have the weight penalty of a bigger battery. So, it helps a bit. In some situations, it might be quite a bit.
Great video and assessment of a lot of the variables involved. My 2 cents: one other major component of loss that was not discussed is the motor itself. I did some measurements with an elevator (straight up, straight down, slow speed, no significant wind drag), using a 1/2HP permanent magnet brushed DC motor. The motor was rated at 90v * 5.4A = 486W. With 745W/HP, that works out to (745 / 2) / 486W = 76.6% motor efficiency. (For ease of math, the motor loss was just over 25%.) That loss applies whether the motor is providing power, or sinking power. In short, I had ~50% loss right out of the motor alone. BLDC (brushless DC motors) can be more efficient than that, potentially past 90%--but still, that's a 20% round-trip loss. In addition, the bulk of motor loss came out of the voltage, not the current--so I basically provided it 90v, and got 45v back at the same regulated motor speed. This also means that the motor controller has to "boost" the generated motor voltage to just beyond the battery voltage in order to recharge the battery. And unfortunately the easiest way to do it (i.e. no extra parts needed) is by PWMing a short circuit across the motor leads, and dumping the resulting "kick" back into the battery through the blocking diodes on the motor driver FETs. (Basically a boost converter, but with the power source and boost inductor being one and the same device.) While cheap and effective, this design is unfortunately also very inefficient. At least it's the thought that counts...
I work as a courier on the evenings and I can burn through breaks very rapidly, so the cost saving in brake pads is a lot more than a few extra battery cells.
Awesome video, thanks. This proves the gut feeling I had surrounding the usefulness of regen B. It also makes sense then that hub motors are losing in popularity to mid-mounted motors, despite the fact that they are not capable of RB. I do think that RB would be a lot more useful in extremely large vehicles, like long haul trucks, mine vehicles and trains. Just think of all those trucks that have to gear down to brake to save their brakes.... There has also been a study of an electric mine dump truck that actually produced more power than it consumed, purely because it was primarily carrying loads down an incline!
An immediate like for this impressive video. Have you considered monitoring the temperature of your battery and charging unit to get an indication as to where some of the power is wasted?
+Richard Wenner I draw it all in Adobe Photoshop, then move things per frame like stop motion animations. Or for the energy meters, I adjust the scale of the coloured bars in Sony vegas video editing software
Slick - its a good result. When we do this I remember an evening I had here in Cardiff with the guy who edited 'Match of the Day' - back in film processing days. The sound on film is physically offset from the image so editing was a fag. He told of nights when the program went out where you could still see the cleaning water on the final broadcast as the film was being fed by hand from final processing directly into the tele-cine. We don't know we are born! :) Thanks
Yeah, your animations are great too :) Have you seen this? Its about MOSFETs temp conditions with better (some, at least) airflow, which you have, - it is sufficient? Cooling of battery (not so tightlyy coupled cells? some active fan too?) may be key to better efficiency, but still its very aggressive discharge/charge, where some pro LiFePo4 are probably better capable also. And BTW, VESC is open-hardware with parts at $50 probably. Its quite curious your 3 VESC4 died as, proper temp handling is core feature of firmware, when limits currents too. ua-cam.com/video/UwVQQ_LY5bc/v-deo.html
THANK YOU! Pretty much the first question everyone asks about my ebike is, "does it have regenerative braking?" I always responded with, "it's not really worth the added complexity and cost..." But video this will be the perfect explanation! Cheers!
What complexity? What cost? Any of the BLDC motor controllers available these days will already have regenerative braking enabled, and so there is no additional cost (for the motor or controller), and the only additional complexity is hooking up a switch to enable the feature.
Great video! You will be a great engineer and are a great teacher. I spent a lot of time thinking about regen braking in bikes, and the data you generated elucidated the real thing is not that great. The electronics efficiency could be a reason, but I think battery is the most starving component at the electric system. Maybe with capacitors, as suggested below you could get a few more miles.
Hi Tom. Excellent project and video, but at 4:09 you show a kinetic energy figure of 3.59Wh and you state that theoretically you should be able to recover ALL of this energy if it were not for the inefficiencies of energy conversion. No problem, that is a given, but since you actually recovered 1.27Wh, is the efficiency of the regen braking not closer to 35%? Available = 3.59Wh Recovered = 1.27Wh 1.27 as % of 3.59 = (1.27/3.59) x 100 = 35.3% Surely the figures can't lie? Regards Mark in the UK
+thecorbies Hi Mark, yes that is correct! When riding, I mentioned it was 14% 'efficient', however what I meant to say was "it recharged 14% of the energy used". So later in the video, I adjusted what I meant to say that it gains 15.7% back into the battery. This 15.7% wasn't meant to be the efficiency of the regen period, but instead, how much battery is recharged during the regen period. Regards Tom
Good point, but in this case the efficiency of the regen braking is much more than 35%: Available = 3.59Wh Recovered = 1.27Wh Max speed 35mph or 15m/sec, consuption is 30Wh/km Final speed 0, average consuption is 15Wh/km Time of breaking 12sec acceleration is - 1.25 m/ss Breaking distance 1.25x12x12/2=90m spent energy for passing breaking distance 0.09x15= 1.35Wh then 3.59-1.35= 2.24Wh is left for returning back Recovered 1.27Wh of 2.24Wh or 56% !!! :) .
Excellent video! I've heard that poor regen efficiency is due largely to internal resistance in batteries. Some have proposed hybrid battery/ultra capacitor set ups. Wonder if you could test a set up like that?
i added regen to my direct drive rear hub motor kit and one thing that no one is talking about is how good it is at braking without ever locking up. normally when braking as you apply both brakes hard the weight goes forward making the back wheel light and making it lock up, by using regen braking you have a brake that wont lock on the back allowing you to use the front without having to modulate the rear.
This gives you a great way to test the effect of aerodynamics. I remember when carrot shaped helmets and closed wheels were a thing in professional cycling. All that does is reduce drag, which you should be able to measure with this setup. I suspect you could improve efficiency quite a bit by reducing energy loss through friction, drag, etc.
I wonder which is the real efficency of the bike on a test bench. I mean without any drag, air friction and so. The motor has one eficiency, than all the transmission and frictions in the wheels. I guess this is less than 50%, and probably dependant on the cruising speed.
Just found this video, great stuff - I'm thinking of doing the E-bike conversion on my old touring bike and researching the options. I can confirm that overcoming air resistance is the main energy loss on a bicycle I have drop handlebars, on a gentle downhill I can easily coast past somebody on an upright bike who is pedalling furiously, so even reducing the air resistance a bit by going down on the drops makes a huge difference - so presumably would make the use of Regen more effective.
Curious. On a vehicle like this what's better: - friction free drive train coasting - regenerative braking Like traditional drive trains have a ratcheting cassette, but if you could disengage the cassette with a clutch instead of the ratchet it may save on friction, allowing you to coast further than you'd gain from regen. As it really seems friction and drag is the biggest enemy to your energy reserves here. Being able to coast without expending battery use on flat sections seems key to making the most of this.
The most glaring issue with this was the noise. Not that the noise itself is an annoyance but it is clearly the main source of the energy that is lost from the main kinetic->electrical-> kinetic loop. For that matter I would wonder if it were instead possible to place a flywheel next to the rear wheel that can return kinetic energy from braking without producing a lot of noise or drag.
Riding a bicycle means that most of of the power and also the energy is used to overcome the losses due to the driving resistances, mainly caused by air turbulences. The thrust needed to overcome the driving resistances goes with v², the drive power therefore goes with v³. You cannot expect to regain a lot of energy by regenerative breaking because the kinetic energy is relatively small.
Does your bike allow you to pedal also? You can recover some of your own input to the battery. Viscous friction wastes more of the energy the faster you go. As someone has mentioned before, a fully enclosed recumbent (velomobile) would give better results.
regenerative braking is perfect for me where there is lots of up and down hills and i have to constantly hold the brake for several hundred meters to not over speed. I just ordered a e-bike kit with regen braking so i will se if i love it or not in a month or so
Thanks for the vid - the controller supplier discontinued the product, but I imagine there are alternatives around, can I ask which motor/bat combo is yours?
great data gathering and analysis. Probably on a recumbent bike the regenerative ratio could be better due to lower wind resistance. Would a setup like this work: 2 batteries A + B + a generator unit: A is the driving battery, B is the gathering battery. A is driving the bike, B is connected to a generator that takes / converts the kinetic energy like a dynamo into electricity. A smart setup could measure A capacity and reaching a certain point of low voltage switches to B while Generator B is disconnected from B and connected to A. Etc. Not sure whether this could work...
This kind of breaking is useful for town taxies (specially in Paris): hybrid cars can use the generator to descelerate not the brakes. Recharging his battery and breaking less w/ physical brakes, these taxies makes 5x more kilometers than others typical taxies!! (By a reducing of brakes replacement and by a better autonomy of his batteries)
Now I'm even more curious on the le mans hybrid cars like the porsche 919. If you watch those cars they are able to gather a huge amount of battery charge from their regenerative braking.
one should also account for different efficiency in other types and makes of battery, generator, and bikes! Might not add up to much difference, but it is a factor. Great video!
The intent of electric assist isn't to be used constantly like on an electric car but to be used for accelerating and a boost up hills. Typically when riding, one would pedal more for cruising and after needing that boost at the red light. The energy input from the cyclist would likely add to the regeneration of battery energy.
Is there a way to put an efficient capacitor in the loop to boost performance? If the maximum recharging rate of the batteries is less than the draw on the system, then that would mean a lot of waste heat. My guess is that if the batteries charge slower than they discharge on this system, a capacitor in between would allow more of the energy to be recovered and either pushed back into the battery slower or used to accelerate the bike on the next leg.
The vehicle systems are efficient, obviously the mass but also the smart BMS. Friend of mine drove Uber for 3 years in a Prius and the brake pads weren't even bedded in.
At such low efficiencies using electric regen, you should explore other methods like compressed air, hydraulic, fly wheel, springs etc, basically any method to get something back quick when you have to slow down and waste energy unnecessarily.
Great video. I think windage loss varies as square of speed. To prove this can you use your setup measure energy used at various speeds over a fixed distance Viz: Do a 1 mile trip there and back at 20MPH and measure energy used. Do a 1 mile trip there and back at 10MPH and measure energy used. Do a 1 mile trip there and back at 5MPH and measure energy used. This will give very useful practical results.
This is interesting but could pedalling help with keeping the battery charged too. I find it interesting in breaking is where we can get some of the energy back!!
The go to person for these thing. Thanks Tom. Thought of a couple of factors that could influence the efficiency. First is speed. E-bikes here as standard are limited to 25 km/h compared to 35 mph? So both pedal assistance and reduced drag would improve overall efficiency at lower speed and, thus slowing down with regerative breaks would give back relatively more distance. Then there’s using a dedicated breaking generator versus the same driving motor (say front wheel) to improve energy recovery. Any thoughts? I could not tell how quick the braking was at say a low speed of 20 km/h. It looked rather too gradual,so may need additionnel friction braking for safety.
Have you done tests with long hill climbs like 500-1000m of elevation? I’m thinking the efficiency might be higher for those, but curious by how much...
You should try a super cap as a load for the breaking. Regenerative breaking works if you do it at peak efficiency point. I doubt the batteries are sinking the current the motor could be generating. Its all about matching impedances. A super cap is really low impedance so the idea is you charge the cap as much as posible and then dump the energy back to the batteries with a boost circuit. The more current the more breaking. You could get an idea of what you need by checking the voltaje open circuit at full speed and then the short circuit current at full speed. Do the math and that should give you your equivalent series impedance. Your load should always match that to have the most power transfered.
This is what I am looking for, a super cap to even out up and down hill rides. I don´t want an electric bike I just like to store the energy wasted if I need to break and use it shortly after. A flywheel might do it but I guess it will be to heavy.
I second that motion. But I'dd add a voltage measuring capability accross the capacitor. When the voltage gets near the capacitors rated voltage (which is often low for super-caps) then the excess current that couldn't be sunk by the now charged capacitor is dissipated as heat over a high wattage resistor. The same principal easy to understand would be a 220uf 50v capacitor slowly being charged. Over the cap, you'd connect a 47v zener. The cap would absorb current until it got to the 47v, and then the other incoming power would be dissipated as heat in the zener - Protecting the capacitor from overvoltage, and allowing the continual un-interrupted load.
@@dancoulson6579 a zener alone would be impractical, but you could get a 30-35v zener and hook it to the gate of a mosfet (or several) with maybe a resistor to take some load off of the mosfet. also theres no need to let the super cap charge at all, as the more voltaje it has the less difference between the two and so less current being sunk. you should really keep it as low as the boost circuit allows. The zener thing could work as a sort of protection for the super cap in case it were to "saturate", it should still give you some magnetic breaking though it would be wasting the energy as heat. if anything the boost converter and the voltage limiter should be capable of handling more power than what you would or could generate typically un average. say a slow breaking or a sudden one where youd be dumping all the kinetic energy from the bike and your body into the regenerative break, to avoid saturating the cap. also i wouldnt use a normal high voltage cap as they have really high esr compared to super caps. you need energy density not voltage tolerance. sure 50v makes all easier but it will be huge in comparison and your boost converter would also be bigger becase it would saturate much faster so it would basically have to be as powerful as your motor breaking at which point you might as well do away with the cap and just use the boost converter as a break. if resistance is constan then the cap at 0v will sink twice as much current as when it reaches 25v (assuming 50v open circuit), and that means half the breaking force, thats why it doesnt matter as much the voltage of the cap as its the impedance. like how a 100C 1Ah lipo cell can do 100A because its really low impedance even though its only 3.7v.
Tom you are right -15% is the maximum of energy which can be returned by ebike regen during the most agressive riding like yours. With other kind of riding the regen will drop till 2-3% and even till zero. I checked this on hillclimbing : spent 99Wh and returened 16Wh, i.e. the same 15%. During another long usuall riding including some hillclimbing I gained only 5%. I deleted my previose post :-) .
Only 15% for a hill ? At a low speed air friction remains low and I expected way more while the majority of the energy is "stored" in potential energy.
First of all I just want to say that I really like your videos! I believe that an error is made when comparing the efficiency of the regen braking accelerating and going up a hill. Pounder that you would go up the hill very very slowly, and then go back down very slowly with the regen breaking, then the wasted energy from drag would be minimal. On the other hand, if you would go up the hill a great speed and down again fast because the regen breaking wasn't powerful enough for the slope, then you would waste more energy to drag. Therefor, I think it is wrong to suppose that the 15,7% would be true for the slope example. Another point is that in theory you would approach your final speed as a limit when time goes to infinity so you would have better efficiency if you accelerated to 25mph instead of 35.
you would deffo save your brakes as you are not using friction pads but rather using electricity and magnetism to slow you down so there is no grinding or scraping sound of the brakes, i think this would be very useful in a long downhill where you dont want to go full speed as its too slippery and rocky but dont want to hold brakes constantly
You can use the front wheel to regenerate energy all the time you moving, just add another motor in the front wheel and another battery, the front wheel will collect energy to the new battery and the new battery will support the other one you just need to control it if it is charging or supporting , then you bike will travel double the distance that you have now
I think you should use this braking system in both of the brakes it will increase the efficiency by 30% if u r using on one . Another thing is if u use an another motor in front wheel so while the vehicle will move that motor will also generate energy while moving along with the wheel. I hope i am correct.
It's much more efficient to coast than it is to brake, a bicycle and a rider is a small mass, there is very little momentum to recover any useful amount of energy from. You may see a difference in charge after a 200 mile descent, but you'd never make it back home again. Best to just have a freewheeling motor so you can still pedal home if necessary.
Well, usually in the RB, you specify on the controller how many amps should be put back in batteries while RB. There is no way you can return electricity back to the batteries the same way you draw electricity out of them. Charging is always slower. So fo safety reasons in order not to burn your batteries, usually you specify about half or less than the actuall amperes you draw out of them. This is also one of the main reasons Rb is so low apart from thermal losses of course
Peddal assist would change this dramatically....you would use far less battery and gain more back ( as a percentage ) when riding down the same hill. The more you rely on your pedaling the higher percentage return on the downhill.
Very interesting! We’re adding solar cells to our electric scooter (Model M365) which will be utilized as a range extender. We noticed that the phone app which communicates with our scooter (Mi Home app) shows an energy consumption as low as 50 watts when traveling at a constant speed of approximately 5 to 7 miles per hour. We are presently using the SunPower monocrystalline solar cells, rated at 160 Watts (See Elfeland SP-22 Folding 160 Watt Solar Panels). If these cells can produce half of their rated output we may have a constant input (while the sun is shining) of power - acting as a range extender for the EV scooter. The EV scooter has a 250 watt motor, and a 36 Volt Lithium battery. We purchased a Genasun Boost MPPT Charger which will add power to our battery, from the solar cells. Marissa Muller here on you tube utilized 240 watts of SunPower solar cells to power her 250 watt electric bike across the country, permitting the solar cells to do 60 to 70 percent of the work, while she add the 30% to 40% via peddling the bike. Marissa notes she chose to limit the sunlight’s contribution of power so that she could get some exercise and thus help improve overall health. It’s a story that inspired us and so we too are studying how we can allow nature to help power our electric scooter. Very nice Ebike by the way! Great video!
Maxwell Super caps may be more efficient for fast charging/discharging. I think Tesla bought Maxwell, so I'm not sure if you can purchase them like before.
What I know about cars is that most of the energy used is from pushing air out of the way. It's probably the same for bicycles. So I think you are incorrect about hills. (Shell used to run a modified VW between LA and Phoenix and back to claim their gas was more efficient. The mountains along the route was most of the trick.) And I have a somewhat relevant test. I drove my Prius up into the mountains using 'Worried About Global Warming Mode", or 'extra wimpy.' Eco Mode+ gentle accelerating, not going too fast... I averaged 99 mpg for the trip. (Normal for me in my Pri-V is 39-43 mpg). I live in Los Angeles, we have 2000 foot mountains in the city, in less than 40 miles on roads I can be at 5000 feet. So with mountains regenerative braking on a bicycle could be not just practical but amazing. If I had a bicycle that could go 40 miles on a charge with pedal assist, then I'm sure that I could get up to Mt Wilson where my charge would be empty, and then recharge going down hill regaining a significant charge to go the last 20 flat miles home. The calc for the downhill section would be 20 miles at a reasonable speed 15-20 mph, full charging along the way. (I don't know how you're regenerating power, but with a generator connected to the wheels. I'd still be going 15-20 mph, it's that steep.) This particular highway, the Angeles Crest, is downhill 99.5% of the way for about 100 miles. The portion discussed above is even steeper. The last 18.6 miles has a 5200 feet elevation gain. (La Canada to Mt Wilson). What would also work is knowing at what battery level I should turn around to have enough juice left to make up the 100' hill to my house.
Wow! That's mostly bad news, but mountains with electric assist should be even better than in a car. I don't have an e-bike yet, but surely someone in Southern California does, get in touch, I've already got the experiment designed.
It's not necessary to know the entire LA metropolitan area to understand conservation of energy. When you ascend any hill, and have an entirely empty battery at the top it, the range you will have at the (any) bottom is only dependent on regen efficiency and power lost during ascend+descend due to drag (wind and rolling). Tom does a great job explaining the physics, but to be honest it seems to me high-school level. The ingenious bit comes from actually making it work and providing some figures. Probably you could model any situation quite well knowing wind and drag resistance functions/curves, motor power efficiency curve and regen efficiency curve. I suspect the motor+esc won't be as efficient at regen at every rpm and regen power.
Yes, but.... Mt Wilson (5700 feet, La Canada is 1200 or 1400 feet altitude. I don't understand how the 5200 foot gain was arrived at. At most 4500'. ( I was nearby today at 1100 feet) This is significantly different from UK hills. I think it's the perfect laboratory for regenerative E-bikes. What you say about speeds, gears, etc.. are all essential controls. Same bike, same speed, same rider, same weather..... But I think 19 miles and 4500' elevation gain will clearly delineate what's air/road/mechanism resistance and what is gained potential energy. The one factor that should make this a terrific route for a re-gen e-bike, is of course the rider is pedaling. It will look like 100% efficiency, but only because the rider is pedaling. Btw the trip down requires braking because it is steep. PE ( I don't like using K) = m*g*m, so let's say bicycle and rider are 80kg* 1500m*9.8 m/s/s= 1,176,000 joules. 36 v x 15,000 mAh battery - 1,944,000 joules ( I looked this up) So. 1. Depending on the rider's contribution the battery alone cannot possibly get him/her to the top of Mt Wilson. 1.1 versus 1.9 is about half, all the frictions and efficiencies have to be very significant, and rider's contribution may be known amount--but I don't know it. And clearly this battery can not possibly be fully charged by the downhill part. (Unless perhaps the rider carries some rocks down, which is quite feasible). I wish I knew more, and had an e-bike. Coasting downhill with a generator and uncharged battery would quickly answer a lot of questions. But 5%? I think it will be significantly more. My Prius, and the Shell gas test, clearly demonstrate that. (Also I don't mind Tom using high school physics, I do think he's missed a few things. As a teacher I've been there too. Just go over the same thing a dozen times and suddenly it becomes not only clear, but obvious. (You really can explain everything in physics to a three year old, just not all at once.) And I do think being in England Tom doesn't have the opportunity our local mountains provide. Of course he's investigating a practical re-gen solution.
wow that is a fabulous bit of research.I have been telling people that regen is not worth it.Partly because I don't believe the batteries can absorb the regenerated power fast enough.They put out say 15 amps on acceleration but cannot absorb 15 amps on recharge or can they?
Well in an electric car you wouldn't even think about pushing it up a hill and then recharging the batteries with regen braking. With a bike you can choose to use your muscles to get you up the hill and use the regen braking downhill to charge the battery if it's getting low. This deserves to be considered or atleast mentioned.
the thing is if you're using pedal assist, the regen would probably keep the battery alive for a lot more time. cause some of the kinetic energy came from you pedaling.
I dont know why they have not come up with a friction mounted generator.One that is separate from the motor but connected to the battery LIKE THE OLD BIKE LIGHT wheel generator's to charge the battery .It could be powered by the rim front or rear in some fashion and regulated for regeneration and always in the generation mode while in motion unless disabled by a lever or switch.A drive system could be developed based on the rim or hub.It also could be incorporated with regenerative braking.
At 35 mph, 3.59 watt hours of energy is probably needed every 10-15 seconds just to overcome drag, so almost half of your kinetic energy was lost to the wind. Also charging efficiency is likely less at high currents. You could reduce this by wearing skin tight clothing, but the best way to test how efficient regenerative braking can be is to test going up and down a hill at speeds below 10 mph - not too steep that your motor struggles.
When calculating rb efficiency you should only consider the energy being used to accelerate and decelerate the total mass and disregard wind resistance etc as the energy recovered by the braking is only harvesting from the kinetic energy available(+ potential lost if going down hill or - pe gained if going up hill). Energy lost due to drag etc will be a function of acceleration as this will determine how long it took to reach terminal speed an so will be greater for slower acceleration and this means that all the energy put in cannot be available for braking recovery. Only the kinetic available at the instant of applying the brakes will be available and a great deal will still be lost due to drag and conversion losses during braking. You should find that your actual energy recovery from the available energy is much higher than it first appears.
It would be interesting to see how much power rolling resistance consumes in general and during regenerative braking. If you check how many watts of power your motor draws at a few velocities, you could determine how much power is required to overcome rolling resistance (or overall resistance due to air drag) and interpolate a curve to get the power dissipated by rolling resistance at any speed. From there you can get a Watt-hour approximation for kinetic energy lost to rolling resistance during regenerative braking. You could also find out what range your system should have at different velocities and what the rolling resistance coefficient is for the road with respect to your bike's tires.
hope near future Tesla owner - do you think that ‘strategic coasting’ is close to regenerative braking? (current BMW can disconnect transmission in its ECO mode) Wondering if Tesla rapid/easy user controlled no-regen can increase distance more than fixed amount of regen
What happens if it goes 88mph tho? ;-) Have you not considering fitting a device/programming the arduino to maintain a constant speed above an adjustable single value or between 2 adjustable values (for example 25mph and 30mph)? You could have the motor only run to maintain a speed above or between 2 set values values and therefore you could potentially be saving battery power that may extend battery life and increase distance. You'd have to calculate for spin up/spin down when the motor runs but that together with the regenerative braking could increase the possible travel distance. EDIT: After more thought, this idea would be like a thermostat on a heating system that only turns on the heating when the heat drops but in this context, the heat is the speed. It could all just be a large game of minimalist giving and taking but efficiency tests are always good things to try and look at the results.
Its not equal, even in theory, because you would include the rolling resistance of the tyres, bearings etc as losses, as well as the motor & controller you are using in the calculation.
Whilst Regen braking is good in theory, surly some folks must have tried putting another light weight motor in the system to power up battery whilst in motion yes?
well done Tom. good vids, a quick question- could a front dynamo hub be wired into that vesc 6 thing of yours and put a sort of continual small trickle charge into the battery and would it be worth it?
Could you please show the wiring that allows the regen braking? Lots of people like me on the internet trying to figure out how to wire their hub motor kits that have "Low Brake" feature, but no wiring diagram. I agree with other comments that point out the excellent braking quality of regen brakes, especially the anti-lock nature. To optimize... Front wheel hub motor! Add an ultracapacitor... That is my dream bike!
usually, regenerative braking isn't put up to 100% braking, it should slow you down alot better at somewhere around 80 or 90% whereas its usually as 30-50%
This is the reason why it makes sense to add a freewheel to electric drive. :-) And it was also one of the reasons why my electric motorcycle didn't have regenerative braking. I really like the simplicity of your project. I am seriously thinking of building another e-bike - mainly because I have almost free access to VESCs (I designed a custom - smaller version). Thank you very much for sharing all your files and so many details of the project!
I was wondering about this too, I wonder how freewheel would compare in efficiency to regen in real use, I suspect that it would do very well in hilly terrain where you can just build speed on the down hill and if you peddle assist. Also, if you run out of range, I wonder how much peddling resistance the electric motor provides that would be eliminated with a free wheel, seems a lot to push that belt and motor along with the bike.
great info and content Tom. if you dont mind...How well does that gear ratio and electric motor handle the uphills around here (i'm from st albans and also went to herts uni). i was contemplating adapting or buying an electric motor/bike for my hybrid to take me to luton and back each day for my commute but i've seen a few of the electric scooter vids on youtube not really like that sort of gradient and struggle to power the uphills.
Seems to me like if you used better battery management (in software, probably) to control the rate of regen, you could probably significantly boost the efficiency of the recharging by controlling the current rate of recharging. Perhaps an intermediary supercapacitor bank and some MOSFETs?
For LiPo batteries, the safe rate of charge is much lower that the rate of discharge - by an order of magnitude or two (10-100x). Without knowing the details of your power system, and for the sake of the argument, I'll guess a safe discharge rate of 1000W and a charge rate of 100W. You would therefore have to decelerate very slowly indeed to avoid having to 'throw away' power to avoid damaging your batteries. My guess that most of the blame for the very poor efficiency scores in your first, flat test were due to this effect. I'm interested in the use of a relatively small bank of super-capacitors to rapidly absorb braking energy, then safely mete it out to the main batteries. What are your thoughts?
Just wondering, in the kinetic energy calculations, you made it seem quite simple. I was just wondering whether you considered the rotational kinetic energy in addition to your translational(linear) kinetic energy. If you were to leave this additional energy out it would make your predictions for the energy regained artificially high. Just wondering.
Great vedeo. Would like to know energy lost by wind. What was the power at constant speed? Also... the efficiency will change depending on acceleration rate.
I know this is kind of redundant now 🙃but couldn't we just put multiple motors in series regenerative breaking and super capacitor to maximized the output also storage that way is just a extra electric generator to power basically anything with inverter 😶
I will have to try your test with my faired recumbent bicycle equipped with a "Copenhagen Wheel" similarly bluetooth enabled. It uses the bluetooth for all control functions, except for pedal input responses, so I have to pedal slowly backwards to enable regen. One factor I suppose you did consider was to begin your test with a battery less than fully charged so that regeneration voltage differentials could effectively deliver energy to the pack. I think the Tesla formulation, beside good aerodynamics, is to have a large enough battery and a motor of sufficient power to do regen quite agressively. My Prius does use regen as best it can, but an 8 mile long mountain down hill route I do quite regularly fills the battery to its software governed 80% level in just a few miles, and then it has no place to store the potential/kinetic energy.The big motor/big battery Tesla can probably use much more of that potential energy returning from its climb up. I was surprised that your E-bike was capable of 35mph. I thought most european e-bikes cut out at 25 kph, and here in the US the legal assist top speed is 20mph(32kph) so I cannot duplicate your top speed test. I do know that my bike is much better aerodynamically than yours.
i think regenerative braking only will save your brakes over time, i like this vid, it showed me lots of interesting stuff and if its worth getting regenerative braking for a bike if i go electric.If you were to pedal while regen is on would you recharge your battery
As a hipster, I clearly noticed there were some codewords in this conversation. It roughly translates to ; Tom, would you like to go fight some eastern-european kangaroos in Japan? Not sure what it means though.
That's great. I have a high power kit 52v 40 amps. Can go pretty fast. My city has no bike lanes so it's safer to go faster, when riding alongside the cars. I've been using the regen brake for the main purpose of assisting the brakes when braking from high speed. It's great. I don't even care about the battery regen, just the added brake power is awesome. I wonder why it's not more popular, some people don't even know their controllers are capable of regen braking. It's a great safety feature. I've been wondering if there is a way to use the brake levers sensor to make a brake tail light for the bike. Any ideas? Thanks.
Is your regen variable or fixed? On my Chevy Volt, sometimes if i max regen brake, I get a burst of high generation for a very short period, vs if I regen slowly over a longer distance, seems to be better. Just curious, great experiment!
Intresting... I wonder if there could be a way to charge the battery while riding? Maybe a setting that that you could set to like 10% or 5% load can be applied to charge the battery?
Interesting to see regenerative braking used on a bike. For railroads, regenerative braking is much more powerful. In the US it is called "Dynamic Braking." Electric and diesel-electric locomotives use it to control heavy trains (10,000 + tons) while descending long grades. The dynamics can take the entire load of the train to maintain a constant speed without need to use its air brakes or drag its brakes all the way down the hill. Since locomotives do not use regenerative braking to recharge batteries, the power from their traction motors is fed into large, air-cooled, resistor banks where the energy is transformed into heat. Standard dynamic brakes can not completely stop a train. Their effect fades out between 3 or 4 mph and air brakes are used for stopping. Just thought you might like to know.
Steel wheels on rails are a lot more efficient than rubber tyres due to a lower rolling resistance. I've been told a railed vehicle of the equivalent weight over the same distance would consume 1/7th the energy of that a rubber wheeled vehicle.
So a lot of that energy could be captured back in comparison.
You would not believe me, but I'm a locomotive engineer and everywhere in the world electirc trains, electric locomotives and diesel electric locomotives have dynamic braking with resistors lol, even diesel hydraulic locomotives and trains have hydro dinamic braking. But apart from that, newer electric trains and locomotives like siemens vectron are putting their ED braking back to grid and it's recuperative braking ;) More on the same subject, all diesel - electric locomotives are hybrids and they have been around for 100 yrs, but now I guess is just recognised as technology evolves, so it can be used better then just a heavy load transmision that can sustain grater forces :)
@@fauzirahman3285 that explains why it was possible for a strongman to pull 1000 ton train, yet pulling 700-1000 cars even on perfectly flat ground would be impossible. Hysteresis losses in the rubber and also the slight rubbing from expanding and contracting contact patch area (strongly depends on tire pressure) would make the force requirements too much.
In fact because there are so many cases where the brake pads in cars slightly rub against the wheels, it might even be impossible to turn 4000 car wheels at once. Even if each pad rubbed against the wheel and created 250 grams of resistive force at the edge of the wheel, it would still take a tonne of force to simply turn all the wheels!
@Jake\Nerd No. The engineer applies the brakes manually. And newer A.C. motors can stop a train completely.,
Mining haul trucks do the same thing. They will have a diesel generator and electric wheel motors. Most of the time they go down hill they'll use the regen to slow them, and like the trains, dump it across a big resistor bank. Sometimes you can hear a high power fan turn on to help cool the resistors.
Part of the sales pitch for regen brakes is the reduced wear on the friction brakes.
Where Regen really access is at saving your brake shoes. Ebikes tend to wear brake pads down faster because you generally travel at higher speeds and they are heavier. Regen breaking mitigates that problem nicely.
I havn' t heard such a concise explanation of energy transfer since college days. Congratulations Tom, very well done!😎
I must say, This is probably one of the best channels on youtube, I would like to thank you!
On a bicycle I’d assume most of the “wasted energy” is overcoming wind resistance. With a fiberglass shell or even a polycarb windshield I bet that your efficiency would go way up. This was a great we’ll put together video, but I’m waiting for a wiring guide to this setup!!! How does the recharge thumb throttle fit into the wiring? Can you ask the motor to spin up while it’s spinning in reverse while the e-brake is active? Would you get more recharge by slowly easing into the thumb throttle, like in electric motor vehicles? I am so interested in your unique approach to ebike engineering, I hope to have my video up soon detailing some of the work I’ve done on my bike, because, although we have similar performance, our approaches are vastly different!
Thanks for the great videos man. Keep it up
The regen is all done within the speed controller (VESC6), I simply wire in two inputs (input ADC1: twist throttle & input ADC2: thumb throttle), then configure in software that I want the ADC1 to activate throttle and ADC2 to activate regen. When the brake is activated, the throttle from the ADC1 is cut off. In terms of how much intensity of regen braking to use, I'm not entirely sure and will need to do some testing on it. Thanks!!
I would say the used energy for bikes really depends on where/how you are driving and the specifics of your bike.
But in general i would assume that at the speeds most people drive every day it would be the friction of the tires.
The percentage of energy used on a bike, to counter wind resistance is **very** high - on an upright bike, it'll be 80% or so.
For typical bikes the crossover between air restance and rolling resistance is at about 15km/h. Since drag increases with the square of the speed (rolling resistance only linearily), drag is quite quickly the relevant type of resistance. For example going 30km/h instead of 20 (so 50% faster), you need almost 3x (+200%) the power. All "non-refundable" unfortunately.
didn't know about those figures specifically, that whats one of the things I was getting at. Since most of us aren't willing to compromise the design of a upright bike to make it more aerodynamic, I wouldn't worry too much about that % of "wasted energy".
With that being said though, EV's like the Nissan leaf have some pretty undesirable industrial design quirks due to the engineers trying to maximize aero. The battery in the first leaf was only 24 kWh, less energy than found in a single tank of gasoline. (81,891 BTU vs. 115,000 BTU)
it's actually pretty remarkable what we've been able to do with EV's in the last decade alone, especially when considering that gasoline cars have so much more stored potential in almost every circumstance. Yet we still are getting comparable ranges and performance specs because of gains in regen efficiencies and aero.
Finally an intelligent and comprehensive explanation. Thank you.
Man! The level of research you've been through is amazing! Thank you!!!
The trick is just ride downhill. Thing is that regenerated energy doesn't have the weight penalty of a bigger battery. So, it helps a bit. In some situations, it might be quite a bit.
Great video and assessment of a lot of the variables involved. My 2 cents: one other major component of loss that was not discussed is the motor itself. I did some measurements with an elevator (straight up, straight down, slow speed, no significant wind drag), using a 1/2HP permanent magnet brushed DC motor. The motor was rated at 90v * 5.4A = 486W. With 745W/HP, that works out to (745 / 2) / 486W = 76.6% motor efficiency. (For ease of math, the motor loss was just over 25%.) That loss applies whether the motor is providing power, or sinking power. In short, I had ~50% loss right out of the motor alone. BLDC (brushless DC motors) can be more efficient than that, potentially past 90%--but still, that's a 20% round-trip loss.
In addition, the bulk of motor loss came out of the voltage, not the current--so I basically provided it 90v, and got 45v back at the same regulated motor speed. This also means that the motor controller has to "boost" the generated motor voltage to just beyond the battery voltage in order to recharge the battery. And unfortunately the easiest way to do it (i.e. no extra parts needed) is by PWMing a short circuit across the motor leads, and dumping the resulting "kick" back into the battery through the blocking diodes on the motor driver FETs. (Basically a boost converter, but with the power source and boost inductor being one and the same device.) While cheap and effective, this design is unfortunately also very inefficient.
At least it's the thought that counts...
I work as a courier on the evenings and I can burn through breaks very rapidly, so the cost saving in brake pads is a lot more than a few extra battery cells.
Awesome video, thanks. This proves the gut feeling I had surrounding the usefulness of regen B. It also makes sense then that hub motors are losing in popularity to mid-mounted motors, despite the fact that they are not capable of RB. I do think that RB would be a lot more useful in extremely large vehicles, like long haul trucks, mine vehicles and trains. Just think of all those trucks that have to gear down to brake to save their brakes.... There has also been a study of an electric mine dump truck that actually produced more power than it consumed, purely because it was primarily carrying loads down an incline!
An excellent video on the efficiency of regenerative braking
An immediate like for this impressive video. Have you considered monitoring the temperature of your battery and charging unit to get an indication as to where some of the power is wasted?
Currently I've only monitored the temperature of the speed controller. But thats a good idea to look into! Thanks
Off track - how do you produce your graphics?
+Richard Wenner I draw it all in Adobe Photoshop, then move things per frame like stop motion animations. Or for the energy meters, I adjust the scale of the coloured bars in Sony vegas video editing software
Slick - its a good result. When we do this I remember an evening I had here in Cardiff with the guy who edited 'Match of the Day' - back in film processing days. The sound on film is physically offset from the image so editing was a fag. He told of nights when the program went out where you could still see the cleaning water on the final broadcast as the film was being fed by hand from final processing directly into the tele-cine. We don't know we are born! :) Thanks
Yeah, your animations are great too :) Have you seen this? Its about MOSFETs temp conditions with better (some, at least) airflow, which you have, - it is sufficient? Cooling of battery (not so tightlyy coupled cells? some active fan too?) may be key to better efficiency, but still its very aggressive discharge/charge, where some pro LiFePo4 are probably better capable also. And BTW, VESC is open-hardware with parts at $50 probably. Its quite curious your 3 VESC4 died as, proper temp handling is core feature of firmware, when limits currents too.
ua-cam.com/video/UwVQQ_LY5bc/v-deo.html
THANK YOU! Pretty much the first question everyone asks about my ebike is, "does it have regenerative braking?"
I always responded with, "it's not really worth the added complexity and cost..." But video this will be the perfect explanation!
Cheers!
What complexity? What cost? Any of the BLDC motor controllers available these days will already have regenerative braking enabled, and so there is no additional cost (for the motor or controller), and the only additional complexity is hooking up a switch to enable the feature.
Great video! You will be a great engineer and are a great teacher.
I spent a lot of time thinking about regen braking in bikes, and the data you generated elucidated the real thing is not that great.
The electronics efficiency could be a reason, but I think battery is the most starving component at the electric system. Maybe with capacitors, as suggested below you could get a few more miles.
Hi Tom. Excellent project and video, but at 4:09 you show a kinetic energy figure of 3.59Wh and you state that theoretically you should be able to recover ALL of this energy if it were not for the inefficiencies of energy conversion. No problem, that is a given, but since you actually recovered 1.27Wh, is the efficiency of the regen braking not closer to 35%?
Available = 3.59Wh
Recovered = 1.27Wh
1.27 as % of 3.59 = (1.27/3.59) x 100 = 35.3% Surely the figures can't lie?
Regards Mark in the UK
+thecorbies Hi Mark, yes that is correct! When riding, I mentioned it was 14% 'efficient', however what I meant to say was "it recharged 14% of the energy used". So later in the video, I adjusted what I meant to say that it gains 15.7% back into the battery. This 15.7% wasn't meant to be the efficiency of the regen period, but instead, how much battery is recharged during the regen period.
Regards
Tom
Good point, but in this case the efficiency of the regen braking is much more than 35%:
Available = 3.59Wh
Recovered = 1.27Wh
Max speed 35mph or 15m/sec, consuption is 30Wh/km
Final speed 0, average consuption is 15Wh/km
Time of breaking 12sec
acceleration is - 1.25 m/ss
Breaking distance 1.25x12x12/2=90m
spent energy for passing breaking distance 0.09x15= 1.35Wh
then 3.59-1.35= 2.24Wh is left for returning back
Recovered 1.27Wh of 2.24Wh or 56% !!! :) .
Excellent video! I've heard that poor regen efficiency is due largely to internal resistance in batteries. Some have proposed hybrid battery/ultra capacitor set ups. Wonder if you could test a set up like that?
i added regen to my direct drive rear hub motor kit and one thing that no one is talking about is how good it is at braking without ever locking up. normally when braking as you apply both brakes hard the weight goes forward making the back wheel light and making it lock up, by using regen braking you have a brake that wont lock on the back allowing you to use the front without having to modulate the rear.
This gives you a great way to test the effect of aerodynamics. I remember when carrot shaped helmets and closed wheels were a thing in professional cycling. All that does is reduce drag, which you should be able to measure with this setup. I suspect you could improve efficiency quite a bit by reducing energy loss through friction, drag, etc.
I wonder which is the real efficency of the bike on a test bench. I mean without any drag, air friction and so. The motor has one eficiency, than all the transmission and frictions in the wheels. I guess this is less than 50%, and probably dependant on the cruising speed.
Just found this video, great stuff - I'm thinking of doing the E-bike conversion on my old touring bike and researching the options. I can confirm that overcoming air resistance is the main energy loss on a bicycle I have drop handlebars, on a gentle downhill I can easily coast past somebody on an upright bike who is pedalling furiously, so even reducing the air resistance a bit by going down on the drops makes a huge difference - so presumably would make the use of Regen more effective.
Curious. On a vehicle like this what's better:
- friction free drive train coasting
- regenerative braking
Like traditional drive trains have a ratcheting cassette, but if you could disengage the cassette with a clutch instead of the ratchet it may save on friction, allowing you to coast further than you'd gain from regen.
As it really seems friction and drag is the biggest enemy to your energy reserves here. Being able to coast without expending battery use on flat sections seems key to making the most of this.
The most glaring issue with this was the noise. Not that the noise itself is an annoyance but it is clearly the main source of the energy that is lost from the main kinetic->electrical-> kinetic loop. For that matter I would wonder if it were instead possible to place a flywheel next to the rear wheel that can return kinetic energy from braking without producing a lot of noise or drag.
Riding a bicycle means that most of of the power and also the energy is used to overcome the losses due to the driving resistances, mainly caused by air turbulences. The thrust needed to overcome the driving resistances goes with v², the drive power therefore goes with v³. You cannot expect to regain a lot of energy by regenerative breaking because the kinetic energy is relatively small.
Does your bike allow you to pedal also? You can recover some of your own input to the battery. Viscous friction wastes more of the energy the faster you go. As someone has mentioned before, a fully enclosed recumbent (velomobile) would give better results.
regenerative braking is perfect for me where there is lots of up and down hills and i have to constantly hold the brake for several hundred meters to not over speed.
I just ordered a e-bike kit with regen braking so i will se if i love it or not in a month or so
Thanks for the vid - the controller supplier discontinued the product, but I imagine there are alternatives around, can I ask which motor/bat combo is yours?
Regen saves on brake pads too 😀
great data gathering and analysis. Probably on a recumbent bike the regenerative ratio could be better due to lower wind resistance.
Would a setup like this work: 2 batteries A + B + a generator unit: A is the driving battery, B is the gathering battery. A is driving the bike, B is connected to a generator that takes / converts the kinetic energy like a dynamo into electricity. A smart setup could measure A capacity and reaching a certain point of low voltage switches to B while Generator B is disconnected from B and connected to A. Etc. Not sure whether this could work...
This kind of breaking is useful for town taxies (specially in Paris): hybrid cars can use the generator to descelerate not the brakes. Recharging his battery and breaking less w/ physical brakes, these taxies makes 5x more kilometers than others typical taxies!! (By a reducing of brakes replacement and by a better autonomy of his batteries)
that products no longer available, which is a shame, since i wanted it for a project
Now I'm even more curious on the le mans hybrid cars like the porsche 919.
If you watch those cars they are able to gather a huge amount of battery charge from their regenerative braking.
one should also account for different efficiency in other types and makes of battery, generator, and bikes! Might not add up to much difference, but it is a factor. Great video!
The intent of electric assist isn't to be used constantly like on an electric car but to be used for accelerating and a boost up hills. Typically when riding, one would pedal more for cruising and after needing that boost at the red light. The energy input from the cyclist would likely add to the regeneration of battery energy.
Is there a way to put an efficient capacitor in the loop to boost performance? If the maximum recharging rate of the batteries is less than the draw on the system, then that would mean a lot of waste heat. My guess is that if the batteries charge slower than they discharge on this system, a capacitor in between would allow more of the energy to be recovered and either pushed back into the battery slower or used to accelerate the bike on the next leg.
The vehicle systems are efficient, obviously the mass but also the smart BMS.
Friend of mine drove Uber for 3 years in a Prius and the brake pads weren't even bedded in.
At such low efficiencies using electric regen, you should explore other methods like compressed air, hydraulic, fly wheel, springs etc, basically any method to get something back quick when you have to slow down and waste energy unnecessarily.
Great video. I think windage loss varies as square of speed.
To prove this can you use your setup measure energy used at various speeds over a fixed distance Viz:
Do a 1 mile trip there and back at 20MPH and measure energy used.
Do a 1 mile trip there and back at 10MPH and measure energy used.
Do a 1 mile trip there and back at 5MPH and measure energy used.
This will give very useful practical results.
It does.
Fd=0.5*Cd*A*rho*V^2
This is interesting but could pedalling help with keeping the battery charged too. I find it interesting in breaking is where we can get some of the energy back!!
i want to add a flatwound brake that can release the energy again and regen into caps so you get most brake energy back
The go to person for these thing. Thanks Tom. Thought of a couple of factors that could influence the efficiency. First is speed. E-bikes here as standard are limited to 25 km/h compared to 35 mph? So both pedal assistance and reduced drag would improve overall efficiency at lower speed and, thus slowing down with regerative breaks would give back relatively more distance. Then there’s using a dedicated breaking generator versus the same driving motor (say front wheel) to improve energy recovery. Any thoughts? I could not tell how quick the braking was at say a low speed of 20 km/h. It looked rather too gradual,so may need additionnel friction braking for safety.
You could use a pelletier and a heat sync in the battery to use the heat to produce electricity
Thanks for the practical test of the system! In addition it would be nice to see how you build the regenerative braking as well :)
this is just top quality content man you're on fire. keep it up :)
I agree with you Dr Müller
Have you done tests with long hill climbs like 500-1000m of elevation? I’m thinking the efficiency might be higher for those, but curious by how much...
Thanks, I think you have vindicated my choice of a non- regenerative e-bike for use in my flat area - the Fens!
You should try a super cap as a load for the breaking. Regenerative breaking works if you do it at peak efficiency point. I doubt the batteries are sinking the current the motor could be generating.
Its all about matching impedances. A super cap is really low impedance so the idea is you charge the cap as much as posible and then dump the energy back to the batteries with a boost circuit.
The more current the more breaking.
You could get an idea of what you need by checking the voltaje open circuit at full speed and then the short circuit current at full speed. Do the math and that should give you your equivalent series impedance. Your load should always match that to have the most power transfered.
This is what I am looking for, a super cap to even out up and down hill rides. I don´t want an electric bike I just like to store the energy wasted if I need to break and use it shortly after. A flywheel might do it but I guess it will be to heavy.
I second that motion. But I'dd add a voltage measuring capability accross the capacitor.
When the voltage gets near the capacitors rated voltage (which is often low for super-caps) then the excess current that couldn't be sunk by the now charged capacitor is dissipated as heat over a high wattage resistor.
The same principal easy to understand would be a 220uf 50v capacitor slowly being charged.
Over the cap, you'd connect a 47v zener.
The cap would absorb current until it got to the 47v, and then the other incoming power would be dissipated as heat in the zener - Protecting the capacitor from overvoltage, and allowing the continual un-interrupted load.
@@dancoulson6579
a zener alone would be impractical, but you could get a 30-35v zener and hook it to the gate of a mosfet (or several) with maybe a resistor to take some load off of the mosfet.
also theres no need to let the super cap charge at all, as the more voltaje it has the less difference between the two and so less current being sunk. you should really keep it as low as the boost circuit allows.
The zener thing could work as a sort of protection for the super cap in case it were to "saturate", it should still give you some magnetic breaking though it would be wasting the energy as heat.
if anything the boost converter and the voltage limiter should be capable of handling more power than what you would or could generate typically un average. say a slow breaking or a sudden one where youd be dumping all the kinetic energy from the bike and your body into the regenerative break, to avoid saturating the cap.
also i wouldnt use a normal high voltage cap as they have really high esr compared to super caps. you need energy density not voltage tolerance.
sure 50v makes all easier but it will be huge in comparison and your boost converter would also be bigger becase it would saturate much faster so it would basically have to be as powerful as your motor breaking at which point you might as well do away with the cap and just use the boost converter as a break.
if resistance is constan then the cap at 0v will sink twice as much current as when it reaches 25v (assuming 50v open circuit), and that means half the breaking force, thats why it doesnt matter as much the voltage of the cap as its the impedance.
like how a 100C 1Ah lipo cell can do 100A because its really low impedance even though its only 3.7v.
As always. Great video and well done tests and explanations :)
Tom you are right -15% is the maximum of energy which can be returned by ebike regen during the most agressive riding like yours. With other kind of riding the regen will drop till 2-3% and even till zero.
I checked this on hillclimbing : spent 99Wh and returened 16Wh, i.e. the same 15%. During another long usuall riding including some hillclimbing I gained only 5%.
I deleted my previose post :-) .
2:14 R2D2? Is that you?
Only 15% for a hill ? At a low speed air friction remains low and I expected way more while the majority of the energy is "stored" in potential energy.
there is a lot more at play than just air resistance, wheel bearings, tire friction, bumps in the road, to name a few
@@justinc2633 Near 50 km/h the air drag is most of energy losses, with my ebike from 40 to 50 km/h the steady required power is almost 2 times higher!
@@IncroyablesExperiences well he does seem to try to be a bit conservative in spending at times perhaps your model was a newer higher grade?
@@IncroyablesExperiences where can i find curves for drag?
Awesome! I was wondering about this myself when riding my friends boosted board, which also has RB. Interesting video Tom :D
Thanks Gavin!
First of all I just want to say that I really like your videos!
I believe that an error is made when comparing the efficiency of the regen braking accelerating and going up a hill. Pounder that you would go up the hill very very slowly, and then go back down very slowly with the regen breaking, then the wasted energy from drag would be minimal. On the other hand, if you would go up the hill a great speed and down again fast because the regen breaking wasn't powerful enough for the slope, then you would waste more energy to drag. Therefor, I think it is wrong to suppose that the 15,7% would be true for the slope example.
Another point is that in theory you would approach your final speed as a limit when time goes to infinity so you would have better efficiency if you accelerated to 25mph instead of 35.
you would deffo save your brakes as you are not using friction pads but rather using electricity and magnetism to slow you down so there is no grinding or scraping sound of the brakes, i think this would be very useful in a long downhill where you dont want to go full speed as its too slippery and rocky but dont want to hold brakes constantly
You can use the front wheel to regenerate energy all the time you moving, just add another motor in the front wheel and another battery, the front wheel will collect energy to the new battery and the new battery will support the other one you just need to control it if it is charging or supporting , then you bike will travel double the distance that you have now
I think you should use this braking system in both of the brakes it will increase the efficiency by 30% if u r using on one . Another thing is if u use an another motor in front wheel so while the vehicle will move that motor will also generate energy while moving along with the wheel. I hope i am correct.
yea i never expected the region to increase the range much. the big thing is the reduced damage to the breaks
It's much more efficient to coast than it is to brake, a bicycle and a rider is a small mass, there is very little momentum to recover any useful amount of energy from. You may see a difference in charge after a 200 mile descent, but you'd never make it back home again. Best to just have a freewheeling motor so you can still pedal home if necessary.
if you added a cap bank you would recover more of the energy from braking.
Well, usually in the RB, you specify on the controller how many amps should be put back in batteries while RB. There is no way you can return electricity back to the batteries the same way you draw electricity out of them. Charging is always slower. So fo safety reasons in order not to burn your batteries, usually you specify about half or less than the actuall amperes you draw out of them. This is also one of the main reasons Rb is so low apart from thermal losses of course
Peddal assist would change this dramatically....you would use far less battery and gain more back ( as a percentage ) when riding down the same hill. The more you rely on your pedaling the higher percentage return on the downhill.
Very interesting! We’re adding solar cells to our electric scooter (Model M365) which will be utilized as a range extender. We noticed that the phone app which communicates with our scooter (Mi Home app) shows an energy consumption as low as 50 watts when traveling at a constant speed of approximately 5 to 7 miles per hour. We are presently using the SunPower monocrystalline solar cells, rated at 160 Watts (See Elfeland SP-22 Folding 160 Watt Solar Panels). If these cells can produce half of their rated output we may have a constant input (while the sun is shining) of power - acting as a range extender for the EV scooter. The EV scooter has a 250 watt motor, and a 36 Volt Lithium battery. We purchased a Genasun Boost MPPT Charger which will add power to our battery, from the solar cells. Marissa Muller here on you tube utilized 240 watts of SunPower solar cells to power her 250 watt electric bike across the country, permitting the solar cells to do 60 to 70 percent of the work, while she add the 30% to 40% via peddling the bike. Marissa notes she chose to limit the sunlight’s contribution of power so that she could get some exercise and thus help improve overall health. It’s a story that inspired us and so we too are studying how we can allow nature to help power our electric scooter.
Very nice Ebike by the way! Great video!
Maxwell Super caps may be more efficient for fast charging/discharging. I think Tesla bought Maxwell, so I'm not sure if you can purchase them like before.
What I know about cars is that most of the energy used is from pushing air out of the way. It's probably the same for bicycles. So I think you are incorrect about hills. (Shell used to run a modified VW between LA and Phoenix and back to claim their gas was more efficient. The mountains along the route was most of the trick.) And I have a somewhat relevant test. I drove my Prius up into the mountains using 'Worried About Global Warming Mode", or 'extra wimpy.' Eco Mode+ gentle accelerating, not going too fast... I averaged 99 mpg for the trip. (Normal for me in my Pri-V is 39-43 mpg). I live in Los Angeles, we have 2000 foot mountains in the city, in less than 40 miles on roads I can be at 5000 feet. So with mountains regenerative braking on a bicycle could be not just practical but amazing. If I had a bicycle that could go 40 miles on a charge with pedal assist, then I'm sure that I could get up to Mt Wilson where my charge would be empty, and then recharge going down hill regaining a significant charge to go the last 20 flat miles home. The calc for the downhill section would be 20 miles at a reasonable speed 15-20 mph, full charging along the way. (I don't know how you're regenerating power, but with a generator connected to the wheels. I'd still be going 15-20 mph, it's that steep.) This particular highway, the Angeles Crest, is downhill 99.5% of the way for about 100 miles. The portion discussed above is even steeper. The last 18.6 miles has a 5200 feet elevation gain. (La Canada to Mt Wilson). What would also work is knowing at what battery level I should turn around to have enough juice left to make up the 100' hill to my house.
Will Nettles Bikes are much worse than cars for the percentage of energy lost to wind resistance - 75-90%.
Stupid air! :)
Wow! That's mostly bad news, but mountains with electric assist should be even better than in a car. I don't have an e-bike yet, but surely someone in Southern California does, get in touch, I've already got the experiment designed.
It's not necessary to know the entire LA metropolitan area to understand conservation of energy. When you ascend any hill, and have an entirely empty battery at the top it, the range you will have at the (any) bottom is only dependent on regen efficiency and power lost during ascend+descend due to drag (wind and rolling). Tom does a great job explaining the physics, but to be honest it seems to me high-school level. The ingenious bit comes from actually making it work and providing some figures. Probably you could model any situation quite well knowing wind and drag resistance functions/curves, motor power efficiency curve and regen efficiency curve. I suspect the motor+esc won't be as efficient at regen at every rpm and regen power.
Yes, but.... Mt Wilson (5700 feet, La Canada is 1200 or 1400 feet altitude. I don't understand how the 5200 foot gain was arrived at. At most 4500'. ( I was nearby today at 1100 feet) This is significantly different from UK hills. I think it's the perfect laboratory for regenerative E-bikes. What you say about speeds, gears, etc.. are all essential controls. Same bike, same speed, same rider, same weather..... But I think 19 miles and 4500' elevation gain will clearly delineate what's air/road/mechanism resistance and what is gained potential energy. The one factor that should make this a terrific route for a re-gen e-bike, is of course the rider is pedaling. It will look like 100% efficiency, but only because the rider is pedaling. Btw the trip down requires braking because it is steep. PE ( I don't like using K) = m*g*m, so let's say bicycle and rider are 80kg* 1500m*9.8 m/s/s= 1,176,000 joules. 36 v x 15,000 mAh battery - 1,944,000 joules ( I looked this up) So. 1. Depending on the rider's contribution the battery alone cannot possibly get him/her to the top of Mt Wilson. 1.1 versus 1.9 is about half, all the frictions and efficiencies have to be very significant, and rider's contribution may be known amount--but I don't know it. And clearly this battery can not possibly be fully charged by the downhill part. (Unless perhaps the rider carries some rocks down, which is quite feasible). I wish I knew more, and had an e-bike. Coasting downhill with a generator and uncharged battery would quickly answer a lot of questions. But 5%? I think it will be significantly more. My Prius, and the Shell gas test, clearly demonstrate that. (Also I don't mind Tom using high school physics, I do think he's missed a few things. As a teacher I've been there too. Just go over the same thing a dozen times and suddenly it becomes not only clear, but obvious. (You really can explain everything in physics to a three year old, just not all at once.) And I do think being in England Tom doesn't have the opportunity our local mountains provide. Of course he's investigating a practical re-gen solution.
wow that is a fabulous bit of research.I have been telling people that regen is not worth it.Partly because I don't believe the batteries can absorb the regenerated power fast enough.They put out say 15 amps on acceleration but cannot absorb 15 amps on recharge or can they?
Well in an electric car you wouldn't even think about pushing it up a hill and then recharging the batteries with regen braking. With a bike you can choose to use your muscles to get you up the hill and use the regen braking downhill to charge the battery if it's getting low. This deserves to be considered or atleast mentioned.
the thing is if you're using pedal assist, the regen would probably keep the battery alive for a lot more time. cause some of the kinetic energy came from you pedaling.
I dont know why they have not come up with a friction mounted generator.One that is separate from the motor but connected to the battery LIKE THE OLD BIKE LIGHT wheel generator's to charge the battery .It could be powered by the rim front or rear in some fashion and regulated for regeneration and always in the generation mode while in motion unless disabled by a lever or switch.A drive system could be developed based on the rim or hub.It also could be incorporated with regenerative braking.
At 35 mph, 3.59 watt hours of energy is probably needed every 10-15 seconds just to overcome drag, so almost half of your kinetic energy was lost to the wind. Also charging efficiency is likely less at high currents. You could reduce this by wearing skin tight clothing, but the best way to test how efficient regenerative braking can be is to test going up and down a hill at speeds below 10 mph - not too steep that your motor struggles.
thanks, a good factual presentation... don't stop now
When calculating rb efficiency you should only consider the energy being used to accelerate and decelerate the total mass and disregard wind resistance etc as the energy recovered by the braking is only harvesting from the kinetic energy available(+ potential lost if going down hill or - pe gained if going up hill). Energy lost due to drag etc will be a function of acceleration as this will determine how long it took to reach terminal speed an so will be greater for slower acceleration and this means that all the energy put in cannot be available for braking recovery. Only the kinetic available at the instant of applying the brakes will be available and a great deal will still be lost due to drag and conversion losses during braking. You should find that your actual energy recovery from the available energy is much higher than it first appears.
It would be interesting to see how much power rolling resistance consumes in general and during regenerative braking. If you check how many watts of power your motor draws at a few velocities, you could determine how much power is required to overcome rolling resistance (or overall resistance due to air drag) and interpolate a curve to get the power dissipated by rolling resistance at any speed. From there you can get a Watt-hour approximation for kinetic energy lost to rolling resistance during regenerative braking. You could also find out what range your system should have at different velocities and what the rolling resistance coefficient is for the road with respect to your bike's tires.
Infinion Air resistance alone, will account for at least 80% of the energy used. Bikes are, ridiculously un-aerodynamic.
hope near future Tesla owner - do you think that ‘strategic coasting’ is close to regenerative braking?
(current BMW can disconnect transmission in its ECO mode)
Wondering if Tesla rapid/easy user controlled no-regen can increase distance more than fixed amount of regen
for city usage which necessitates more stop and go figure it would probably extend range a bit more, especially with pedal assist
all you need is a low force regen brake and to get a dog. The dog will pull you and charge your battery. You won't ever need to charge it at home.
What happens if it goes 88mph tho? ;-)
Have you not considering fitting a device/programming the arduino to maintain a constant speed above an adjustable single value or between 2 adjustable values (for example 25mph and 30mph)?
You could have the motor only run to maintain a speed above or between 2 set values values and therefore you could potentially be saving battery power that may extend battery life and increase distance. You'd have to calculate for spin up/spin down when the motor runs but that together with the regenerative braking could increase the possible travel distance.
EDIT: After more thought, this idea would be like a thermostat on a heating system that only turns on the heating when the heat drops but in this context, the heat is the speed.
It could all just be a large game of minimalist giving and taking but efficiency tests are always good things to try and look at the results.
Its not equal, even in theory, because you would include the rolling resistance of the tyres, bearings etc as losses, as well as the motor & controller you are using in the calculation.
i have a 48v e bike,rear wheel drive, sometimes i think front wheel drive hub would be more fun, because you could wheelspin everywhere in the mud.
Whilst Regen braking is good in theory, surly some folks must have tried putting another light weight motor in the system to power up battery whilst in motion yes?
Maybe RB would be more useful in an urban commute setting with lots of red lights (aka repeated braking).
well done Tom. good vids, a quick question- could a front dynamo hub be wired into that vesc 6 thing of yours and put a sort of continual small trickle charge into the battery and would it be worth it?
You traveled 100m which is work to move 100kg of mass out of the electric energy that you used and didn’t recover
Could you please show the wiring that allows the regen braking? Lots of people like me on the internet trying to figure out how to wire their hub motor kits that have "Low Brake" feature, but no wiring diagram. I agree with other comments that point out the excellent braking quality of regen brakes, especially the anti-lock nature. To optimize... Front wheel hub motor! Add an ultracapacitor... That is my dream bike!
usually, regenerative braking isn't put up to 100% braking, it should slow you down alot better at somewhere around 80 or 90% whereas its usually as 30-50%
This is the reason why it makes sense to add a freewheel to electric drive. :-) And it was also one of the reasons why my electric motorcycle didn't have regenerative braking.
I really like the simplicity of your project. I am seriously thinking of building another e-bike - mainly because I have almost free access to VESCs (I designed a custom - smaller version). Thank you very much for sharing all your files and so many details of the project!
I was wondering about this too, I wonder how freewheel would compare in efficiency to regen in real use, I suspect that it would do very well in hilly terrain where you can just build speed on the down hill and if you peddle assist. Also, if you run out of range, I wonder how much peddling resistance the electric motor provides that would be eliminated with a free wheel, seems a lot to push that belt and motor along with the bike.
great info and content Tom. if you dont mind...How well does that gear ratio and electric motor handle the uphills around here (i'm from st albans and also went to herts uni). i was contemplating adapting or buying an electric motor/bike for my hybrid to take me to luton and back each day for my commute but i've seen a few of the electric scooter vids on youtube not really like that sort of gradient and struggle to power the uphills.
Seems to me like if you used better battery management (in software, probably) to control the rate of regen, you could probably significantly boost the efficiency of the recharging by controlling the current rate of recharging. Perhaps an intermediary supercapacitor bank and some MOSFETs?
For LiPo batteries, the safe rate of charge is much lower that the rate of discharge - by an order of magnitude or two (10-100x).
Without knowing the details of your power system, and for the sake of the argument, I'll guess a safe discharge rate of 1000W and a charge rate of 100W. You would therefore have to decelerate very slowly indeed to avoid having to 'throw away' power to avoid damaging your batteries. My guess that most of the blame for the very poor efficiency scores in your first, flat test were due to this effect.
I'm interested in the use of a relatively small bank of super-capacitors to rapidly absorb braking energy, then safely mete it out to the main batteries. What are your thoughts?
They should be able to handle it for short bursts where there's not enough time to build up heat etc.
Just wondering, in the kinetic energy calculations, you made it seem quite simple. I was just wondering whether you considered the rotational kinetic energy in addition to your translational(linear) kinetic energy. If you were to leave this additional energy out it would make your predictions for the energy regained artificially high. Just wondering.
Great vedeo. Would like to know energy lost by wind. What was the power at constant speed? Also... the efficiency will change depending on acceleration rate.
I know this is kind of redundant now 🙃but couldn't we just put multiple motors in series regenerative breaking and super capacitor to maximized the output also storage that way is just a extra electric generator to power basically anything with inverter 😶
I will have to try your test with my faired recumbent bicycle equipped with a "Copenhagen Wheel" similarly bluetooth enabled. It uses the bluetooth for all control functions, except for pedal input responses, so I have to pedal slowly backwards to enable regen.
One factor I suppose you did consider was to begin your test with a battery less than fully charged so that regeneration voltage differentials could effectively deliver energy to the pack.
I think the Tesla formulation, beside good aerodynamics, is to have a large enough battery and a motor of sufficient power to do regen quite agressively. My Prius does use regen as best it can, but an 8 mile long mountain down hill route I do quite regularly fills the battery to its software governed 80% level in just a few miles, and then it has no place to store the potential/kinetic energy.The big motor/big battery Tesla can probably use much more of that potential energy returning from its climb up.
I was surprised that your E-bike was capable of 35mph. I thought most european e-bikes cut out at 25 kph, and here in the US the legal assist top speed is 20mph(32kph) so I cannot duplicate your top speed test. I do know that my bike is much better aerodynamically than yours.
This is a home made bike, top speed need not apply ;)
He could have gone even faster with a lower gear ratio!
i think regenerative braking only will save your brakes over time, i like this vid, it showed me lots of interesting stuff and if its worth getting regenerative braking for a bike if i go electric.If you were to pedal while regen is on would you recharge your battery
Excellent description and analysis. And lovely English countryside. Where is this?
Given that the distance travelled is 69 miles and the energy discharged is 420.0Wh at what percentage can we calculate that the CNC finished?
+JusticeFPV HAHAHAHA soon ;)
As a hipster, I clearly noticed there were some codewords in this conversation.
It roughly translates to ; Tom, would you like to go fight some eastern-european kangaroos in Japan? Not sure what it means though.
RenzVC Good dective work, we are indeed nearing kangaroo wrestling season 😉😂
That's great. I have a high power kit 52v 40 amps. Can go pretty fast. My city has no bike lanes so it's safer to go faster, when riding alongside the cars. I've been using the regen brake for the main purpose of assisting the brakes when braking from high speed. It's great. I don't even care about the battery regen, just the added brake power is awesome. I wonder why it's not more popular, some people don't even know their controllers are capable of regen braking. It's a great safety feature.
I've been wondering if there is a way to use the brake levers sensor to make a brake tail light for the bike. Any ideas?
Thanks.
Would be particularly useful in a city where you’ve got lots of junctions and traffic lights.
Is your regen variable or fixed? On my Chevy Volt, sometimes if i max regen brake, I get a burst of high generation for a very short period, vs if I regen slowly over a longer distance, seems to be better.
Just curious, great experiment!
Intresting... I wonder if there could be a way to charge the battery while riding? Maybe a setting that that you could set to like 10% or 5% load can be applied to charge the battery?