Pulse and Glide results

[itchynackers]

I have a e-bike kit 2807 front motor with 36V sla's and have read all the topics here on "pulse & glide". There is certainly confusion as to whether or not it works. Since I've been tracking data (along a 4.86mile course in non-windy condtions), I have concluded that for my setup (and I suspect most others' setups) it is not more efficient. These were my conditions to see whether or not the technique was more efficient: 1) same route, 2) similar weather conditions, 3) same average speed through the course (same arrival time at destination).

Summary:
Pulsing to max speed (whatever it is) and gliding to a lower speed, then repeating, is more efficient that maintaining the max speed.

However, if you intend to arrive at the same time to your destination, then simply using a constant throttle to maintain your speed is more efficient (P&G technique uses approximately 13% more energy).

This may seem intuitively obvious to some, but apparently not others. Take it for what it is worth.

Thanks,

Adam

 

[dogman dan]

Pulse and glide works better with gearmotors that freewheel. I find the most efficient way to use a direct drive is to slow down to a desired speed that allows you to pedal assist. Set the trottle to for instance 18 mph, and then pedal to 19 or 20. The watts used will drop like crazy, but it will still take little effort to go 20 mph. As things get harder, like wind or a hill, throttle up, but keep the speed the same. I can nearly double my range by going 15 mph and pedaling.

 

[swbluto]

Now the question is... is it more efficient with free-wheeling motors? (Geared hub motors, non-hubs set up with a freewheel, etc.)

The hypothesis is since you're not incurring no-load losses from the motor when you're freewheeling, you'll be more efficient by "turning off" the no-load losses for a percentage of your ride. I'm guessing the ideal would be to accelerate fast to minimize the motor time window, and then glide for a long time. But, the heat is i^2, so more acceleration = significantly more heat, so there'd be an optimal value somewhere or maybe not...

 

[John in CR]

Adam thanks for going to the trouble and sharing the results. Were they virtually the same, or was pulse and glide noticeably less efficient?

I ask because if the results were virtually the same, then, assuming I'm thinking this through correctly, that pulse and glide is better in terms of motor/controller efficiency, but you just give the gain back to the wind due to its geometric relationship.

 

[spinningmagnets]

This is very useful info John, but I think the data "might be" measurably different if the e-bike had a very smooth and optimally shaped velo fairing.

That being said, I have no plans to build/buy a smooth teardrop velo-fairing,...so the point is moot in that "real world" you insist on living in. Don't be annoyed at me posting this, its just the beer talking...(Ron searches for that half of a spicy burrito he set down somehwere...)

 

[swbluto]

This is very useful info John, but I think the data "might be" measurably different if the e-bike had a very smooth and optimally shaped velo fairing.

That being said, I have no plans to build/buy a smooth teardrop velo-fairing,...so the point is moot in that "real world" you insist on living in. Don't be annoyed at me posting this, its just the beer talking...(Ron searches for that half of a spicy burrito he set down somehwere...)

Since the drag coefficient and frontal area are degree one constants in the drag equation, you'd see a reduction in the difference of the results proportional to the difference of the effective drag coefficient (drag coefficient * frontal area). So the difference would still be around 13%, with some minor difference due to lower electronics losses, but the losses in both cases would be noticeably less so the actual difference would be less, though the percentage difference wouldn't be much different.

 

[John in CR]

This is very useful info John, but I think the data "might be" measurably different if the e-bike had a very smooth and optimally shaped velo fairing.

That being said, I have no plans to build/buy a smooth teardrop velo-fairing,...so the point is moot in that "real world" you insist on living in. Don't be annoyed at me posting this, its just the beer talking...(Ron searches for that half of a spicy burrito he set down somehwere...)

SpinningMags, you definitely lost me. Please excuse me while I plug myself back into the matrix.

 

[icecube57]

The GM motor I had coasted extremely well and there would often be times I would ride with a friend and not pedal for quite sometime ... 1/4 mile or more and I was just coasting and slowly overtaking him by just coasting and he was putting genuine effort pedaling along with his motor to keep up with me. I have more weight personally and in my battery pack and rig therefore more inertia to aide in this accomplishment. But i think that will change with this X5 which has way more cogging that I can actually feel. I can feel every pole transition with this 5303 at low speed and pulse and glide defintely wont work on this motor.

I have cruise control enable on my controller and i like to play with it sometimes. I will barely turn my throttle and hold it to where its using like 30-40w at like 15mph and pedal along. When I stop pedaling it doesnt quite maintain the speed but it almost gives the effect that you have like a tailwind pushing you and you can coast an incredibly long time. Even better stop pedaling and have it pull you along at a slow 10ish mph on level ground. No real pulling power just a very subtle mild.... well very weak assist. I use this alot when im hypermiling.... for no apparent reason with a 1.4kw pack lol. It just reduces pedaling effort drastically and effectively.

It has been noted on the forum that if you disconnect you phase you coast better/farther than leaving the controller plugged in but that when you are going on non power rides and want it to feel more like a regular bike.

 

[madact]

Crudely put, to double the power at the same speed, you have to double the torque. In an idealised model, to double the torque in an electric motor, you have to double the current - this is close enough to the practical reality as makes little difference. Now if you double the current, the resistive losses are increased according to P = I^2 * R, so you actually quadruple the power lost to heat by doubling the power.

Say for example you have scheme where you deliver 100W continuously to the wheel, with 25W heat losses - vs delivering 200W to the wheel for half of the time (pulse-glide) and switched off half the time, with (according to the previous relationship) 100W heat losses while it's running. That gives the pulse and glide an average power delivery of 100W and average heat losses of 50W, as opposed to 100W delivered and 25W losses for the continuous case. 150W / 125W = 120%, so in this example, the pulse and glide technique uses 20% more energy than continuous.

Obviously, the exact figures will vary with the setup and duty cycle (on time vs off time), but a 13% disadvantage to pulse-glide in a real-world test is very much expected.

In general, continuous running at low power is the most efficient way to run an electric motor (unless you are using a very crude resistive controller on a DC motor, like old school R/C cars). Note also that maximum efficiency is not achieved at the same speed at a given voltage as maximum power.

The pulse-glide technique is, however, very effective for small internal combustion engines (I've competed in fuel efficiency competitions with 5cc to 50cc engines, and I can tell you it works), on account of the very different power and efficiency curves you get with those devices. The lesson here is that what's good for the internal combustion engine isn't necessarily good for the electric.

 

[Tiberius]

In general, continuous running at low power is the most efficient way to run an electric motor (unless you are using a very crude resistive controller on a DC motor, like old school R/C cars). Note also that maximum efficiency is not achieved at the same speed at a given voltage as maximum power.

The pulse-glide technique is, however, very effective for small internal combustion engines (I've competed in fuel efficiency competitions with 5cc to 50cc engines, and I can tell you it works), on account of the very different power and efficiency curves you get with those devices. The lesson here is that what's good for the internal combustion engine isn't necessarily good for the electric.

Exactly.

Its a technique developed for ICE engines, because they can be more efficient at full throttle than partial throttle.

But it's different for electric motors. The efficiency of an electric motor goes up with speed and down with throttle. So, assuming the controller is relatively efficient, P&G isn't attractive because in one part of the cycle you are running the motor at full throttle. There is also the question of how much you allow the speed to vary over the cycle. If its a large variation, then you risk entering two inefficient combinations - full throttle and low speed at one end and high aerodynamic drag at the other.

Constant speed looks like a better bet for electric vehicles.

Nick

 

[itchynackers]

Adam thanks for going to the trouble and sharing the results. Were they virtually the same, or was pulse and glide noticeably less efficient?

John, it was noticeable, in that, the results from my watt-meter were fairly consistent. The P&G always used more energy (all else being as equal as practicable).

I've basically been tracking all of my outings to find the most efficient way to maximize unassisted range. The slowest I can go (comfortably without losing balance) is 10mph. So, at 10mph the motor is most efficient. Anything faster, without exception has been less efficient. The course I use has very little grade change and pavement slopes no more than 1%. The route was re-paved with a blacktop surface this year, and is very smooth. This is as close to minimizing other variables as I can get.

Adam

 

[John in CR]

Great, that's one thing I can scratch off the need to try list. While I appreciate what the electronics guys are say above, that analysis leaves out the fact that our max speed generally coincides in the region of maximum efficiency for our motors, so for riding at a substantially lower average speed, I thought P&G might offer a benefit for the same reasons it does for ICE. Apparently the electric motor efficiency curve is too flat, so I can drop the idea.

Again thanks for the effort, because that P&G stuff is a royal pain. I tried it once for short errand ride, so I tip my hat to you for the follow through and sharing.

John

 

[madact]

This page has a pretty good explanation of the differences and power curves: http://www.fastelectrics.com/elecmotorbasics.htm - I might go into the theoretical basis of why these curve look the way they do when I have a bit more time.

The electric motor performs best at a high rpm with low load and low throttle, so your best bet to get the motor operating at it's most efficient is when you're (more or less) WOT on the flat and hunched down into a tuck of whatever sort you can manage on your steed of choice. Of course, Amps / km may suffer from wind resistance, and whether you hit the "sweet spot" at WOT on the flat will depend on gearing, air resistance and motor size, but if you're in the mood for testing, WOT vs low throttle would be a good one. Note however that WOT is NOT your friend for efficiency while accelerating - kick off with the pedals, start ramping up the throttle at 1/2 top speed and hit WOT at close to your top speed, and you should be good. I'd love to see results of WOT vs low-speed with this riding style for acceleration...

 

[liveforphysics]

Its a technique developed for ICE engines, because they can be more efficient at full throttle than partial throttle.

Bingo.

An ICE engine always has the lowest BSFC's (brake specific fuel consumption) at the RPM and throttle position that it makes peak torque. For performance engines, this means high RPM (~7,500-11,000rpm) at full throttle is the most efficient conversion of fuel into mechanical power. For most commuter engines, this means low-mid RPM (~2,500-4,500rpm) at full throttle is the most efficient conversion of fuel into mechanical power. If your engine were sized to be just powerful enough to maintain the speed you desire, then full throttle at torque peak would be the optimal method to use for peak fuel economy. However, they are sized to be more powerful (to enable hill climbing, rapid acceleration to freeway speed, fun, etc.) This means if you tried to operate at this point, the car would accelerate past the speed you wish to travel (additional wind loss, laws, etc.) This is the reason for the pulse and glide, the glide point is used to regulate speed that would otherwise creep up to undesired levels during the full throttle acceleration.

For an electric motor, particularly something with lead batteries (Peukert effect), the least power losses are incurred by simply maintaining a constant speed level with a constant throttle position. However, as John mentions, this can do ugly things to controllers if it means trying to run a high speed capable setup at low speeds.