Regenerative braking confusion

[Pentoad]

I had an electric bike with a brushless 3 phase hub motor and a controller which didn't have regen. Despite this, when the bike went down hills and the motor spun faster than the battery voltage could propel it, I noticed the battery voltage increasing and it was difficult to make the bike go any faster. That was due to the motor producing a greater voltage than the battery and the body diodes of the mosfets in the controller acting as a bridge rectifier and putting the energy back into the battery.

How is it then that an electric car can recharge itself even at lower speeds? I have studied electric car inverters quite a bit and they do not have a dc boost converter in them which would boost the low voltage generated by the motor when the car is at a low speed and step it up to charge the battery.

[richard.cs]

I have studied electric car inverters quite a bit and they do not have a dc boost converter in them which would boost the low voltage generated by the motor when the car is at a low speed and step it up to charge the battery.

Oh but they do, it is just not a separate device. The motor when spinning is a voltage source in series with an inductance (the windings). The MOSFETs/IGBTs/whatever that are already part of the driver can be switched in such a way as to short out a winding and the series voltage builds up current in the inductance of the motor winding (whilst also producing braking torque). When the switch is opened the voltage across the inductance rises and the stored energy is released at higher voltage. Exactly the behaviour or a buckboost converter. Yes the induced voltage from the motor is a.c., but the switching frequency is much higher such that you can ignore that inconvenient fact.

Consider turning on two bottom fets when  the motor is rotating, current flows and energy is stored in the inductance. Turn off one of the fets (the one sinking current!) and you get a positive voltage spike produced on its drain, and the top fet then conducts that into the supply (whether through the body diode or by being tuned on as a synchronous rectifier).

Edit: buck->boost

[langwadt]

I have studied electric car inverters quite a bit and they do not have a dc boost converter in them which would boost the low voltage generated by the motor when the car is at a low speed and step it up to charge the battery.

Oh but they do, it is just not a separate device. The motor when spinning is a voltage source in series with an inductance (the windings). The MOSFETs/IGBTs/whatever that are already part of the driver can be switched in such a way as to short out a winding and the series voltage builds up current in the inductance of the motor winding (whilst also producing braking torque). When the switch is opened the voltage across the inductance rises and the stored energy is released at higher voltage. Exactly the behaviour or a buck converter. Yes the induced voltage from the motor is a.c., but the switching frequency is much higher such that you can ignore that inconvenient fact.

Consider turning on two bottom fets when  the motor is rotating, current flows and energy is stored in the inductance. Turn off one of the fets (the one sinking current!) and you get a positive voltage spike produced on its drain, and the top fet then conducts that into the supply (whether through the body diode or by being tuned on as a synchronous rectifier).

you probably meant boost converter

it acts like a buck converter from battery to motor, and as a boost converter from motor to battery

[richard.cs]

Yes, that is what I meant. Thanks. :-)

[David Hess]

I think they rely on the motor's inductance to act as the boost inductor.  So each half bridge alternatively shorts the motor output and then directs it to the battery.

[max_torque]

A buck conveter looked at from the output side to the input side, ie backwards, looks just like a.........  yup, a boost converter!

"regen" for a brushless motor controller is effectively free, all that the controller has to do is to put on a lower average voltage (by turning down the duty cycle of it's pulse width modulation) than the back emf that motor motor is making at that moment, and the direction of current reverses. Instead of the battery feeding current to the motor, the motor feeds current to the battery, and the inductance of the motor phase windings is what drives that current into the battery against the higher potential difference.

Most cheap motor controllers for bikes deliberately don't do regen (they either just disable the gate firing when you lift off the throttle or they never let the PWM index fall to a value below the back emf) for two reasons:


1)  on low inertia, slow moving objects (like electric bikes) there really isn't much to gain from regen, it's better to just "freewheel" along at zero net torque, because regen has a loss associated with it
2) Cheap controllers tend to be poorly heat sinked and have low rated switching elements, not doing regen gives those elements a "rest" when you back off

[langwadt]

A buck conveter looked at from the output side to the input side, ie backwards, looks just like a.........  yup, a boost converter!

"regen" for a brushless motor controller is effectively free, all that the controller has to do is to put on a lower average voltage (by turning down the duty cycle of it's pulse width modulation) than the back emf that motor motor is making at that moment, and the direction of current reverses. Instead of the battery feeding current to the motor, the motor feeds current to the battery, and the inductance of the motor phase windings is what drives that current into the battery against the higher potential difference.

Most cheap motor controllers for bikes deliberately don't do regen (they either just disable the gate firing when you lift off the throttle or they never let the PWM index fall to a value below the back emf) for two reasons:


1)  on low inertia, slow moving objects (like electric bikes) there really isn't much to gain from regen, it's better to just "freewheel" along at zero net torque, because regen has a loss associated with it
2) Cheap controllers tend to be poorly heat sinked and have low rated switching elements, not doing regen gives those elements a "rest" when you back off

and you also add the complexity of staying with the safe limits of the battery charging

https://youtu.be/1pm1RtCuE3A he only got 3.5% extra range from regen, though with a lot of hills that could be different

[richard.cs]

In a very mountainous area there's an additional advantage to regenerative braking, your brakes don't catch fire. Most effective if when the battery gets full it dumps the energy to a shunt resistor rather than stopping regen.

My parents once encountered a man on a mountain road tying a tree branch to his bike to drag on the road and save his brakes.

[langwadt]

In a very mountainous area there's an additional advantage to regenerative braking, your brakes don't catch fire. Most effective if when the battery gets full it dumps the energy to a shunt resistor rather than stopping regen.

My parents once encountered a man on a mountain road tying a tree branch to his bike to drag on the road and save his brakes.

if you live on top of a hill you can charge your bike on the way down for the trip home

[David Hess]

In a very mountainous area there's an additional advantage to regenerative braking, your brakes don't catch fire. Most effective if when the battery gets full it dumps the energy to a shunt resistor rather than stopping regen.

Then it becomes *dynamic* braking which is ubiquitous in diesel-electric locomotives.  Sometimes it is the dynamic brakes which catch on fire though:

[Pentoad]

Thanks for clearing that one up, I would love to have a bike controller with regen, more for saving my brakes and the small amount of extra efficiency. I only ever charge to 4.08 per cell anyway so there would be loads of headroom to store extra power on a long downhill stretch. Does anyone know where I could get a regen controller?

[Jeroen3]

In a very mountainous area there's an additional advantage to regenerative braking, your brakes don't catch fire. Most effective if when the battery gets full it dumps the energy to a shunt resistor rather than stopping regen.

My parents once encountered a man on a mountain road tying a tree branch to his bike to drag on the road and save his brakes.

if you live on top of a hill you can charge your bike on the way down for the trip home

A 700 meter hill? It will be tough to get back!
There just isn't enough energy. You have to weigh more.
This guy did maths: https://gocarlite.com/electric-bicycle-regenerative-braking/
And also the explanation why even the drag torque of the system invalidates it's effectiveness. Even for braking.

Maybe you can replace one of your disk brakes for eddy-brakes? (ignoring prohibitive energy costs)

[Rerouter]

I used to ride my old DC motor scooter up a long consistent hill each weekend, 3km flat to the hill, 4km up the hill, then a wavey 2km to her place, it was at about 25% charge when I arrived at her place on average (1 light out of 4), Getting it back to the top of the hill on the return leg took it to near flat, then by the time I was at the base of the hill, was back to a single light, and able to use it the whole 3km back home.

So there can be benefits, but you really need a route that helps it, I personally do not remember how long lived the batteries where, too young.

[chickenHeadKnob]

Back in 2008 Justin Lemire-Elmore gave an engineering analysis of regen in this thread on the endless-sphere forum:
https://endless-sphere.com/forums/viewtopic.php?t=7891

It is an excellent analysis and worth reading.