Newbie question about regen braking

[Lu.Sochr]

Hello to all,

I am gladly joining to your forum, I've been reading it for some time, because so far I always did find very usable information from you guys for my problems:-) But now I can not find right answers, so I decide to post question myself.

I have e-bike with 901 kit (HBS motor), which is very great and recently I throw my heavy SLAs away and switch to 36V15Ah GM LiFePo4 pack which is also great:-) But I was not sure if regen braking is or is not damaging battery pack, so I rather turned regen off.

I now that regen in fact is not so much efficient in order of increasing range, but I very liked that braking effect, I was rarely using mechanical brakes on my bike when regen was turned on (regen was set only on 1% in magic controller), braking effect was very perceptible.

So finally my question is: does regen damaging battery pack or not? Pack has separate wires for charging and discharging, does not mind that current would go back over discharging wires instead of charging ones? And do anybody of you know how much amps this motor giving back during e-braking on 1% settings (continuous current is set to 22amps and peak 35amps).

And one last question: what is happening if I do not have sufficient speed during e-braking, so voltage from the motor is not higher than battery voltage, I believe in this situation amps could not go to the battery and are transformed to heat in controller, is it right? or does controller have some kind of step-up voltage converter???

I apologize for my not so perfect english (I am from Czech Republic), but this is the best I can do:-)))

I very looking forward to hearing your thoughts:-)

Lubos

[Leslie]

I cant answer you in full, but I can fill in some information gaps.

Regen goes back through the discharge wires.  The charge wire has a separate circuit on the BMS and uses much lighter components that is not suitable for high current.  Only use this charge wire for charging and its unwise to attempt modification to route regen through this wire and circuit.  One is likey to damage both controller and BMS trying to block regen say, with a diode and route it back to the pack throgh the charge wires to attempt utilizing HVC. 

There is battery protection within the controller although limited in its reliability..  It cuts the controller out if the voltages goes over you pack voltages.  Yuan Yao posted @60v ages ago, so I cant tell you this varies when you configure your controller via software at 36v or 24v.  This can push a 48v system up to 60v with high current using a BMS suited only for slower charging.  It might just cause a fully charged battery cell to go out of balance and maybe a cells voltages  could go over their specs.  Also avoid regen on a fully charged packs.

Regen varies from battery to battery, you get a little more regen amps on higher capacity batts per regen percent, And you will notice more amps on lower volt packs.  I had one internal controller on a 48v system with regen set at 85%, this controller died after some use.  Howevermy regen didn't seem over the top.  There could of been a weak connection I missed when beefing up the controller PCB.

Because regen is not constant non the GM controller, it inst as bad as say, a high rate charger @ 10amps for over an hour, but you may exceed the batteries peak limits occasionally. It also depends on how heavy you and your bike is.  If your SLA's survived your set regen for a considerable time, I could advise you that a similar back-resistance and charge current should be fine on a LiFePo4 pack.  SLA's do not take kindly to high rate charging or discharging compared to the LiFePo4 batteries.

Set regen percentages is not formal from one bike/rider, system, battery chemistry/voltage/capacity to another.  I would set it so both regen and mechanical brakes have to be used to just manage to lock the wheel to a skid, and Regen is not too obtrusive to the braking experience.  You can at best get 85% efficiency on braking, which is little return considering brakes are not so often used and regen only works for as little as 8% of your ride.

Regen has benefits in city runs where you do a lot of stopping and starting,  otherwise riding momentum out till you stop is much more efficient.
Regen is ultimately limited to your controller and BMS shunt,so there is some considerable protection.  The difference from motor regen output, to battery voltages is what dictates the amount of current flow. Youre lucky to deliver more than 1C and this is fine in small spurts.  High regen isnt practical to ride down hills on as it will slow the bike too much, but on huge steep hills where heavier longer braking time frames might become needed the higher percentages regen may become an issue.



Last but not least, you seem to be going towards a more custom design., Make sure there is no weaknesses in any of your connections, and maybe attempt to limit either the power cable distance between the pack and controller, or increase the wire gauge between your pack and controller as much as practical.  Where you have a voltage drop when you accelerate you can get voltage spikes before the controller shunt and BMS shunt during regen.  If there are any weak connections this may cause damage to your controller. Anything to reduce the resistance between your motor and pack will be beneficial and use the software to do power limitation.  Even adding another proper shunt to a controller and BMS may increase reliability to your system..  I never trust regen on 6 fet controllers that can not handle 82v pack, but this is my idea that voltage and current spikes exist on the lower end controllers or poorly soldered.controllers on the PCB connections. Because IMHO there is a considerable V drop in the phase wires after the 6 fets during acceleration and spikes during regen than there is on the 12 fet controllers. 

The more parallel fets, and the more you rely upon software, and larger shorter controller/BMS shunts, the lower the resistance between your pack and controller exists, this will ensure that the battery has a great effect via connections at ironing voltage spikes out during all of the PWM/Hall switching and regeneration operations..

I figured one solution out.  Which is to have the right brake lever working regen and motor cut off, and the left brake lever switch was just connected to the throttle signal wire and ground.  So in operation, one brake worked regen and motor cuttoff while the other just disabled the throttle, so the motor could not operate when either brake lever was activated.  One small issue is the brake grounded the throttle signal, did not take cruise control offline.  However the advantage was I could choose either to use regen or not while still having the safety of motor cut out on both brakes.

[Bikemad]

I figured one solution out.  Which is to have the right brake lever working regen and motor cut off, and the left brake lever switch was just connected to the throttle signal wire and ground.  So in operation, one brake worked regen and motor cuttoff while the other just disabled the throttle, so the motor could not operate when either brake lever was activated.  One small issue is the brake grounded the throttle signal, did not take cruise control offline.  However the advantage was I could choose either to use regen or not while still having the safety of motor cut out on both brakes.

Les, if the throttle signal and cruise wires were connected to the same brake switch through separate diodes, it should disengage the cruise mode as well. Because the cruise control does not engage at zero throttle, pulling the lever would hopefully not active the cruise function, even if the throttle was held wide open at the time.

A simple crossover switch could be used to switch regen on or off using this method so you could temporarily disable the regen for downhill braking on a fully charged battery.

Did you have a 330 ohm resistor in the throttle signal wire (as required for the Cycle Analyst), or simply short the throttle hall sensor output directly to ground?

Alan
 

[Lu.Sochr]

Thanks very much for your replies.

I am not sure If I understood everything you said because I am not familiar with what shunts and fets are, but so far I have understood that it is okay if regen current flows back to the battery pack via discharging wires (I had no intention of doing modifications that regen current would flow via charging wires).

So if regen current won't be higher than 1C (which is 15Amps in my case), I can use regen and not be afraid that lifespan of my battery pack will be significantly reduced, do I understand it right?

I believe that my wire gauge is sufficient and lenght of power wires as minimal as posibble, so this should not be a problem.

So only one question remains unanswered: I need to ride very fast in order to have voltage from motor higher than battery pack, but most of time during ride I do not have so much speed so regen voltage is almost always lower than battery pack voltage(according my turnigy watt meter), but regen braking is still occuring, so where does the current goes? I believe it cannot go to the battery pack because voltage needs to be higher than battery right?

So I think I should reverse connection of my wattmeter in order to find out how much amps is flowing back to the battery pack during regen and if it will be less than 15Amps, I will leave regen turned on and otherwise I will turn it off:-)

Thanks in advance for confirmation that I am doing things right:-)))

[Leslie]

To Alan

Yer I thought about a resistor but then I thought this is what the Hall does anyway, and most importantly the voltage + wire to the throttle is a tiny bit lower than the standard 5v rail as there is already a resistor on the throttle red wire in my controller and even the load of my voltage probes initiates a v drop. Providing I used the correct wires its all good,

Quite possibly and I am more sure about this, with the GM controllers, what I've done by shorting out the hall output, would 0V out the hall output signal initiating a hall failure code and beeps would occur. So id say add a standard diode to the brake switch 5v wire would be more advised, just enough to allow .8v through to the throttle signal via the brake switches. None of the 5v wiring are resistor unprotected on my controller and this ensures along with the resistance added by the hall sensors such ground shorts don't draw too much current and ruin the throttle hall or 5V rail regulators

May I ask where the supposed diodes go on your idea?  Before the switch and hall, on the output ground side on the switch and hall output.  This is all a little bit complicated to contemplate as to what is going to occur.  I had a long response to this idea but it all fell apart, but I am interested.  I get the cruise wont initiate at 0v throttle part but am confused as to how giving 5v from the the cruise wire to the controllers throttle signal wire in one image in mind of what you suggest is going to work?

[Leslie]

Thanks very much for your replies.



So if regen current won't be higher than 1C (which is 15Amps in my case), I can use regen and not be afraid that lifespan of my battery pack will be significantly reduced, do I understand it right?

I believe that my wire gauge is sufficient and lenght of power wires as minimal as posibble, so this should not be a problem.

So only one question remains unanswered: I need to ride very fast in order to have voltage from motor higher than battery pack, but most of time during ride I do not have so much speed so regen voltage is almost always lower than battery pack voltage(according my turnigy watt meter), but regen braking is still occuring, so where does the current goes? I believe it cannot go to the battery pack because voltage needs to be higher than battery right?

So I think I should reverse connection of my wattmeter in order to find out how much amps is flowing back to the battery pack during regen and if it will be less than 15Amps, I will leave regen turned on and otherwise I will turn it off:-)

Thanks in advance for confirmation that I am doing things right:-)))

Sorry I should of made it easier to read.

A motor rotating at an average speed can output very high volt spikes at low current if not kept down by the internal resistance of your battery or any components within your controller.  The more resistance you add between the motor phase wires the lower the volts go but more current is allowed to pass.  This called is loading an inductive device.  If you were to view this regen action on an oscilloscope with out a battery you would notice a high voltage sine wave.  At the top of the peak of the sine wave (highest volts) is less wide, and at the bottom (lower volts) the sine wave is much wider.
This is just an indication and means to imagine the workings so to say not a true indication of the current flow. So the more load you place upon an inductive output by using a battery or resistor the thicker current is at the expense of lower volts.  Regen voltages should only go as high as the pack allows for it.

Motors and generators, transformers, or anything that converts magnetism into electricity via a former coil, has the ability to transform momentum (volts) and torque (current) into voltage and current,  and transform a high voltage (speed), into lower voltage with high current / or / high current (torque), into  higher voltages with lower current.  One can indeed slow an ebike faster than it accelerates using regen alone.   

Reversing the watt meter may be a good idea.

But as an easy guide, without all that messing about, if your bike slows faster than it accelerates using regen alone, then this would not be a good sign.  With GM controllers you may experience the most torque at low mid to mid speed.

So feel the inertia differences at about 20kph..  If your experiencing a type of whiplash (lol) when using regen alone and acceleration is not so pronounced, your are getting more regen than you are discharging, it is easy to slow a bike faster than it accelerates using regen, thus transforming any average speed to very high current in a shorter braking time frame..  Make this inertia effect for regen less than the effect of acceleration and this should be safe for your battery.

[Bikemad]

May I ask where the supposed diodes go on your idea?  Before the switch and hall, on the output ground side on the switch and hall output.  This is all a little bit complicated to contemplate as to what is going to occur.  I had a long response to this idea but it all fell apart, but I am interested.  I get the cruise wont initiate at 0v throttle part but am confused as to how giving 5v from the the cruise wire to the controllers throttle signal wire in one image in mind of what you suggest is going to work?

Les,

A diode is fitted in series with each of the two signal wires (throttle signal and Cruise signal) and the brake switch contact.
Each wire has its own diode that will allow the current to flow to ground via the brake switch, thus preventing any back voltage or grounding via the other signal wire from affecting the independent operation of each circuit.

I'll try and post a diagram at a later date when I get a chance to draw one.

Alan
 

[Bikemad]

This is what I meant:



I have some scribbled down diagrams somewhere of a similar circuit that I devised for dual motors that would allow independent regen linked to each of the brake levers.

Both brake levers would instantly cut the throttle, but front and rear regen would only work with the respective brake levers.

Alan
 

[Lu.Sochr]

Hi to all again,

I would like to revive this older thread because I have some new knowledge about regen braking. I recently mounted speedict unit, which is able to measure reverse current (till that moment I only had turnigy wattmeter which measures only forward current) and I noticed that if I start to regen brake, current (between 10-20amps) begins to flow immediatly into the battery (so far everything OK). But after approx. one second current stop flowing back to the battery although regen is still one and bike is braking. So I wondered where does the energy go???

I believe it has to be burned in motor windings and controller. Am I wrong???could someone clarify that to me?

[e-lmer]

on my bike, when I go over 25mph the raven cuts out.

I suspect the Magic Pie has some over voltage protection builtin.

It also cuts out when I roll to slow.

[Lu.Sochr]

But in my case controller does not cut out. Braking still continues till I stop, only ackward thing is, that current is not flowing into the battery. According my temp sensor temperature of motor winding is rising durring braking, so that means energy is transformed into heat directly in windings. Why does not it flowing into the battery instead?