Topic open for Discussion. Burnt Out Controller with Regenerative Braking

[Perbear]

I was told by Li Ping that the BMS on his batteries will prevent regen braking to protect the battery from the higher than 5a regen current. I assumed that was also correct for the GM LiFePo4 batteries as well.

If you open a GM Lithium battery it has a BMS with a FET for charging and dual FET for discharge. These FETs have built-in protection diodes against reverse voltages. So during regen the discharge FET will conduct current to the charge FET via this diode. You will get around 0.7 V voltage drop over this diode (unless it is a schottky type which has much less). Then current flow until it exceeds the maximum allowed by the BMS. When the BMS senses that the charge current is higher than the limit it will open the charge FET to protect the cells from damage. Then with no load, the voltage generated will increase above the maximum ratings and the magic smoke will start leaking...

To make regen work on Lithium batteries, you need a BMS that is designed for regen, and use battery cells that is able to handle the maximum regen current.  In addition, you should also have a braking resistor as often used on AC variable speed drives. This resistor handles the regenerative load when the battery is full. I am currently designing my own BMS using the new Texas Instruments bq78PL114 BMS chip where I implement this function. My BMS is specially designed for regenerative ebikes/velomobiles so it has a built-in regulator for cycle lights as well (fixed voltage or fixed current for LED).

Maybe I can talk GM into producing and selling it (if it works as planned  ).

 

[rolf_w]

To make regen work on Lithium batteries, you need a BMS that is designed for regen, and use battery cells that is able to handle the maximum regen current.  In addition, you should also have a braking resistor as often used on AC variable speed drives. This resistor handles the regenerative load when the battery is full.

If I understand correctly and if the 2 discharging and the 1 charging FETs of the BMS are all connected to the same battery terminal then it would be possible to connect a braking resistor to the charging socket of the battery, as the charging FET of the BMS is switched on by the controller if the battery is fully charged? that would be an elegant solution to prevent the burning of the controller! r

[Perbear]

If I understand correctly and if the 2 discharging and the 1 charging FETs of the BMS are all connected to the same battery terminal then it would be possible to connect a braking resistor to the charging socket of the battery, as the charging FET of the BMS is switched on by the controller if the battery is fully charged? that would be an elegant solution to prevent the burning of the controller! r

It is a bit more complex than this, but yes, it is possible to have the resistor switched into the circuitry when the battery is full. The problem is that you also have freewheeling diodes in parallel with the FET's making matters a bit complex. In my BMS I have added a logical function that connects the resistor to the speed controller (ESC) only when the BMS turns of the charge and precharge FETs due to fully charged battery. To add such solution to an existing BMS without schematic is IMHO both difficult and dangerous.

[rolf_w]

The problem is that you also have freewheeling diodes in parallel with the FET's making matters a bit complex.

As far as I understand the circuits of the GM BMS and BAC there is never a condition where regenerative current is actively managed by the built in MOSFETs if the back emf is higher than the battery voltage - regen current is then always conducted by the free wheeling diodes and therefore regen current is completely uncontrolled in this condition!? The BLPMDC-motor always generates a back emf proportional to the rpm (with permanent excitation there is no control over emf). If the back emf is larger than the battery voltage all the freewheeling diodes start to conduct (the bridge in the controller and the twin discharge MOSFET in the BMS). The induced current which creates the braking torque in the motor (in this case generator) is the charging current of the battery. If the battery is not connected, there would be no load and no braking torque, thus the e-velo speeds up and the voltage would rise continuously in sync with the rpm. If the battery is full, there is no way to cut off the regen current other than disconnecting or stopping the generator/motor!
If the back emf is smaller than the battery voltage (braking at normal travel speeds) charging the battery is not obvious as the back emf is too low. How to switch the bridge so that it works (PWM ) can be found e.g. here http://www.4qdtec.com/bridge.html
r

[Leslie]

I thought of an idea.  An electronic switch that has large capacitors and keeps the controller engaged until the a breaker switches on the brake switches.  It only needs to happen in a couple of seconds,  This is by no means an idea that is a sellable device.

Ive had a couple of controllers burn out and learned to avoid it.  I need Regen for the braking capabilities because I drag some heavy SLA's and shopping around.  I make sure I install a big fat 50 amp fuse and place the switch where it wont be bumped.  It is a temptation to switch the bike off down hills on those controllers that need to be switched off to disable regen.  unlees you got the new one with issue solved just don't do it.

I learned the hard way.  Once you have a good controller going just leave it and it should live for a long time. 
Stop the bike before switching off and on.

[Leslie]

The transistor gate drivers and fets put a reverse phase signal in against the forward operating phase that can control the regen, and another fet is installed to milk the regen into the batt input. It is initiated when you decrease the throttle.  Well on my controller.


If you look at the gate drivers there are two per channel and both different and faced opposite, and one of these these could well be triggered by what look to be a nearby dual inverted opamps but the part numbers are beyond my spectacles.  When the motor outputs voltages above the input voltage and by the throttle decrease the opamp switches the reverse phase gate into gear. Possibly depletion mode drivers, so when the hall is on or off respectivly the gate driver is a depletion or normal causing the opposite to happen.  How this is works variable to the throttle I have no idea but is quite intellegent as you can controll the amount of regen with the throttle.

On the new controllers I like the idea of full braking regen on the brake levers but doesnt this lose variable braking and regen systems that the old controller had.  I guess you cant have your cake and eat it so to say.

I'll admit still haven’t analysed the fet bridge properly on my controllers as yet but I swear some half of the fets outputs go straight to the battery input.

The tvs diodes do filter the current into the right cables also to clamp spikes from inductive switching..  You can turn these hubs into a really good wind generator just by placing diodes on the phase wires.  But to rev up the current a reverse to phase input can really pull out the juice at lower RPM.