Braking: regenerative or not?

[DaveThomasPilot]

Chris,

If there's no current, there's no braking.  The motor will freewheel.

If the motor is regenerating (has non-zero RPM) and  has no load (battery fully charged) the EMF (voltage) will increase until current DOES flow (or it reaches Motor Kv * Motor RPM). 

If this voltage is greater than about 4.2 volts per series cell, excessive current CAN start flowing into the battery.  At this point,  the life of the battery life will be compromised since lithium plating will occur.

So, it's really no different than applying a voltage across a lithium cell.  If you apply 4.2 volts on a fully charged Lithium Ion cell, no current will flow.  However, if you increase the voltage, current WILL start flowing.  The higher the voltage, the more current.  High enough current and the cell can explode.

Saying it another way--he motor RPM dictates the voltage.  The resulting current depends on the load.  A fully charged battery won't draw any current, unless the voltage is high enough to cause lithium plating.  That should be avoided, but I guess it's better than causing a crash by isolating the battery from the motor.

A relatively simple snubber network could address this--I'd be very surprised if that's not in the motor drive circuitry.

 

[chriscalandro]

so lets say, when going down hill, with a dead battery, the pwm signal is telling the motor to reverse, but because of gravity and mass and so on, even though the motor controller is telling the motor to run in reverse, it is still pushing forward, against the signal, generating voltage.  The more dead the battery, the stronger the motor will fight the forward motion.  

In this case, the motor acting as a generator causes resistance in the wheel, slowing you down.

 

As the battery charges, that resistance will decrease, and more current will divert away from going to the battery, but will put a higher load on the motor controller trying to push the wheel backwards.  The net current generated is still the same, but now less goes to the battery, and the motor generates less current, as there is less draw.

 

at some point this entire thing will switch where the battery becomes fully charged, is no longer pulling a load, and the motor itself will be pulling the load while it actively starts pushing backwards to break rather than the resistance caused from the load the battery was pulling causing it to break.  This means that any electricity being generated from the motor will go straight to driving the motor in reverse. At this point, the load is the motor itself.  The battery is not actively charging, and the power generated from mass / gravity / etc / will be fed into the motor evening it out.  only when current generated = current used will the motor be freewheeling - meaning the motor is being driven backwards using exactly the same amount of power it's generating while being pushed forwards, and because you can't actually do that, for that instantaneous moment where all things are equal no electricity will be generated and no electricity will be used.  Other than that it's a constant balancing act between what the motor is using and what the motor is generating.

 

again, basic rules of a circuit says that current will only flow when there is a differential.  In the above freewheel case - there is NO differential.

 

Now... as mentioned above... the battery charging is based on what voltage actually feeds the battery.  A big motor will generate big electricity which means there needs no be a way to filter it.  That is where those capacitors come in, and a regulator to regulate the voltage. (60v-ish for a 55v battery)  If you happen to increase that voltage, yes, that battery will start charging again, and will overcharge.  But as long as you have good filtering, and a good regulator, that should not be an issue.

 

The graph above where the guy stopped very quickly showed a large spike in current generation, but only a small spike in voltage gain.  That would be the voltage regulator doing it's job.

[DaveThomasPilot]

Again, not sure if you're disagreeing with anything I said, but I disagree with this:

"and the motor itself will be pulling the load while it actively starts pushing backwards to break rather than the resistance caused from the load the battery was pulling causing it to break.  This means that any electricity being generated from the motor will go straight to driving the motor in reverse. At this point, the load is the motor itself"

A generator  (or motor acting as a generator) will spin freely under no load.  There are some parasitic losses (bearing friction and parasitic current paths), but the resistance is negligible compared to what's required for braking down a hill.

So, yes a series regulator can be added between the motor and the battery to prevent (or regulate) current flow into the battery.  But, if current isn't allowed to flow from the motor to somewhere, there will be no resistance to the wheel turning.  No braking.  That's what I mean by freewheeling.

In a series, step down voltage regulator, the load sees at least as much current as is drawn from the source.  In a switching step down  regulator, there is MORE current at the load than provided by the source.   Power IN = Power Out - efficiency losses.  Voltge In * Current In = Voltage Out * Current Out - efficiency losses.

A shunt regulator would work.  I.  Shunt regulators steer the current to someplace other than the load. But, the "snubber" network I've been referring to much simpler trivial than a regulator.  It's basically a simplistic shunt regulator.  Think of it as a lossy voltage clamp.  A regulator of sort.

A regulator needs  to maintain a constant voltage (or current)  under varying load (voltage regulator ) or changing voltage (current regulator).  That much isn't required to address this issue.

The H bridge could also be used as noted previously, if sized to handle the power.

[DaveThomasPilot]

Chris,

We might not disagreeing.  I think there must be something in the EUC  power conversion circuity that prevents current from flowing from the motor to the battery at voltages that are damaging the battery.

You're calling it a regulator.  I'm being more specific than that, as the circuit that is typically labelled a regulator (series voltage regulator) won't do the job.

Being more general , people observe that for the motor to brake current must flow.  That current has to flow somewhere.  The battery works as place to dump the current (and resulting energy), if it's not fully charged.

If the battery is fully charged, the energy must be dissipated as heat.  Snubber network, shunt regulator, H bridge.  (Or lithium plating in the battery).  But, it's a lot of thermal load compared to what happens normally.

Perhaps we should relabel this thread

"How to brake without breaking."  Or maybe to brake and break, or not to brake.  That is the question

[chriscalandro]

We're totally on the same page. 

[esaj]

The people (jayjay23, electric_vehicle_lover, Mystamo, lizardmech, zlymex and others) who probably know most are active in these threads:

 

 

 

 

I probably should split the posts from this current thread to another, as we've veered way off from the original topic, thanks to me   I'll do it later, as right now I'm a bit busy with work stuff...

[lizardmech]

The energy always has to go somewhere, conventional brakes convert energy to heat. There's two main issues, regen braking sends too much current to the batteries damaging them or potentially cause a BMS shutoff. The other issue is hill descent while fully charged.

The overcurrent is relatively easy to fix, you need two buffers made from supercapacitors or high C prismatic lifepo4. One to protect the battery packs from sudden drain, another to capture regen energy and slowly release it back into the primary batteries.

For hill descent you need to have a large resistor that current is sent through if both the buffer and battery pack is full. I have yet to try a long downhill ride with a full EUC, I think it will still brake, the batteries will probably become very hot and resistance will increase until a fire or BMS shut down. Even if you can get the engine to convert the energy to heat that  is a bad idea as a BLDC motor will destroy it's magnets at around 100C, resulting in you losing your brakes.

[Chriull]

Somehow this discussion reminded of something in the past:

Here i found an article about an electric trike implementing regenerative breaking with a dc-dc converter and supercapacitor as buffer. Since he used a trike he did not implement dynamic braking (burning the energy at a resistor) but relied on additional mechanical brakes;

 

And here a nice article with an overview of the different braking techniques for BLDCs: http://ieejournal.com/Vol_3_No_2/Different%20Braking%20Techniques%20Employed%20to%20a%20Brushless%20DC%20Motor%20Drive%20used%20in%20Locomotives.pdf

And dynamic breaking we had once in esajs battery pack thread: (with an BLDC controller implementind dynamic breaking)

 

[Chriull]

On 27.1.2016 at 1:50 PM, DaveThomasPilot said:

...

But, this requires that the switches between the battery and motor (speed control) can block current flow in the "reverse" direction.   FETS don't do this--they have an intrinsic diode that always conducts when the source is more positive than the drain.

Bad things can happen when the intrinsic diode conducts high current, so generally a discrete diode is place in parallel with the FET if current flow in the reverse direction is desired, eg, for "regenerative braking".  These diodes must block the full battery voltage (plus inductive spkes) in one direction and carry the full motor current when conducting.  So, they add cost and take up board space.

So, I'd say if the EUC is braking, the energy is going into the battery.  Diodes could be included in the design to prevent this, but then you'd just get freewheeling with no braking.

 

If just these intrinsic Diodes are used, this really could/would be a "big" problem. They have a forward voltage of about 1-1.3V (for the mosfets used in the E+/P motherboards) and with a breaking energy of 1-2 kW this would lead to currents about 15-35A. Times the 1-1.3V thats about 15-35W to dissipate for the Mosfets - could be more than they have to dissipate while in motor driving mode!

Here the use of some low forward voltage power diodes (schottky diodes) could help to get the dissipated power lower.

And maybe ninebot already did this - at my motherboard at the heatsinks one row of the 6 TO220 thingis can be identified as MOSFET. The other 6 have nothing written on them and could be power diodes and not a second row of mosfets in parallel? Also cranium has made the photos identifying the Mosfets just from the lower row of TO220s: 

There are still many if's and when's - but that's maybe the reason for all the reports of dying EUCs while aggressive breaking....

 

 

[dmethvin]

4 hours ago, Chriull said:

If just these intrinsic Diodes are used, this really could/would be a "big" problem. They have a forward voltage of about 1-1.3V (for the mosfets used in the E+/P motherboards) and with a breaking energy of 1-2 kW this would lead to currents about 15-35A. Times the 1-1.3V thats about 15-35W to dissipate for the Mosfets - could be more than they have to dissipate while in motor driving mode!

Here the use of some low forward voltage power diodes (schottky diodes) could help to get the dissipated power lower.

The spec sheets on some high-power Schottkys say they have a forward voltage drop of about 0.7v at max rated current so if they used something like that it would cut the power in half. Plus that's the worst case, with lower currents the forward voltage drop (thus the power) should be less.

Still those are just crazy big currents if you had to dissipate them for any length of time and makes you wonder if heat from the MOSFETs and diodes are a weak point. My Firewheel has a much wimpier heat sink than the Ninebot but on the good side it's attached to an aluminum mounting plate exposed to the air and convection from the wheel. The problem is that they only used a thermal pad between the MOSFETs and the heat sink, and nothing between the heat sink and aluminum plate. I want to fix that when I rebuild.