Regen braking: confirmed!

[SpEEdyBL]

A few weeks ago I posted some eagletree data and noted the current being put back into the battery in the second post of this thread. http://www.rc-monster.com/forum/showthread.php?t=18458

Regenerative breaking is very simple the way I see it. If you recall that motor torque is proportional to current, and that torque is negative during breaking, your are going to get negative current, or in other words, current that is charging the pack. When you charge a pack, the voltage increases.

If it doesn't make sense how a motor would create negative torque on its own without using power, think of what happens when you physically try to spin a motor faster than its kv*volts. It begins to push against you. You can drastically reduce the kv of a motor just by adding a resistor or equivalently, switching the esc on and off the same way it switches with different throttle amounts. This is just a guess based on general knowledge of circuits. I don't know the specifics.

[tom255]

Originally Posted by tom255

Why you cant brake without power, ofcourse you can by just shorting windings on motor or put some capacitance (capacitors) to windings thats it, maybe its not so effective. Maybe you cant change brake force without power source, somehow you just need change chorting impulse lengt and opposite phase switching, that basicaly any ESC do.

I will speak with guy who made a tests on ESC to check Ezrun 150A, does energy recuperation braking works there?

Ok that guy made a test and still No true regen brakes in ezrun 150.
He made test device/stand. In one side motor spinning other motor which was he tried to brake. Yeah, he noticed regen spikes (Voltage and negativ current) but those spikes are realy short, impulse lenght was 0.002s-0.007s so no true regeneration there, just some spikes thats all.
Seems ezrun 80a make same spikes

[Arct1k]

Patrick - Please could you put us out of our missery with a nice technical explanation!

[BrianG]

I wonder if Patrick even looks in this forum. After all, it's the CC section...

[Arct1k]

Maybe move the thread to CC area?

[BrianG]

lol, I suppose it could be moved, but it initially had nothing to do with a CC product - it became that way when one person asked "Does the MMM do this?".

[jhautz]

Originally Posted by BrianG

jhautz: why you get less runtime with motor braking is still puzzling to me. Without a consistent and repeatable test method, results are simply objective. Not trying to shoot ya down, just my $0.02.

I hear ya, but thats just about as consistent and repeatable as I will be able to get in a real world test. This is something that I just observed when I went from mechanical brakes to motor brakes in my truggy. I was trying to just simplify the system as the mechanical brakes werent giving me any lap time benefits on the high traction indoor track. So when I took the mechanicals off I noticed I used a little more mah. I wanted to understnad what was happening. it could have been differences in the track layout or the traction from one day to the next so I did my little experiment to try and eliminate all of the other variables except the mechanical/motor brake difference. What I saw confirmed my suspision. Not a highly scientific or overly controlled experiment, but the results were clear enough that I am assuming what I am seeing is real.

I'm not arguing that regen braking isnt happening, I'm just wanting to understand if it means anything at all to the runtime. my suspision is that the motor brakes take energy to work. Of course... I have no real proof of that except for my little experiment and what I personally have observed.

I didnt mean to totally derail your thread Brian, but at this point I do hope that we get some input from someone who truely understands this all. I had the same discussion with Mike and we didnt come up with a real answer either.

FYI: I went and found my notes from my test. I know its not a huge sample size or a super controlled test but it seems like pretty simple clear result to me. I dont think I would have the time to put together something that would be statistically significant and completely controlled. Its enough to convice me personally.

Setup was:
RC8T, 1515/2.5D, MMM, 5s lipo, 15/50 gearing.
Lipo Pack #1 was a 5s 4300mah Neuenergy 25c
Lipo pack #2 was a 5s 5000mah Zippy Flightmax 30C


Track was a medium/small sized track. (125ft x 75ft with a 100ft straight) indoor high traction clay surface with alot of quick burst straights and sharp 180 corners. The layout required alot of braking to get around it quickly.

mah usage on a 5 minute run with mechanical brakes:
Pack #1 =1122mah Pack #2 =1146mah

mah usage on a 5 minute run with motor brakes:
Pack#1 =1241mah Pack #2 =1296mah

Pack #1 results 1241-1122=119 119/1122= 10.6% increase in mah used with the motor brakes.

Pack #2 results 1296-1146=150 150/1146= 13.1% increase in mah used with motor brakes.

[lutach]

Originally Posted by jhautz

I hear ya, but thats just about as consistent and repeatable as I will be able to get in a real world test. This is something that I just observed when I went from mechanical brakes to motor brakes in my truggy. I was trying to just simplify the system as the mechanical brakes werent giving me any lap time benefits on the high traction indoor track. So when I took the mechanicals off I noticed I used a little more mah. I wanted to understnad what was happening. it could have been differences in the track layout or the traction from one day to the next so I did my little experiment to try and eliminate all of the other variables except the mechanical/motor brake difference. What I saw confirmed my suspision. Not a highly scientific or overly controlled experiment, but the results were clear enough that I am assuming what I am seeing is real.

I'm not arguing that regen braking isnt happening, I'm just wanting to understand if it means anything at all to the runtime. my suspision is that the motor brakes take energy to work. Of course... I have no real proof of that except for my little experiment and what I personally have observed.

I didnt mean to totally derail your thread Brian, but at this point I do hope that we get some input from someone who truely understands this all. I had the same discussion with Mike and we didnt come up with a real answer either.

FYI: I went and found my notes from my test. I know its not a huge sample size or a super controlled test but it seems like pretty simple clear result to me. I dont think I would have the time to put together something that would be statistically significant and completely controlled. Its enough to convice me personally.

Setup was:
RC8T, 1515/2.5D, MMM, 5s lipo, 15/50 gearing.
Lipo Pack #1 was a 5s 4300mah Neuenergy 25c
Lipo pack #2 was a 5s 5000mah Zippy Flightmax 30C


Track was a medium/small sized track. (125ft x 75ft with a 100ft straight) indoor high traction clay surface with alot of quick burst straights and sharp 180 corners. The layout required alot of braking to get around it quickly.

mah usage on a 5 minute run with mechanical brakes:
Pack #1 =1122mah Pack #2 =1146mah

mah usage on a 5 minute run with motor brakes:
Pack#1 =1241mah Pack #2 =1296mah

Pack #1 results 1241-1122=119 119/1122= 10.6% increase in mah used with the motor brakes.

Pack #2 results 1296-1146=150 150/1146= 13.1% increase in mah used with motor brakes.

The ESC probably has to use some of the power to actually stop the motor.

[BrianG]

The only way to do a real test would be some type of automated setup that places a consistent load on the motor for the whole run, and applies brakes with the same intensity and duration. There could be slight variations in each run (exact grip of terrain, acceleration ramp, how slow you brake before accelerating again, etc) which will skew the results. The definition of a good test is comparing using a consistent and repeatable process.

Any regen braking will be minimal when compared to pack capacity to be sure. I would be surprised if regen braking exceeded 5mAh to be honest. For one, the voltage generated while braking has to exceed battery voltage for reverse (charging) current to flow. If braking gently, the generated voltage may not be high enough to generate any charge current, or just slightly over the battery voltage and won't generate appreciable amounts of charge current. Two, the duty cycle (time braking vs time running) of braking vs running is tiny, probably in the 5% range or most likely less. Three, our motors are most likely way more efficient when in "motor mode" as opposed to "generator mode".

Again, this isn't a thread to say we should all use motor brakes to get more runtime, this thread just says "see, this ESC does have regen braking". I personally prefer motor braking because it is "fade-free", simpler, and layout is cleaner. I really wouldn't care if motor braking used more power than mech brakes as long as it didn't exceed 5%.

[SpEEdyBL]

Originally Posted by jhautz

I hear ya, but thats just about as consistent and repeatable as I will be able to get in a real world test. This is something that I just observed when I went from mechanical brakes to motor brakes in my truggy. I was trying to just simplify the system as the mechanical brakes werent giving me any lap time benefits on the high traction indoor track. So when I took the mechanicals off I noticed I used a little more mah. I wanted to understnad what was happening. it could have been differences in the track layout or the traction from one day to the next so I did my little experiment to try and eliminate all of the other variables except the mechanical/motor brake difference. What I saw confirmed my suspision. Not a highly scientific or overly controlled experiment, but the results were clear enough that I am assuming what I am seeing is real.

I'm not arguing that regen braking isnt happening, I'm just wanting to understand if it means anything at all to the runtime. my suspision is that the motor brakes take energy to work. Of course... I have no real proof of that except for my little experiment and what I personally have observed.

I didnt mean to totally derail your thread Brian, but at this point I do hope that we get some input from someone who truely understands this all. I had the same discussion with Mike and we didnt come up with a real answer either.

FYI: I went and found my notes from my test. I know its not a huge sample size or a super controlled test but it seems like pretty simple clear result to me. I dont think I would have the time to put together something that would be statistically significant and completely controlled. Its enough to convice me personally.

Setup was:
RC8T, 1515/2.5D, MMM, 5s lipo, 15/50 gearing.
Lipo Pack #1 was a 5s 4300mah Neuenergy 25c
Lipo pack #2 was a 5s 5000mah Zippy Flightmax 30C


Track was a medium/small sized track. (125ft x 75ft with a 100ft straight) indoor high traction clay surface with alot of quick burst straights and sharp 180 corners. The layout required alot of braking to get around it quickly.

mah usage on a 5 minute run with mechanical brakes:
Pack #1 =1122mah Pack #2 =1146mah

mah usage on a 5 minute run with motor brakes:
Pack#1 =1241mah Pack #2 =1296mah

Pack #1 results 1241-1122=119 119/1122= 10.6% increase in mah used with the motor brakes.

Pack #2 results 1296-1146=150 150/1146= 13.1% increase in mah used with motor brakes.

What are you using to measure the mAh? If you are using an eagletree data logger, which measures reverse current as positive, it WILL say that your are using more mAh than you actually are when using motor breaks. My eagletreee says my pack uses about 4,000mAh during a run when it takes less than 3,600mAh from the charger. From this it is quite obvious that there is regenerative breaking, because some mAh are being reused and are thus being counted more than once.

Was my previous post not read? It seemed to have been skipped over. The ESC implementing low throttle when the motor is spinning fast is how motor breaks actually work. You may think that wherever position of the trigger correlates to throttle but that is not the case with car escs. This was the issue with the old MGM compros. The throttle input from the reciever was exactly correlated to the throttle outputed by the ESC causing the motor to "break" when going from high throttle to low throttle.

[jhautz]

Originally Posted by SpEEdyBL

What are you using to measure the mAh? If you are using an eagletree data logger, which measures reverse current as positive, it WILL say that your are using more mAh than you actually are when using motor breaks. My eagletreee says my pack uses about 4,000mAh during a run when it takes less than 3,600mAh from the charger. From this it is quite obvious that there is regenerative breaking, because some mAh are being reused and are thus being counted more than once.

Was my previous post not read? It seemed to have been skipped over. The ESC implementing low throttle when the motor is spinning fast is how motor breaks actually work. You may think that wherever position of the trigger correlates to throttle but that is not the case with car escs. This was the issue with the old MGM compros. The throttle input from the reciever was exactly correlated to the throttle outputed by the ESC causing the motor to "break" when going from high throttle to low throttle.

mah put back in the battery with the charger. Was my first post not read. lol

[Pdelcast]

Sorry to resurrect a really old thread... but I saw that people were asking me to jump in and "pay my respects" so to speak... (And I'm having trouble getting to sleep tonight!!)


Regenerative braking does occur in all Castle ESCs. And yes, the actual braking action is achieved by just shorting the windings together through the FETs.

What happens is this -- the ESC shorts the windings of the motor, and forces the motor to start acting like a generator. Remember that a turning motor generates a specific voltage -- back EMF. This voltage is actually the voltage induced in the windings by the moving magnets.

Because the motor windings are shorted, the voltage drops to a very low level (usually less than .1V) and current rises very high, very quickly (often hundreds of amps.) As energy is generated (by the drag created by the voltage difference) current rises, and energy is circulated through the windings and the FETs - -- And a large magnetic field (with a LOT of energy) is stored in the winding.

After a short time, the FETs turn off -- and this is when the regeneration occurs. The current that was flowing through the windings suddenly has nowhere to go. Inductors (like a motor winding) abhor a change in current, so the stored energy (in the motor winding magnetic field) forces the voltage to rise until the winding current can continue to flow. The current flows from the battery negative, up through the body diodes of the low side FETs, through the motor winding, back through the body diodes of the high side FETs, and into the capacitors (and battery...)

This is similar to how a boost converter works:

http://en.wikipedia.org/wiki/Boost_converter

(stolen from the Wiki:) The key principle that drives the boost converter is the tendency of an inductor to resist changes in current. When being charged it acts as a load and absorbs energy (somewhat like a resistor), when being discharged, it acts as an energy source (somewhat like a battery). The voltage it produces during the discharge phase is related to the rate of change of current, and not to the original charging voltage, thus allowing different input and output voltages.

But instead of a supply, there is a magnet passing the coil that creates the current source.


Hope that clears it up for you all!

[lutach]

Awesome post Patrick. I don't know how accurate some of the data recorders are, but they usually show some sort of upward spike in the voltage. I also noticed that sometime an amp spike can occur as well, does that mean it's actually amp going back into the pack?

[Pdelcast]

Originally Posted by lutach

Awesome post Patrick. I don't know how accurate some of the data recorders are, but they usually show some sort of upward spike in the voltage. I also noticed that sometime an amp spike can occur as well, does that mean it's actually amp going back into the pack?

Yes, that amperage is going back into the pack during braking. Many data loggers do not have the ability to measure negative currents, so some just show zero current (like the Phoenix ICE data logger -- it does not have reverse current measurement, so it shows zero current during regen.) And some will show positive current, even though the current is negative. It depends on how the current sensing is achieved.

[BrianG]

For example the Eagletree 150A uses a Hall-Effect sensor that is unidirectional. It would be nice to upgrade the HE sensor to a bidirectional version of the IC, but my email to ET asking about that is as of yet unanswered.