Spun front axle and flew over the bars @ 25mph

[ZapPat]

Hey... wait a minute...!..
Are we talking about regen THAT RETURN THE POWER TO THE BATTERY OR REGEN THAT SHORT THE MOTOR TO DISSIPATE THE ENERGY INTO IT IN HEAT ???
Cause if the Methods's motor heat so much, i would guess that it's because of the controller would just short every phases instead of conducting to put back the energy into the cells? right?
Doc

Regen implies back to the battery in my mind.
Otherwise I would call it a crow-bar brake
Regen currents on the motor that gets hot (mind you this is a different setup that what we are talking about here) are around 20A @ 100V so 2KW regen. No motor can regenerate 2KW without creating heat regardless of how efficient the firmware is.
-methods

I have done away with regen on my "real" ebike.
Since I disabled regen my motor temps have been much lower. I dont have definitive numbers but with regen the motor got so hot that I could smell it on hard rides. Now it only gets hot enough to burn my hand
-methods

OK, I see you have a kelly controller (not a golden motors or similar), so I would suspect they must use some good regen firmware code - not something that's hacked together. So yes I agree that you will generate heat when doing 2kW regen for sure... but how much more than while just driving 2kW into the motor? What regen percentage do you get on one of your runs that makes the motor so hot? I could swear I saw someone post a graph of motor amps+volts+temp vs time, and we saw the temp go up fast during the regen time (going downhill after going up). But I didn't see the same graph done while just coasting back down (no regen) to compare to.

A quick question if you don't mind: How do you like kelly's regen implementation? After using it for a while, what do you find are it's good and bad points?

 

[methods]

I'm a tad confused here methods - If I understand the accident scene, you managed to go over your max speed with your controller disconnected from the batteries, then turned it on, thus forcing regen through the FET diodes? I kinda saw it differently from reading your first post, because you talked about just playing with your throttle causing you to go into regen...

No.
There are too many conversations happening at once here.

The controller was hooked up as normal
The controller was never turned on or off
I was going 25mph down hill
I do not know what speed the maximum speed I can attain at 40V is
The maximum speed I can attain fresh off the charger is about 22-24mph
I had the throttle on full blast
I released the throttle instantly
This caused the motor to regenerate at 14.7A
This caused the axle to spit out.

Exactly as I explained in the original post.

-methods

 

[methods]

Hi Methods,

Not sure this can be right. With a "standard" controller you can only put current into the battery if the wheel is turning faster than the battery could turn it. You can't go straight from forward thrust to regen.

Glad you're still with us, though.

Nick

As I clearly stated several times I was going 25mph down hill on a low battery.
Full power on a fresh charge is about 22-24mph for my 12S Lipo (Nine Continents Kit)
My battery was down to about 40V so 25mph is easily enough to regenerate.
Like I said, I regenerated 14.7A peak at the moment of impact.

Read my original post

-methods

 

[methods]

OK, I see you have a kelly controller (not a golden motors or similar), so I would suspect they must use some good regen firmware code - not something that's hacked together. So yes I agree that you will generate heat when doing 2kW regen for sure... but how much more than while just driving 2kW into the motor? What regen percentage do you get on one of your runs that makes the motor so hot? I could swear I saw someone post a graph of motor amps+volts+temp vs time, and we saw the temp go up fast during the regen time (going downhill after going up). But I didn't see the same graph done while just coasting back down (no regen) to compare to.

A quick question if you don't mind: How do you like kelly's regen implementation? After using it for a while, what do you find are it's good and bad points?

ARGH!
Too many topics at once.

Ok, first: I think regen is a waste of time and money on ebikes. They don't carry enough mass to make it useful. All of your energy goes into the wind, road resistance, etc. On your best day you *could* get back 5-10% (more likely less than 5%) which is not worth it when you consider the heat that is going back into the motor, the wear and tear on the mechanical parts, and the wear and tear on the electronics.

I have taken a bike to the top of the largest hill in my area and used regen all the way down and only regenerated a measly 100mah. I can blow that in about 3 seconds off the line with my 100V 200A controller.

The lighter the setup the less effective the regen will be. Justin probably got some good regen numbers (5%ish) because he was riding that tank. . . Someone on a sub 50lb bike is not going to benefit much at all.

That said, to answer your question, I think that Kelly has a fairly good implementation.
The original 208 version has a problem where you cant get regen below 5% which is annoying. If you set the regen max current too high it would cut off before coming to a full stop.

Version 209 has a throttle regen that is a fun toy. Acts just like an engine brake. Problem is that they did not allow a way to defeat it (coast). I have seen implementations where you could "snap" the throttle off to coast or "roll" the throttle off to engine brake (1500 Lead Acid scooter).

Overall I like Kelly as a company. I am sure that if I really wanted to get into it they would take suggestions but I am personally done messing with Regen. Maybe on my next Motorcycle or car project.

I do respect your choice to make your own controller.
Much luck with that.
I used to do a lot of from-scratch DIY but not anymore. I now prefer to just enjoy the endless supply of new gadgets that stream in from overseas.

-methods

P.S. Regarding temperature: I suspect that my abnormally high regen temperatures were due to the novel ABS regen that I was using. Any time that I was not "on the gas" I was "on the regen". There was no such thing as coasting. (version 209). I did manage to reach upward of 10% on many trips but I also burned a lot more Ah due to the lack of coasting. In the end I did not get good mileage.

 

[ZapPat]

ARGH!
Too many topics at once.

Sorry for picking your brain like that!

Version 209 has a throttle regen that is a fun toy. Acts just like an engine brake. Problem is that they did not allow a way to defeat it (coast).[...]
P.S. Regarding temperature: I suspect that my abnormally high regen temperatures were due to the novel ABS regen that I was using. Any time that I was not "on the gas" I was "on the regen". There was no such thing as coasting. (version 209). I did manage to reach upward of 10% on many trips but I also burned a lot more Ah due to the lack of coasting. In the end I did not get good mileage.

Ahh, this helps explain the temp increase indeed - thanks for that precision. Not good to have regen always on - annoying and can even reduce global efficiency! Do they have a better mode for their regen (that would let you coast when wanted that is)?

I have seen implementations where you could "snap" the throttle off to coast or "roll" the throttle off to engine brake (1500 Lead Acid scooter).

Great idea - Noted!

Thanks for the answers and your patience, methods!

 

[pwbset]

I think regen is a waste of time and money on ebikes.

Word. Did a 1,600 vertical ft decent experiment this summer and got back like .35ah. Not bad as things go, but that's an extreme example. I imagine Justin coming down the backside of those trans-CA hwy passes though and bet that would have been worth it! :lol:

http://endless-sphere.com/forums/viewtopic.php?f=2&t=6553

 

[Ypedal]

ok.. so now i'm a tad confused as well... :lol:

I have used all sorts of hubs at just about every voltage possible. Going down a big hill, past the no load speed, throttle on or off makes no difference, not any significant difference at least..

The force applied to the axle flats would be the oposite from acceleration, vs using the brakes on the rim, that i can see causing problems. but not what i originally pictured as sudden hard regen action...

Hmmm.. interesting. It's late and i've had a few beers.. so maybe it will make more sense in the morning.. off to bed i go.

 

[ZapPat]

OK, I'm obviously not the only one confused about a regular controller going into regen as soon the throttle is droped from full to nothing.

Usualy the FETs internal diodes will automaticaly force regen as soon as the motor voltage starts creeping up over the battery voltage (minus two diode drops). So in theory we shouldn't ever be able to get enough speed to get lots of regen, since the diode-caused regen will slow us down. This is the issue of confusion here... :?

Methods: is your controller still alive?

 

[ZapPat]

I think regen is a waste of time and money on ebikes.

Word. Did a 1,600 vertical ft decent experiment this summer and got back like .35ah. Not bad as things go, but that's an extreme example. I imagine Justin coming down the backside of those trans-CA hwy passes though and bet that would have been worth it! :lol:

I believe you guys about the questionnable usefullness of regen on a lite ebike as to recuperating energy, but I actually want it more for breaking purposes, since my V brakes don't really like the extra weight of the conversion. For people living in hilly areas, with heavy bikes (or bodies), or maybe even people in the city doing a lot of start and stop, regen breaking seems like a good alternative to only using regular breaks. Plus it costs me almost no extra money to add to my design anyways...

Another reason for implementing regen in my controller design is that it is modular, so I will want to eventualy move up to making one big enough for a motorcycle and car too. Call me optimistic if you wish! At any rate, regen with this type of heavy EV is almost a must.

 

[methods]

but not what i originally pictured as sudden hard regen action...

Let me be clear, that is exactly what I am saying I experienced.
A *SUDDEN* and *HARSH* regen action that (helped to) spin out the front axle.
The only clues I have are that the controller regenerated 14.7A at that time and my ass flew over the handlebars.

Let me pose a question for those who are uneasy with the idea of a 36V 20A standard controller producing regen:
If I am going 25mph down hill
Suppose I am going about 4 - 5mph over the WOT motor speed
Suppose I then snap off the throttle
If there was no regen effect, what forces could possibly cause the bike to fiercely decelerate, spin out the axle, and throw me over the handle bars? :?

I know it sounds weird to say that this particular controller behaved in this manner but... TRUST ME .... I run a 120V 220A kelly at 100V on a 5305. . . I know what regen feels like and I know my volts and amps. . . This controller hit 14.7A of regenerative current back into a 40V Lipo battery which was enough to pop the axle and and throw me over the handlebars.

Remembering of course my above statements about the other factors: Walmart bike, no torque arm, unknown nut status (though I tightened them TIGHT), etc.

I can see that I am going to have to reassemble the bike and reconstruct the exact situation with my Eagle Tree logger.

* Weld up the drops
* Wire in my datalogger with VOLTAGE, CURRENT, and SPEED
* Run the battery down to the same level
* Go down the same hill at the same speed
* Snap off the throttle

Then, when the controller does the same thing and I have graphs we can accept it as fact :wink:

-methods

 

[Ypedal]

[ what forces could possibly cause the bike to fiercely decelerate, spin out the axle, and throw me over the handle bars? :?

that right there is the mystery... i have not experience this effect ... at that speed.. the throttle has no impact of that nature on any of my bikes...

 

[methods]

Going down a big hill, past the no load speed, throttle on or off makes no difference, not any significant difference at least..

Incorrect.
The DV/DT is what you have to consider, not the steady state.

I was going down hill past the no load speed with the throttle at WOT
This created nearly no resistance between the axle and the fork
I SNAPPED the throttle off which INSTANTLY switched the motor from NO LOAD to MAXIMUM AVAILABLE REVERSE LOAD
It is the ***fast*** change from no-load to load that causes the spike in force.

now. . .
You could possibly be right. I may find that when I snapped that throttle off the controller blew and shorted the controller out internally. That would certainly explain the harsh regen engagement. I dont think the controller is blown though. I am not about to hold the axles with my hand just to test it out

-methods

 

[methods]

[ what forces could possibly cause the bike to fiercely decelerate, spin out the axle, and throw me over the handle bars? :?

that right there is the mystery... i have not experience this effect ... at that speed.. the throttle has no impact of that nature on any of my bikes...

I would agree with that statement for the 72V 40A controllers, I never saw any difference.
With the 36V 20A controller I get regen when I let off.

What size batteries have you use with what motor combinations?

This is a particularly high pole count motor (which should not make a difference I dont think)

I dont want to start a pissing contest because I can see that many of you have owned many more bikes than I have but I am not a jackass. . . I know what I know.
This controller did exactly what I said it did.

I guess I am starting to get a little irritated

-methods

 

[Ypedal]

Hey.. Instant Start or Pedal First controller ?

 

[methods]

I just thought the exact same thing as I left the house for work
I came back inside to type this message

This is a *peddle first* controller
Any time that I am not on the gas it has some degree of drag
The lower the battery voltage the worse the drag
Whenever the throttle is on WOT there is no drag

When I popped that throttle off the regen came on hard bro.
I am going to run that experiment so we can put this behind us

Like you know who says: "One test is worth a thousand opinions"

-methods

 

[ZapPat]

I dont want to start a pissing contest because I can see that many of you have owned many more bikes than I have but I am not a jackass. . . I know what I know.
This controller did exactly what I said it did.
I guess I am starting to get a little irritated
-methods

I don't think anyone here is questioning your experience, methods (I certainly am not). 8)
I'm just a tech-loving guy who can't understand what happened inside the controller to have this happen. I thought since you started this thread that you would be open to debating the probable cause(s) of the accident.

I would agree with that statement for the 72V 40A controllers, I never saw any difference.
With the 36V 20A controller I get regen when I let off.

Interesting... It may be then that the sensorless technique they use has something to do with it? I have heard of one sensorless design that involves sending small bi-directional pulses into the motor's windings to determine it's position (using inductance variations or something similar). The other sensorless techniques I heard of are passive (just sensing phase voltage or current), so they shouldn't be able to be the cause of such a malfunction.

Anyways, if you do retry this (with more safety features of course), I'll be very curious to see the results! And be carefull - Savvy ebikers are a rarety! :wink:

 

[Tiberius]

Hi Methods,

Please don't get irritated just because I'm confused. I did read the OP and the thread, and I thought I understood it until you said it was NOT a regen controller.

14.7 A is a large current just for motor overrun, and there is still the problem of working out why it responded that way to the throttle.

Then I, along with others I think, thought it may have been a case of the controller blowing and shorting the motor. That would have given the reverse thrust, but it wouldn't have scored a reverse current on the cycle analyst.

Now I learn that its a sensorless controller, I can offer another theory. When you closed the throttle suddenly, it got its phase control and PWM wrong and somehow for a brief period acted like a regen controller and stepped up the motor emf enough to give that current pulse into the battery.

Another theory is that the 14.7 A is an erroneous reading. I don't know what rate the CA samples at, or what smoothing it does, but if it was a fast pulse it could have got it wrong. Or if there was a huge current pulse flowing elsewhere round the system, it could have created a false reading.

These are only theories, but there is a mystery to be solved.

Nick

 

[methods]

I thought since you started this thread that you would be open to debating the probable cause(s) of the accident.

Yep, of course. Always need to stay open to an argument.
I just get defensive :lol:

I cant claim to know anything about how they implemented the controller design in the sensored vs. sensorless versions. I have designed simple dual H-Bridge rectumfrier PWM dirvers for DC motors and I have done a lot of stepper motor work with micro controllers and 8086 drivers but not brushless. I understand how they work but I have never reverse engineered one or designed on.

btw:

so they shouldn't be able to be the cause of such a malfunction.

I am not sure we are talking about a malfunction here.
Under normal driving conditions i would see 1A-3A of regen when I let off the gas
Seeing 14A when letting off the gas going down hill on a low battery does not strike me as odd

Can someone please chime in who has a 36V sensorless controller and can confirm what I say about the controller having "drag" when off the throttle? Even better, who has a Nine Continents kit with the higher KV motor in the 20" wheel (or I guess the lower KV in a 26") so we can compare apples with apples.

When I get home tonight I will set up the motor in a clamp and fire up the motor.
I will run a few tests with the logger on to at least baseline my clams that the motor regenerates on a regular basis.

-methods

 

[methods]

Hi Methods,
Please don't get irritated just because I'm confused.
Nick

Those sound like good theories.
I am just an irritable type of guy btw: :lol: Nothing personal.

I like that set of feedback. Sounds like you may be on to a few things:

The CA smoothing is set to 5 right now btw. Any sort of spike would no have been caught and if it was it would have been averaged.

I experienced the "strong regen effect" for what seemed like a long time before I went over. Time changes when you are about to crash so maybe it was only a few hundred miliseconds but I would guess that I was under hard regen for 2 - 3 seconds.

We can determine if the controller popped with a simple test tonight.

Like I said to pat, I am not convinced that anything bad happened at all. I have seen regen on this controller from day one. Given I only rode the bike for maybe 30 miles it is possible that it was a fubar controller on the way out but it felt fine to me. I remember being irritated that it did not freewheel as well as i would like at top speed. It would tend to slow down to around 18mph then freewheel.

but it wouldn't have scored a reverse current on the cycle analyst.

Why do you say that?
I can think of several failure modes that would result in current being run backward into the battery. Say for instance some of the fets just shorted. Would that not tie the motor directly to the battery and allow current to flow from the motor to the battery thereby scoring a reverse current? Please explain.

Again, we should just run a test and stop guessing

I will do it tonight.


-methods

 

[Tiberius]

Those sound like good theories.
I am just an irritable type of guy btw: :lol: Nothing personal.

Understandable after a nasty accident. Just don't get too rude about physicists, ok?

 

[Tiberius]

but it wouldn't have scored a reverse current on the cycle analyst.

Why do you say that?
I can think of several failure modes that would result in current being run backward into the battery. Say for instance some of the fets just shorted. Would that not tie the motor directly to the battery and allow current to flow from the motor to the battery thereby scoring a reverse current? Please explain.

Back into the battery is uphill, and the current won't do it except in special circumstances. You can get some regen current if the motor is going fast enough and the body diodes of the FETs act as rectifiers. If a FET shorted it would be more likely to screw the rectification up than enhance it. Ie., more likely to dump the current through the controller than through the battery.

I'm trying to work out what a stuck commutation would do (equivalent to a top and bottom FET shorting at the same time). It would be like connecting an ac motor to a dc power source. It might just produce the effects you saw.

Nick

 

[pwbset]

Can someone please chime in who has a 36V sensorless controller and can confirm what I say about the controller having "drag" when off the throttle?

With a 4011 @ 37v and a 24-72v Clyte pedal first I quite often will be full throttle and hit a brief downhill/dip in the road etc. and when I "snap off" the throttle the motor most definitely "regens" or whatever you want to call it. I frequently used that to keep speed in check during long descents and/or get a brief slowdown to back under no load speed on the flats/dips. It's never been a dramatic over-the-handlebars force, but it definitely slows me down noticeably. At 37v I can ride an 8-12% descent at about 18-22mph and barely use the brakes at all. It's quite a nice feeling actually.

.02

 

[dogman dan]

It's weird how it can feel, I find with the throttle off, I can descend a really big hill well above 30 mph, but with the trottle full on, it gets difficult to go above 31 mph. Same thing if I just barely turn the throttle. Once energized, the motor seems to have a definite top speed. On the same hill, on a bike with no motor, I can easily hit 45 mph without the motor drag. This is with a WE brushed motor. But I have never felt that kind of deceleration when shutting off the throttle, just from getting a water bottle caught in the front wheel.

 

[methods]

Understandable after a nasty accident. Just don't get too rude about physicists, ok?

Where did this come from?
My wife is a physicist so it is open season.
I also work with the largest group of egg-heads on earth. . . LLNL
I am an engineer so physicists will be the butt of all jokes :wink:

-methods

 

[methods]

It's never been a dramatic over-the-handlebars force, but it definitely slows me down noticeably.
.02

Thanks for the data point!

FWIW: I went over the bars because the front wheel came off and the forks hit he ground, not due to the strength of the regen. The regen was just strong enough to pop the axle.



***\\

I came home at lunch and did a quick test.
Tightened the axle in with a monkey wrench
Spun the wheel by hand and hit the gas
The motor spun up then quickly spat out of the drops and disconnected.

So the controller is fine.

My guess is that when I run the final experiment we will see identical reverse currents.
btw: I also read the peak speed off the CA, it was 29.9mph but I dont think I was going that fast before the crash.

-methods