Are euc's exceptionally vulnerable or dangerous?

[enaon]

39 minutes ago, blox1130 said:

Wrong way has demonstrated, and I can certainly feel the same, that braking distance is equivalent.

I agree, initial tests done by Freestyl3r shows that braking distance is just fine, once the tool is ready for use it will be clear that it is all about technique I think. But the problem remains, we should be careful with braking, the batteries do not like it. 

molicel p45b may be the answer when they get available.

Edited October 7, 2022 by enaon

[blox1130]

4 hours ago, enaon said:

I agree, initial tests done by Freestyl3r shows that braking distance is just fine, once the tool is ready for use it will be clear that it is all about technique I think. But the problem remains, we should be careful with braking, the batteries do not like it. 

molicel p45b may be the answer when they get available.

I think some of the problems with that could be alleviated with an emergency brake resistor. Some power could be diverted to it during peaks or when voltage gets too high (like when the battery is fully charged). Most of the time it wouldn't be used, so it doesn't burn up haha

[mrelwood]

20 hours ago, enaon said:

I thought is was obvious from monitoring voltage and amps on braking action, have you ever noticed something different? 

But you said:

20 hours ago, enaon said:

All braking action is using regen alone.
The energy has to go to the batteries,

Based on what you've later said, you haven't measured how much of the braking energy goes back to the batteries. Only that some level of regen exists. Which is well known and no-one argues against.

If you short the motor phases, the motor becomes rather stiff to rotate, but you can still rotate it by hand. Isn't the maximum resistance felt this way the maximum braking torque the motor can provide by regen alone? Even I ask for stronger braking from my wheel from time to time. Where does the energy for that come from?

Edited October 7, 2022 by mrelwood

[blox1130]

The braking force during regen is or should be stronger at higher RPMs. I dunno in EUCs if all deceleration is due to regen (maybe its even reversing the motor?), but i do know from my VESC controller that regen alone, if set to high amps, is very effective at reducing speed to the point that I don't use mechanical brakes.

Also to note, it should be way easier to measure and plot values for things like regen/motor current/voltage if I could build an EUC with a VESC controller. VESC firmware comes preinstalled with a balance input for making DIY EUCs and also allows for the measured values to be extracted into a log file.

That's got me thinking too: what if we just took one of the boards from these EUCs like the master, and made our own chassis and battery? Or swap a master's board to a T4 with a battery mod?

Edited October 7, 2022 by blox1130

[Planemo]

Completely aside to EUC's, it's possibly noteworthy that my Li-Po powered RC cars definitely don't use regen but brake incredibly hard and very controllably too. So RC are certainly doing it just by reversing polarity I guess. This is what made me assume that EUC's work the same way until I read on here some time back they they use regen as well. Like I say I have no confirmation either way and I have never done any tests personally. Curious though!

[enaon]

it is using a DC motor most probably, those motors cannot regen, the back emf can not exceed the input voltage. 

I cannot really explain it, I lack some knowledge and language skills, but generally the EUC is using regen to keep us balancing at all time. At any given speed the motor produces a fixed amount of back emf related to that speed, if the controller provides greater voltage, using PWM,  then current is flowing to the motor and we accelerate, if it provides lower voltage, the current flows to the batteries and we decelerate. 

what is important I think to know is that braking is using the batteries as energy storage, and the batteries are not ok with getting as much current in as they can deliver. 

Edited October 7, 2022 by enaon

[Planemo]

6 minutes ago, enaon said:

it is using a DC motor most probably

So do EUC's... dont they?!

Possibly you meant to say brushless, and yes, the RC stuff all uses brushless 3 phase now, same as EUC.

6 minutes ago, enaon said:

 and the batteries are not ok with getting as much current in as they can deliver. 

I totally agree with that as we well know that max safe charge is nowhere near what the packs are capable of max safe discharging.

[enaon]

41 minutes ago, Planemo said:

So do EUC's... dont they?!

Possibly you meant to say brushless, and yes, the RC stuff all uses brushless 3 phase now, same as EUC.

I meant DC series motor, those cannot be generators, they can ok, but Iet's say they can't to keep it simple. Ours are AC in principle. If your rc car is a 3 phase then I guess it uses regen to stop the motor, then reverses the input to spin backwards. This implies that the tire will skid, not really comparable to our case. You cannot do both regen(battery input) and reverse(battery output) at the same time. 

Edited October 7, 2022 by enaon

[samzed]

Wearing bright colors to increase visability increasing EUC safety.

[enaon]

two more things on battery and braking:

when braking on full charge, voltage will go up, regen etc, but:

1. Kingsong will never report anything over 84.5, they cap the voltage, so be careful if you see 84.5, it could be way higher.
2. Gotway caps at 100.8, use freestyler's euc-dash and switch to extended packets to see the true voltage.*

* do this, even if you do not care about braking when full, if your wheel reports 100.8, true top voltage is very important to keep a track of battery health. 
 

Edited October 7, 2022 by enaon

[Chriull]

18 hours ago, enaon said:

it is using a DC motor most probably, those motors cannot regen, the back emf can not exceed the input voltage. 

As the inverter (3 half bridges) works with the motor coils during normal driving operation  as buck converter, during regen breaking it works as boost converter.

In https://forum.electricunicycle.org/topic/7549-current-demand-versus-battery-voltage/?do=findComment&comment=104078 is the schematics for the buck converter with the 3 phase bldc motor simplified as dc motor (coil and back emf)(low side fets not regarded).

With the low side fets one can "short" the coil(s) to ground (battery minus) and the high side mosfets are the rectifying diodes of the boost converter.

So voltages well above battery voltage are generated and regen breaking is possible.

18 hours ago, enaon said:

I cannot really explain it, I lack some knowledge and language skills, but generally the EUC is using regen to keep us balancing at all time. At any given speed the motor produces a fixed amount of back emf related to that speed, if the controller provides greater voltage, using PWM,  then current is flowing to the motor and we accelerate, if it provides lower voltage, the current flows to the batteries and we decelerate. 

During normal driving the wheel continously balances the rider by "small" accelerations and decelerations. With the used bldc motor this happens by the commutation. The direction of the magnetic field of the outrunner ("tyre") with the magnets while turning has some actual revolution speed. The turning magnetic field created in the coils can be a bit "before" the outrunners magnetic field to accelerate or a little bit behind to decelerate. Both of this operations need power from the battery! 

This mechanics is used partially for (slight?) breaking - normal (stronger) breaking maneuvers in the logs are regen breaking. (Increasing battery voltage and negative battery currents - as kingsong wheels report it)

18 hours ago, enaon said:

what is important I think to know is that braking is using the batteries as energy storage, and the batteries are not ok with getting as much current in as they can deliver. 

Afair in my ks16s logs breaking current never/seldomly exceeded 15A. Would be 3.75A per cell and a bit above 1C for the 3.5Ah.

No idea if this is a value due to current limiting by firmware (as @blox1130mentioned available for vesc controllers) or just this wheels "natural" limit for breaking maneuvers as i performed them. I never did (had to) really try what's really possible...

Breaking torque will allow some proportional maximum motor current. Together with the back emf created at the actual speed one has the generated voltage. Invers proportional to this generated voltage in relation to battery voltage is this breaking max motor current in relation to regen battery current.

I do not really see some "hard overlean" limit as during driving mode with i motor max = (U battery no load - U back emf)/(R coil + R battery) for regen breaking operation.

Just the boost converter can not limitless boost the voltage up as losses will greatly increase.

And the body diodes will have to dissipate quite some power! Imho that's the reason why the "ancient" wheels died like flies during hard breaking. Seems that by now the mosfets get actively switched so that instead of the conducting body diodes forward voltage only the conducting mosfets rdson is relevant for power dissipation.

At higher speeds boost factor is low so one needs almost as high battery currents as the motor breaking current(~breaking torque)

At lower speeds one needs high boost factors, so with lower battery currents higher breaking currents are achievable. No idea how notable losses hinder this advantage?

At very low speeds and standstill presumably breaking is realized like the above described self balancing during normal riding operation by the generated magnetic field (vector).

I'm looking forward for new findings with the help of the "breaking tool" - not too much experiments with measurements have been made (to my knowledge).

[blox1130]

4 hours ago, Chriull said:

Afair in my ks16s logs breaking current never/seldomly exceeded 15A. Would be 3.75A per cell and a bit above 1C for the 3.5Ah.

No idea if this is a value due to current limiting by firmware (as @blox1130mentioned available for vesc controllers) or just this wheels "natural" limit for breaking maneuvers as i performed them. I never did (had to) really try what's really possible...

 

That's gonna be the amps pushed to the battery after its been through the controller. At 72v x 15 amps that honestly sounds about right for that wheel. You can get more by simply both 1. Going faster 2. Slowing down faster.

On a Vesc controller, you can independently change either the amount of regen amps done by the motor, or the regen amps that gets pushed to the battery. It's not a 1:1 correspondence though, as the motor amps are always or almost always much higher and more important than battery amps. A 15 amp setting for motor regen is really really weak braking but 15 amps for battery regen is ok.

[enaon]

have not tested the 16s but I have one here so I will, but on the s18, 3p system(mine is 4p, I am not sure if the board somehow takes advantage), it very easy to do 20amps, I have my alarms at 15 and the go off very often. 

 

it is refreshing to see someone, from USA too, using the nominal voltage instead of peak 

Edited October 8, 2022 by enaon

[0000]

11 hours ago, Chriull said:

The turning magnetic field created in the coils can be a bit "before" the outrunners magnetic field to accelerate or a little bit behind to decelerate. Both of this operations need power from the battery! 

Coming at this from a layman's point of view, could you then interpret this power requirement for deceleration as a kind of moving target activation cost (speed/RPM and braking magnitude dependent) where under braking you would need to exceed this constantly changing current requirement before any regen/current going to the batteries occurs? So basically regen doesn't automatically occur under all braking/decel conditions as a simple way to put it.

11 hours ago, Chriull said:

At higher speeds boost factor is low so one needs almost as high battery currents as the motor breaking current(~breaking torque)

At lower speeds one needs high boost factors, so with lower battery currents higher breaking currents are achievable. No idea how notable losses hinder this advantage?

At very low speeds and standstill presumably breaking is realized like the above described self balancing during normal riding operation by the generated magnetic field (vector).

Attempting to play interpreter again, is it fair to say then that heavy braking/decel at higher speed/RPM tends to result in higher stress on the batteries (regen current) and heavy braking/decel at lower speed/RPM tends to result in higher stress on the controller components, specifically the mosfets performing the buck converter function?

One last thing I was kinda stuck on thinking about how a traditional automobile brakes/decelerates and how that relates to EUC, but maybe got a better handle on it now:

With cars, it is the clamping force applied at the brake pads against the brake rotors creating friction some distance away from the axis of rotation which allows the tires to take advantage of friction with the road and slow the vehicle down. The result of shedding all of this momentum is a large build-up of heat in the pads and rotors as we know. On a smaller EUC scale, I guess the "brakes" then can be likened to the current orienting the magnetic field in the coils, and the amount of current used during deceleration (also acceleration) determines the heat losses with Rcoil, but also secondarily in Rcontroller (as buck conversion occurs), and finally in Rbattery when threshold for regen is surpassed. Does all of that sound right?

Edited October 8, 2022 by Vanturion

[Chriull]

13 hours ago, Vanturion said:

So basically regen doesn't automatically occur under all braking/decel conditions as a simple way to put it.

Regen can happen once the controller switches the inverter in "boost converter" mode and actually braking occurs.

Imo in "buck converter" mode no regen is possible.

This combination is in german called "synchronwandler" (https://de.m.wikipedia.org/wiki/Synchronwandler) - did not find the english translation for this...

14 hours ago, Vanturion said:

Attempting to play interpreter again, is it fair to say then that heavy braking/decel at higher speed/RPM tends to result in higher stress on the batteries (regen current)

Yes, for the same breaking torque (~motor current) battery current increases with speed.

14 hours ago, Vanturion said:

and heavy braking/decel at lower speed/RPM tends to result in higher stress on the controller components, specifically the mosfets performing the buck converter function?

Imo the significant losses happen by the coil resistance. With actively switching mosfets their power dissipation should be comparable to normal driving burdens?

14 hours ago, Vanturion said:

On a smaller EUC scale, I guess the "brakes" then can be likened to the current orienting the magnetic field in the coils, and the amount of current used during deceleration (also acceleration) determines the heat losses with Rcoil, but also secondarily in Rcontroller (as buck conversion occurs), and finally in Rbattery when threshold for regen is surpassed. Does all of that sound right?

Main heat losses occur on R coil and R battery which are both in the range of tenth of Ohms.

Mosfet resistances are hundredth of ohms - so about a decade less.

The main difference between car (friction) brakes and EUC regen braking is that with car (friction) brakes 100% of the braking energy is converted to heat, with EUCs some part is used to increase the battery charge and the rest is mainly converted in heat at the motor coils and battery.

[enaon]

would you say it is safe for the batteries to apply the brakes on near full state of charge? They will accept over 1c charge, at over 4.25 voltage/cell, is that ok?

Edited October 9, 2022 by enaon

[Chriull]

14 hours ago, enaon said:

would you say it is safe for the batteries to apply the brakes on near full state of charge? They will accept over 1c charge, at over 4.25 voltage/cell, is that ok?

@RagingGrandpaafair once posted some li ion "danger graph" and they should nicely handle mistreatment up to some ?or even above 4.5V?.

They likely won't forget this torture, but for some short periods they should not overreact...

Edited October 10, 2022 by RagingGrandpa
(img url)

[enaon]

15 minutes ago, Chriull said:

@RagingGrandpaafair once posted some li ion "danger graph" and they should nicely handle mistreatment up to some ?or even above 4.5V?.

They likely won't forget this torture, but for some short periods they should not overreact...

I follow ragingGrandpa closely, I have seen his comments on 4.5 volts, but at over 1c charge when they are already full, I also think they will not forget the torture.

it bugs me that both ks and gw caps the voltage reporting, I want to see how up the voltage gets when braking at full. 

Edited October 9, 2022 by enaon

[techyiam]

21 minutes ago, Chriull said:

Regen can happen once the controller switches the inverter in "boost converter" mode and actually braking occurs.

Once the inverter switches to boost converter mode, wouldn't the output be AC current, since the communtation function is lost? Would a full bridge rectifier be required in order to charge the battery. ? 

[Chriull]

16 minutes ago, techyiam said:

Once the inverter switches to boost converter mode, wouldn't the output be AC current, since the communtation function is lost? Would a full bridge rectifier be required in order to charge the battery. ? 

The top side mosfets body diodes perform the rectifying. Or power dissipation wise much better solved by correctly switched mosfets.

[0000]

10 hours ago, Chriull said:

This combination is in german called "synchronwandler" (https://de.m.wikipedia.org/wiki/Synchronwandler) - did not find the english translation for this...

Ah thank you, I was still thinking of EUC controller operations as equivalent to an ebike controller before in that when you would actuate the regen throttle to decelerate (or lean back in the case of EUC), regen occurs immediately and the batteries would act as the main energy sink for slowing, with the rest of the energy being "bled off" as heat loss. It is as you say of course, more complicated and nuanced. Happy to be corrected of this error.

Yandex translates synchronwandler as synchronous converter - an edited quote I found helpful which just summarizes the functionality from the points you already made:

Quote

A main feature of the synchronous Converter is that the energy flow in both directions can occur: the ability of the sensing level of the Synchronous Converter allows power flow in both the load from as well as the load back to the source. The amount of the output voltage varies depending upon on the choice of the input and can output higher or lower voltage than the input voltage.

Ok, I'm going to try for maximum learning here, hopefully it's not annoying..

10 hours ago, Chriull said:

Yes, for the same breaking torque (~motor current) battery current increases with speed.

Neglecting any kind of field weakening effect (not even sure if that's relevant to any EUC motor controllers anyway), would it follow that maximum available braking current is the inverse of maximum available acceleration current as the spinning motor approaches back emf voltage parity? At least, after the constant current portion of the rpm vs current/torque motor performance curve. Edit - although this is probably more of a theoretical question I guess as very large braking torque could result in the tire breaking traction.

10 hours ago, Chriull said:

Imo the significant losses happen by the coil resistance. With actively switching mosfets their power dissipation should be comparable to normal driving burdens?

Sounds reasonable to me, in the case of acceleration vs deceleration, the mosfets are just doing what they always do, switching. Then again, I lack of hands on testing experience with these kinds of components which puts me squarely in the useless speculation category. I was and am just curious if any controller components including the mosfets take some kind of additional penalty beyond normal acceleration operations while in buck or boost converter modes.

10 hours ago, Chriull said:

Main heat losses occur on R coil and R battery which are both in the range of tenth of Ohms.

Mosfet resistances are hundredth of ohms - so about a decade less.

That follows with what I've seen messing around with the ebikes.ca motor simulator where they tend to estimate an ebike controller at .03 ohms, battery packs at .2 ohms and motors, depending on the winding, usually over .2 if not .3 ohms.

Yes, I definitely botched that explanation of heat losses. Another try for fun:

Heat losses are always present in EUC operation primarily through Rcoil and Rbatt and are a function of current, during both acceleration and deceleration operations, with a minor or comparatively negligible amount occurring in the motor controller and other wiring. Regen only occurs above a dynamic power threshold which is both RPM and current dependent and requires the controller to be in boost converter mode.

Edited October 9, 2022 by Vanturion
wording

[blox1130]

On 10/8/2022 at 11:02 AM, enaon said:

 

it is refreshing to see someone, from USA too, using the nominal voltage instead of peak 

I guess it's helpful to just use the number of cells in series, so everyone knows instantly what is meant. But yeah I generally use nominal because the "max" isn't a true set value and depends on where you want to stop charging. I wish manufacturers made it clear when listing their voltage if that's the maximum or not, cause I buy my own chargers.

 

[Chriull]

11 hours ago, Vanturion said:

Neglecting any kind of field weakening effect (not even sure if that's relevant to any EUC motor controllers anyway),

No idea - at least it was never mentioned from anyone

11 hours ago, Vanturion said:

would it follow that maximum available braking current is the inverse of maximum available acceleration current as the spinning motor approaches back emf voltage parity?

I would not see a reason for any such "hard" real limit for breaking. But this does not have to mean anything - not too much discussions were in this direction.

[blox1130]

@VanturionThere's no reason EUCs couldn't use field weakening. Some of the people who built onewheels with vescs have posted videos showing a top speed increase.

[0000]

12 hours ago, Chriull said:

I would not see a reason for any such "hard" real limit for breaking. But this does not have to mean anything - not too much discussions were in this direction.

Probably when the brakes "break" (my inner pedant was murdered some time ago in this conversation)

In all seriousness though, I guess what I was getting at with this line of questioning is if there was a limit in initiating a hard deceleration in which something does break or trying to understand what component(s) would be the limiting factor while "hard" braking. I've never gotten anywhere near the limit while braking to break traction yet or data logged a ride to tell what's going while decelerating, so yeah, just curious about these things.

I'm also vaguely aware of there being a saturation point based upon the design of a particular motor in which additional current yields diminishing returns with regard to torque and am curious how and if that comes into play in any realistic EUC braking scenario.

Edited October 10, 2022 by Vanturion