VESC vs "Native" controller

General topics and discussions about the VESC and its development.
buran
Posts: 12
Joined: 28 Jan 2017, 18:02
Location: Russia

VESC vs "Native" controller

Postby buran » 27 Aug 2017, 18:09

Hello guys!

I have VESC 4.12 and tried to use it with my geared motor-wheel (12'',350W, 48V, gear ratio 4.4). Also I have "native" FOC controller for this motor-wheel.
"Native" controller has very modest current-limits, like 10A continous and 15A peak, so my idea was to "tweak" motor with VESC.
But results are very strange.

Native controller can perform 35km/h of max speed while VESC only 28km\h. At the same time torque of VESC is definetely feels much higher than torque of native controller. I cannot measure torque in numbers, but let's say that torque of VESC higher at the same ratio as VESC's loss of max speed (i.e. 20%). At least it will be physically correct :)

So my question is - how this can be true? If I understand well, a ratio between torque and maxspeed fundamentally depends only on physical properties of motor - number of windings and thickness of copper. Can this situation it be consequence of bad motor parameters detection? If so, what is the best practice to determine this params some way to verify results from bldc tool?

Thank you

devin
Posts: 253
Joined: 08 May 2017, 01:55
Location: San Francisco, California, US

Re: VESC vs "Native" controller

Postby devin » 27 Aug 2017, 18:31

buran wrote:Hello guys!

I have VESC 4.12 and tried to use it with my geared motor-wheel (12'',350W, 48V, gear ratio 4.4). Also I have "native" FOC controller for this motor-wheel.
"Native" controller has very modest current-limits, like 10A continous and 15A peak, so my idea was to "tweak" motor with VESC.
But results are very strange.

Native controller can perform 35km/h of max speed while VESC only 28km\h. At the same time torque of VESC is definetely feels much higher than torque of native controller. I cannot measure torque in numbers, but let's say that torque of VESC higher at the same ratio as VESC's loss of max speed (i.e. 20%). At least it will be physically correct :)

So my question is - how this can be true? If I understand well, a ratio between torque and maxspeed fundamentally depends only on physical properties of motor - number of windings and thickness of copper. Can this situation it be consequence of bad motor parameters detection? If so, what is the best practice to determine this params some way to verify results from bldc tool?

Thank you



Hi,

In addition to the physical parameters of the motor, there are the motor amp limit and battery amp limit settings which determine how many watts will go into the motor at different rpms & depending on the throttle position (these parameters affect the duty cycle % or %ON TIME of the mosfets leading from the battery under different loads and conditions).

Could you please tell me the FOC detection of the winding resistance in ohms, as well as your current battery amp limit and motor amp limit settings?

-Devin

devin wrote:Source: http://vedder.se/forums/viewtopic.php?f=6&t=883&start=50

Here is an interesting graph. I asked an anonymous rider to send me an acceleration log documenting loaded full throttle acceleration from standstill on an electric skateboard using a vesc. I'm not sure exactly which motor type was used or the motor amp limit and battery amp limit settings used, but it clearly shows around 90a motor amps produced in the motor from only 10a battery amps close to full throttle standstill:

Image

This second graph taken by the same rider on the same board clearly shows (@ log 48 sec) that at about 28mph, the same 10a battery amps produces only about 28a motor amps instead of about 90a as in the 1st graph, a result of back emf voltage opposing the battery voltage more and more at faster and faster speeds.

Image

The electrical wattage in both cases -- the 90a motor amps @ 10a battery amps -- and the 28a motor amps @ 10 battery amps -- is the same.

Simply if he wants to draw 500w electrical (10a battery amps x 50v) full throttle at all physically possible speeds including close to standstill, his motor amp limit setting needs to be at least about 90a motor amps.

buran
Posts: 12
Joined: 28 Jan 2017, 18:02
Location: Russia

Re: VESC vs "Native" controller

Postby buran » 27 Aug 2017, 19:53

Hi Devin, thank you for your attention!

Resistance in Ohms in FOC mode was something like 0,05Ohm (0,0514 once and 0,047 another time).
I have setup battery limits to 15A (48V * 15A ~ 750W)
So I have P = 750W, P = I^2*R, R ~ 0,1 Ohm ( detected resistance by 2 as I need phase to phase resistance) => 750 = I^2 * 0,1 => I = sqrt(7500) ~85A
This number a have set as a motor limit: 85A

Small remark:
I have thought about bad limits, but (if I understand process correct) lower limits simply cause lower power, and as power = torque * speed, it means that if my low max speed caused by bad limits, i.e. by low power, than (as torque/speed ratio is constant for given motor) it means that torque must be less than torque of native controller too, but it is not true.

devin
Posts: 253
Joined: 08 May 2017, 01:55
Location: San Francisco, California, US

Re: VESC vs "Native" controller

Postby devin » 27 Aug 2017, 20:27

maybe the original controller was BLDC (not FOC?) and the VESC is FOC? In that case it would be the difference between BLDC and FOC. We did some tests on an electric skateboard and found that on a VESC, the skateboard had a higher maximum speed when in BLDC mode than FOC mode on a VESC.

also isn't there some timing advance parameter that determines when the commutations happen relative to the rotor position?

or perhaps you would get the original max speed by switching to BLDC mode?

or try adding some more watts? :)

buran
Posts: 12
Joined: 28 Jan 2017, 18:02
Location: Russia

Re: VESC vs "Native" controller

Postby buran » 28 Aug 2017, 06:29

maybe the original controller was BLDC (not FOC?) and the VESC is FOC?

they are both FOC, at least manufacturer of "native" controller say so.

In that case it would be the difference between BLDC and FOC. We did some tests on an electric skateboard and found that on a VESC, the skateboard had a higher maximum speed when in BLDC mode than FOC mode on a VESC.

Yes, you are right (and this is another question, why?) I have the same situation on my both wheels (DD and geared, both have hall sensors). VESC in BLDC-mode in both cases works with higher max rpm (no load) than in FOC-mode, but difference was not too big, about 5-8%.

also isn't there some timing advance parameter that determines when the commutations happen relative to the rotor position?

As I have hall sensors in my wheel - I thought that commutation sync will not the problem, but may be you are right and i must look deeper.

or perhaps you would get the original max speed by switching to BLDC mode?

nope, cannot achieve it even in BLDC mode :(

or try adding some more watts? :)

draft try was unsucsessfull (max speed was the same), but may be i miss something, will try to do it more carefully

rew
Posts: 940
Joined: 25 Mar 2016, 12:29
Location: Delft, Netherlands.

Re: VESC vs "Native" controller

Postby rew » 28 Aug 2017, 06:31

buran wrote:Hi Devin, thank you for your attention!

Resistance in Ohms in FOC mode was something like 0,05Ohm (0,0514 once and 0,047 another time).
I have setup battery limits to 15A (48V * 15A ~ 750W)
So I have P = 750W, P = I^2*R, R ~ 0,1 Ohm ( detected resistance by 2 as I need phase to phase resistance) => 750 = I^2 * 0,1 => I = sqrt(7500) ~85A
This number a have set as a motor limit: 85A
This is really an odd calculation.

If the motor is advertized as 750W, then they mean that you can put in 750W of electrical power or that it will put out 750W of mechanical power. The difference between the two should be about 10% making the motor about 90% efficient. So we can estimate the motor can handle about 75W of power.

You have made a calculation that will put 750W of losses into the motor. I don't think the motor will like you for that....

buran
Posts: 12
Joined: 28 Jan 2017, 18:02
Location: Russia

Re: VESC vs "Native" controller

Postby buran » 28 Aug 2017, 08:24

Yes, actually you are right. Epic fail :) Unfortunately it have no any influence (as far as my engine is still alive :)) on the final situation.
Engine marked as 350W nominal. I have set 750W for the sake of tweak :) and because of knowledge that in uphill moving native controller "eats" 15A from the battery, so I decided that 750W is ok for this motor at least in a short experiments.
So, if we calculate 750W as an input power and will assume 10% as an heat loss, we dissipate 75W => 75 = I^2*0,1 => I = 27A.
I had set 25 and 30 limits for motor current and it had no result in a max speed value.

May be there are some other reasons ?

devin
Posts: 253
Joined: 08 May 2017, 01:55
Location: San Francisco, California, US

Re: VESC vs "Native" controller

Postby devin » 28 Aug 2017, 13:56

rew wrote:This is really an odd calculation.

If the motor is advertized as 750W, then they mean that you can put in 750W of electrical power or that it will put out 750W of mechanical power. The difference between the two should be about 10% making the motor about 90% efficient. So we can estimate the motor can handle about 75W of power.

You have made a calculation that will put 750W of losses into the motor. I don't think the motor will like you for that....


the rpm at which you get max mechanical power out per electrical power in varies with voltage and kv...

assuming 100% duty cycle, you get about 85% conversion of electrical to mechanical power at about 85% of no load rpm.

Image

at 50% of no load you have half of max torque and half of max rpm giving max power output.

at 85% of no load you have only 15% of max torque, but 85% of max rpm... since the electrical consumption at this rpm (thanks to back emf) is so low, and the output wattage (torque times rpm) is still considerable, this particular rpm gives best electrical to mechanical conversion efficiency.

if the motor is rated to 350w it likely means it can handle 350w continuously at stall, which would equate to about 60 motor amps assuming 0.1ohm lead to lead resistance via ohm's law.

devin
Posts: 253
Joined: 08 May 2017, 01:55
Location: San Francisco, California, US

Re: VESC vs "Native" controller

Postby devin » 28 Aug 2017, 14:18

devin wrote:the rpm at which you get max mechanical power out per electrical power in varies with voltage and kv...

assuming 100% duty cycle, you get about 85% conversion of electrical to mechanical power at about 85% of no load rpm.


if we assume the motor is 100kv and the lead to lead resistance is 0.1ohm & battery voltage 50V

0% rpm = 0rpm = 500A motor amps @ 100% duty = 25000W electrical

100% rpm = 5000rpm = 0A motor amps @ 100% duty = 0W electrical

therefore:

most efficiency: 85% rpm = (85/100)(5000rpm) = 4250rpm = 75A motor amps @ 4250 rpm @ 50V @ 100% duty gives peak electrical to mechanical conversion efficiency

most mechanical power: 50% rpm = (50/100)(5000rpm) = 2500rpm = 250A motor amps @ 2500 rpm @ 50V @ 100% duty gives peak mechanical power

most torque: 0% rpm = (0/100)(5000rpm) = 0rpm = 500A motor amps @ 0 rpm @ 50V @ 100% duty gives peak torque

most rpm: 100% rpm = (100/100)(5000rpm) = 5000rpm = 0A motor amps @ 5000 rpm @ 50V @ 100% duty gives peak rpm

Simply at 100% duty cycle (full throttle), if the magnitude of the mechanical load causes the motor to slow down to 85% of no load rpm, you will be at peak electrical to mechanical conversion efficiency (and motor size and kv is considered very well balanced with the mechanical load).

At 100% duty cycle (full throttle), if the magnitude of the mechanical load causes the motor to slow down to 50% of no load rpm, you will be at peak mechanical output of your motor at that operating voltage (this condition causes maximum power to weight ratio of your motor/vehicle combination, but at the sacrifice of efficiency.)
Last edited by devin on 28 Aug 2017, 15:20, edited 1 time in total.

devin
Posts: 253
Joined: 08 May 2017, 01:55
Location: San Francisco, California, US

Re: VESC vs "Native" controller

Postby devin » 28 Aug 2017, 15:16

was the "native" controller capable of 100% duty cycle? i've heard from anonymous sources the VESC only goes up to 95%.


Return to “General”

Who is online

Users browsing this forum: No registered users and 2 guests