Revolt 120 pro eTrial overheating problem

jonnydrive

100 W
Joined
May 11, 2015
Messages
174
Location
Italy
Hi,

after my first experience with DIY EV bike (https://endless-sphere.com/forums/viewtopic.php?f=28&t=73027&hilit=cyclone+1680+overheating), I moved on with a new project, an electric trial motorbike.

I choose Revolt 120 pro with 30 Kv managed by a Kelly controller KLS7230S for my new project.
I just finish to assembly all components obtaining what you see in following pictures.

20171021_144810.jpg

View attachment 1




If someone is interested, I will post the full building log as soon I fix my actual overheating problem.

Here are some specs:
Engine: revolt 120 pro 30kv directly connected to the wheel (datasheet says 6kw - 15kw peak)
Controller: kelly Controller KLS7230S 72volt 300A max
Battery: 2x Hobbiking 16S3p 60v 15ah

After I completed all settings required on the Controller and double checked the wirings, I did my first ride test today.
The result: after 30 seconds of ride the controller stopped working
The problem: engine overheating! more than 170°C
After the walk of shame, I checked the temperature on the controller, it was still too high, maybe too much considering a flat riding of less than a minute with 10°C degree.

3.PNG


The following checks has been already done
- checked the wiring to avoid wrong readings that come from engine termistor
- reset controller and made a new Identification Angle procedure (to identify the right hall sensor phase)
- checked all the parameter using Kelly manual http://kellycontroller.com/mot/downloads/KellyKLS-SUserManualV1.10.pdf
- reduced the max controller output to 50% (less than the engine rated power)
- the engine spinned smoothly, also the torque was awsome. Should I assume that the controller settings were correct?
- the engine was really overheated, I checked it out touching it

Here are my configurations
5.PNG

6.PNG

7.PNG

8.PNG


Am I missing something to check? what are the possible reasons of this kind of overheating?

Any suggestion is appreciated!

Thanks!

Stefano
 
Since a touch check showed the motor really was hot, then gearing is often a reason for overheating; running the motor at too slow a speed for the current/voltage being put thru it, so it cant' build up enough back-emf to keep internal currents low so it overheats. This will show up as excessive power draw for the speed you're going under the terrain and wind conditions you have.

Wrong phase/hall combo can also do this; called a false positive, it can make the motor run in the correct direction, but will be far advanced or retarded in timing phase vs halls, so the controller signals are sent early or late. If the kelly is not autodetecting/autolearning type, or it is not correctly detecting them on it's own, you may have to manually determine the order.

There is an ES wiki article and at least one forum thread about how to determine the wiring for a brushless motor, by that title.
 
Thanks Amberwolf for the clarification.

I'am going to work with the phase/hall sensors to see if it is in a "false positive" configuration. I hope this is the problem, otherwise I need to replace the motor :cry: .

Stefano
 
id also reccomend turning down the phase current to ~10% until you get it working smoothly... will help stop you from cooking hte motor by accident
 
I guess you probably have the halls combinations correct from what you describe.

The phase current that was recommended to lower heating from Revolt in my case was max 200A so phase current percent would be lowered to 66
Gearing recommended from Revolt in my case was minimum ratio of 9front/45rear in 26" wheel

Here's what i did to get my Revolt to work (i guess you know it all but here it comes):
I tried auto-identify first, that did not work on my motor so i had to switch the hall wires manually until i found the good hall wires combination.

Google a commutation table so you can see what the correct hall sequence should be. From the diagnostic screen on the Kelly program you can see the status for each hall sensor.

If you turn the motor slowly by hand in the correct direction you will see the halls status change - if it follows the table then you can move on to the phase wire combinations.
If it's not according to the table then you can use the table to see which hall wires should be switched to get the correct sequence.

Once the hall sequence is correct you have to get the phase wire connections correct:
set the phase current to 20A so you don't burn the motor when you try the different combinations
Lift the wheel / secure that the motor can rotate freely.
There are six possible combinations between the phase wires on controller and motor so it is fairly quick work, one of the combinations will give rotation and good torque on low speeds without stuttering or screaming when you give throttle.

According to Fany at Kelly there is no manual tuning of the phase angle in the Kelly program. If your hall placements are as bad as mine then you need to replace the hall sensors to ever get any good performance - good luck!

PS. if you get trouble with warranty from Revolt then you would probably need to show that halls are bad. If you have a friend with an oscilloscope then it's possible to prove it by measurement, i'd really like to see that done.
 
larsb said:
I guess you probably have the halls combinations correct from what you describe.

Hi larsb thanks a lot for the super-fast reply!
unlike your case, I have been able to auto-configure the Kelly controller (lucky me!)

Alexey still suggest me to try other hall/phase combinations, I doubt that will be the solution!

Here are the tests that I made (I have checked all the possibile six phase combinations, assuming that the auto-configure on controller is able to identify the related and correct hall sequence/angle):
http://www.cerberotech.com/etrial/test_plan.PNG

here the results:
http://www.cerberotech.com/etrial/t1.mp4
http://www.cerberotech.com/etrial/t2.mp4
http://www.cerberotech.com/etrial/t3.mp4
http://www.cerberotech.com/etrial/t4_1.mp4
http://www.cerberotech.com/etrial/t4_2.mp4
http://www.cerberotech.com/etrial/t5.mp4

All the tests produced the exactly same results in term of engine speed and sound (obviously in some cases the rotation direction changed).
The vibration that you hear in the video came from the frame not from the motor.

All the tests had the same result because every time I set up the controller for auto-configure, otherwise the result are in the following video:
http://www.cerberotech.com/etrial/Wrong_phase.mp4

larsb said:
Here's what i did to get my Revolt to work (i guess you know it all but here it comes):
I tried auto-identify first, that did not work on my motor so i had to switch the hall wires manually until i found the good hall wires combination.

Google a commutation table so you can see what the correct hall sequence should be. From the diagnostic screen on the Kelly program you can see the status for each hall sensor.

If you turn the motor slowly by hand in the correct direction you will see the halls status change - if it follows the table then you can move on to the phase wire combinations.
If it's not according to the table then you can use the table to see which hall wires should be switched to get the correct sequence.

As you suggested I have compared a commutation table found at the following link to the controller phase value in the Controller under "Monitor" section.
https://e2e.ti.com/blogs_/b/motordr...ontrol-3-phase-bldc-motors-using-hall-sensors

comm_table.PNG


The results, as reported in the following video are right, the controller follow the correct sequence on the table slowly by turning the engine in the correct direction.
I have no Idea how to check the correspondence between the Hall Sensor position (the first column in the table) with the motor phase (the second column in the table), but I think that it doesn't matter, the engine will manage the current on phase according to the Hall position (yes if the angle is wrong it is a problem, with kelly no way to manage that).
http://www.cerberotech.com/etrial/phase_check.mp4

Do you think that is enough to consider the correct phase/halls settings?

larsb said:
According to Fany at Kelly there is no manual tuning of the phase angle in the Kelly program. If your hall placements are as bad as mine then you need to replace the hall sensors to ever get any good performance - good luck!

Thanks for this precious info, I am struggling to find the correct entry in the controller settings to change the phase angle, it isn't !

larsb said:
The phase current that was recommended to lower heating from Revolt in my case was max 200A so phase current percent would be lowered to 66
Gearing recommended from Revolt in my case was minimum ratio of 9front/45rear in 26" wheel

Yes, I have reduced the Current Percent to 50% (150A), I feel a considerable performance decrease, unfortunately engine still getting hot (it requires quite more time).

Said that, I see the following opinions:
- Get refund because the engine do not respect the paper performance (but nobody redound you for taxes and shipping costs)
- Disassemble/verify the phases with an oscilloscope and eventually replace the hall sensors (void the warranty)
- Provide an active cooling system to keep temperature low to the maximum accepting the energy waste

I will update soon...
 
I had trouble viewing the videos, from what you describe it seems all is correct as far as halls and phase combinations goes. What noload amps do you have at full throttle? I saw 20-30A in the video, that is a lot. 1200-1800w for turning motor without load indicates that something is off.

Does motor turn easily by hand? if it does then it's not an installation or bearings issue

Next step could be to measure the volts produced between phases (1-2, 1-3, 2-3) when you use the motor as a alternator.
Drive the motor with an external source like a drill press.
Rpms must be fixed or measured for each combination. Then you kan calculate kV value from each phase voltage, they should be the same and correspond roughly to your theoretical kV.
If they are not the same then motor is faulty by short circuit or incorrect winding.
 
larsb said:
I had trouble viewing the videos, from what you describe it seems all is correct as far as halls and phase combinations goes. What noload amps do you have at full throttle? I saw 20-30A in the video, that is a lot. 1200-1800w for turning motor without load indicates that something is off.

Does motor turn easily by hand? if it does then it's not an installation or bearings issue

Hi larsb,
As you suggested I make a new no-load fullspeed test, the resultrs are:
- controller, as you noticed, get high amp per phase but the meeter is very irregular
- current at constant full speed drained from the battery is 3.8A (it rises during the acceleration up to 8a for a few seconds), running at 60v it is 200-240watt
- the motor turn easy by hand, without irregular resistance.
- afer 5 minutes at full speed no-load the temperature rise to 55 C
- after 10 minutes at full speed no-load temperature rise to 65 C

As usual I made a video :D http://www.cerberotech.com/etrial/No_load.mp4

I am start thinking that there is no wrong in my motor, but it is simply NOT build to real manage the power that it is sell for.

larsb said:
Next step could be to measure the volts produced between phases (1-2, 1-3, 2-3) when you use the motor as a alternator.
Drive the motor with an external source like a drill press.
Rpms must be fixed or measured for each combination. Then you kan calculate kV value from each phase voltage, they should be the same and correspond roughly to your theoretical kV.
If they are not the same then motor is faulty by short circuit or incorrect winding.

It is clever, I will made this test!

Thanks a lot for the support

Stefano
 
but it is simply NOT build to real manage the power that it is sell for.

I believe the same from the experience of my motor. If you have a no-load amps at 3.8A then you have lower losses than in any of my two motors.

Your gearing seems to be low, i would have thought that you need only low speed for trials?
I would look at getting more gear ratio, that lowers the load on the motor a lot.
 
Yep,
I sized the engine considering a maximum speed of 50Km/h. What I need for trial motorbike is a lot of torque.

My target is 50Nm, I found this magic formula: Torque = (Power * 9,55) / rpm

with revolt 120 pro:
rpm kv*Volt = 30kv*60volt=1800rpm
with 10Kwatt theoretically this engine can put 53 Nm (also in some revolt datasheet it seems possible)

Yes, IF the engine not overheat after the first full throttle acceleration on flat! what will happen when I will run the trial on a real mountain path? a real trekking of shame? :cry:

I made a mtb with only 2kw able to climb any slope without overheating at 15 Km/h, from 5x powered engine I refuse to reduce the gear ratio

How is going your motor rewinding? Have you already finished? I am looking forward if your mod give the right vitesse to this motor!
 
I played the small motor game a bit, and ran into what you're running into.

If you only have a couple pounds of copper and iron, it's just not going to handle much continuous power.

The best mass you can add to a performance vehicle is copper, iron, and magnet.
 
The winding is quite a lot of work, i do one coil at a time, it takes around 1.5hrs so i will be finished in two weeks i think.
I might have to rework some coils, the wire is so stiff that shorts created during the winding are probable.

@livefor...: nice to see you in a thread, read a lot from you here on ES.

As the Revolt motors lack substantially in copper they would be better bought without winding. I hope that result for my motor is anyway OK after rewind. All the RC flight motors specs should be reduced a lot, i think the RV120 in best case will be 4kW continuous when they are not placed in a 6kW propwash that cools them.
 
Thanks for the kind words.


Rewinding can only offer a small incremental continuous power/torque improvement.

If you're at the brink of it being enough motor for you, it could be worth it.
 
Hi larsb,
I am not brave as you are, I am not going to rewind the motor by myself :D
But I may know a small company here in Italy that can rewind the engine, they operate with industrial application engine, but I may ask them if they can rewind smaller engine basing on a given specs.

In the meantime, I am quite surprised by Alexey, a revolt guy, he provides me support by mail and fast replay. He still suggests me to try to swap the halls because the engine must run smoother than what he saw in my videos. I'm going to do that to be 100% sure that I'm using the right hall/phases configuration.

However, a doubt grow in my mind: revolt provide phase and hall cables very short, I had to stretch the hall cable to reach the cable housing on the motorcycle seat. Is it possible that stretch of 50cm to the hall cable can, in some ways, produce a misalignment between phases and halls?

Another thought in mind: is it possible that revolt engines are born in aeronautical environment, so the use of a lower copper fill is to provide a good air flow to the engine during the fly (I imagine the engine directly connected to the airscrew so that the air flow steps parallel to the stators) and only with this configuration it is possible reach the power on the datasheet?
 
That your hall sequence follows the commutation table proves that halls are OK. You can try swapping the phases to see if that gives any improvement just as a proof for Revolt.

When you have such a low noload amperage shows that the motor is not running with wrong setup. One problem you might have though is interference in the signals from hall sensors, i have run both Adaptto and kelly controller and it seems the kelly is more sensitive to interference from the magnet fields.

the wire or connections can be damaged if you pull it tight. I got that after a short time on my first motor because the routing was too tight.
But i don't think that's the issue here, i assume overspeced motor, possibly in combination with offset hall sensors is the reason for the overheating.
 
Another interesting point for RV-160. I let the motor run freely for 1 minute at 80% throtle, and I already feel some heat in rotor casing, but only in rotor. Like 5C difrence from ambient temperature. Winding stayed same temperature after test. So there is some parasitic current in rotor.
 
Karolis said:
Another interesting point for RV-160. I let the motor run freely for 1 minute at 80% throtle, and I already feel some heat in rotor casing, but only in rotor. Like 5C difrence from ambient temperature. Winding stayed same temperature after test. So there is some parasitic current in rotor.

Only the motor is overheated at 30 seconds?How about the controller case temp?

1,We can try to swap phase wires or hall sensor wires to do the Identification angle operation again.

2,We think the motor may not be balanced well.The hall phase timing may not match with each other.
The tolerance of hall signals or hall/phase timing may be huge so that there is a big deflection between waveform of hall and phase timing.
If you have a oscilloscope,please try to capture the waveform of hall/phase timing.
I attached the instructions.
I had experienced with this issue before.
 

Attachments

  • Hall Phase timing 120 degrees.pdf
    71.8 KB · Views: 127
  • Hall Phase.pdf
    7.9 KB · Views: 122
fany said:
Only the motor is overheated at 30 seconds?How about the controller case temp?

1,We can try to swap phase wires or hall sensor wires to do the Identification angle operation again.

Hi Fany, thanks for your support!

The controller still cold when the motor overheats. For example:
- Motor temp. = 123°C
- Controller temp. 25°C.

I am going to swap the hall sensors and check if there is any improvement.

fany said:
2,We think the motor may not be balanced well.The hall phase timing may not match with each other.
The tolerance of hall signals or hall/phase timing may be huge so that there is a big deflection between waveform of hall and phase timing.
If you have a oscilloscope,please try to capture the waveform of hall/phase timing.
I attached the instructions.
I had experienced with this issue before.

If opt 1 do not provide any improvement I might be able to find an oscilloscope and I will check the motor as you suggested.

Many thanks

Stefano
 
jonnydrive said:
fany said:
Only the motor is overheated at 30 seconds?How about the controller case temp?

1,We can try to swap phase wires or hall sensor wires to do the Identification angle operation again.

Hi Fany, thanks for your support!

The controller still cold when the motor overheats. For example:
- Motor temp. = 123°C
- Controller temp. 25°C.

I am going to swap the hall sensors and check if there is any improvement.

fany said:
2,We think the motor may not be balanced well.The hall phase timing may not match with each other.
The tolerance of hall signals or hall/phase timing may be huge so that there is a big deflection between waveform of hall and phase timing.
If you have a oscilloscope,please try to capture the waveform of hall/phase timing.
I attached the instructions.
I had experienced with this issue before.

If opt 1 do not provide any improvement I might be able to find an oscilloscope and I will check the motor as you suggested.

Many thanks

Stefano

I made some tests swapping phase and hall cables (about 20 combinations of 36 possible).

Here what I saw:
- swapping halls does not provide any change (the sound and the engine speed are exactly the same), also "no-load full throttle test" give the same results on each swap
- swapping halls provide small changes only in "no-load full throttle test", the metered value are between 4.8A - 5.5A.
- for any combination ,I made the "no-load full throttle test" twice, every time re-doing the auto-configure option on controller, some times I saw a small change the results during the full throttle test.
- In my previous post I said that no-load amps are 3.8A but it was only one shot test. In this the new series of tests, I kept the exactly same condition: same ammeter position, exactly same controller configuration, battery fully charged (and kept under charger for all the time of the test)
 
fany said:
2,We think the motor may not be balanced well.The hall phase timing may not match with each other.
The tolerance of hall signals or hall/phase timing may be huge so that there is a big deflection between waveform of hall and phase timing.
If you have a oscilloscope,please try to capture the waveform of hall/phase timing.
I attached the instructions.
I had experienced with this issue before.

Under the precious suggestion of fany, I have proceeded my test with the option 2.
Unfortunately I do not have an oscilloscope, so, I tried to figured out how to check hall sensors.

My way is the best ones? sure not, but I think it works.

First I made an ultra-semplified oscilloscope using Arduino Uno (10 euros on Amazon), here the picture:

d1.png


The connections is quite simple:
- A0,A1,A2 are the input pin that came from the hall sensors of the engine
- A3 is connected to the potentiometer to trigger in real time the samples speed
- A4,A4 are used only for test the correct VCC and GND

After I coded the following in to Arduino.
d5.png


here you can find the code http://www.cerberotech.com/etrial/BLDC_halls_checker.ino

Then I spun the engine at constant speed using a drill connected directly to the shaft with a piece of water pipe (fixed with clamps).

I connected the engine hall sensors to the Arduino breadbord (the white ones) in this way:
The engine Hall wires to A0,A1,A3 (using the green clamp in the first photo)
The +VCC and GND wires from the engine to the GND and +VCC of Arduino (+5volt)

The Arduino IDE provides a "serial plotter" function (remember to configure this to a 115200 boud) and give you back the the data.

Here the results:
d4.png


d3.png


I made various test changing the drill the speed and triggering the potentiometer to correct the sampling rate of Arduino.

I think that something is wrong with one hall sensor (the blu one).
It both overlap the red sensor signal and give an out-of-range signal (it may be a sort of shortcut?)

d6.png


I contacted Revolt sending this results. I will update this post as soon I will receive an answer from them.

Stefano
 
Hi
Interesting that you can work with Arduino.

I seen the graphs and seems it shows a noise of about 2 Vpp.
If it is true you can configure the controller changing the threshold of the signal to exclude the noise.

I hope that this is so, because it would be the easiest solution.

Best wishes
Nuccio
 
elo said:
Look at this posts, the problem can be the big bearing....

https://endless-sphere.com/forums/viewtopic.php?f=30&t=75494

Hi elo,
I think that is not a mechanical problem, no load current is in an acceptable range. The problem is that the motor is unable to manage the power under load.
There is no way that my engine can manage 6kw of constant power, now I am trying to get the engine replaced with a good ones from Revolt.
 
Back
Top