KT motor controllers -- Flexible OpenSource firmware for BMSBattery S/Kunteng KT motor controllers (0.25kW up to 5kW)

[Iambuilderman]

I have several motors, all selected from AliExpress for minimal width (I want 7 speed gears as well) and maximum performance vs price ratio:

1. A 1500 watt no-brand 3T motor with ferrofluid: comes from a totAl package of motor + controller + wheel + accessories that many sellers like pasionebike sell. Great value motor, phase wires are quite thin (2.5 mm2?). Currently connected to the modified kt controller. Handles up to 80 amps before phase wires start to heat rather quickly. Easily hits 72 km/h with stock kt firmware.

2. A 3000 watt rated 3t motor from the seller Suringmax. Also filled with some ferrofluid. This one is connected to my sabvoton. I run it at 8000 watt max (130 Amp peak). I think it is quite similar to the mxus 3k motor. I think MQCON would rate these type of motors at 2000 watt. It has a temp sensor and a spare set of hall sensors + cable. I quite like it, although it was not perfectly balanced (little shake).

3. A no brand 48v motor rated by the factory at 2000 watt. This is my secret find on AliExpress. I think the factory and seller don’t know what they are selling. It looks the same as the 1500w no brand motor, except that it has 50 mm magnets (55mm casing), over 6mm2 phase wires and runs very fast at 72v (3t). So my own rating for this motor is 5 kW, based on its size and ability to shed heat compared to similarly sized motors. It is very basic, with 1 set of hall sensors, no temp sensor and a steel inner structure. But it has large magnets and lots of copper, so I am planning to slam some serious power through it. Of course also with ferrofluid added.

So in conclusion: a bit out of the box motors compared to the general public perhaps?

 

[Vbruun]

That i am not sure if voltage divider would help. Can you describe how are powering the T9? Are you disconnecting the battery with a connector or maybe a switch? Those small DC/DC converters are not really good in protecting against voltage spikes and that also could be the reason to kill the sensor.

I run it from the main battery switched by a mosfet that is turned on by the light output of the controller.

The voltage conversion is done by one of these https://www.aliexpress.com/item/32910554251.html?spm=a2g0s.9042311.0.0.27424c4dToEae3

 

[Iambuilderman]

Question (“Let’s talk about FETs”): how many peak battery volts have people successfully run their Toshiba TK150E09NE at (rated 85v dss)? I am wondering, because one of my first 18 fet KT controllers came with IRFB3206GPbF rated at 60V VDSs and ran fine for years at 58.8v with regen enabled. So without any margin for voltage spikes.

Currently I run my 18 fet with tk150e09ne at 58.8v max, which leaves a healthy margin of 25V for VDSs. But given the experience with the irfbs, these Toshiba’s might actually survive a lot higher voltage in practice?

If so, it might be enough to only change the electrolytic and potentially also Cbb capacitors for higher capacity and voltage ones to turn a modern iteration of the 18frt 48v kt controller into a 72v 18 fet one, saving a lot of time, effort and swearing on changing the mosfets as well? I might actually try it myself soon if I can gather the nerves...

 

[sdobbie]

Can someone please explain to me the hall sensor angles shown and how I can set the proper values? The motor runs fine but will not start unless I pedal first. My motor specific angle is set so that there is no drag when I turn the wheel so I would like to set the hall sensor values.
I flashed in diagnostic mode and recorded the uint8_t_60deg_pwm_cycles[0] to 5 as stancecoke advised but I am not sure how to use the data.

With the motor at full speed I get 10,0, 8, 8, 0, 10

With the motor stopped I get
195,0, 144, 240, 0, 101

 

[Tepa76]

Hi,

Just received Erider T9 torq sensor to be later added for my Kunteng 6fet controller. Because the rumours of quality issues and broken sensors i did following founding’s.

1. Cadence output has different topology than regular PAS sensor.
T9 Cad output is sourcing 4.1 volt at high state and grounded at low!
Kunteng has also sourcing input to 5v via internal pullup.. right? This might lead to issues? (how is this in Lishuis? pullup to 3.3v?)
This might be solved by removing pullup flag from input and also swich from falling to rising edge might be considered (not necessary).

I had also a parallel project where i use a ESP32 (3.3v) with Zwift game for deliver cad and torq... there i add a n-fet on that input and feed T9 cadence signal through it. Works but needed to change for rising for proper edge detection.

2. Issue and probably to the worst thing is these cheap buck converter modules!
Mine works decent at on desc.. getting only @20mV ripple out through it, but things go horrible after installed these on bike.
There is high level high frequency spikes that comes through it. I even try to improve this by adding a linear post regulation, but without success.. these spikes come through still.

600mA (HT7463) https://www.ebay.de/itm/DC-DC-Buck-...var=491696696057&_trksid=p2060353.m2749.l2649
version is little better than
2A (MP4560) https://www.ebay.de/itm/2A-Mini-DC-...var=562960492188&_trksid=p2060353.m2749.l2649

600mA on left:

This is something which varying on every installation depending in many scenarios..
My goal is to find better solution for harsh environment. Have anyone tried, Traco or similar converters?

Example this: https://www.tracopower.com/int/model/tec-2-4819wi

ps just found this one.. at least more caps and diodes on board.
https://www.aliexpress.com/i/32965982441.html?spm=2114.12057483.0.0.6c217b4bpJJysc
https://www.ti.com/lit/ds/symlink/t...73112&ref_url=https%3A%2F%2Fwww.google.com%2F

br,
Tep76

 

[Vbruun]

Hi,

Just received Erider T9 torq sensor to be later added for my Kunteng 6fet controller. Because the rumours of quality issues and broken sensors i did following founding’s.

1. Cadence output has different topology than regular PAS sensor.
T9 Cad output is sourcing 4.1 volt at high state and grounded at low!
Kunteng has also sourcing input to 5v via internal pullup.. right? This might lead to issues? (how is this in Lishuis? pullup to 3.3v?)
This might be solved by removing pullup flag from input and also swich from falling to rising edge might be considered (not necessary).

I had also a parallel project where i use a ESP32 (3.3v) with Zwift game for deliver cad and torq... there i add a n-fet on that input and feed T9 cadence signal through it. Works but needed to change for rising for proper edge detection.

2. Issue and probably to the worst thing is these cheap buck converter modules!
Mine works decent at on desc.. getting only @20mV ripple out through it, but things go horrible after installed these on bike.
There is high level high frequency spikes that comes through it. I even try to improve this by adding a linear post regulation, but without success.. these spikes come through still.

600mA (HT7463) https://www.ebay.de/itm/DC-DC-Buck-...var=491696696057&_trksid=p2060353.m2749.l2649
version is little better than
2A (MP4560) https://www.ebay.de/itm/2A-Mini-DC-...var=562960492188&_trksid=p2060353.m2749.l2649

600mA on left:

This is something which varying on every installation depending in many scenarios..
My goal is to find better solution for harsh environment. Have anyone tried, Traco or similar converters?

Example this: https://www.tracopower.com/int/model/tec-2-4819wi

ps just found this one.. at least more caps and diodes on board.
https://www.aliexpress.com/i/32965982441.html?spm=2114.12057483.0.0.6c217b4bpJJysc
https://www.ti.com/lit/ds/symlink/t...73112&ref_url=https%3A%2F%2Fwww.google.com%2F

br,
Tep76

Very interesting work you have done. I Think the final piece for this firmware would be to not have it fry the erider torque sensor as it is quite a nice sensor when it works.

As far as I can tell, the ripple current is still well within the specified voltage input for the erider (8-12v), so why do you think it still might be an issue?

Have you considered turn on/off voltage spikes in your work?

 

[Valopallo]

Troubleshooting help needed:

I changed hall sensors to my geared hub motor(52V battery/S06S controller). Last night I flashed S06S with OSEC and tried to find the right motor specific angle. I started the controller and the wheel started to spin on it's own without throttle. I tried different motor specific angle @ BluOSEC options increasing and decreasing in steps of 15 from the default 213. The wheel never stopped but the behavior changed with different parameters. I want to understand did the wheel spin because of the wrong motor specific angle parameter or some other parameter?

Then the motor made an awful sound and before I got to shut down the controller my battery blew a fuse. Changed the fuse and the battery is fine. Controller doesn't show any visual damage on pcb but I can feel that when I plug in the motor, the wheel has a resistance without power. I believe I blew a mosfet from the controller because without controller the wheel rotation and the phases are fine.

How to do the troubleshooting on board with a multimeter/blueterm before I connect the controller to the battery again? I have spare mosfets and other S06S components.

 

[geofft]

Troubleshooting help needed:

I changed hall sensors to my geared hub motor(52V battery/S06S controller). Last night I flashed S06S with OSEC and tried to find the right motor specific angle. I started the controller and the wheel started to spin on it's own without throttle. I tried different motor specific angle @ BluOSEC options increasing and decreasing in steps of 15 from the default 213. The wheel never stopped but the behavior changed with different parameters. I want to understand did the wheel spin because of the wrong motor specific angle parameter or some other parameter?

Then the motor made an awful sound and before I got to shut down the controller my battery blew a fuse. Changed the fuse and the battery is fine. Controller doesn't show any visual damage on pcb but I can feel that when I plug in the motor, the wheel has a resistance without power. I believe I blew a mosfet from the controller because without controller the wheel rotation and the phases are fine.

How to do the troubleshooting on board with a multimeter/blueterm before I connect the controller to the battery again? I have spare mosfets and other S06S components.

I managed to blow several mosfets during testing of the firmware, in every case the mosfet had gone full short D-S so was easy to locate with a multimeter in circuit. Usually both lowside and highside fets were damaged.

However, with shorted fets the fuse would instantly blow on reconnection of the battery to the controller - yours sounds like it isn't doing this, so your issue may be different. I would be inclined to do resistance checks of the fets in circuit and replace any that look different to the others in any significant way.

This is my rather simplistic method of testing, maybe you'll get some more 'in depth' suggestions....

 

[Valopallo]

I tested the mosfets on board with mutlimeter and diode testing mode; positive lead to drain and negative lead to source. I found two mosfets had continuity between their drain and source so I changed those to the new ones. I connected the motor cable to check if the phases are still resisting the wheel with no power and there seems to be no resistance. That means the mosfets were bad and now they are good.

How to find the right motor specific angle so that I don't accidentally do this same thing again?

 

[geofft]

How to find the right motor specific angle so that I don't accidentally do this same thing again?

....stancecoke describes a method on page 170 of this thread.

 

[Vbruun]

It seems that it might be best to use the stancecoke approach at a modest voltage..

 

[Valopallo]

It seems that it might be best to use the stancecoke approach at a modest voltage..

Please explain what do you mean. I'm using full 14s(58.8V) battery to find the motor specific angle with S06S.
Does the parameter gear ratio = 100 poise a risk of some sort when finding out the right angle?

Also, how to make the throttle react optimally? What to look for BluOSEC or Blueterm?

 

[Vbruun]

It seems that it might be best to use the stancecoke approach at a modest voltage..

Please explain what do you mean. I'm using full 14s(58.8V) battery to find the motor specific angle with S06S.
Does the parameter gear ratio = 100 poise a risk of some sort when finding out the right angle?

Also, how to make the throttle react optimally? What to look for BluOSEC or Blueterm?

I just commented that because it seems that several people have blown fets while looking for the angle. Thought that lower voltage woul probably be safer.

For the throttle reaction, I think the settings are called kP and kI. The need to be a Lot lower than default. I have had good luck simply experimenting with the settings until I found some I liked.

 

[Valopallo]

It seems that it might be best to use the stancecoke approach at a modest voltage..

Please explain what do you mean. I'm using full 14s(58.8V) battery to find the motor specific angle with S06S.
Does the parameter gear ratio = 100 poise a risk of some sort when finding out the right angle?

Also, how to make the throttle react optimally? What to look for BluOSEC or Blueterm?

I just commented that because it seems that several people have blown fets while looking for the angle. Thought that lower voltage woul probably be safer.

That's probably true. I just blew 4 more Are there any other ways to find out motor specific angle that are not as risky as this one? Also at last test I think I found the best angle near 255, but there seems to be some kind of pulsing effect where motor speeds up gradually even the throttle is full open. What could cause this because I'm pretty sure this is what keeps my mosfets blowing right at the moment when I seem to find the optimal angle?

 

[stancecoke]

I strongly recomment, to make sure, that the motor runs properly with the original firmware first, to find the right hallsensor and phase wire combination. Then the motor will run with the default settings in the open source firmware. You can fine tune the angle afterwards but with maximum +/- 15 from the default setting.

If you can't do that, for what reason ever, I strongly recomment to use a lab power supply for testing, with a max amp setting of 3A e.g.

In the Lishui project I've implemented a self learning function, like many very cheap China-controllers like the Brainpower have. This makes the start for newbies much easiser

but there seems to be some kind of pulsing effect where motor speeds up gradually even the throttle is full open.

This may be the speed limiting function. The speed oszillates with no load....

regards
stancecoke

 

[Vbruun]

So for poor people, er can power the controller from the charger?

 

[Valopallo]

So for poor people, er can power the controller from the charger?

I have a spare 42V charger and I'm stepping it down with LM2596HV -module to 24-36V. It has a limit of 2.0A and the LM2596HV has 3A. Can it survive?
My battery blew it's 30A fuse twice so maybe lower voltage could prevent the mosfet burn right? I wonder if it matters that I test the motor specific angle with 36V because I'm going to eventually use it with 14S battery and higher voltage?

So If I change the controller parameters as undervoltage to 20V, overvoltage to 30V and max current to 3A, with max phase current to 6A, I could use the 24V input and I should see the same load difference but as in proportionally lower values than with 15A max current and 52V battery, probably safely finding the right motor specific angle. Right?

 

[sylvain_wm]

Hi,
I have a problem with my fatbike.
The max speed I got with a 48V battery was 35km/h.
I decided to get a 72V controller with a 72V battery to increase the max speed, but it doesn't work at all.
The max speed remains at 35km/h.
Does anyone have an idea?
Here is my config.h
View attachment config.h
Another problem : the "battery current max" doesn't work. I set the "assist level 5" to 12 to limit the current.

Have a good day.

 

[atkforever]

What is the speed when unloaded (ie wheel in the air)?

I draw about 15A to remain at 38kph on flat so maybe your current limit is too low

 

[Valopallo]

This may be the speed limiting function. The speed oszillates with no load....

The gear ratio and speed limits were set to 100. Are there some other limiting factors?

I can only input 24V-42V (2.2A) from old charger and LM2596HV if I don't want to use my 52V battery.

EDIT: Ok. I think I found the correct motor specific angle. I used 24V from charger and LM2596HVS and was able to find an angle that was smooth and needed very little load by hand to respond to the low voltage.

I was about to try this parameter with 52V and once again the controller burned a fuse from my battery even though I didn't even had a chance to press the throttle. Something brakes the mosfet(s) and tries to kill the battery. What component I should test/repair to find out the cause?

 

[sylvain_wm]

What is the speed when unloaded (ie wheel in the air)?

I draw about 15A to remain at 38kph on flat so maybe your current limit is too low

Hi, the speed when unloaded is the same (35km/h).
When I reach 35km/h the current decreases and I have no power anymore.
The max current doesn't change anything.

 

[Vbruun]

This may be the speed limiting function. The speed oszillates with no load....

The gear ratio and speed limits were set to 100. Are there some other limiting factors?

I can only input 24V-42V (2.2A) from old charger and LM2596HV if I don't want to use my 52V battery.

EDIT: Ok. I think I found the correct motor specific angle. I used 24V from charger and LM2596HVS and was able to find an angle that was smooth and needed very little load by hand to respond to the low voltage.

I was about to try this parameter with 52V and once again the controller burned a fuse from my battery even though I didn't even had a chance to press the throttle. Something brakes the mosfet(s) and tries to kill the battery. What component I should test/repair to find out the cause?

Are you sure that your controller is a 48v version?

 

[Valopallo]

This may be the speed limiting function. The speed oszillates with no load....

The gear ratio and speed limits were set to 100. Are there some other limiting factors?

I can only input 24V-42V (2.2A) from old charger and LM2596HV if I don't want to use my 52V battery.

EDIT: Ok. I think I found the correct motor specific angle. I used 24V from charger and LM2596HVS and was able to find an angle that was smooth and needed very little load by hand to respond to the low voltage.

I was about to try this parameter with 52V and once again the controller burned a fuse from my battery even though I didn't even had a chance to press the throttle. Something brakes the mosfet(s) and tries to kill the battery. What component I should test/repair to find out the cause?

Are you sure that your controller is a 48v version?

Caps are rated 63V so I'd imagine yes.

The model is actually KT24/36SVPR-GLT2F but I have been in that impression that this can handle 59V because the caps are 63V. Am I wrong?

EDIT: I changed the mosfets once again and tried with 42V straight from the charger. I noticed that bluOSEC show a load of 8.5A without throttle, without brake and withtout motor connected. I noticed this also the first time it burned mosfets. Only when I used 24V I did not saw that idle current draw and was able to spin the motor. Even then it felt like there was some sort of limit the controller would reach before resetting and starting over. How to troubleshoot?

I think the motor can't handle 59V because of it's windings are rated to 36V. Could this cause the idle load?

 

[Vbruun]

No. The motor should not see any current when the controller is idle.

I don't know what is wrong, but buy a 48v rated controller and Gry again if you want to get this to work

 

[Valopallo]

EDIT: Both controllers burned mosfets because of the freaking angle! just right when I was about to go and test the bike. I hate this hobby. Let's continue.