KT48 (KT48ZWSRKT-SJT02L) controller problem

[markusb]

Hello guys,
This is my first post here.

Got a 48V/1500W kit 2 days ago. Mounted it and took a ride. No problems.
Yesterday I was gonna go for a ride but when turning the system on the battery indicator on the LCD3 did not light up and motor did not work.
After some measuring and faultfinding I noticed the 5V from the controller was missing. It was only 0.3V.
Opened the controller and investigated. Its the KT48ZWSRKT-SJT02L controller in this kit.

The 3W dropper resistor (270R) was discolored but still measured 270R, about 3.5V out from the LM317 reg, and 5V on the output of 7805 regulator was missing.
After some poking around and feeding the 5V rail with a CV/CC power supply I noticed the MCU (BS003F3P6) on the vertical board got hot. Pulling 0.5A at 1.6V from the power supply. I measured 4.5ohms across VDD and GND on this chip.

Removed the whole MCU board and the 5V is ok again.

Tried the controller using the power supply (50V/1.5A max) with this little board unsoldered and it does spin the motor but only to about 15-20km/h, than it either stops driving (freewheeling) or it just suddenly locks up the wheel pulling 1.5A and kicking the power supply in to current limiting.
Get error 03 which has to do with hall signaling.
Removing the hall signal connector completely makes the motor just tick when giving throttle.

This sudden breaking problem worrys me a bit. Im thinking about what could happen if I connect the battery and try this with no current limiting. Maybe the whole controller shorts. Maybe all FETs turns on the same time and thats what makes the motor stop so suddenly and triggering the CC on my PSU...

What is the purpose of this little board, do you know? I see alot of pictures on here of the KT48 controller but none of them has this little board in them.

Im speaking to the chinese about this now and hopefully they will help me.
Maybe they send a new MCU board for replacing, or hopefully a new controller as there seems to be strange behaviour in it now.

I was trying to find a schematic of the KT controllers on the web but only found a simplified schematic/block diagram here on ES.
Is there a more complete schematic where the function of this MCU can be seen?

 

[AHicks]

No idea what the issue might be, but I think it pretty safe to say customers are not expected to do board level diagnostics/repair. My experience with the controllers has been excellent. Never a real failure.

Point being, I would be after a replacement controller. My luck with vendors being mixed, if they wanted me to jump through a bunch of hoops for a replacement, I don't have the patience for that kind of thing. I'd prefer to just pay for a replacement, from somebody else....

 

[markusb]

Just to be clear, it was not the seller who asked me to diagnose this problem. I did it because I have electronics background and was hoping it was an easy fix like a bad solder or something simple.
Did not want to just order a new and wait the whole summer for it to be delivered, but it looks like thats about to happen.

I have spoken to the seller on ali and they are checking whats in stock and how they can help me.

Non the less Im still curious about this MCU and its function and most of all, why it shorted the second day using the kit. Maybe there are other problems I havent found yet.

 

[Comrade]

about 3.5V out from the LM317 reg, and 5V on the output of 7805 regulator was missing

So with the board removed you have 15V out of the LM317 and 5V out of the 7805?

 

[AHicks]

Just to be clear, it was not the seller who asked me to diagnose this problem. I did it because I have electronics background and was hoping it was an easy fix like a bad solder or something simple.
Did not want to just order a new and wait the whole summer for it to be delivered, but it looks like thats about to happen.

I have spoken to the seller on ali and they are checking whats in stock and how they can help me.

Non the less Im still curious about this MCU and its function and most of all, why it shorted the second day using the kit. Maybe there are other problems I havent found yet.

Is this a 35a controller? Have you tried sourcing through Amazon?

 

[amberwolf]

Is this a sensorless-capable controller? If so, that board may be what reads the phases to generate hall signals for the controller brain. (this used to be done with a comparator chip, usually LM339, but would be more tunable using an MCU, so perhaps they changed their methodology).

Alternately, if this controller is capable of reading a torque sensor, this board could be for that, instead of having the main MCU do it.


The best bet to find out what it is for is to trace where the signals to and from the board header go on the main board. Some of them probably go to the main MCU, and some of them to I/O (wires on board to/from outside world).

Once you know what it is for, then if it is a function you can live without, you can bypass it and use the controller until your new one arrives (if one is necessary).

As for the why of the failure...that depends on it's function.

For instance, if it's a sensorless-conversion board, then it has essentially direct input from the motor phase wires. This allows the potential for very high voltage spikes as feedback from the motor to overload any protections it may have, and exceed the limitations of the MCU inputs, which can cause it to fail in a way that makes it drag it's power supply down. (it's less likely for this to happen when using something like the LM339 comparator for this, as it's inputs are more tolerant of higher voltages).

 

[docw009]

My S06 controller has a vertical board that was used for lights, but there was no MCU. just a transistor and passive components.

Hope they send you a new controller. I was never inclined to repair these things. I always have a spare on hand. I've lost two due to water damage,

 

[markusb]

about 3.5V out from the LM317 reg, and 5V on the output of 7805 regulator was missing

So with the board removed you have 15V out of the LM317 and 5V out of the 7805?

Yes, its the MCU thats shorting the 5V to ground. I measure a 4.5 ohm resistance between its power supply legs.

 

[markusb]

Is this a sensorless-capable controller? If so, that board may be what reads the phases to generate hall signals for the controller brain. (this used to be done with a comparator chip, usually LM339, but would be more tunable using an MCU, so perhaps they changed their methodology).

Alternately, if this controller is capable of reading a torque sensor, this board could be for that, instead of having the main MCU do it.


The best bet to find out what it is for is to trace where the signals to and from the board header go on the main board. Some of them probably go to the main MCU, and some of them to I/O (wires on board to/from outside world).

Once you know what it is for, then if it is a function you can live without, you can bypass it and use the controller until your new one arrives (if one is necessary).

As for the why of the failure...that depends on it's function.

For instance, if it's a sensorless-conversion board, then it has essentially direct input from the motor phase wires. This allows the potential for very high voltage spikes as feedback from the motor to overload any protections it may have, and exceed the limitations of the MCU inputs, which can cause it to fail in a way that makes it drag it's power supply down. (it's less likely for this to happen when using something like the LM339 comparator for this, as it's inputs are more tolerant of higher voltages).

Yes, this is the ZW version (dual-mode) but the controller does not work with hall sensor cable removed at the moment.
It throws error 03 instantly, indicating hall sensor issue. If this board is generating hall signal from BEMF that would explain this problem.
With hall sensor cable connected it runs to 21.5km/h (I can creep up slowly to find the limit) but any more throttle input makes the motor output power on and off in a aprox 1 sec rate to keep wheel speed at 21.5km/h. Sometimes the wheel locks up instantly and triggers CC on my PSU. And sometimes it does not start spinning the motor from a stand still. It just jerks back and forth. Turning the wheel a little makes it work again.
The controller acts like it needs that board to go past 21.5mh/h even with hall sensors connected. Very strange.

Im still struggling to understand what made it break. If I remember correctly the problem showed as soon as I powered the controller on. The battery gauge never went up.

This little MCU board has a 339 on the back side. I will trace the pcb some day and see if its connected to motor phases in any way.

 

[markusb]

My S06 controller has a vertical board that was used for lights, but there was no MCU. just a transistor and passive components.

Hope they send you a new controller. I was never inclined to repair these things. I always have a spare on hand. I've lost two due to water damage,

Yes, I have that little board to. Its much smaller than this MCU board.

I have spoken with Flora at CSC and they will help me out. The only downside is the long wait for the spare.. summer will be almost over by the time I have it.