New KT Controller only puts out 4.5V

[the_horak]

Hey All. I have a 30a 36/48V KT controller and I have disconnected everything to try and figure this out. The low voltage supply with everything disconnected is only 4.5v. With the PAS attached I am down to 4.3V. With the throttle connected and turned all the way, I am only getting 3.5V for the signal. What do I do here? Can I fix this or is the controller shot?

 

[the_horak]

Am I wrong to expect 4v out of the throttle? My controller says 1 - 4.2v for speed signal, but I'd be happy with just 4 at the moment. Part of me wants to disconnect the ground and see if my throttle input increases.

 

[fill]

Disconnect everything even the motor cable a burnt out hall sensor could be draining the 5v supply, then if still not a clean 5v logic voltage level check the controller pcb for the voltage regulator component/circuit actually drops down to a clean steady 5v in the first place, and if not look for the failed component/s? if you get a clean 5v on the pcb trace back to the 5v wires to determine where the voltage drops.

 

[docw009]

On the KT's, I believe the Vcc supply uses a diode off the 5V rail from the regulator. That's used to power the Halls, pedal assist sensor, and throttle, so the max you will see outside the controller is 4.4-4.5 volts. Nevertheless, it shouldn't drop much, as all the sensors are ICs. so something may be wrong with the PAS if you lose 0.2 volts,

 

[amberwolf]

Aside from teh diode which will drop about half a volt or more (depending on the current drawn thru it--more current, more drop across it) most of the 5v regulators on these don't supply much current. They often use the low power tiny version of the 7805 that *with a proper heatsink* can supply as much as 100mA, but which in these controllers probably has no heatsinking and will be warmer than ambient already just because of the heat from everything else inside the controller, and so would have to be derated. By how much, I don't know--they're probably not genuine brand-name parts so the spec sheets for those wouldn't apply, and may not be as capable as the genuine parts in the first place. So I'd just cut that in half.

So 50mA to run everything on the bike that needs 5v--the controller MCU, assorted logic circuits outside the MCU, pullups for the motor halls, power for all the sensors (mtoro halls, throttle, pas, ebrakes if you have them, etc). If you have a misbehaving device that takes more power than it should, it could drop the voltage on the rest of the bus by overloading the 5v supply, and it doesnt' take much to do that.



That said: Most of the halls used in throttles will output half of whatever their supply voltage when no magnet is present, or at half-throttle. The magnets in the throttle don't seem to usually allow full output, but only "most" of the output, just not placed right to give the full output range. Some throttles only go up to about 3.5v even with a full 5v input, and some go as high as 4.1-4.2v with 5v input.

You can test the throttle with a USB charger and just the throttle, no other ebike parts. Connect the USB charger's output to the throttle's ground and 5v input, and a multimeter from ground to throttle signal output. Youll probably read around 0.8-1v at throttle off, and around 3-4v at full throttle.

But whatever you get there, you should get the same on the controller assuming you ahve the same supply voltage (you'll have eto measrue that). If hte controller supply is lower, the throttle will be lower proportionally.

 

[the_horak]

Disconnect everything even the motor cable a burnt out hall sensor could be draining the 5v supply, then if still not a clean 5v logic voltage level check the controller pcb for the voltage regulator component/circuit actually drops down to a clean steady 5v in the first place, and if not look for the failed component/s? if you get a clean 5v on the pcb trace back to the 5v wires to determine where the voltage drops.

I am getting around 4.5V with nothing connected but the display, as I currently am not set up for it to go without the display.

On the KT's, I believe the Vcc supply uses a diode off the 5V rail from the regulator. That's used to power the Halls, pedal assist sensor, and throttle, so the max you will see outside the controller is 4.4-4.5 volts. Nevertheless, it shouldn't drop much, as all the sensors are ICs. so something may be wrong with the PAS if you lose 0.2 volts,

I was wrong. I drop from 4.35 to 4.31 when I plug in the PAS. I misread the first reading and thought it was 4.5V. I'm 4.3V from the supply with nothing connected but the display. Good to know about the diode, though. Maybe my voltage is okay...

Aside from teh diode which will drop about half a volt or more (depending on the current drawn thru it--more current, more drop across it) most of the 5v regulators on these don't supply much current. They often use the low power tiny version of the 7805 that *with a proper heatsink* can supply as much as 100mA, but which in these controllers probably has no heatsinking and will be warmer than ambient already just because of the heat from everything else inside the controller, and so would have to be derated. By how much, I don't know--they're probably not genuine brand-name parts so the spec sheets for those wouldn't apply, and may not be as capable as the genuine parts in the first place. So I'd just cut that in half.

So 50mA to run everything on the bike that needs 5v--the controller MCU, assorted logic circuits outside the MCU, pullups for the motor halls, power for all the sensors (mtoro halls, throttle, pas, ebrakes if you have them, etc). If you have a misbehaving device that takes more power than it should, it could drop the voltage on the rest of the bus by overloading the 5v supply, and it doesnt' take much to do that.



That said: Most of the halls used in throttles will output half of whatever their supply voltage when no magnet is present, or at half-throttle. The magnets in the throttle don't seem to usually allow full output, but only "most" of the output, just not placed right to give the full output range. Some throttles only go up to about 3.5v even with a full 5v input, and some go as high as 4.1-4.2v with 5v input.

You can test the throttle with a USB charger and just the throttle, no other ebike parts. Connect the USB charger's output to the throttle's ground and 5v input, and a multimeter from ground to throttle signal output. Youll probably read around 0.8-1v at throttle off, and around 3-4v at full throttle.

But whatever you get there, you should get the same on the controller assuming you ahve the same supply voltage (you'll have eto measrue that). If hte controller supply is lower, the throttle will be lower proportionally.

Using a 5V supply, I got .85V and 4.32V on the throttle. Like you said, having a lower supply of 4.3V will drop my throttle output.

 

[the_horak]

I think I am going to leave things as they are for now. I ordered another identical controller for $55 and I am going to bench test that one to compare.

If I tear into this now, I won't have a bike. I may not be comfortable with everything just yet, but I can definitely ride the bike. If the controller is bad, it's not dead and I have a new one already coming.

With all the custom work and trial/error I did to put this build together, I know there were times when I made mistakes connecting the 5V peripherals. Maybe I burnt something a little...

If the new controller's performance is identical to the current one, then I can return it or I'll have a backup.

Side question: can I connect a 5V supply to the throttle's Input and Ground and then connect the throttle's Ground and Signal to the controller? Simulate a full 5V supply?

 

[docw009]

Side question: can I connect a 5V supply to the throttle's Input and Ground and then connect the throttle's Ground and Signal to the controller? Simulate a full 5V supply?

I believe you can, but there is a point where it won't work.

For example, I've never got the throttle input to work by just putting the throttle voltage to the signal, although other people they have done it. I think it has to be somewhat lower.