GNG mid drive *sometimes* does not respond to throttle?

Eam0nn

10 µW
Joined
Dec 29, 2020
Messages
6
Hello, I'm hoping someone more knowledgeable can diagnose this issue, I have a mid drive that does respond to throttle, but sometimes the motor will make a bad noise on throttle input, which gets worse on more throttle (which I try not to do...). It does correct itself but before I mount this on my frame I would like to know what it is I've done wrong. Please see this video of the problem:

https://www.youtube.com/watch?v=4B8ub5ItF-o

Additional information:

It's a GNG 3000w motor attached to a Kelly KLS-S 72V, 120A controller (restricted in software to 30A) w/ a 72V battery pack.

Here are the controller settings: https://imgur.com/a/OUibgM3

Thanks in advance for any help whatsoever!
 
If you spin the motor freewheel backwards by hand (it will also spin the motor), and try to throttle again, does it start working again? If so it is hall related fault, it is known cyclone and kelly controllers don't really like each other (well the cyclone motor/halls are just really bad noise wise). I am surprised you have it working to the extent that it is.

Is this a combination of motor+controller from the shop? If not kelly usually recommend you run the motor halls of +12v controller output and not the +5v one.

Btw if "current percent" means phase current, you can fire it up to 2x-2.2x whatever motor current is set to. If you increase bat current to 50a later you will be golden with 120a phase. The kelly shows everything in % so you would have to do your own calculations.
 
If I manually move it backwards it does indeed start working again. This isn't the controller from the gng website, I wanted something smaller to fit the frame better (and potentially be quieter) and thought this might do the trick, perhaps not such a good idea in hindsight.

I've looked over the diagram for the controller in the manual and I don't see where I have the option to hook up the hall sensors to 12v instead of 5v, can you clarify? diagram screen grab here https://imgur.com/a/8tKvrz3

Current percent is defined as follows:

(3) Current Percent : Phase Current Percent. Range: 20~100 Functional description: The max motor current is (TheValue*PeakCurrentoftheController). Suggestion: Factory default is 100%

So I think you're right, I wasn't sure so for testing I set them both very low to be safe. I have two battery packs of 10s3p 21700s for 72v so I think when the time comes I will be keeping the battery current limited to no more than 40a, but I will increase the Current Percent.

Do you think I should just abandon this controller and get the one sold on the gng website (Or another one known to work well with the cyclone)? Thanks for your input
 
Eam0nn said:
If I manually move it backwards it does indeed start working again. This isn't the controller from the gng website, I wanted something smaller to fit the frame better (and potentially be quieter) and thought this might do the trick, perhaps not such a good idea in hindsight.

I've looked over the diagram for the controller in the manual and I don't see where I have the option to hook up the hall sensors to 12v instead of 5v, can you clarify? diagram screen grab here https://imgur.com/a/8tKvrz3

Current percent is defined as follows:

(3) Current Percent : Phase Current Percent. Range: 20~100 Functional description: The max motor current is (TheValue*PeakCurrentoftheController). Suggestion: Factory default is 100%

So I think you're right, I wasn't sure so for testing I set them both very low to be safe. I have two battery packs of 10s3p 21700s for 72v so I think when the time comes I will be keeping the battery current limited to no more than 40a, but I will increase the Current Percent.

Do you think I should just abandon this controller and get the one sold on the gng website (Or another one known to work well with the cyclone)? Thanks for your input

There is no "option" you will be manually cutting and connecting wires.

(11) 12V:12V Source Red

All of these controllers supply a 12v rail. Hook them up to the to the motor hall, instead of the hall + lead (originally: 5V supply output,<40mA.Purple) going into the motor.

Basically substitute a 12v wire coming in to the motor halls instead of the original 5v one. So the halls in the motor will have +12v input instead of +5v.

Don't connect the 5v wire to the 12v one controller side obviously, just change the motor hall input to 12v. You will need to bring a wire from the controller output (11). The original hall +5 wire coming from the controller you will have to cut and tape it up.

I don't know how to write it in a different way.
 
No need, that was perfectly clear. I've changed it over to the 12v wire and it happens less often and when it does it seems to recover more easily, will keep investigating...

Thanks very much for your help! Really, would be lost without it.
 
Eam0nn said:
No need, that was perfectly clear. I've changed it over to the 12v wire and it happens less often and when it does it seems to recover more easily, will keep investigating...

Thanks very much for your help! Really, would be lost without it.

Maybe it will happen a tiny bit less often when the bike is moving, but that is about it with the kelly, unless there are some hall settings in the controller, it won't go much further for this motor.

If you were planning on it, with phaserunner/asi bac it is just as bad. Only confirmed cyclone hall compatible controllers are yuyangkings (square wave one that cyclone ships with them) and whatever GNG ships with them.
 
If it matters, the really important place to have the 12v is on the *signal line pullups* for the halls, but this requires that the controller accept (and output) 12v on those lines, which is not common. I think some Kellys and perhaps some Sabvotons have this option (but may not all be built this way even for a specific model). For those that don't, you'd probably have to build a circuit inside the controller to translate the 12v down to open-collector to work with the existing 5v. (partly because the controller already has 5v pullups in there, and partly because it might use direct MCU lines to read these signals, and putting 12v on those would probably destroy the MCU).

If you want to try it and have a 5v hall-signal-output controller, you can use a simple NPN transistor circuit anywhere inside the controller casing, one for each of the hall signals. You might need a more complex one if the controller doesn't like the "inverted" signal set (it might not). The collector of the transistor goes to the controller hall signal input. The emitter goes to ground. The base goes to a 1kohm to 5kohm resistor, that then goes to one end of anohter 1kohm to 5kohm resistor, that then goes to the 12v power supply. The point the two resistors connects to goes to the hall signal from teh motor. If this circuit doesn't work on a particular controller, there are non-inverting buffer chips taht have open collector outputs that would work.


The reason the 12v on the hall signal lines (regardless of the hall power supply voltage) works better than 5v, is because it makes a larger difference between the "on" (0v) and "off" (+V) states, so it's easier for the controller to read the signals thru all the RF and other noise that's induced into the hall signal wires by the motor.


For those that aren't aware, the typical motor hall is what is called "open collector", and does not output any voltage at all. All it does, when a magnet passes, is to ground the output signal wire.

So the controller has pullup resistors in it that go from each hall signal input wire to the internal 5v power supply. In some controllers it's to the 12v line instead, but not many do this, even though it should work better in the noisy motor environment.

It also doesn't usually matter that the halls are only powered by 5v (though most of them can be powered by 12v just as easily, but generally don't perform significantly differently either way), while using 12v (or more) on the signal pullups.


Note that not every hall sensor of every kind works this way...but the typical "honeywell ss41 or ss411 / clone" sensor that's found in most "ebike" motors does.
 
It was one of the kelly reps that suggested the 12v input thing for the halls. Maybe there is also a setting about what voltage hall signal input to expect.
 
@amberwolf, thanks for the info, but that sounds well outside of my skill level and definitely well outside my comfort zone .

@Tommm the only hall settings I've found are relevant when the speed sensor setting is set to 4, 'line hall', I have it set to 2, 'hall', not sure what the difference is between the two. I may try it just to see then see if those hall settings make any difference

hall settings: https://imgur.com/a/7ELjnv6
 
amberwolf said:
For those that don't, you'd probably have to build a circuit inside the controller to translate the 12v down to open-collector to work with the existing 5v. (partly because the controller already has 5v pullups in there, and partly because it might use direct MCU lines to read these signals, and putting 12v on those would probably destroy the MCU).
Cool post/information! On the idea of an add-on circuit to translate the voltage, would this create a timing issue? I guess even if it did, it's probably too small to matter.
 
i doubt there would be enough delay to make a difference except at really high motor speeds, but you'd have to test that.
 
Back
Top