(Solved) Sabvoton can't find MXUS 3k hall angle

Yermommy

10 W
Joined
Mar 17, 2018
Messages
98
Does anybody have experience pairing a sabvoton 72150 with a mxus 3k? I had my bike up and running for about 3 months using hall angle degree of 53 but I took it apart recently to change the connectors on the battery to XT90's however the controller lost its programming when I turned it back on. No matter how many times I run the angle test, it almost always gives me a "fail3" or "fail4" error. I've googled everything about these two errors and all suggestions just say "change the phase or whall combinations." I've tried all 36 combos and nothing runs smoothly. If I twist the throttle, the motor makes a loud clicking/banging noise. I can show a video if that would help. Even the hall angle test runs kind of funny. It jolts the wheel in one direction, spins it very slowly, jolts, then returns the "fail3" or "fail4." I've tested using both sets of hall sensors coming from the motor and nothing works. My old combo was swap green+yellow phase and hall but this seems to not make a difference.

Any help would be greatly appreciated, I feel like I'm running out of things to try. Thanks!
 
I don't have a Sabvoton yet, but I read about a guy who got it working by lowering the motor curent to something like 16A then testing it. After the test I guess you can put it to whatever you like.
 
eee291 said:
I don't have a Sabvoton yet, but I read about a guy who got it working by lowering the motor curent to something like 16A then testing it. After the test I guess you can put it to whatever you like.

Thanks for the idea but I've already tried a multitude of different amp values ranging from .5 all the way up to 40...
 
Some discussion here on similar Sabvoton issues but not any solution
https://endless-sphere.com/forums/viewtopic.php?t=92558

Perhaps its a software issue.

Have you tried a factory reset?
Is there a way to reflash the firmware?

It appears that these controllers have their fair share of problems though.
 
lionman said:
Some discussion here on similar Sabvoton issues but not any solution
https://endless-sphere.com/forums/viewtopic.php?t=92558

Perhaps its a software issue.

Have you tried a factory reset?
Is there a way to reflash the firmware?

It appears that these controllers have their fair share of problems though.

Yeah, I checked in with that thread and didn't find a solution. I tried restoring the firmware and didn't have any luck either.

Does anybody know if its possible to calculate the hall angle degree manually or is it different for every motor of the same make/model?
 
I would check with mulimeter phazes of motor how many Ohms. All should be equal. Try turning while measure.
Check hall: Connect 5v to red and Gnd to black in hall connector. You can use ATX or battery (can make some AA battery in parell)
while slowly turning it should show the same value on all hall.
Check cable around axle if its in good condition.
 
also remember from my experience that you need to connect e-lock to battery because it will make fail. Make sure you don't have water in motor
 
Thanks again for all the suggestions. I’ve tried hooking up my old Kelly kls7212s and it gives me the same error as sabvoton: hall fault. Ive measured voltage as 3.2v going to the hall sensors (both sets) so I really don’t understand what the problem could be. When I run the hall angle test, it moves one time then immediately trips the hall fault (error 13) on the sabvoton.

Any other ideas short of replacing the hall sensors?
 
I just don’t understand how there can be a hall fault when all 3 halls are reading 3.2v and the power connector is getting 5v. There’s clearly not a hall fault, yet it won’t stop giving me the error no matter what set I try. I even opened the motor and tested continuity between each hall sensor and the return pin inside my sabvoton. This is frustrating.
 
Have you checked manually checked the hall sensors toggle on/off when rotating the motor slowly by hand?

Those 10 dollar motor/controller testers are very useful for trouble shooting hall sensors and their wiring!
 
Punx0r said:
Have you checked manually checked the hall sensors toggle on/off when rotating the motor slowly by hand?

Those 10 dollar motor/controller testers are very useful for trouble shooting hall sensors and their wiring!

I just completed an additional test and have discovered that my green hall sensor on both sides is fried and read 0v in all positions :shock:
Ordered a 10 pack of unipolar sensors from Amazon, will be replacing all 6 sensors on Tuesday.
 
Halls were definitely fried. I ordered these as replacements: https://www.amazon.com/gp/product/B00IFD0F7M/ref=oh_aui_detailpage_o00_s00?ie=UTF8&psc=1

They work, but I still can't get my hall angle :(
I'm getting the "fail 3" error now instead of "fail4" which is actually a good thing. Apparently "fail3" just means that I need to retry all 36 phase/hall combos. "Fail4" indicates hardware failure which was my first problem.

Not sure why I need to do all 36 combos if the whole point of the hall angle test is to find it without doing these but whatever I guess...

Going to work on the hall/phase combos on Thursday and I'll post an update.
 
Ok I figured it out! My Mxus needed all phase wires to match but swap green and yellow hall wires. Runs like a champ now! I also set the hall angle to 120 but I don’t think it really did anything. Please note that I never got the angle test to succeed...I just wired it up and tried the value of 120 then spun it up no problem. Thank you to everybody who contributed and helped me find this solution.
 
although this thread is (solved) I'm still not sure about the hall angleof the mxus 3k turbo.
I couldn't find any "official" information about it and even in this thread different values were mentioned.
Does anyone know for sure and were did you get this information from ?
With the sabvoton 72150 hall angle test I get many different test results depending on the sensor combinations.. so I guess the sabvoton compensates "wrong" wiring by using a hall angle that works ?

So is the best way to get the best results by setting the hall angle to 120 (hoping it's right) and trying all the 36 combinations and then use the best of the three "fine" combinations ? (As far as I know three combs. should work but one might work best)
 
The MXUS 450x "3k" / etc motors use a 120 degree hall setup (vs 60), for both sets of halls in the motor.

What a controller that does autodection often calls "hall angle" is not necessarily the same as the 120 vs 60 hall setup meaning--it may be using the degrees of advancement or retardation of the timing signals instead.

The wiring color order for either (or both) phases and halls could be different even between two motors of the same model and winding, etc., because the people wiring it up may or may not have done every motor the same. (same goes for controllers).


A controller that does proper autodetection doesn't need you to change phase or hall wire order, because it will figure this out on it's own. Even crappy cheap controllers can do this pretty reliably, so controllers that can't do this probably have poorly written firmware with more issues than just this. ;)
 
amberwolf said:
The MXUS 450x "3k" / etc motors use a 120 degree hall setup (vs 60), for both sets of halls in the motor.

What a controller that does autodection often calls "hall angle" is not necessarily the same as the 120 vs 60 hall setup meaning--it may be using the degrees of advancement or retardation of the timing signals instead.

The wiring color order for either (or both) phases and halls could be different even between two motors of the same model and winding, etc., because the people wiring it up may or may not have done every motor the same. (same goes for controllers).


A controller that does proper autodetection doesn't need you to change phase or hall wire order, because it will figure this out on it's own. Even crappy cheap controllers can do this pretty reliably, so controllers that can't do this probably have poorly written firmware with more issues than just this. ;)

Thanks a lot for your reply amberwolf,
that explains a lot. As the sabvoton manual says that the qs205 v3 (real hall angle 120) has a hall angle of 245 +- 10, I thought that has to be also the hall angle of the mxus 3k turbo (also 120). So I set the software to 245 and tried the 36 combinations. The result was that three of them worked good and right direction at no load (0,3 kw/ 55 kph). But when I did the hall test all of them had 295 (0,2 kw/55 kph). The 295 performed way better on the street. So I assume 245 isn't the ideal angle for the mxus or maybe also not for the qs205 ? Maybe its worth mentioning that I didn't get more than max. 80 dc amps out of my sabvoton yet, but that might be another issue and I just made an separate thread about that.

Btw. I've also done the 36 comb. test with hall angle 120 but the result were not that good I guess. The three best of them had no load/0,15 kw/50 kph. With the angle test 176/0,2 kw/55 kph.
 
amberwolf said:
A controller that does proper autodetection doesn't need you to change phase or hall wire order, because it will figure this out on it's own. Even crappy cheap controllers can do this pretty reliably, so controllers that can't do this probably have poorly written firmware with more issues than just this. ;)

Is there a possibility that electrical noise can be a factor when running autodetection? I only have one controller that autodetects, and I recall it took a few times to get it to work, and I had thought possibly there was some noise interference that might be affecting it. That was just a guess.
 
Certainly. Hall signals are never clean, especially if they run in the same cable as the phase wires. Separating them to a different, isolated/shielded cable will help, but they'll still have noise from the switching of the phase currents inside the motor and where they pass thru the axle/etc in the same spot as the phase wires (if they do), and at high currents they may also respond to the stator fields and not just the magnet fields of the rotor.

If the pullups in the controller can be changed to use 12v or higher instead of just 5v, it greatly increases the signal-to-noise ratio, and can make a difference in controller's ability to pick out the real signal. Depends on the controller design how easy or hard this would be to do.

Most of the halls used in these motors can handle 20-30v on their open-collector outputs, but they have fairly low current-sinking ability so they don't "ground" as hard as they could--if a PCB were used inside the motor, larger / lower resistance transistors could be used, triggered by the hall sensor, to provide a better-grounded signal. (or signal transceivers made for this could be used.

Alternately, if a current-loop were used instead (like in a MIDI cable for music stuff), even more noise-isolation could be created; an opto-isolator loop is used to do this in MIDI, but there are other methods. The opto-isolator, if used on both ends, would also help prevent controller damage when motor cable damage occurs that shorts phase wires to hall signals, though it doesn't prevent damage to the hall voltage supply if that gets shorted).


Anyway, there's plenty of ways to clean up hall signals to improve operation.


Can even be done partly in software in the controller, using some form of FFT or other signal processing to compare the signals on the phase wires to those on the halls and subtract the phase signal noise from the hall signals, but for this the hall signals would need to be read as analog signals, not as digital ones.
 
Back
Top