Grin Phaserunner with Cyclone 3000 - hall sensor problem

Having a problem getting the hall sensor working properly between a cyclone 3000 and Grin Phase runner.

Kit used is as follows:

https://www.ebikes.ca/shop/electric-bicycle-parts/controllers/phaserunner-bk.html
https://www.ebikes.ca/shop/electric-bicycle-parts/batteries/b5216-lim-dt.html
https://www.ebikes.ca/shop/electric-bicycle-parts/throttles/t-lhtwist-slim.html
https://lunacycle.com/cyclone-mid-drive-3000w-planetary-kit/

Everything is working properly with the exception of the hall sensors. Wondering if you could provide any troubleshooting advice.

What is working:
- The motor runs (in sensorless mode)
- The hall sensors each toggle as the motor turns on the dashboard. I also confirmed by metering them directly as the motor ran.
- Therefore I don't think there is a fundamental problem with the sensor and the signal is detected by the phase runner.
- No problem connecting the phaserunner to the software

What is not working:
- During the second part of the autotune test the motor spins but after the test there is a "hall sensor error" message.
- The motor then runs in sensorless mode (as I would expect if it detects an error)
- When starting from a stop there is a short period (1 seconds perhaps) of stuttering before the motor starts running smoothly. As expected in sensorless mode.

Troubleshooting attempted:
- I confirmed with a meter the phase runner dashboard accurately reflects hall position voltage reading (it does)
- I tried all 6 combinations of the (Blue, green, yellow) hall sensor wiring options. Same problem each time.
- After each change I ran the spin test of the autotune function
- Each time I got the same "hall sensor error" the came up a few seconds after the test ended.
- Each time I confirmed all three sensors were toggling on the phaser runner software dashboard

Questions:
- Is there a relationship with the motor phase wiring. e.g. there are 6 different combinations of main power phase wires (36 total possible options when the hall sensors are considered).
- I have not tried changing the motor phase wires as the connection is soldered and significantly more time consuming to switch.
- Does only one of the 36 possible combinations work? Would like to confirm this is a possible problem resolution before taking on rewiring for the remaining 30 potential combinations.

I'm somewhat at a loss as to what else to try (except all 36 possible phase and hall connection options). Any suggestions? I haven't been able to find much documentation on how to troubleshoot this problem.

I don't know if any of these will help:
https://endless-sphere.com/forums/search.php?keywords=phaserunner+hall+cyclone&terms=all&author=&sc=1&sf=all&sr=posts&sk=t&sd=d&st=0&ch=300&t=0&submit=Search

One of them says that the phaserunner doesn't work with the Cyclone motor because the motor creates noise on the hall lines (which you won't see with a voltmeter, but could with an oscilloscope).


If it is just noise, you may be able to add capacitors to ground (or an R-C filter) to each of the hall signal lines, where they exit the motor, and fix it.

This topic:
https://endless-sphere.com/forums/viewtopic.php?f=28&t=90446

RC filter might help. I had another idea to stick small piece of thin aluminum-foil tape onto hall sensors. Maybe that will be just about enough to catch that stray flux and pass the real one. Haven't gotten around pulling mine apart again to try.

The hall combinations should not matter. The controller will figure it out when it spins the motor in sensorless during autotune. But if something prevents a repeatable hall sequence then autotune won't work.

Since your sensors seem to toggle ok when turning by hand it should be possible to manually figure out the correct hall sequence by rotating the motor slowly. You also need to figure out the correct hall phase offset angle. At this point sensorless probably sounds like a pretty good idea.

I have a brand new cyclone motor with a non working hall. I opened the motor and didn't see anything obviously wrong. I did notice that the halls were very loose in their little slots in the stator. I wonder how bobbing up and down a millimeter or two affects their behaviour.

After replacing a hall I was having similar problems to the OP with the phase runner.



The left half of the screen is when the motor is being driven and the pwm waveform is being picked up by the hall wires and superimposed while the halls are pulled high. The right side is after releasing the throttle and the motor coasts to a stop.



Much cleaner after adding an rc filter. According to the asi software this retarded the hall signal by about 2 electrical degrees when spinning at top speed. The motor ran a bit smoother and did not trigger hall faults in the controller. But it is still a bit rough. Possibly from the funkiness from the hall on the blue trace. I might try carefully resoldering the broken leg of the original hall and installing it. At least the halls are super easy to access.

i had the same problem with the halls on a cyclone 3000w motor. but i saw those 'false' triggers as on the blue plot on the last picture on all hall signals and on each low-high and high-low transition.

what i think happens is that the halls triggers also on the metal of the rotor lamination between the magnets. but didn't investigate further for now. i might look at this more in the future.

i think i still have the rotor pulled from the motor, if you want some pictures i can make them for you.

Just checking... have you tried using 12vdc input power for the hall sensors? With some controllers (Kelly)this was required to fix an issue with the Cyclone.

https://endless-sphere.com/forums/viewtopic.php?f=28&t=98050

Note Fany's reply.

district9prawn said:

After replacing a hall I was having similar problems to the OP with the phase runner.

The left half of the screen is when the motor is being driven and the pwm waveform is being picked up by the hall wires and superimposed while the halls are pulled high. The right side is after releasing the throttle and the motor coasts to a stop.

Much cleaner after adding an rc filter. According to the asi software this retarded the hall signal by about 2 electrical degrees when spinning at top speed. The motor ran a bit smoother and did not trigger hall faults in the controller. But it is still a bit rough. Possibly from the funkiness from the hall on the blue trace. I might try carefully resoldering the broken leg of the original hall and installing it. At least the halls are super easy to access.

Click to expand...

Wow, that's an amazing amount of noise. But most of it is above the logic threshold level so may not necessarily cause a problem. Most controllers have a RC filter on the hall signal inputs. If the halls are toggling more than once per cycle, this will definitely cause issues. The alignment of the sensors vs. the magnets can sometimes cause unwanted triggering. There are different sensor sensitivities available too. You might want ones that take a higher flux to operate.

I replaced the genuine honeywell hall that was giving the false triggers with the original. It sounds a bit better but still a bit rough. Not sure why selecting "hall start sensorless run" in the asi software makes no difference.

I tried on a vesc and it seemed to be fine. Still rough on halls at high speeds in foc but not surprising that positional accuracy with interpolated halls is not quite good enough.

More importantly it runs fine at high speed with halls in 6 step mode. Now to sell the motor

district9prawn said:

After replacing a hall I was having similar problems to the OP with the phase runner.

The left half of the screen is when the motor is being driven and the pwm waveform is being picked up by the hall wires and superimposed while the halls are pulled high. The right side is after releasing the throttle and the motor coasts to a stop.

Much cleaner after adding an rc filter. According to the asi software this retarded the hall signal by about 2 electrical degrees when spinning at top speed. The motor ran a bit smoother and did not trigger hall faults in the controller. But it is still a bit rough. Possibly from the funkiness from the hall on the blue trace. I might try carefully resoldering the broken leg of the original hall and installing it. At least the halls are super easy to access.

Click to expand...

Can you share a schematic for the RC filter(s)?

I have heard that Luna has a pre-programmed ASI BAC 800 for the Cyclone, any thoughts on what they are doing?

So has anyone had any luck getting the PR to run the cyclone ? I couldn’t even get mine to pass the spin test.


Sent from my iPhone using Tapatalk

Ronan978 said:

So has anyone had any luck getting the PR to run the cyclone ? I couldn’t even get mine to pass the spin test.


Sent from my iPhone using Tapatalk

Click to expand...

attachments have working settings. basically I used mac motor default settings and then run autotune.

Is there a problem running the GNG Electric 2019 model Cyclone in sensored mode with the Phaserunner?

Skirmish said:

Is there a problem running the GNG Electric 2019 model Cyclone in sensored mode with the Phaserunner?

Click to expand...

I would like to know this as well, the stock controller lasted a week before it gave up the ghost at 52V.
As far as i know the GNG motor is just a ordinary cyclone motor.