Pairing controller to new 1kW DD motor

iaval

10 mW
Joined
Nov 25, 2019
Messages
27
Hello all,
I have just mounted my 1st direct drive motor, rear end with cassette mount, and now trying to figure out how to make it work with my controller.
For the purposes of extended functionality (relatively common and cheap, regen brake, power/not speed throttle, community support, works with up to 24s li-ion), I decided to use ESC from Xiaomi m365 scooter for this build.
However I seem to have hit dead end with trying to find correct phase/hall configuration.
Out of 36 combinations (see attachment), I found only two, which were working fine, but both rotated in reverse direction.
The rest either produced no rotation, rotated slowly (180rpm-ish) and stopped with load, rotated slightly faster but drawn in excess of 4A at freespin, or worked mostly fine and fast, but in reverse.

My assumption for this, is that the m365 controller is built for one inversed/two straight phases at 120°.
The seller of this motor assured me it's 120°, but I couldn't confirm the configuration before purchase.

Since the motor is cassette mount, it's bound to be difficult to disassemble.
Aside from adjustments in the firmware, is there a way how to "reverse" this reversed rotation while keeping all other parameters the same? If disassembly of the motor is necessary, what modifications should be made, and are those adjustments mostly possible with only one side opened?

Thanks for any input
 

Attachments

  • Phase-Hall-Wire-Combinations.ods
    15.5 KB · Views: 27
Anyone?
Could this be related to timing? There have been some posts here on ES mentioning that 120° motor may not be possible to reverse.
So far I wired up generic chinese controller, which makes the wheel spin and the bike work, somewhat. But with no regen support and only speed (voltage) throttle, it's quite a downgrade. Not to mention that without regen, one of the brake levers is now redundant.

So suggestions how to fix this (solutions) are welcome
 
E-HP said:
https://endless-sphere.com/forums/viewtopic.php?f=3&t=103829#p1517399

I don't want to sound rude, but you're now 7th person to suggest reversing phases and hall signals after I indicated that I went through all 36 combinations twice, which would have already included the suggested swapped phase/hall combo, and concluded it's not the working configuration, in this case, the brake does not regen properly :?
 
iaval said:
E-HP said:
https://endless-sphere.com/forums/viewtopic.php?f=3&t=103829#p1517399

I don't want to sound rude, but you're now 7th person to suggest reversing phases and hall signals after I indicated that I went through all 36 combinations twice, which would have already included the suggested swapped phase/hall combo, and concluded it's not the working configuration, in this case, the brake does not regen properly :?

Interesting. Based on your posting history, you posted a question last year, received two responses and with no acknowledgement from you. Your other three posts are on this thread with no other responses. Can't explain why everyone is giving you the same advice, but it sounds like you are sure you did everything right. Maybe just toss the controller.
 
E-HP said:
iaval said:
E-HP said:
https://endless-sphere.com/forums/viewtopic.php?f=3&t=103829#p1517399

I don't want to sound rude, but you're now 7th person to suggest reversing phases and hall signals after I indicated that I went through all 36 combinations twice, which would have already included the suggested swapped phase/hall combo, and concluded it's not the working configuration, in this case, the brake does not regen properly :?

Interesting. Based on your posting history, you posted a question last year, received two responses and with no acknowledgement from you. Your other three posts are on this thread with no other responses. Can't explain why everyone is giving you the same advice, but it sounds like you are sure you did everything right. Maybe just toss the controller.

Not sure what is the relevance of my posting history to this topic, but okay. If you want to know why I haven't followed on the last topic, I can explain in PM with no issue, but in this thread it would be off topic.

The idea is to try to figure out where might be the problem, and being told for the 7th time something that was already suggested and tested is not particularly helpful, as then it's about repeating what was already written.

This thread seems to suggest that certain 120° motors might not have a working reverse configuration:
https://endless-sphere.com/forums/viewtopic.php?t=7230

About tossing the controller - I don't have a problem with that, but show me a €60 alternative that can handle high voltages while having the same feature set, and I'm sold :)
BTW, vesc was up to the consideration, however its limitation to 12s posed an impenetrable barrier
 
iaval said:
Not sure what is the relevance of my posting history to this topic, but okay. If you want to know why I haven't followed on the last topic, I can explain in PM with no issue, but in this thread it would be off topic.

The idea is to try to figure out where might be the problem, and being told for the 7th time something that was already suggested and tested is not particularly helpful, as then it's about repeating what was already written.
how would anyone know that? mind reading?

people look at posting history to see if you provided any other details in another thread. The only one I found, is where you asked for assistance, you received two responses, but you never closed the loop. another wasted thread.
 
E-HP said:
iaval said:
Not sure what is the relevance of my posting history to this topic, but okay. If you want to know why I haven't followed on the last topic, I can explain in PM with no issue, but in this thread it would be off topic.

The idea is to try to figure out where might be the problem, and being told for the 7th time something that was already suggested and tested is not particularly helpful, as then it's about repeating what was already written.
how would anyone know that? mind reading?

people look at posting history to see if you provided any other details in another thread. The only one I found, is where you asked for assistance, you received two responses, but you never closed the loop. another wasted thread.

Not mind reading. By deduction.
Either by looking at the provided table, in the OP, reversing the two working configurations produces low RPM, no torque and malfunctioning brake.

Or by deducing from the fact that all possible configurations were tested and therefore swapping any one of them would produce result that was already tested before.

I had no idea about the emotional baggage one carries attached with their forum account, maybe next time I'll just open a new account, a clean start so to speak. I haven't done so to save hosting resources - when one exists, no reason to make another, not to mention it's often against rules. But the previous responses makes me rethink the decision. As I said before, I have no problem following up on that topic, just not here
 
Maybe that's a good idea, but probably won't help. It's not emotional baggage. It's behavioral baggage. Yes, people look at history at times and may do so to decide whether to invest the time to respond. If they see someone post a question, people respond and the thread ends there, that thread is no longer useful to anyone without the followup. No value. But it's nice that you got 7 people to help you :bigthumb:
 
Speaking of value, to ensure better end for this thread,
being treated as above example, if I once find a solution, what incentive would I have to post a solution here, for future reference, and not keeping it to myself?

Just a rhetorical question.

But also dragging and schooling someone over a year old thread, that's...deep. Being a mod on a different forum, I'd have never steep this low, not unless this was done repeatedly and over the course of several weeks/months and only to provoke the crowd.
Oh well, different tribe :)
Shall we delve deeper into the topic how iaval disregarded forum netiquette to follow up on their threads, or can we return to mission impossible? :)
 
E-HP said:
Based on track record, 0%, so definitely rhetorical.

Track record and behavioral baggage of a SINGLE post mind you.
A year old at that.
What can I do about it now? Or is iaval's reputation tarnished forever? 8)

Seems that in your case, you only need 1 post to judge a person in its entirety.
 
Well the other fellow gave up. Even so, I tried to help you a second time, albeit not wasting time to write as many words, then I'm supposed to deduce from a spreadsheet that 7 people gave you the same advice. You get credit for responding, so maybe other benevolent members will be willing to give it a go.
 
Behavioral baggage includes (but is not limited to):
- not responding in timely manner, after serious bike accident leaving me with permanent knee damage
- prioritizing other life priorities afterwards

I am guilty and have sinned.
BTW, the above is true. Who'd have guessed I have to prove myself over that, but well, different tribe.

Why haven't I mentioned that before?
Because I consider this line of questioning quite ridiculous and condescending (seasoned member towards newb).
But, here we go regardless.

Let me head over that post to make an update...
Edit: Post updated, good night everyone. I need another break
 
First, never ever go about the wiring config using the try all 36 combos route. It's totally unnecessary and greatly increases the likelihood of bad connections during the process leading to faulty conclusions. eg Look out for those common cheap hall connectors where one of the blades slides off to the side instead of into the connector.

Instead do it the simple way, which doesn't require a flow chart either-
Assuming the controller isn't made for a proprietary motor where they use some trickery (like messing with the timing using the hall placement) to force you to buy replacement controllers from them, then since the motor spins in a circle with equally spaced magnets it doesn't matter, which phase fires first (ie A-B-C-A-B-C... timing is the same as B-C-A-B... and C-A-B-C...) so there has to be 3 valid combos each for forward and reverse (assuming neutral timing which is standard). That means just connect the hall plug. Then try the 6 possible combinations of phase wires. One will be valid with a nice smooth start and little sound. If you don't get a good spin up with those 6 then something else is wrong, and I'd guess that the something else is a bad hall sensor or something wrong with a hall wire.

If that valid combination spins in reverse, simply swap 2 of the hall wires (not neg or +5V obviously) and find the new valid phase combo for forward of the remaining 5 possible. Just relax and be systematic when doing the combos swapping only 2 wires for each try. Alligator jumpers simplify the process just be sure to avoid shorts. Many motors have what I call "false positive" wiring combos that will spin the wheel, but have a rough start and draw lots of unnecessary current, so be alert for false positive results.

Before declaring that wiring is conquered, it's good practice to connect an ammeter between the battery and controller to measure the no-load current. That's wheel in the air spinning at WOT (Wide Open Throttle). Quite low no-load current means you got it right.

I've used dozens of different motors and almost as many different controllers, and this simple quick method has worked every time.

A good cheap tool for trouble-shooting is an ebike tester, with which you can test throttles, motors, halls, and even controllers, though I've never used one to test a controller.
 
John in CR said:
Assuming the controller isn't made for a proprietary motor where they use some trickery (like messing with the timing using the hall placement) to force you to buy replacement controllers from them, then ...

Based on how the scooter works (on the videos), I wonder if there is some trickery to overcome to get the controller to start. On the videos, the throttle is inoperable, even after turning on the scooter, until you push start it, which activates the throttle.
 
E-HP said:
John in CR said:
Assuming the controller isn't made for a proprietary motor where they use some trickery (like messing with the timing using the hall placement) to force you to buy replacement controllers from them, then ...

Based on how the scooter works (on the videos), I wonder if there is some trickery to overcome to get the controller to start. On the videos, the throttle is inoperable, even after turning on the scooter, until you push start it, which activates the throttle.

I didn't catch that the controller was from a stand up scooter. I seriously doubt that it's sufficient to drive a 1000W ebike hubbie, especially when they have it set up to require a rolling startup. The rolling start requirement probably killed your 36 combo test before, and you just got lucky in the wheel placement to even get the 2 reverses. You should have had 3 valid forwards and 3 valid reverses.

If you want to try anyway, use the method I laid out earlier, but spin the wheel before pulsing the throttle and try fwd and reverse spin before swapping wires. If you get a valid reverse, stop and swap 2 halls and the valid forward will be one of the remaining 5 phase combos.
 
Hi John, and thanks for your reply:

John in CR said:
First, never ever go about the wiring config using the try all 36 combos route. It's totally unnecessary and greatly increases the likelihood of bad connections during the process leading to faulty conclusions. eg Look out for those common cheap hall connectors where one of the blades slides off to the side instead of into the connector.

I went for the 36 combos route simply because I couldn't find even one working combination.
I actually own the connector you're describing, and it was used in the testing process, together with WAGO terminal blocks, to decrease the likelihood of bad connection

Instead do it the simple way, which doesn't require a flow chart either-
Assuming the controller isn't made for a proprietary motor where they use some trickery (like messing with the timing using the hall placement) to force you to buy replacement controllers from them, then since the motor spins in a circle with equally spaced magnets it doesn't matter, which phase fires first (ie A-B-C-A-B-C... timing is the same as B-C-A-B... and C-A-B-C...) so there has to be 3 valid combos each for forward and reverse (assuming neutral timing which is standard). That means just connect the hall plug. Then try the 6 possible combinations of phase wires. One will be valid with a nice smooth start and little sound. If you don't get a good spin up with those 6 then something else is wrong, and I'd guess that the something else is a bad hall sensor or something wrong with a hall wire.

If that valid combination spins in reverse, simply swap 2 of the hall wires (not neg or +5V obviously) and find the new valid phase combo for forward of the remaining 5 possible. Just relax and be systematic when doing the combos swapping only 2 wires for each try. Alligator jumpers simplify the process just be sure to avoid shorts. Many motors have what I call "false positive" wiring combos that will spin the wheel, but have a rough start and draw lots of unnecessary current, so be alert for false positive results.
This is why I did the 36 combos twice, to be sure that the connection error is ruled out. Nevertheless, I did what you're describing, started out with very nice, smooth accelleration, only going in reverse, and then reversed 2 halls and tried all 6 phase combos, and none of them rolled the same way as the one in reverse, jumping all over the place.
Before declaring that wiring is conquered, it's good practice to connect an ammeter between the battery and controller to measure the no-load current. That's wheel in the air spinning at WOT (Wide Open Throttle). Quite low no-load current means you got it right.
Yep, but I haven't used ammeter at all, instead I first used regulated PSU with current limit, to prevent damage.

I've used dozens of different motors and almost as many different controllers, and this simple quick method has worked every time.
Good for you, but in my case it hasn't actually solved the problem

I didn't catch that the controller was from a stand up scooter. I seriously doubt that it's sufficient to drive a 1000W ebike hubbie
I modded the controller extensively, and it has been tested beyond 1kW in the past, with no issues, running on the scooter motor.

especially when they have it set up to require a rolling startup. The rolling start requirement probably killed your 36 combo test before, and you just got lucky in the wheel placement to even get the 2 reverses. You should have had 3 valid forwards and 3 valid reverses.
Rolling startup is not required with the modified firmware which I loaded. Well not exactly, the wheel has to spin at least 0.01rpm for the controller to activate, but there's no set limit. I accounted for this limitation in my initial testing, and came up with the result above.

Based on the information I have gathered, the controller can be used to drive 60deg 1 inversed/3 straight and 120deg all straight motors. This 1kW rear hub should be 120deg, but not sure about its phase orientation.
Given the test results so far, it might be needed to open the motor. In your experience, is it possible to perform any modifications needed with just one side open, e.g. not having to go through the cassette mount, but rather the other side?
 
John in CR said:
Before declaring that wiring is conquered, it's good practice to connect an ammeter between the battery and controller to measure the no-load current. That's wheel in the air spinning at WOT (Wide Open Throttle). Quite low no-load current means you got it right.

I'm getting okay-ish 1.25A at about 200rpm on all 3 "working" configurations rolling forward, with current limit on throttle raised to 15A, noload current goes up to 4-7A, which seems quite high. Generic controller managed to get the wheel spinning at 600rpm, and no-load current was 1.5A
 
OMG! You did the 36 combo thing again, since you found the 3 forwards.

What's the 4-7A stuff. What does it stabilize at going to full throttle? Somewhere in that range wouldn't surprise me at 600rpm+, especially if rim and tire are mounted already.
 
In fact I did it 3 times, 108 combinations :)
Afterwards, fingers had to heal from lifting wago springs so many times.

The 4-7A is nonsense, I've hit a wall that I'm not sure how to go around. Generic chinese controller manages full 600rpm at a little over 1.5A, so that can't be right.
Turns out it was not the right setup. The right one however behaves most strangely. It won't turn more than 200-ish RPM, any more current applied manifests itself as deep noises in the motor, more current going through the phase wires, but no more RPM. This however must be the correct setup, as all other parameters match up - brake functions correctly, accelleration works up until the 200rpm threshold, power consumption below 200rpm is also just right.

ESC uses power control algorithm for throttle, so it accellerates indefinitely, with no speed cap. So by default it'll hit those 200rpm even at the lowest throttle setting. That's why I initially misdiagnosed it as non-working setup.

With this done, I guess the FW needs to be reverse engineered some more, as 200rpm as backemf limit can't be right for this type of motor.
 
Getting back to this, as I promised back in Oct, to supply closure when ready.
The 2.0 build was never finished. I scrapped it for many reasons, weight being the main one.
I went back to the drawing board with it, and opted for 2.4kg DD motor, front drive, and 180-cell battery.

Yesterday, the 3.0 build was finally finished. It's absolute joy to ride. Fast, silent, but also safe. 68cm handlebar sure makes a difference, along with centre-loaded design, rather than rear-heavy.
At 35km/h, the efficiency is about 6Wh/km. Very pleased with it. Current throttle instead of voltage lets me choose just the right amount of assist needed for the length of planned trip. And 1.6kWh battery allows for a lot of range, 200-300km with reserves, so no more range anxiety
 
Back
Top