HELP! Jerky driving with Grin All Axle hub motor

Dear Forum,

today I finished installing my “Ready to Role Grin All-Axle Hub Motor Kit” onto my beloved cargo bike. I'm quite proud of myself since this is my first built and I did everything on my won but my first test ride was very disappointing :| . I'm so sad right right now... , I hope to get a little bit of help from you guys:



Here is my setup:
Grin All Axle Front Hub Motor (Direct Drive)
Controller Grinfineon C4825L10
Cycle Analyst V. 3.14
24 Pole King Meter PAS Sensor
Throttle + AUX-Device (2 Buttons)
36V battery 13,5Ah, Hailong case

While starting to pedal the motor acts very jerky, stutters before reaching a certain speed and you feel a very hard “break” as soon as you turn on/off the throttle. Also the CA doesn’t show any speed information even though I double-checked that I connected everything properly (with the L1019 plug there is no way of connecting some cables the wrong way). I am wondering how to optimise the driving characteristics!? The PAS sensor (King Meter 24P) seems to be working fine since it shows the cadence plus the P and D signs in the display of the CA go up and down, so this might not be the problem. Also the regen brake acts perfectly fine. Maybe this poor driving characteristics is what you get from the above listed setup? This would be very disappointing...

I also noticed that the Controller-LED starts flashing constantly (indicating sensorless mode?) after spinning the wheel a couple of times. How do I configure the controller so it switches into sensored mode?
I changed nearly every setting, did all the basic setup in the CA but it seems to me that this is a "hardware" problem and it doesn't change as long as the controller can't read the hall sensors form the hub motor...

Hope someone can help since Grintech Hardware is not very common here in Europe...

Greets
Oliver

The CA is a sophisticated device with a complex setup process. The consistent recommendation is to start with the setup manual, from page 1, go thru ALL settings, as many, many of them are dependent on other settings for proper function. DO NOT skip ahead assuming some settings do not affect your issue. CHECK THEM ALL.

Your PAS has a higher than normal number of magnets. This could be a problem, or may require some very specific settings.

One thing to test would be to bypass the CA and connect directly to the controller, if feasible. Always simplify the system for best diagnostic success.

I was hoping that all the components works well together because I bought it as a Ready to Roll Kit from GrinTech.

I spent almost 2 days redoing all settings over and over again without a change. The controller stays in sensorless mode with the unsatisfying driving characteristics.

I will try to skip the CA but the only cable I can connect directly to the controller is the throttle because I have the new version with L1090 Cable...

Is there any way I can check if the speed sensor or magnets inside the motor are damaged?

is the red led light on the controller flashing a code like 3 flashes then pause and 3 flashes pause? the controller flashes a code if somethings wrong.

No,it is a constant flashing.I ordered two motors with the exact same components. Both have the same problem, so it is quite unlikely that both sensors are damaged.

If I spin the wheel the P isn’t blinking.

I was thinking that maybe the CA is waiting for a signal from an external speed sensor and therefore doesn’t read the signal from the motor but I don’t know how to change that.

How do I know that I have the DP and not the DPS-Version? Maybe Grin sent me the wrong CA-Version...

I went to grins site but its down. is the controller supposed to be constantly flashing? look in the manual that came with the controller and it will tell you. my phaserunners are just red constant

Did Grin supply your King PAS sensor?

Do you have assurance it is compatible with the CA?

I seem to recall the list is pretty short?

I've seen people reco the Sempu BB sensor, over Thun for example.

There also may be voltage issues, 36V is pretty low these days? Just guessing, a noob myself

The controller is supposed to be red constant. This is the sign for sensored mode. My controller is flashing but as soon as you push the throttle or start pedaling it is constantly red, indicating sensored mode. This is also written down on the Grin page.

"If the controller LED is flashing on and off in a steady rythm that means that it is running in sensorless mode and does not have a set of valid hall signals. When the controllerif it is running in sensorless mode. When the hall sensors are plugged in and properly mapped, t Once the throttle is engaged, then then the LED will either stay steady ON if it is running in sensored mode, or it will blink at a faster 2Hz rate if it is running sensorless. A repeated flash sequence followed by a pause indicates a status or fault condition."

I think the controller might be mapped the wrong way but I don't know how to check or change that.

john61ct said:

Did Grin supply your King PAS sensor?

Click to expand...

Yes I ordered the PAS sensor together with the other parts in a Ready to roll bundle.

john61ct said:

Do you have assurance it is compatible with the CA?

Click to expand...

Yes it definitely does work together with the CA. There is even a Video online with Justin showing how to install and configure the King PAS sensor on the CA!

goatman said:

There also may be voltage issues, 36V is pretty low these days? Just guessing, a noob myself

Click to expand...

It could be that there might be a volt issue, but the controller does support everything from 36-52V. I don't know if that is the problem...

You need to go thru the setup process of the Cycle Analyst. Don't be discouraged, because you bike isn't running right simply because it's not sufficiently configured yet.

The first problem is not the CA, it's that the controller isn't reading the hall signals from the controller.

I would recommend taking the CA out of the loop, if you haven't done this yet, and connect only throttle to controller, battery to controller, and controller to motor.

If the flashing red persists, then there is a problem either in the controller or the motor, or the wiring between. With the L1019 connectors, there is no wiring that you yourself can alter, without disassembling either the controller or the motor to unsolder hall wires and resolder them in different orders until it works. I don't recommend doing this, because it should be wired to work "out of the box".

At that point, you should contact Grin Tech to see if they have seen this with this combination already, and know what is wrong, and can get you an exchange part for whichever one it is, or whcih one you need to send in for repair.


Once that's fixed, it will fix the speedo issue for a CA-DP connection (which uses the motor halls for speedometer) and may take care of the other issues.


Note that with the Grinfineon present model, there is no way to "map" hall signals other than physically changing the wire order. Until they are wired correctly and operating as they should, the controller will continue to operate in sensorless mode.

Also, even if it were just "mapping" the CA would still be getting a valid speed signal from one of the halls. Since it is not, it implies that either none of the hall signals are connected, or power or ground to them is not connected, or it just happens that the one hall signal that the CA uses for speed signal isn't connected. Until the sensors are wired correclty and operating as they should, the CA will not have a speed signal, and any of it's functions that depend on that signal will not operate as intended.

BTW, if this is the kit:
https://www.ebikes.ca/shop/ready-to-roll-kits/direct-drive-rtr/grin-all-axle-ready-to-roll-kit.html
it shows it comes with a Baserunner, rather than the Grinfineon, and doesn't have any options for a different controller, which is strange, especially since they picture it with a Phaserunner. Did you just contact them directly for that change?

amberwolf said:

I would recommend taking the CA out of the loop, if you haven't done this yet, and connect only throttle to controller, battery to controller, and controller to motor.

Click to expand...

If I do that the flashing persists. I assume then that the controller is mapped incorrectly. That's harsh, having spent days already with configuring the Cycle Analyst over and over again and racking my brain what I have done wrong.

amberwolf said:

At that point, you should contact Grin Tech to see if they have seen this with this combination already, and know what is wrong, and can get you an exchange part for whichever one it is, or whcih one you need to send in for repair.

Click to expand...

I've done that already and gotten a short response in which they were asking for my order details so they can double check if the default configuration has been done correctly. It's just very disappointing since I was hoping to get my bike going for a bike trip the next days. Also it is quite expensive to send parts from Europe to Canada and back and I have been saving money for quite a while to get one of Grins famous bike motors.

amberwolf said:

Note that with the Grinfineon present model, there is no way to "map" hall signals other than physically changing the wire order. Until they are wired correctly and operating as they should, the controller will continue to operate in sensorless mode.

Also, even if it were just "mapping" the CA would still be getting a valid speed signal from one of the halls. Since it is not, it implies that either none of the hall signals are connected, or power or ground to them is not connected, or it just happens that the one hall signal that the CA uses for speed signal isn't connected. Until the sensors are wired correclty and operating as they should, the CA will not have a speed signal, and any of it's functions that depend on that signal will not operate as intended.

Click to expand...

That makes sense to me. Thanks a lot for the explanation.

amberwolf said:

BTW, if this is the kit:
https://www.ebikes.ca/shop/ready-to-roll-kits/direct-drive-rtr/grin-all-axle-ready-to-roll-kit.html
it shows it comes with a Baserunner, rather than the Grinfineon, and doesn't have any options for a different controller, which is strange, especially since they picture it with a Phaserunner. Did you just contact them directly for that change?

Click to expand...

It is rather that kit here: https://www.ebikes.ca/shop/electric-bicycle-kits/direct-drive/grin-all-axle-motor-kit-advanced-pas.html
I don't exactly know the difference between the "DIY" kit and the "Ready to Roll" version but in the link above the Grinfineon is listed as a normal option to order. Maybe that is a unusual combination to order and the parts don't work well together. Since I'm running my system on 36V I don't need such a powerful controller such as the Phase- oder Baserunner...but that was maybe not the best thing to do

OleyT said:

If I do that the flashing persists. I assume then that the controller is mapped incorrectly.

Click to expand...

No, if it was just mapped incorrectly then the CA would still get a speedo signal.

I have a couple of the 4825s from the previous non-L1019 connector version almost a year ago I guess, with andersons/jsts; I had trouble at first figuring out the flashing indicating the sensorless vs sensored, and some other issues, too, and for me the sensorless problem was just figuring out the correct hall order and making good connection for them all. They're for my SB Cruiser cargo trike, though they ended up not being powerful enough regen braking as my primary issue with them (I have a DIY solution for that over here: https://endless-sphere.com/forums/viewtopic.php?f=30&t=105711 that I'm still working on), though not quite enough startup/acceleration power is another issue, when I'm really loaded down. I doubt either of those would be an issue for you.


One other thing about sensorless with these is it is not silent--it reverts to the trapezoidal type of motor drive, so the motor has a buzz or a whine--it's not loud (except under high loads) but it's noticeable. The sensored mode is silent. There's a "whirr" when running it off ground, but it's completely inaudible when riding with even the slowest quietest tire noise. So taht's another way to tell which mode it's running in.

Sensorless also has the possibility under certain load scenarios, especially at startup, to be unable to get moving easily, and may be jerky or "grindy" until it gets started, where sensorless is just smooth and goes when you tell it to.

That's harsh, having spent days already with configuring the Cycle Analyst over and over again and racking my brain what I have done wrong.

Click to expand...

Yeah, it's not much fun if you don't know where to start, and end up starting with the most complex stuff. I spent...I dunno--months? trying to find a problem with a system only to eventually discover that it was a hidden setting in the OEM Limits section fo teh CA--you can't even see that from the onboard menus.... And no idea how it got set that way, but it did, and changing it to default in the PC setup program fixed it, but only after Teklektik pointed out where my problem might be. Plenty of other stuff like taht I've run into with assorted builds. :/

I've done that already and gotten a short response in which they were asking for my order details so they can double check if the default configuration has been done correctly. It's just very disappointing since I was hoping to get my bike going for a bike trip the next days. Also it is quite expensive to send parts from Europe to Canada and back and I have been saving money for quite a while to get one of Grins famous bike motors.

Click to expand...

Hopefully it is something that they can walk you thru fixing, but I suspect it's an issue you can't fix there (without disassembling things).


One other thing I thought of that *might* be the problem, but is unlikely...are you certain the L1019 connector is *fully* plugged in? If it is not, maybe the phase pins make a partial connection (which would make the motor response jerky as connection is made and broken), and the halls don't make a connection at all (or the sensors do but not the 5v to power them, etc).

This can happen with some of these types of connector if it is rotated so the keys don't line up, so they can't fully plug in.

It can also happen if they have a "sealing" snapring/detent in the connector moulding that makes it hard to push past that point. I haven't worked with the L1019 yet, but similar ones can be difficult to snap together.

It is rather that kit here: https://www.ebikes.ca/shop/electric-bicycle-kits/direct-drive/grin-all-axle-motor-kit-advanced-pas.html
I don't exactly know the difference between the "DIY" kit and the "Ready to Roll" version but in the link above the Grinfineon is listed as a normal option to order. Maybe that is a unusual combination to order and the parts don't work well together.

Click to expand...

No, any combination of those things would work fine together (or Grin wouldn't offer them, at least not without a compatibility note). Other generic ebike stores might, but Grin knows better than that.

The only reason to have a compatibility problem is if one of them is not wired correctly to match with the other ones; with the old connectors this would be easy to fix (but not water-resistant, etc), but with the new ones it has to be done inside the controller (or by cutting into the cable housing).

FWIW, the reason I thought you had (almost) the kit I linked was because you called it a Ready to Roll kit in the first post.

Since I'm running my system on 36V I don't need such a powerful controller such as the Phase- oder Baserunner...but that was maybe not the best thing to do

Click to expand...

Well, the PR/BR are pretty expensive compared to the Grinfineon, so unless you need the features they have there's no reason to get them. If they could handle more power I'd be interested in them, but ended up going the DIY route linked above instead.

The Grinfineons are just fine for most applications, just not as small for the power or as feature-rich / configurable as the PR/BR. Should work fine for your bike, as long as they make enough power for your usage.

I blew a 20 amp grinfinion opened it up and saw what I thought was a jst for the halls??? maybe you have one in yours??

You mean a JST inside the shell, from the L1019 cable to the PCB itself?

If they do that vs a direct soldered connection, it leaves open some possibilities to fix it yourself:

The connector is unplugged, partially or completely, and easy to fix.

Some or all of the contacts in the connector are backed out on one side or the other (common with JST, and one reason the L1019 is better). Also easy to fix.

The JST of one or the other is miswired (most likely completely reversed, as that's the easiest way to miswire one), and that's also easy to fix.

I never saw this before, I don't know, just thought id share, it was crammed inside shell so maybe its misaligned? but who knows

So is it just a random JST laying in there, not connected to anything? Or is it on one of the cables that passes outside the controller? If so, which one? Or something else entirely?

no it was jammed inside the case. when i opened up the controller and pulled the board out i was surprised to see it. its the only controller ive seen with a jst inside

amberwolf said:

One other thing about sensorless with these is it is not silent--it reverts to the trapezoidal type of motor drive, so the motor has a buzz or a whine--it's not loud (except under high loads) but it's noticeable. The sensored mode is silent. There's a "whirr" when running it off ground, but it's completely inaudible when riding with even the slowest quietest tire noise. So taht's another way to tell which mode it's running in.

Click to expand...

Yes, you can hear a clear audible buzz as soon as the motor starts spinning when the wheel ist off the ground and it pitches up the faster the wheel spins.

amberwolf said:

Sensorless also has the possibility under certain load scenarios, especially at startup, to be unable to get moving easily, and may be jerky or "grindy" until it gets started, where sensorless is just smooth and goes when you tell it to.

Click to expand...

That is exactly the problem. I could live with the sensorless driving but it is very rough at the start but as soon as I have reached a certain speed the motor spins fine. I assume that this is not even good for the motor and the front fork because it gets quite hefty kicks while accelerating and stuttering. It doesn't stutter much when the wheel is of the ground.

amberwolf said:

One other thing I thought of that *might* be the problem, but is unlikely...are you certain the L1019 connector is *fully* plugged in? If it is not, maybe the phase pins make a partial connection (which would make the motor response jerky as connection is made and broken), and the halls don't make a connection at all (or the sensors do but not the 5v to power them, etc).

This can happen with some of these types of connector if it is rotated so the keys don't line up, so they can't fully plug in.

It can also happen if they have a "sealing" snapring/detent in the connector moulding that makes it hard to push past that point. I haven't worked with the L1019 yet, but similar ones can be difficult to snap together.

Click to expand...

Thanks for the advice. I double checked it. The connector is fully plugged in. The have a snapring which can be screwed together.

amberwolf said:

The only reason to have a compatibility problem is if one of them is not wired correctly to match with the other ones; with the old connectors this would be easy to fix (but not water-resistant, etc), but with the new ones it has to be done inside the controller (or by cutting into the cable housing).

Click to expand...

What bothers me also is the fact that I ordered two exact same kits from Grin with the exact same components and mounted them on the same bike. Both bikes doesn't work quite right. I am not sure how likely it is that Grin sends out two sets/controllers with the same problem. That's why I was so sure in the beginning that I was missing out on something during the installation progress. I swapped the controllers but no effect.

goatman said:

no it was jammed inside the case. when i opened up the controller and pulled the board out i was surprised to see it. its the only controller ive seen with a jst inside

Click to expand...

It's still not clear what you mean, which is why I asked the specific question(s).

It sounds even more now like you mean it was just a loose JST connector with no wires on it laying in there, just rattling around loose in the case, not connected to anything at all, no pins, no wires, but I don't see why anyone would do that.

It would be helpful if you would specify exactly what it is that you mean.

if you look at the picture, that jst connector has YBG. it was connected inside the controller. I blew some mosfets. unscrewed the end caps of the controller. 4 phillips screws per end cap. then unscrewed the 3 screws holding the mosfet heat sink to the controller case and pulled the board out. while pulling the board out I saw the JST connector. it was connected and fine. it was mosfets that I blew.

it sounds like grin might have a bad batch of motors or bad batch of controllers.

grin wires all the motors and controllers the same as bafang.

I would open up the controller and see if there is a jst in there maybe its the wrong sequence. take a picture of it share so we can see whats going on. ill bring another picture

OleyT said:

That is exactly the problem. I could live with the sensorless driving but it is very rough at the start but as soon as I have reached a certain speed the motor spins fine. I assume that this is not even good for the motor and the front fork because it gets quite hefty kicks while accelerating and stuttering. It doesn't stutter much when the wheel is of the ground.

Click to expand...

It is a normal thing with sensorless operation with many controllers, but it sounds like yours is rougher than most (and more than my similar Grinfineon that is under a higher startup load).

What bothers me also is the fact that I ordered two exact same kits from Grin with the exact same components and mounted them on the same bike. Both bikes doesn't work quite right. I am not sure how likely it is that Grin sends out two sets/controllers with the same problem. That's why I was so sure in the beginning that I was missing out on something during the installation progress. I swapped the controllers but no effect.

Click to expand...

I wouldn't expect any two "random" sets to have the same problem...but they make the all-axle hubs only as they are ordered. So maybe both you have were wired up by the same person that made the same mistake with both? Or maybe both cables came from a batch that has the same problem, etc.

I don't know if Grin builds up the controllers or just gets them in completed with all the new wiring and such, but I would guess the same kind of mistake could happen there, too, regardless of where they are built up, but getting two in a row the same is unlikely unless there's a larger batch that have the same problem.


Of course, it could be some completely different problem...but it really sounds like either the halls aren't wired up either in controller or motor, or they aren't getting power (or ground), or some similar issue preventing the controller from getting signals from them.


If it weren't for the type of connector, it'd be easy to test this with a multimeter set to the 20VDC range, motor connected to controller, black lead on black hall wire or negative battery lead, and red lead on any of the three hall signals yellow blue or green, then manually turn the wheel, and the signal should toggle between something close to 4v-5v and something close to 0v-1v. Then test each other signal wire in turn. If none toggle, check the red wire for about 5v.

But I can't think of a way for you to do this without opening up either motor or controller, or else making a battery and "pullup resistor" test rig to do it at the motor connector without controller connected, and then testing the controller at it's connector separately. If you want to try this test, I'll write up the steps.

goatman said:

if you look at the picture, that jst connector has YBG.

Click to expand...

I don't see a picture, but when I quote you I see a link to an external imgur site jpg. Unfortunately those don't show up for me, on any browser, most of the time (randomly, sometimes some do; I have no idea why).

But anyway, if it's a YBG with nothing else, my guess is that it is phase wires for something. If they come out of the PCB to the JST and connect to another JST that goes somewhere else inside the controller, I don't know for certain that this is what they would be for, but one possibility is a "Sensor board" for a "sensorless" (but not really) controller.

I've seen a number of controllers advertised as sensorless, with no hall wire connector, that have a little board inside that has wires going to both the phases and the halls. This board translates the phase signals as "distorted" by the attached motor into position signals for the controller (whcih isn't really sensorless, as the MCU doesn't know how to work without halls signals, so the board is there to "fake" them). They don't generally work as well as a "real" sensorless (especially FOC) controller, but they do work well enough for many systems and setups.

amberwolf said:

I don't know if Grin builds up the controllers or just gets them in completed with all the new wiring and such, but I would guess the same kind of mistake could happen there, too, regardless of where they are built up, but getting two in a row the same is unlikely unless there's a larger batch that have the same problem.

Click to expand...

I’m wondering how Grin wants to trace back the controller configuration/built.They asked me for the order number to check that but I haven’t gotten an answer back.

amberwolf said:

If it weren't for the type of connector, it'd be easy to test this with a multimeter set to the 20VDC range, motor connected to controller, black lead on black hall wire or negative battery lead, and red lead on any of the three hall signals yellow blue or green, then manually turn the wheel, and the signal should toggle between something close to 4v-5v and something close to 0v-1v. Then test each other signal wire in turn. If none toggle, check the red wire for about 5v.

Click to expand...

I think I’ going to wait another day for Grin to answer. I think their opening hours are from Tue-Fri.Maybe I should even try to call them. If they don’t answer, can’t help me or don’t want to replace the controller then I definitely would try the multimeter testing with opening the controller. Then there is not really much to loose. I haven’t used a multimeter before though and never seen a controller from the inside...Thanks a lot for the offer to guide me through that.

It is definitely a tough pill to swallow having spent all my money for this motor set and now having to open up the controller/motor by myself... :?

OleyT said:

I think I’ going to wait another day for Grin to answer. I think their opening hours are from Tue-Fri.Maybe I should even try to call them. If they don’t answer, can’t help me or don’t want to replace the controller then I definitely would try the multimeter testing with opening the controller. Then there is not really much to loose. I haven’t used a multimeter before though and never seen a controller from the inside...Thanks a lot for the offer to guide me through that.

It is definitely a tough pill to swallow having spent all my money for this motor set and now having to open up the controller/motor by myself... :?

Click to expand...

I would first test without opening anything. That's the part I can walk you thru.