Motor sputter at higher RPMs...

[izeman]

Lebowski if you know how to solder.

Or ask Animalector to solder one for you...

touché andy did a fantastic job which i could never do.

 

[bchampig]

I replaced all of the hall sensors with genuine Honewell SS41 sensors. I tested the bike out and the problem is still there. Time for a new controller....will just get a Lyen 12FET. What a horribly frustrating experience, I just wanna ride!

-Brian

 

[bchampig]

Just wanted to give an update about my controllers woes. I ended up ordering a Lyen 12 FET from Edward. I swapped it into my setup, and it is working great. No issues with motor sputter. I'm still using the CA3. I wish I knew what the deal was with the PV controller. I guess it was just a faulty unit since Mike from LR uses a PV with the same setup.

Regards,
Brian

 

[fechter]

Thanks for the report back. Hopefully you can get the PV replaced under warranty and try again.

 

[sparkiesweden]

I initially has the same problem with a MAC motor and a KH612 based controller (similar to PV), seems to be a false phase/hall combo. Swapped green/yellow on both halls and phase.

 

[bchampig]

I initially has the same problem with a MAC motor and a KH612 based controller (similar to PV), seems to be a false phase/hall combo. Swapped green/yellow on both halls and phase.

I thought that may be the case, but I tried many of the hall phase combos. I found two that spun the motor fine on the test stand, but under load, the motor sputtered. One of these combos was what Mike at LR suggested, and another combo was found through experimenting.

I'm going to junk the PV controller since I wasted way too much time with it. It may have been faulty, or I just didn't find the magic hall/phase combo, but I'll scrap it before I get tempted to waste more time on it. I'll stick with Lyen or the tried and true Cyclone controllers from now on.

-Brian

 

[sparkiesweden]

Yeah, I know that frustrated feeling of not getting anywhere, when you don't know where the problem really is.

Will be trying the Phaserunner soon, guess it will (initially) create some of that frustration (again) :-D

 

[80for20]

I've run into the same problem. After talking to Mike I'm just going to revert back to an old 18fet Lyen. It's a trapezoidal. We'll see if that fixes things.



Sent from my iPhone using Tapatalk

 

[bzhwindtalker]

We're having the same issue with our engines, causing the same symptoms and incompatibility with sin controllers.
Basically the sensors are not giving a correct signal and cause the controller to "back fire". That can lead to bigger issues like over current faults errors.
I want to try lower sensitivity halls to see how we can solve this, I think they may be too sensitive and pick up some sort of harmonic field.
View attachment A0002DS.BMP

 

[sparkiesweden]

The "trick" to fully utilize sine (FOC) controllers is to run them sensor(hall)less, especially on high (e)rpm's.

Some controllers have an option to just use the halls at start and then switch to sensorless.

 

[Alan B]

There are a few characteristics of sinusoidal commutated controllers that are potentially an issue here. They are generally based on low end processors but use higher capability three phase PWM generation, so they can generate smoothly rotating motor voltages from table lookups but cannot do the heavy duty calculations required for full field oriented control. So they do things in simplified ways that have certain effects on the result. The rotor position is the key variable that is needed but difficult to know accurately.

Sine commutated controllers interpolate the hall sensors in time to estimate the rotor position. For this they need accurate and noise free hall signals. This depends not only on the hall sensors themselves, but also on the motor's magnetic field quality at the hall locations. Not all motors have good magnetic design for this, some have various issues like noise or shifts in timing with RPM.

Another characteristic of sine controllers is they control voltage by setting PWM values, they don't precisely control the currents. This works pretty well at low RPM but as the RPM increases the inductance of the motor causes the phase shift between voltage and current to increase. The goal is to control the fields in the motor, which come from current (and keep these precisely at 90 degrees from the rotor). Controlling the voltages instead of the currents causes errors which increase with RPM.

Field Oriented Control, aka Vector Controls instead use measurements of the motor currents and much higher computational power to compute the rotor position with a number of feedback loops. If these feedback loops are working well the result is precise and efficient control over a wide RPM range. But there are many tuning parameters to adjust and the assumptions about the motor application that were made when the code was designed can affect how well it works, and mid drives with varying collections of gears, clutches and chain lag can be difficult to tune or compensate for. Oscillations may creep in and make the system unstable. More programmable controllers are capable of working around this, but those are rare in the ebike world.

Selecting motors with better magnetic design, making the bike transmission system stiffer and avoiding powertrain oscillations can help stabilize the system which can help any controller perform better.

 

[80for20]

Thanks for this Alan.

The last paragraph, especially, seems relevant to me. One of the scenarios that causes my "jittering" is when I go over a speed bump and have the throttle engaged. FYI, the bike is a hardtail so your comment about stiffer transmissions (in my case a wobbly chain because of the bounce) may make sense.


Sent from my iPhone using Tapatalk

 

[Alan B]

Could be the feedback loops or motor parameters need adjustment. The disturbance from a speed bump will tend to trigger any control issues, as well as possibly perturb the inputs.

Many ebike controllers know nothing about the motor, load, etc. All the assumptions made in the firmware may or may not be correct for your setup. The more the controller correctly knows about the actual system and application the better it can control.

 

[80for20]

UPDATE: So it appears that the fix for my particular problem was simply to change the Max Throttle Out setting on the CAV3 to 3.3V. This seems to have fixed the issue (hopefully for good).

Thanks to both Mike and Vad (from PowerVelocity) for staying on top of this and getting to the resolution.


Sent from my iPhone using Tapatalk