I think i'm having Gyroscope Problems

[seage]

Hi everyone, 

I'm a beginner and have no experience with any other wheel (and close to none with this one) so i wanted to ask in here with people who are more experienced than I. I have the ks 16-s and over the past maybe 2 or so sessions I've noticed my pedals dipping a lot. I initially attributed this to the pressure in the tire, as i was riding quite low, then filled it up. But it feels like its a bit...extra? Yesterday when I went out I noticed it. When spinning around a basketball pole that was on uneven ground. If i spun in the direction of the downhill it would tilt me forward pretty hard then stable out when i stood still. And if i went uphill it would put me on my heels. But this was a very slight incline, it was no hill (Enough that kids could still play bball here), but today when trying to mount, it just tipped me forward. I keep thinking its human error, but I'm not quite sure now. Its also a pretty abrupt dip, its not slow. I've made 2 videos. One just walking it in the basement.  And one of me walking in my garage, set to medium riding mode. My garage has a slight incline so when i walk up it, the wheel tilts back, and down the wheel tilts forward. I don't remember feeling this before...

The app is updated to the latest it will let me download. In the first video it is set to experienced mode and the tire is around 48psi and the second is set to medium with the same tire pressure. The handle broke yesterday so we did a fix that stops it from locking (I'm getting a new one soon). But the wheel feels like its dipping WAY more than it used to and feels like its getting worse. Is this just something that comes with higher air pressure that i need to get used to? Its just strange because the tilt is even heavy when i'm just walking it ,and its something I didn't notice at all in my first few days with the wheel. 

Thanks everyone! 

[meepmeepmayer]

The pedal dipping in turns is typically due to a bad calibration. Definitely do a proper calibration and see if it changes anything.

Keys to a good calibration are:

• Do not tilt the wheel sideways during the entire process. Front/back lean (= just the pedal angle) can be whatever you want it to (keep it horizontal for this). Just no sideways lean, that "twists" the sensor which leads to bad calibrations.
• Keep the wheel still during the entire process. Lean it against a hard wall, not hold it in your hand or lean it against something soft.
• I always suggest to da a crazy pedal angle calibration (like 45°) simply to confirm the method you used worked. Then you can calibrate again, this time to horizontal.
• You have to find out yourself how to do a calibration with a KS

What is shown in the second video is a bit more strange. Almost like tiltback enables? Bordering on a bad board behavior. But it still might be explained as dipping in turns, albeit a very crazy one.

What you can do here, is NOT use Wheellog or any other third party app (that might accidentally change some setting, like tiltback). Only use the official KS app, and set all settings to some normal number, while finding out what's going on. I suggest setting the ride mode to hardest to minimize intended pedal dipping.

Also see if the dipping appears if you trolley the wheel in a straight line (instead of curves like in your videos). Yes: there's a problem. No: more likely a bad calibration. Make sure you don't start with a turn.

I could speculate more what the issue might be here, but for now you should do a calibration and see if it changes anything.

edit:

• Dipping has absolutely nothing to do with the tire pressure.
• Dipping has absolutely nothing to do with the incline. A wheel does not know what incline it is riding and therefore can not react to it. It only knows the pedal angle and keeps it horizontal. So any perceived incline-dependent behavior must be something else (which may indirectly be incline-dependent, but the wheel doesn't know).
• Also, as a general idea, a EUC's pedals should always be horizontal. That's kind of how it works. So unless it's short and mild temporary dipping from a soft riding mode, any dipping is not intended and officially bad.

[seage]

35 minutes ago, meepmeepmayer said:

The pedal dipping in turns is typically due to a bad calibration. Definitely do a proper calibration and see if it changes anything.

Keys to a good calibration are:

• Do not tilt the wheel sideways during the entire process. Front/back lean (= just the pedal angle) can be whatever you want it to (keep it horizontal for this). Just no sideways lean, that "twists" the sensor which leads to bad calibrations.
• Keep the wheel still during the entire process. Lean it against a hard wall, not hold it in your hand or lean it against something soft.
• I always suggest to da a crazy pedal angle calibration (like 45°) simply to confirm the method you used worked. Then you can calibrate again, this time to horizontal.
• You have to find out yourself how to do a calibration with a KS

What is shown in the second video is a bit more strange. Almost like tiltback enables? Bordering on a bad board behavior. But it still might be explained as dipping in turns, albeit a very crazy one.

What you can do here, is NOT use Wheellog or any other third party app (that might accidentally change some setting, like tiltback). Only use the official KS app, and set all settings to some normal number, while finding out what's going on. I suggest setting the ride mode to hardest to minimize intended pedal dipping.

Also see if the dipping appears if you trolley the wheel in a straight line (instead of curves like in your videos). Yes: there's a problem. No: more likely a bad calibration. Make sure you don't start with a turn.

I could speculate more what the issue might be here, but for now you should do a calibration and see if it changes anything.

edit:

• Dipping has absolutely nothing to do with the tire pressure.
• Dipping has absolutely nothing to do with the incline. A wheel does not know what incline it is riding and therefore can not react to it. It only knows the pedal angle and keeps it horizontal. So any perceived incline-dependent behavior must be something else (which may indirectly be incline-dependent, but the wheel doesn't know).
• Also, as a general idea, a EUC's pedals should always be horizontal. That's kind of how it works. So unless it's short and mild temporary dipping from a soft riding mode, any dipping is not intended and officially bad.

@meepmeepmayer If i ever meet you, im buying you the drink of your choice in excess. Lol. I didn't know that the side to side angle had any bearing on the forward and backwards calibration. I had calibrated it a few times, but each time the sideways angle was a bit more than necessary so that the wheel wouldnt fall over. This time i followed your advice to the letter and stood it as straight as could be. (I did the 45 degree thing first). And set it. And suddenly its going straight with no dips! I made a little ramp, and the pedals stayed strraight up and down it, backwards and forwards. I stood on it and all the weird jerkiness of it disappeared. 

But hey, thank you so much for all your information. I'm trying my best to learn as much about this thing under my feet and I feel like a complete idiot a lot. Im not used to being so far out of my element and completely new to both to activity and the technology, lol. I assumed this was happening due to human error, or my lack of experience, and that was KIND of the reason. I dont know if i would have figured out that the lean messes up the calibration though. I've uninstalled wheelapp for the time being since i wasnt even using it anyways. And sticking to default everything on the wheel until i look into everything and learn how it works. At least now i know what to look for in case there are any other issues that arise due to my lack of knowledge XD. 

Thank you so much, meep. I really appreciate the help you always give me here. 

I get the feeling this is really going to help my riding, because ive been riding while the wheel dips me backwards and forwards pretty much since my first video that I posted. I wonder how its gonna feel now that its sturdy....

[meepmeepmayer]

Thanks!

It works now? Cool I was worried of a lengthy repair saga as with so many issues posted here. Lucky for you, it isn't. Nothing worse as a new rider with a new wheel.

49 minutes ago, seage said:

And suddenly its going straight with no dips!

The sensor inside the wheel is essentially (or maybe literally?) the same one you'd find in some smartphone as gyrometer/accelerometer. Imagine those smartphone apps that show a virtual level "water" surface if you rotate your smartphone around . That means: the sensor is just not very good.

So imagine a tired, bored and at the same time rushed Chinese factory worker putting newly-built wheels into some stands and starting the initial calibration on 25 wheels at a time. If a wheel is unluckily sideways tilted then, bad calibration, and you get the dipping. That's how I picture it.

The sensor can also "drift" over time (as it is not very good) and then the wheel starts dipping in curves, and another clean calibration will fix it again. So a less-bad factory calibration may still develop dipping soon (like it did for you).

The "not sideways" thing was learned from the frustrations of many riders (not so much me) dealing with the cryptic quirks of their new wheels (wondering if the feature was intended, some manufacturers at some point might even have claimed that). The drift thing was told to us by the Inmotion CEO. That the sensors are not very sophisticated is the conclusion

And yes, the dipping is VERY disconcerting as a new rider. Just when you start to get it, it seems the wheel is trying to kill you. My ACM did it too, luckily I got that fixed fast.

49 minutes ago, seage said:

But hey, thank you so much for all your information. I'm trying my best to learn as much about this thing under my feet and I feel like a complete idiot a lot. Im not used to being so far out of my element and completely new to both to activity and the technology, lol.

You're selling yourself short, just like with the learning to ride But I guess EUCs can seem mysterious at first.

Indeed EUCs are ultra simple. There is a tilt sensor (caring just for the forward/backward direction tilt) and there's a motor. If the wheel begins to tilt, it engages the motor to drive in the direction of the tilt in order to counterbalance it and get back to horizontal. That's literally the entire way a self-balancing device works. Sure, how exactly it reacts to a beginning tilt is a firmware question, but I presume the design curve there is very simple, too.

Wheels can also detect a fallen over wheel (and disengage the motor) if the sideways tilt is too bad, but that's about it. You couldn't ride a wall of death because then the wheel would think it has fallen over and switches the motor off, despite you happily riding. Even the speed meter is just counting the tire revolutions per minute.

So there's only the tilt sensor and attached electronics (motor, battery) controlled by the sensor tilt, tyring to get back to level.

That means a wheel does not know where you ride, what the incline is, if it is even on the ground (lift it up while on and watch it go crazy from the tilt and even control it by tilting it), or anything really. It just keeps itself upright. Literally all it does. It's stupid but very effective at doing one thing. So the pedals always being nice and horizontal is the essence of a working EUC.

That's why wheels have no fixed top speed, incline, whatever - just a power ceiling. However much power the motor and battery can supply to keep the balance if a tilt appears at any time.

[LanghamP]

I have to recalibrate my KS16S every few months or so. Pedal dip always gets worse after time, with the pedals eventually pointing down quite far, and then it's time to recalibrate. I wonder what would happen by simply reversing the wheel on half the rides.

In my opinion the biggest drawback to pedal dip is while turning and the leading corner catches on the ground. Everything else you can conform to except catching your pedals on something, and it is for this reason that I even do calibration.

To confirm your wheel calibration was correct, simply grasp the wheel by the handle in closed postion and roll the wheel forward, then gently do a 180 turn. The wheel should stay dead level, yes? But in practice I've never gotten my KS16S perfect although it's very close; I mean I've used two movable walls with electronic levelers and I've still not gotten it perfect. If you stop rolling the wheel midway through the turn the pedal stays dipped forward, then slowly rights itself (why?!). Moving the wheel forward then causes the pedal dip to go away. Pedals dip backwards when braking yet suddenly tilt forward when you start turning?

I do notice there's no pedal dip if you blast through a corner without slowing down or speeding up; the pedals stay flat at higher speeds. This makes me believe the rate of turn has something to do with pedal dip.

In my opinion, since pedal behavior is consistent across all wheels (calibration just makes the behavior more mild), there's something in the algorithm that makes pedal dip unavoidable. Maybe newer wheels will have updated algorithms?

[meepmeepmayer]

12 minutes ago, LanghamP said:

In my opinion, since pedal behavior is consistent across all wheels (calibration just makes the behavior more mild), there's something in the algorithm that makes pedal dip unavoidable. Maybe newer wheels will have updated algorithms?

I don't believe so. A good enough sensor will give the right numbers however it is positioned. And the wheel will use them directly. What could go wrong?

Also, I have never had to recalibrate my ACM in 5000+km and its nearly 2 years of age, it is rock-solid horizontal ever since the first good calibration.

But maybe you're right.

And the manufacturers might have poured oil on the fire by having their firmware have special quirks just to mitigate the shitty sensors. So different wheels may be more or less susceptible to the same sensor drift/dip. Maybe the 16S is more susceptible? Or it depends on something random like the positioning of the board, wheel geometry, something unexpected like this, whether the sensor is easily confused or not?

[LanghamP]

2 hours ago, meepmeepmayer said:

I don't believe so. A good enough sensor will give the right numbers however it is positioned. And the wheel will use them directly. What could go wrong?

Also, I have never had to recalibrate my ACM in 5000+km and its nearly 2 years of age, it is rock-solid horizontal ever since the first good calibration.

But maybe you're right.

And the manufacturers might have poured oil on the fire by having their firmware have special quirks just to mitigate the shitty sensors. So different wheels may be more or less susceptible to the same sensor drift/dip. Maybe the 16S is more susceptible? Or it depends on something random like the positioning of the board, wheel geometry, something unexpected like this, whether the sensor is easily confused or not?

All my wheels are a few years old but then again so is your ACM. And all five of my wheels, from different companies, exhibit the same pedal dip under the same conditions.

It kinda pisses me off.

Here's some advertising for vibration gyros. Is this what our EUCs use?

https://www5.epsondevice.com/en/information/technical_info/gyro/

If this is so, wouldn't you need two of these not just one? Perhaps six for redundancy. One for left and right tilt, and one for front and back, and it seems if you rotate the wheel then both gyros (not just the left and right title) are affected, due simply to the gyros not being dead center in the wheel.

It seems very complicated yet all wheels exhibit this pedal dip in a turn behavior. It's consistent, well except for your ACM.

[Seba]

16 minutes ago, LanghamP said:

Here's some advertising for vibration gyros. Is this what our EUCs use?

Yes, our wheels use so-called MEMS IMUs. MEMS is an abbveriation of "MicroElectroMechanical System", and IMU is an abbreviation of "Inertial Measurement Unit". IMU is what we used to call "gyroscope", but in fact it consists of two sets of sensors - three accelerometers and three gyroscopes. In the process of calibration only accelerometer part is calibrated, because this is the sensor which determines wheel position in regard to local gravity direction. Gyroscope is measuring rotation rate. Signals from both sensor sets are then merged by EUC controller so it knows its linear and angular movement. The same IMUs are also used in drones, for example.

This tiny chip marked with red arrow is the most important sensor in every EUC (this image shows KS-18L/XL controller board):

[Unventor]

My first wheel was an Inmotion V8. It would dip down a fair bit when doing tight turns especially if I were decelerating during the turn. 

Now my KS18L do not have this behaviour as severe but it do still happen a little. But the major difference is the KS18L feels much more intuitive and stable, which leads to more confident in the wheel and how anticipate its behavior. 

This is why I love my KS18L so much. It ozzes quality to me. How a KS16S is in this context and its normal model behaviour is I can't say. It might be different. 

Right now I face a scrap or major repair job due inner and outer shell is far more damaged after the tumble the V8 did when I had an accident with clost to full speed aka 30ish kmh. At first it behaved were oddly kinda like standing on a surfboard on a wave passing under it as you padel put to sea. So I face repair cost of 200 to 250ish £ and moving all parts to a new shell. 

So I had briefly looked at KS16S or KS14D as a short distance low speed and practice wheel to train ambidextrous start or step up if you like and backwards riding and penduling or idling. That is why your thread catched my eye.

@US69 I know you have a lot experience with different brands and wheels. Do the KS14D, KS14S, KS16S behave different to the KS18L as dip down during turning?

I don't think my V8 was special or faulty as I tested another V8 that felt the same way.