Jump to content

Could a modified EUC be ridden on a "Wall of Death"?


Jonathan Tolhurst

Recommended Posts

2 hours ago, Tomek said:

but the gyro doesn't record that perceived rotation change

I am no expert of gyros, but I beleive these inexpencive ones uses a gyro that actually measures the angualr rate of change. It must be integrated (summed) to find the actual angle. Biases in this integration chain will result in the measured/computed angle to drift from the correct angle over time.

Btw: Everyone calls these "oreientation sesor" in the EUC for a gyro, but it is actually multiple gyros and accelerometers packed in the same casing (such as the MPU 6050). It is important to recognize the difference between a gyro and an acelerometer. They depend on quite different physical phenomena and therefore measures quite different things. They are used in combination (aka sensor fusion) for computing more accurate values.

Link to comment
Share on other sites

  • Replies 63
  • Created
  • Last Reply
1 minute ago, Smoother said:

Or a trained monkey.  No, I'm not volunteering.

:D Creative idea. Just learn a chimp basic EUC skills, and wait and se what happens! Would actually be quite interesting - I think (he might throw it in your head though, so best to wear a helmet)

Link to comment
Share on other sites

4 hours ago, esaj said:

The gyros are multi-axis, for example the MPU-6050, which is used in many wheels, has 6 axes (3-axis gyro + 3-axis accelerometer):

https://www.invensense.com/products/motion-tracking/6-axis/mpu-6050/

The MPU-9250 is 9-axis (3-axis gyro + 3-axis accelerometer + 3-axis magnetometer). I don't know if that's used in wheels, the magnetometer axes are reference to the magnetic poles of the Earth (so kinda like a 3-dimensional compass).

The reason for using gyro + accelerometer seems to be that gyros tend to "drift" during use, but the accelerometer data can be used to offset that, as used in "sensor fusion" algorithms to enhance the resolution/precision. The accelerometer can also be used to sense which way is down (at least when stationary), because the gravity will pull the axes vertical to ground downwards at constant acceleration (if a single axis isn't exactly vertical, two or more axes will receive a fraction of the acceleration), my best guess is that's what's used in the wheels to detect the "0-position" at start up.

FXYJ9W1IVO3T137.ANIMATED.MEDIUM.gif

The above is a microcontroller detecting the direction of gravity using a 6-axis accelerometer/gyro, not at a very good precision (could also be due to the low resolution of the led-matrix), but I guess the principle is the same.

 

Quite informative. Thanks 

Link to comment
Share on other sites

On 23/02/2017 at 0:53 PM, Frode said:

My guess is that the truth is somewhere in between, that it will slowly readjust the vertical (slow with respect to typical movements for a typical user). I would not count on it not turning off when it reaches the turn-off tilt angle. I would rather expect it to do exactely that - turn off (because the tilt angle changes too fast for it to adjust to a new "vertical"). But I don't know.

 

On 23/02/2017 at 1:12 PM, Frode said:

:D Creative idea. Just learn a chimp basic EUC skills, and wait and se what happens! Would actually be quite interesting - I think (he might throw it in your head though, so best to wear a helmet)

Given the vid from EUCextreme in a skater park, the problem which fast/sharp turns potentially pose in particular on a slanted surface, and the fact that the earth surface angle turns 1º every four minutes, I tend to think that for lateral orientation the g-force reference perceived downforce must be used all the time. Hence I would bet against the wheel turning off. Just ordered the chimp :)

Link to comment
Share on other sites

6 minutes ago, Mono said:

Given the vid from EUCextreme in a skater park, the problem which fast/sharp turns potentially pose in particular on a slanted surface,

This is the very reason why I think the vertical cannot be re-aligned too fast. In such situations the system must be able to keep track of the vertical or the result will be a quite unpredictable behaviour of the EUC - I beleive.

Well, we'll wait for EUC Extreme. I don't doubt for a second that he will try to prove me wrong. And he might even succeed in that :D

Link to comment
Share on other sites

On 24/02/2017 at 1:46 PM, Frode said:

In such situations the system must be able to keep track of the vertical or the result will be a quite unpredictable behaviour of the EUC - I beleive.

Why do we need to use vertical alignment for the lateral-tilt-turn-off functionality to work as we expected? AFAICS g-force perceived downforce should be perfectly sufficient for this functionality to work as expected.

Link to comment
Share on other sites

3 minutes ago, Mono said:

Why do we need to use vertical alignment for the lateral-tilt-turn-off functionality to work as we expected? AFAICS g-force should be perfectly sufficient for this functionality to work as expected.

Everything would become self refering. You and your EUC would probably start ocillating.

Link to comment
Share on other sites

4 minutes ago, Frode said:

Everything would become self refering. You and your EUC would probably start ocillating.

I don't see how this answers the question, which was only referring to the lateral tilt turn-off functionality.

Link to comment
Share on other sites

On 24.2.2017 at 3:20 PM, Mono said:

I don't see how this answers the question, which was only referring to the lateral tilt turn-off functionality.

I might have been misnderstanding you. What I say is that the EUC's vertical reference (what the EUC think is upwards on the casing) should not be aligned too fast to the real world vertical. I now think you agree to that.

Remember though that the g-force (towards the earth center) is 9.81 m/s^2 and it works all the time. The resultant "g-force" (vertical) measured by the accelerometers is the vector sum of the circular acceleration and the real g-force. The measured vertical will drift from what is upwards on the EUC's casing depending on where you are in the ball. How that will feel and wether it is controllable by the driver is not an easy question to answer without trying it.

Link to comment
Share on other sites

1 hour ago, Frode said:

I might have been misnderstanding you. What I say is that the EUC's vertical reference (what the EUC think is upwards on the casing) should not be aligned too fast to the real world vertical. I now think you agree to that.

Remember though that the g-force (towards the earth center) is 9.81 m/s^2 and it works all the time.

oups, sorry, I used g-force in the same meaning as Tomek used it, namely to mean all acceleration forces.

Link to comment
Share on other sites

16 minutes ago, Mono said:

oups, sorry, I used g-force in the same meaning as Tomek used it, namely to mean all acceleration forces.

If I have calculated this correct, then in a ball with 3 meter radius, the vector sum of the g-force and your centripetal force will be 45 degrees related to vertical if you have 20 km/h and you have climed up to the middle of the ball (i.e. you are laying horiontal). Then your EUC is at the brink of turning off if the turn off angle is 45 degrees. (If that speed is fast enough to keep you there is another question).

My prediction is also that if you try to go for a loop like the motorcycles in the ball, the EUC will stay vertical (vertical relative to the earth) and you will end up slamming into the ball's vertical part of its "sides". A vertical faceplant - you and the EUC are vertical - but youur face hits somwhere in the upper half of the ball. :P

Link to comment
Share on other sites

8 minutes ago, Frode said:

My prediction is also that if you try to go for a loop like the motorcycles in the ball, the EUC will stay vertical (vertical relative to the earth) and you will end up slamming into the ball's vertical part of its "sides". A vertical faceplant - you and the EUC are vertical - but youur face hits somwhere in the upper half of the ball. :P

Then our predictions disagree. 

How do you explain the vids showing EUCs going up and down slanted skater walls?

Link to comment
Share on other sites

Just now, Mono said:

Then we disagree on our predictions. 

How do you explain the vids showing EUCs going along slanted skater walls?

Look at the EUC in this video at the first seconds from 0:23. This is how an EUC behaves.

 

Link to comment
Share on other sites

1 minute ago, Frode said:

Look at the EUC in this video at the first seconds from 0:23. This is how an EUC behaves.

Right, seemingly it can ride along a curved wall with an angle of more than 45º degree. I don't see how this is in alignment with your prediction.

Link to comment
Share on other sites

You can also studdy these clips:

 

Just now, Mono said:

Right, seemingly it can ride along a curved wall with an angle of more than 45º degree. I don't see how this is in alignment with your prediction.

Yes, but it will stay vertical. So what happens when it hits the part that is 90 degrees? (or actually before that)

Link to comment
Share on other sites

1 hour ago, Frode said:

So, the EUC is not vertical between 23 and 26 seconds?

You mean in the example where it fails? When riding up a hill the forward-backward alignment remains horizontal (which is what we see between 23 and 26 seconds), and it is not possible to directly go up a wall with 80º inclination. That doesn't mean it is not possible to go along a curved wall which has 80º inclination. For example at 0:27 and 0:39 we see that the EUC rides perfectly fine with quite non-vertical alignment which seems to exceed 45º. 

It is anyway immediately obvious that the contact patch of the EUC must be kept "below" the CoG all the time, where "below" is defined in terms of downforce (meaning perceived acceleration), not in terms of vertical alignment with the earth surface.

Using failure examples is also somewhat problematic: we have many examples where riders fail to ride an EUC on a perfectly flat surfaces. They don't tell us anything about whether and to what extend it is possible to ride an EUC on a perfectly flat surface.

Link to comment
Share on other sites

3 hours ago, Frode said:

If I have calculated this correct, then in a ball with 3 meter radius, the vector sum of the g-force and your centripetal force will be 45 degrees related to vertical if you have 20 km/h and you have climed up to the middle of the ball (i.e. you are laying horiontal).

sounds about right to me. One might need to go a little faster to actually climb up to the middle, because at 45º one might still not have enough friction and slide down the wall.

Quote

Then your EUC is at the brink of turning off if the turn off angle is 45 degrees. (If that speed is fast enough to keep you there is another question).

this is where we disagree, as I have come to conclude by now that the angle with an overall perceived downforce is what turns the wheel off rather than the (estimated) angle with the centre of gravity of mother earth. Do you agree that most wheels cannot measure this latter angle directly?

Link to comment
Share on other sites

1 hour ago, Mono said:

You mean in the example where it fails?

In the example where i lets it roll on its own down the slope.

The point is that an EUC will always be vertical in the forward/backward direction. It can of course tilt sideways. It will be possible to clime a wall sideways. A loop will however not be possible.

Link to comment
Share on other sites

14 minutes ago, Mono said:

Do you agree that most wheels cannot measure this latter angle directly?

No, I don't agree. This is exactely what a gyro can do. It can measure how the EUC is oriented independent of what the accelerometers senses. The gyro do not care where the earth is. It don't care if there are an earth at all.

Actually, what I beleive is that the gyros are used for the measurements of orientation and the accelerometers are used for compensating the gyros for drift in relation to the earths vertical over time (due to earth rotation, missalignment at startup, integration bias etc., time constand in the order of seconds at least)

Link to comment
Share on other sites

2 minutes ago, Frode said:

The point is that an EUC will always be vertical in the forward/backward direction.

That might be so, even though I haven't seen enough evidence to be entirely convinced. I don't see however why this would be prohibitive to ride the wall of death.

Link to comment
Share on other sites

1 minute ago, Mono said:

Exactly, hence it cannot measure the angle to the CoG of the earth directly. 

Not directly, but via the (compensated) gyros. I think it does it pretty well to. I have never felt that my KS-16 have tilted forward or backward due to me accelerating, runing up a slope or its like. I have however felt it slowly tilting forward when turning at very slow and sharp turns. I don't really understand why this happens, but my guess is gyro drift for some reason (I havent tried to figure it out). I have also felt it slowly tilting forward when running on av very bumpy gravelroad (lots of small bumps making the KS-16 shaking a lot). I think (but don't know) this might be due to the accelerometers measuring a slightly forward tilted downward force which will make it tilt a little forward, but since this will not correct the forward tilted force (since it is not due to the EUC being tilted backwards), the EUC will continue to tilt forward (and I have to reduce speed bacause it feels like it will end with me being thrown off).

Well, are you going to demonstrate me being wrong about looping? :D

Link to comment
Share on other sites

21 minutes ago, Frode said:

I have however felt it slowly tilting forward when turning at very slow and sharp turns. I don't really understand why this happens, but my guess is gyro drift for some reason (I havent tried to figure it out).

This is a behaviour some wheels have and others do not. I don't quite see how and why that would be related to forward-backward accelerations. 

Quote

I have also felt it slowly tilting forward when running on av very bumpy gravelroad (lots of small bumps making the KS-16 shaking a lot). I think (but don't know) this might be due to the accelerometers measuring a slightly forward tilted downward force which will make it tilt a little forward, but since this will not correct the forward tilted force (since it is not due to the EUC being tilted backwards), the EUC will continue to tilt forward (and I have to reduce speed bacause it feels like it will end with me being thrown off).

You explain change of the forward-backward tilt from short-term accelerations, which seems a good hypothesis to me. It seems however to contradict your original claim that forward-backward tilt is determined based on the vertical earth surface alignment with only long-term corrections for the earth movement.

Quote

Well, are you going to demonstrate me being wrong about looping? :D

Possibly, but most likely not directly :P

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...