Diff between gen 1 and 2

[KSagal]

Originally Posted by Bob.Kerns

Quick response -- I'll try to do better later.

To arrest a person's fall, you have to apply force.

Thanks to F = m * a, the only way to do that is by acceleration. Velocity won't do it. You can analyze the situation identically whether v = 0 or v = 12.5. To visualize the latter situation, think of someone videoing someone from another Segway.

One more way to slice it: The contact patch on the tire can provide two components of force. The force up, if not aligned with your center of mass, produces a torque, leading to face plant city.

But add in the horizontal force, and you can balance everything with a torque around your CG in the other direction. But F = ma! So the whole system (you, Segway, your groceries) accelerates.

In any static situation (static in the sense that the situation is not changing -- there can be constant motion), the sum of all the force vectors, and the sum of all the torques (also a vector quantity), must be zero. Otherwise, you'll have acceleration or rotation. Rotation = face plant.
The only place velocity yields a force is with viscosity, e.g. wind resistance, or similar electromagnetic phenomena, where kinetic energy is being dissipated in a way that increases with velocity. (Simple friction doesn't depend on velocity).

The Segway speed limiter stops the lean by simply accelerating the powerbase a bit more than enough to balance your lean -- bringing it and your feet forward, thus eliminating the lean!

No matter how much you TRY to lean at that point -- it has control, because you don't have anything to push on, but it does. You try to bring your CG forward -- it brings the contact patch forward to match, and a little more so it can bring the speed back down to 12.5 (or whatever speed it thinks you should be limited to).

I think you just agreed with me. If you consider a 'lean' as a specific amount of weight out of balance, and a person were to lean that much out of balance but not attempt to go more out of balance (lean a certain amount and hold it, as opposed to continuing to lean where they will eventually fall) then it is possible for the segway to apply forward force to the platform, via the friction on the road with the contact patch of the tires, and equalize the lean or out of balance condition, and when it equalizes, the segway will be moving at a constant speed, and there will be no acceleration (which would arrest the fall) or rotation (which is the continuation of the fall).

The only other time you can lean forward with no acceleration that I can think of would be when you do so in front of a very large pile of jello. But that goes into your viscosity argument. ( As I skydiver, I have experienced 200mph+ winds upon myself, and the viscosity of air at any reasonably atainable speed is just not going to stop this ole' fat boy in mid fall...)

[Bob.Kerns]

Originally Posted by KSagal

I think you just agreed with me.

Nope. But we're narrowing in on it.

Originally Posted by KSagal

If you consider a 'lean' as a specific amount of weight out of balance, and a person were to lean that much out of balance but not attempt to go more out of balance (lean a certain amount and hold it, as opposed to continuing to lean where they will eventually fall) then it is possible for the segway to apply forward force to the platform, via the friction on the road with the contact patch of the tires, and equalize the lean or out of balance condition, and when it equalizes, the segway will be moving at a constant speed, and there will be no acceleration (which would arrest the fall) or rotation (which is the continuation of the fall).

The only other time you can lean forward with no acceleration that I can think of would be when you do so in front of a very large pile of jello. But that goes into your viscosity argument. ( As I skydiver, I have experienced 200mph+ winds upon myself, and the viscosity of air at any reasonably atainable speed is just not going to stop this ole' fat boy in mid fall...)

The only ways to equalize the lean or out-of-balance condition are:

• Continue accelerating so the upward & horizontal components of the force from the contact patch sum to a vector that goes through the CG, or
• Remove the lean. The vertical vector goes through the CG, and no horizontal forces; hence, no acceleration. And no lean.

(Of course, you could do something between -- remove some of the lean, and accelerate less.)

So you can "equalize the lean" by removing it, or by accelerating. No other choices are available, outside of plowing into a block of jello, as you so aptly put it.

I figured as a skydiver, you could relate to how much faster you'd have to go for wind resistance to be a significant factor. The wind resistance increases as the square of the velocity, so you really don't see much effect at all at low speeds. At 10% of terminal velocity, you'll only see 1% of the force. That works out to be about 0.57 degrees of lean. (I'm not sure what value of terminal velocity to apply to someone standing on a Segway -- I'd guess 125 MPH would be close, so something in the ballpark of a half degree of lean would be about all you'd obtain from that.

[KSagal]

Originally Posted by Bob.Kerns

Nope. But we're narrowing in on it.




The only ways to equalize the lean or out-of-balance condition are:

• Continue accelerating so the upward & horizontal components of the force from the contact patch sum to a vector that goes through the CG, or
• Remove the lean. The vertical vector goes through the CG, and no horizontal forces; hence, no acceleration. And no lean.

So you can "equalize the lean" by removing it, or by accelerating. No other choices are available, outside of plowing into a block of jello, as you so aptly put it.

I figured as a skydiver, you could relate to how much faster you'd have to go for wind resistance to be a significant factor. The wind resistance increases as the square of the velocity, so you really don't see much effect at all at low speeds.

So you are saying that it is possible to accelerate and not increase in speed?

If not, how is it possible to maintain a speed of 10 mph on a segway?

[Bob.Kerns]

Originally Posted by KSagal

So you are saying that it is possible to accelerate and not increase in speed?

If not, how is it possible to maintain a speed of 10 mph on a segway?

No. You understand correctly (from the definition of acceleration as the rate of change in velocity) that that would make no sense.

I'm not sure why you think I said that, so I can't usefully respond to that first part.

As for the second -- it's possible by standing up straight with no lean, once you reach 10 MPH. Briefly lean forward a tiny amount to pick up some speed if you lose a bit of speed (say, a gust of headwind). Briefly lean back a tiny amount of you pick up a bit (overcorrection, or a tail wind).

So your feedback control is around that balance point -- and it's vertical, not leaned. (Again, ignoring wind resistance).

You might want to re-read my previous post -- it looks like I edited it after you read it; I wasn't quite quick enough.

[KSagal]

Originally Posted by Bob.Kerns

No. You understand correctly (from the definition of acceleration as the rate of change in velocity) that that would make no sense.

I'm not sure why you think I said that, so I can't usefully respond to that first part.

As for the second -- it's possible by standing up straight with no lean, once you reach 10 MPH. Briefly lean forward a tiny amount to pick up some speed if you lose a bit of speed (say, a gust of headwind). Briefly lean back a tiny amount of you pick up a bit (overcorrection, or a tail wind).

So your feedback control is around that balance point -- and it's vertical, not leaned. (Again, ignoring wind resistance).

You might want to re-read my previous post -- it looks like I edited it after you read it; I wasn't quite quick enough.

Wait a minute! Can I go back and fill in holes I left in my posts that you pointed out? I did not know that was in the rules!

I still cannot completely agree. If you could lean to get to speed, then stand directly above the CG to maintain speed, you are describing a perpetual motion machine. I will accept that you need to lean more to get to speed than you need to lean to maintain speed, but I cannot accept that you are backing off the lean to full vertical, or you would coast to a stop.

There has got to be a bit of weight in front of the cg for the seg to move... Of course, a segway at speed is not level. I suppose if you consider vertical in this case to be a line from the contact patch of the tires, thru the cg, and continuing up, than you could stay in line with the cg and still move, but that line would not be the same a vertical, it would lean forward a bit...

Edit: Correction. A line from the contact patch, thru the cg of a forward leaning segway would not lean forward, but backward, as the contact patch is in front of the axle. A person standing perpendicular to the platform would indeed be leaning forward of the cg...

Edit 2: I just realized that a person standing in front of that backward leaning line (Contact point which is in front of the axle thru the cg and up) could be standing vertical. Since the segway is leaning forward, they could be leaning backward relative to perpendicular to the platform but stil vertical and forward of the line...

[KSagal]

You were supposed to come around to me, not me toward you!

Sheesh! Don't think for a moment I will not be just as arrogant about other stuff....

Nicely done...

[Bob.Kerns]

Originally Posted by KSagal

Wait a minute! Can I go back and fill in holes I left in my posts that you pointed out? I did not know that was in the rules!

Sorry -- I was trying to improve the message for your benefit before you read it, since I didn't think of some details to add until after I hit "Submit". Sorry to make you read it twice.

Originally Posted by KSagal

I still cannot completely agree. If you could lean to get to speed, then stand directly above the CG to maintain speed, you are describing a perpetual motion machine. I will accept that you need to lean more to get to speed than you need to lean to maintain speed, but I cannot accept that you are backing off the lean to full vertical, or you would coast to a stop.

There has got to be a bit of weight in front of the cg for the seg to move... Of course, a segway at speed is not level. I suppose if you consider vertical in this case to be a line from the contact patch of the tires, thru the cg, and continuing up, than you could stay in line with the cg and still move, but that line would not be the same a vertical, it would lean forward a bit...

OK, with this "perpetual motion" argument, you're taking us into very detailed territory, because we're talking about small forces -- like wind resistance, and rolling resistance of the tires. We've already established that wind resistance requires some lean. The tires are a more complex picture, but a tiny amount of lean is necessary there to counter the torque the motors have to put out to overcome it. (So it's not so complex at that level, just when you try to work out the forces involved at the tire).

Again, tiny fractions of a degree. Very different from your statement that you lean a certain amount, and it picks up speed until your speed balances you out. You'd need a LOT of speed for wind and rolling resistance to to balance out even a 2 degree lean, and you can lean a lot more than that. Until it decides not to let you anymore.

BTW, it's only a perpetual motion machine if you actually get free energy out of the system. It's an *ideal* system, if you ignore small loses, and get continuous motion with no energy input when you don't lose any energy either.

There are no ideal systems, either! At least, none that you can observe, though at a subatomic level they can exist until observed.

But unlike perpetual motion machines, ideal systems are a useful simplification for analyzing the overall behavior of a system. Sooner or later, though, anyone with an engineering frame of mind is going to want to look at how the system departs from the ideal, just as you did.

But it helps to keep the thinking straight, and focus on the ideal first, and then add in additional factors, preferably one at a time.

So normally, we ignore wind resistance, rolling friction, and jello, until we've worked out the idealized behavior. And then we say, "but what about...?", and find out that "vertical" isn't quite "vertical" because there are other small torques and horizontal forces.

They still all add up to zero net force, and zero net torque, for constant velocity and angular velocity. You just have a few more terms in the equation.

[Gihgehls]

If you have a gen1 machine this is really easy to demonstrate. First you have to improvise a counter balance, like I have in the picture below. This removes operator response from the equation.



Now, if you walk the Segway to a level area, it should hold nearly still. It will actually tilt a bit and correct itself without moving forward because the tiny forces required at this stage aren't enough to overcome the rolling resistance of the tires. Now, if you push the Segway forward it will move forward for quite a ways before eventually stopping due to rolling resistance, but lets look at what is happening.

The Segway does its best to stay level. This is rule #1. It doesn't care about keeping a steady speed or holding still. Its main concern is staying level. I'll try to explain what is happening by breaking down the process into discrete steps, sort of how the CU boards work. (in actuality, the sampling rate is different than the motor command refresh, so this is only an estimation)

In the first tick of the clock, we push the handlebars forward. Segway feels the tilt.

In the second tick, the wheels move forward just enough to correct the tilt, bringing the platform level and moving the entire system forward.

Now the Segway (and case of waterbottles) has forward momentum and the wheels are stopped. So in the next tick, the momentum (instead of my hand) tilts the platform forward again. Segway sees this. (so, for a fraction of a second, the wheels were locked and the entire machine rocked forward, with the rolling resistance absorbing some of the energy.)

Fourth tick, the wheels are commanded again to move foward, but not as far as before. Rolling resistance has slowed the whole unit down in between ticks. The correction has been made and the plaform is level again. The wheels hold firm.

Next tick, forward momentum carries the whole mass forward again, pivoting again on the locked wheels, some energy again being absorbed into overcoming the tires.

This goes on until we are back at the first step, with the segway holding mostly still, making small corrections that don't take it anywhere.

[KSagal]

Nice.

I do not see how waterbottles on a pushed segway address if a person who rides a segway wants to keep it moving at 10 mph needs to lean forward or not...

But I do enjoy hearing you talk, too.

Have a nice day.

[Bob.Kerns]

Nice picture and demonstration.

It's not quite true that the Segway tries to stay level, though that's a good starting model. In fact, the faster you go, the more it tips back (speed limiter). Pushing the handlebars forward won't make it go, either, if you're standing on it. It'll end up making you lean back.

I've been trying to figure out a good way of describing it. The best I've come up with is that it tries to maintain a particular angle, depending on speed. That captures the immediate Segway part of it.

But when you add the human into the equation, it gets more subtle. Because as I said, if you push on the handlebars, it tips forward a bit, you tip back a bit. And then, it moves forward a bit, moving you back even more. And all you did was move forward a few inches and come to a stop, with a different weight distribution (more spread out), but still aligned with your CG over the contact patches.

Actually moving forward involves pushing back with your feet to rotate your weight forward. It's a lot like trying to describe how you enter a spin or a jump in figure skating, actually. In both cases, it's not that you just rotate. In order to rotate, you have to do a sequence of actions, with different moments of inertia. We do this so instinctively while walking or segging that it's really hard to think about what we do.

In fact, I'm not entirely sure exactly how we do it. I understand the space of possible ways, but what we specifically do is subtle. Walking is a LOT more complicated than it looks from a physics standpoint, too. In fact, it's more complicated than anything we've discussed here so far.

Or standing. Since my legs don't work well, I do most of my balancing with my upper body, which is how you learn to do it as a skater. But rotating around any axis is not a simple action. It's a sequence of actions that result in a change of position. You can't just create angular momentum.