Could a Gyroscopic inertial thruster ever work?

[KerikBalm]

Let me sumarize:

Hurr durr, I don't understand what I'm watching

video 5

My Hypotheses is that Newtonian Dynamics does not describe behaviour of the offset Gyroscope, as demonstrated by video 5. This has been dismissed as being due to a misunderstanding of the purpose of the experiment resulting in the heavy object being exchanged for the light one. Having done the experiment myself I know that this is not so.

Video 5 shows an offset gyro rotating about a plausible CG.

The purpose of the experiment is to show that it rotates about the CG, the only misunderstanding is on your end.

You claim to do the experiment, but offer no proof.

So I'll just claim to have done the experiment 1000x more than you, and generated data disproving your position

Edited March 31, 2014 by KerikBalm

[Z-Man]

All of those questions are theoretical. The problem is that it does not happen that way, there is no significant deviation from vertical and the gyro mass moves immediately, it is not accelerated by the string.

You mean like here? https://www.youtube.com/watch?v=8H98BgRzpOM

Well, it depends on the parameters. Check K^2's formula for the É << Ω case. The maximum deviation in your initial conditions is roughly dÉ/Ω. Plug in your own numbers and see what you get. In the video, É is a bit below 1/s, Ω for the ca. 1m string would be 3/s, which would give a maximum deviation of about 0.3 d. Looks about right.

And what the formula neglects, as stated, is the torque caused by z and r not being colinear. Well... such a torque is quickly swallowed up by a gyro, which reacts with an inclination change. You can observe it in that video. Watch how quickly the oscillation of the string dies down. That is not normal friction, that is the wheel eating up torque.

In the other realm of É >> Ω, the formula predicts z=-d or precession around the gyro's center of mass. I managed to make a video. We have a high, reasonably open stairwell at work where I was able to easily set up a very long gyro pendulum. I did not measure the length, but it were 4 storeys, so at least 10 m and Ω > 1. É for the toy gyro is about 3/s at the best of times. The gyro had a mass of 84g (yes, I used it as the weight previously. I already had it on a rope. Sue me.), mass of the string is about 1g. Video:

http://youtu.be/N2trsfBqxPA

As you can see, before it gives in to friction, the gyro always precesses pretty much about its own center of mass. Definitely not around the suspension point.

Difficulties:

Crappy gyroscope with too much friction, winds down too fast.

The string was more elastic than anticipated, causing the gyro to bounce up and down.

Crappy gyroscope. What can I say, it already was 5 Euros over the budget I usually allow for internet debates. And it was the only one the toy store had.

And incidentally, the É >> Ω is also pretty much equivalent to a gyroscope precessing on a very light tower you always demanded. So there.

[Brotoro]

My Hypotheses is that Newtonian Dynamics does not describe behaviour of the offset Gyroscope, as demonstrated by video 5. This has been dismissed as being due to a misunderstanding of the purpose of the experiment resulting in the heavy object being exchanged for the light one. Having done the experiment myself I know that this is not so.

The Gyroscope pendulum is another example where the analysis which is correct by the current paradigm does not predict the actual outcome.

To start right at the beginning with a spinning gyro suspended from two strings. One string is burnt through leaving the gyro suspended, point of suspension directly above the point of attachment, distance from there to the Centre of mass d. The gyro couple transfers the reaction to support the weight and the gyro remains horizontal.

The gyro precesses, which means that it turns about its own centre of gravity. It cannot displace itself. There must be an external force applied in direction of the displacement to accelerate the mass from rest. As the gyro is attached externally by the string, the displacing force must be provided by the string.

For the string to provide a horizontal component of force, it must deviate from vertical. This should be visible, initially the gyro stays still then gradually accelerates as the subtended angle increases.

It is only after the string has provided this acceleration that the gyro can move in an orbit.

There is a way to demonstrate this conclusively. At some point the ratio of d to L will be such that the horizontal force will be too small to displace the mass and the gyro will remain in position with the string circling around it. So, if L then is reduced, would the point of rotation move away from the Gyro?. Would it be a gradual process? Can L be fixed at a length such that the point of rotation would be at d/2? When does K^2's prediction become valid and z become determinate?

All of those questions are theoretical. The problem is that it does not happen that way, there is no significant deviation from vertical and the gyro mass moves immediately, it is not accelerated by the string.

Some thoughts rather than abuse would be different. Maybe make an effort to understand rather than chant the mantra, Momentus is stupid to think that known science could just possibly be wrong.

Momentus .

So... I took out my gyroscope. It is a pretty crappy gyroscope, because it's just a toy. And I tied a string to one end of the axis. Then I spun it up... held it with its axis horizontal and the string vertical...and let go of the other end.

What I saw (after repeated attempts for verification), is that when released, the gyroscope will precess and therefore rotate around its center of mass. This rotates the endpoint of the axis where the string is attached away from being directly below the support point...so the string is no longer vertical. This results in a horizontal force (provided by the string's tension). So the gyroscope started moving sideways (roughly perpendicular to the direction of its axis). It was also precessing, of course, so it continued to deflect the string, which continued to provide the horizontal push that sort of 'chased the gyroscope around the circle' making a larger deflection until things stabilized with the gyroscope circling the center point as it precessed, and the string providing the centripetal force needed.

My gyroscope doesn't run long enough or smoothly enough for this all to be pretty, but I saw nothing mysterious about what it was doing.

Edited March 31, 2014 by Brotoro
typos and clarification

[Dodgey]

For the last time video 5 does not demonstrate your point. Asserting that it does without providing evidence does nothing to reinforce your point. In fact it does heaps to damage your reputation and credibility. Once again you claim that you have done the expeirment yet refuse to provide evidence to support this. You have even gotten to the point of saying we are just chanting a mantra simply because we don't agree with you, without considering you might be wrong. I give you one last chance to present your evidence, not video 5 as you can not determine the center of mass with the supplied data, your evidence from the expeirment you conducted before I discount you as a troll. Make your case and prove it.

[K^2]

You mean like here? https://www.youtube.com/watch?v=8H98BgRzpOM

...

In the other realm of É >> Ω, the formula predicts z=-d or precession around the gyro's center of mass. I managed to make a video. We have a high, reasonably open stairwell at work where I was able to easily set up a very long gyro pendulum. I did not measure the length, but it were 4 storeys, so at least 10 m and Ω > 1. É for the toy gyro is about 3/s at the best of times. The gyro had a mass of 84g (yes, I used it as the weight previously. I already had it on a rope. Sue me.), mass of the string is about 1g. Video:

http://youtu.be/N2trsfBqxPA

That is awesome. The behavior in your video is something I honestly did not even think of as a possible mode until I derived it from the equations of motion. It's great to see it in action. And you clearly see that pendulum oscillation is still there as a separate mode.

That MIT video is also a great demonstration. Though, it's very hard to get a good estimate for displacement from it. I really want to try and get a setup for a cleaner run. But I do need to get a better gyro. I'm going to see if I can salvage something or even buy one.

In both cases, it does look like there is some interference between pendulum mode and gyro precession. I still want to know how destructive that actually is, and since I see no way to get a good analytic result, I think I'll go with simulation. If I get anything interesting out of it, I'll post some animations. I'll see if I can match initial conditions for these two videos as well.

[Momentus]

Cheap gyro, cannibalize a "power ball" and use the gyro inside. Some models also have a built in rev counter. http://www.ebay.co.uk/sch/sis.html?_kw=powerball+gyroscope+exercise+neon+pro

[Z-Man]

That's actually not a bad idea. I still have one, and as these gimmicks go, it's not getting much use Still, you'd need good bearings and a way to get it up to a decent rotation.

But I do need to get a better gyro. I'm going to see if I can salvage something or even buy one.

I hear ya. Pro tip: Find one that does not explode.

In both cases, it does look like there is some interference between pendulum mode and gyro precession.

At least in my case, a lot of it is probably due to the friction in the bearings, that causes the gyro axis the string is attached to to rotate a little. The force from the string would thus be applied off-axis. And, well, there are those horrible vibrations.

In the MIT experiment, the rope is sturdy and short and gets wound up. That causes a considerable amount of torque around the z axis counteracting the precession, which in turn pushes the wheel rotation axis downwards.

And yes, even ignoring those practically relevant disturbances, analytical progress is difficult. Maybe switching to a frame co-rotating with the gyro precession helps? Simulation is probably the best approach so one can see what you can try to prove analytically in the first place.

[Brotoro]

I didn't previously see Z-Man's videos (I was playing with my gyro and posting my comments when he posted them), but that was the behavior I was expecting to see. But my gyro was on a shorter string and ended up in a different mode where the gyro's attachment point was circling around below the suspension point with the gyro's axis pointing radially outward from the attachment point.

[Z-Man]

Yep, that's what I observed too with a shorter string, and it's what was predicted. I did not make videos of that because the stable circling radius changes a lot with the gyro's precession frequency and thus rotation speed, and the rate of change there was simply too big to get anything meaningful out of it.

[M Drive]

I've conducted the pendulum test!

I'm extremely tired/sleepy now, so just a short question. Is a machine like this able to "swing", like a human swings on a literal play ground swing?

[K^2]

Is a machine like this able to "swing", like a human swings on a literal play ground swing?

Yes, definitely. In fact, if you replace string with a rigid rod and remove gravity, a machine that can move around to any place in a circle can be built. This is why gravity is important here, and why only average deflection would show deviation from expected behavior.

[M Drive]

*Yawn* Finally awake again.

So as you've already guessed, I got the machine swinging. What I don't know is if the motion the machine does is actually useful towards a swing or if the increased swing speed (momentum I guess) is due to what I'm trying to achieve, gyroscopic propulsion.

http://www.youtube.com/watch?v=S3BaN2tWAPg

A "cycle" in my machine is really just that. Gyros in starting position > start rotating scaffold and precess them forward > stop/slow rotation and have them move back to their original position.

I did two things. Doing cycles continuously as I did previously, one after another in quick succession. I haven't analyzed the video to draw any conclusions yet. Still don't know what software to even use (need something that counts frames and lets me skip forward frame by frame).

However, the other thing I did was simply to start doing cycles to get it to swing gently, then only do cycles as the machine was swinging forward. This caused the swing to get bigger pretty fast, and eventually swing as much as ~1.5 meters from maximum back to maximum forward position, if not even more. At this point I could fit 2-3 cycles during the forward swing (while just letting it rest during the backward swing).

It seems like the motion you need to do to increase swing speed as a human is pretty complex, even if every kid knows how. At the bottom of the pendulum motion, the legs (center of gravity?) needs to be the furthest from the pivot. Not to mention the pendulum experiment I did had 4 ropes, instead of a swings' 2. But, if you say this is all irrelevant and that a shifting center of gravity will cause it to swing, then I'll accept that. I suppose I should do a "control experiment" with the gyroscopes off too.

[Sillychris]

The swingset is kind of anchored to the earth...

And the swinger is kind of anchored to the swingset...

[K^2]

M Drive, in the most general terms, what you need to do to be able to swing on a swing set is alter moment of inertia about the pivot in a cycle that matches frequency of oscillations. You might be able to do that with gyros using a simpler cycle than without. If that's what you are getting, it's still really cool, but doesn't get you anywhere closer to propulsion.

In order to start moving in free space, total momentum of the system must change. I hope you won't argue with that. Your idea is that we can alter momentum without applying an external force. Id est, momentum need not be perfectly conserved. That's not actually all together impossible, but it would imply broken symmetries which we ought to have observed with far, far more sensitive experiments that have been done already. But if you wish to verify your device's operation, ultimately, you still need to show a change of net momentum.

And that's where gravity and pendulum come in. Gravity is an external force, so it results in change of momentum. Indeed, if you drop something, it accelerates. The string on which the object is suspended, also is a force, and also results in momentum change. If the string is perfectly vertical, the two cancel each other out. Net momentum change of object is zero. If the suspended mass is swinging, then on average the momentum changes are canceled. But if there is an average deflection, there is a net momentum flow. So if your system is capable of maintaining average deflection, then we know that it would be able to accelerate in vacuum.

Swinging back and forward, in contrast, only tells you that your system has been able to build up angular momentum. Which as I've said, is already pretty cool when you can do it with a very simple system, but as you yourself point out, it's something little kids can do on a swing. So it's not groundbreaking. Nor does it contradict the well established principles of conservation of momentum and conservation of angular momentum. (Both are transferred via pivot and gravity, rather than created/destroyed.)

[M Drive]

First off I'll just let you have the videos I took yesterday. Just because.

https://mega.co.nz/#!hUAUCZTT!eReEZf5hYHyRoow0Fw4wMkTjzule6mXflUPQjhy6DwU

https://mega.co.nz/#!NZYHSRSK!IO9I_4r9svlZpjIKHAJbfdLdTTLbIcFsoJDDxwwm504

https://mega.co.nz/#!Ac41VQia!EA2LUUhzNJ9dC_K4lpxFVRtp9ChINJkNeQmz_eE6nuI

In video 9 I keep doing cycles over and over. Same at start of video 10, then after a while I start swinging it. And video 11 is self-explanatory. Watch out for my friends butt crack though.

I'm sort of busy right now, so a more proper response to the above one later. Thanks for your help though, K^2!

[N_las]

Hi M Drive,

I tock the video behind the first link (00009.mts) and tracked the laser point with the free program 'Tracker'. The video had approx. 2440 frames. The autotracker didn't work good, so i tracked it manually every 10th frame. In some frames the laser point wasn't visible. I then had a list with 233 x-positions of the point. Using statistics, the median is 12.0±8.6 pixels to the right of the line. So it is only 1.4 ÃÆ’ from zero. This means there is no evidence in the first video, that there is an average deflection. Weak evidence for a deflection would be, if the median was more than 3 ÃÆ’ from zero.

My methode isn't perfect, several things could create small errors.

- Framerate is to low to allow for a perfect tracking (instead of a point, there is a laser streak, so no accurate position, but maybe good enough)

- The camera was't completely static, i adjusted the zero x-position to be at the drawn line on the paper. But at the end of the video the line was at a different position.

- I could give a better result, if i would track the point at every frame, not every 10th frame. But that would be to much work.

- Perspective! A x-position value for a point at the far end of the paper is really a different x-position value for a point at the near end of the paper. But i ony measured pixels.

- I didn't watched the other videos

If this first video would have shown a deflection of more than 3 ÃÆ’, than it would be worth to try to be more thorough. But as it is I have more real life work to do.

Edited April 8, 2014 by N_las

[K^2]

That is a more interesting pattern of movements than I expected. I'll see if I can set up a small program to track that dot automatically to verify N_las' results.

[M Drive]

The first video (00009.mts) is flawed anyway. Not only do I manually stop the machine at least twice in it, it also seems the camera crapped out early on. That video should be much longer. I have scheduled more experiments at the hall I was in on Thursday though. I'll be more thorough then.

I appreciate what you did though, N_las. I'll give you something relatively short and easily (manually) analyzed next time. How about a video where the dot starts off static, then I do cycles for 10-15 seconds, then stop and just wait until the dot stops moving again?

"That is a more interesting pattern of movements than I expected"

I hope you're referring to the zoomed out 'swing' video (00011.mts). I mean, it doesn't seem to have any problems gaining swinging momentum. And what I aim to try on Thursday is only do cycles as it swings back. Logically, if it's acting like a human on a swing, then it should gain swing momentum that way too (unfortunately I didn't try, but hindsight is always 20/20). But if it doesn't work and just keeps going into standstill I think it's.... well it's at least cause for more curiosity.

I'm also going to do some "control tests" with the gyroscopes off and see if I can get it to swing in either direction.

Edited April 8, 2014 by M Drive

[N_las]

It looked at the first video again.

- I tracked every 5th frame

- took the perspective into account

- corrected for the small camera movement at the beginning.

- I always tracked the 'middle' of the laser-streak

I don't now the length of the paper, so I use '% paper-length' as unit of length.

Now the median of the dot position in x-direction is (3.6±0.7) %paperlength from the marked line. Thats a difference of 5 ÃÆ’ from zero. So it looks like the median position is clearly not on the marked line. It would be nice if someone could do another analysis, if the values are in the same ballpark.

3.6% the length of the paper isn't much, but one can't ignore that. A few questions MDrive:

- Is the drawn line on the paper perpendicular to the paper? Or did you just draw a roughly straight line?

- Is the drawn line the position the dot has at rest? It would be nice if the video started with the point at rest position with no movement, and would last until the dot is at rest poistion again with no movement. Or couldn't the tremble be stoped?

- You control the device through a cable. What was the deflection if you picked up your controls? Could you observe the dot moving from that, before you started the machine?

The deflection from my analysis does go to the right (from the cameras point of few). I don't no how exactly your machine is orientated, but if there is a deflection, in which direction would you expect it to be? Does a deflction to the right conform with your observations from the machine movement on the track?

EDIT: Ok, dosen't matter if the video is flawed.

Edited April 8, 2014 by N_las

[N_las]

The first video (00009.mts) is flawed anyway. Not only do I manually stop the machine at least twice in it, it also seems the camera crapped out early on. That video should be much longer. I have scheduled more experiments at the hall I was in on Thursday though. I'll be more thorough then.

I appreciate what you did though, N_las. I'll give you something relatively short and easily (manually) analyzed next time. How about a video where the dot starts off static, then I do cycles for 10-15 seconds, then stop and just wait until the dot stops moving again?

That would be great. The dot should start static, and should stop moving at the end of the video.

If you use paper with crossed lines on it (like for math class) than it is much easier to take the perspective into account.

The paper in the first video was folded and a little bend. It shouldn't have much of an effect, but try to avoid it if possible.

It is not so bad, that the dot leaves the paper from time to time, but the dot should never leave the camera picture.

Do another video, with the machine off, but hold your controls and stay in the same position. Now your friend should give the machine a bit of momentum, and film the movement of the dot as the machine moves just like a pendulum. From that we can do an analysis how strong the influence of the your control cable is.

If you can make the videos, i will do the analysis again. After that, i can send you the files and diagrams.

Edited April 8, 2014 by N_las

[K^2]

There are times in that video when machine isn't running. Average on that gives a more reliable neutral position. Albeit, at the cost of increasing the error.

[M Drive]

Like I said, the video 00009.mts is flawed and isn't really worth analyzing, unfortunately. I start the machine at 12 seconds in, only something goes wrong at 24 seconds in and I have to stop it. At 37 second in I start it again (never really waiting for the dot to stay still), and run it until 1:26, only to manually stop it again a few seconds later.

There's also the question of the center of gravity and how it affects the dot. If you look at 00011.mts you'll see how the machines CoG moves forward as I start doing cycles. However, because it is strung up by 4 strings, the strings will want to stay horizontal, eventually (as that's the lowest energy state). So when I activate the machine the gyros move forward, causing the "wagon" part to move back, causing the strings to... (pardon my bad english) go from vertical to a slight angle. This is corrected as the strings will "want to" remain vertical, sooooo... I don't know if that would affect the dots placement, meaning it would stay more to the right side of the line than the left, even if no propulsion occurred.

When a cycle is completed, the center of gravity shouldn't have shifted. But that doesn't mean the dot will stay stationary after a cycle, nor does the dot measure the center of gravity, as it's attached to the wagon part. It'll always need some time to come to a complete stop.

Physics is hard.

K^2: What do you think about doing cycles during it's back-swing versus it's forward swing? Doing them during the forward swing increases the swing momentum. What if doing them during the backwards swing doesn't increase swing momentum? What would you think?

Edited April 8, 2014 by M Drive

[J.Random]

Gyroscopic inertial drives don't work. Russians have tested it. We were stupid enough to launch an actual satellite with this thing on board.

[M Drive]

I'm back from performing new experiments. I'm hungry as hell, so I'll make it short. I was able to accelerate the machine in one direction only, the direction of travel (forward), which is the direction it is designed to go. This means it was easy to get it into a swinging motion, as long as you kept doing cycles as it swung forward.

This is the interesting part though. If I did cycles as it swung backwards the swinging would stop, decelerate. I don't think even a kid on a swing could do that.

I did several attempts at passing the pendulum experiment, but it was hard to notice with your naked eye if the dot stayed more on one side than the other.

I recorded everything, and here it is:

https://mega.co.nz/#!MQIhwbCJ!l_4nMFWCw-HHnLOF7SCqI9MWuFIIDqE3qRyPMWWIve8

https://mega.co.nz/#!cAYgXb5J!T_Q4cberuHMGOoN2-O2k8P0jf1bPZ-QaAM9uKXlLG9g

https://mega.co.nz/#!lcIDVShZ!_hNiVzMxwZHmOdXx6uTYSyL5kLd-aFqhrWbD4m7PEEI

https://mega.co.nz/#!JUJE1Y5Q!sw2z0oPuSluyh9h1vsDu0tmACkfy5ZQmKY1gt6IfPPs

https://mega.co.nz/#!QN5wRZxI!VKgp6kcluSbFD7tekyBuWFNFplyOEHanhqvLrI_h2W4

https://mega.co.nz/#!cQQTwZob!x6V4LIisdZqTIsX3_XteGfmEpawlLyf_ZE7QHiPXLQ0

https://mega.co.nz/#!4JZEgJLZ!iABc7watO2BWIYc2SDPM7dgVLgQnPPsbR2tZZOrSrQ8

https://mega.co.nz/#!lZpGnSgD!AJeYBnxdkmo7-aqIJL5rmiZGKP_7CFvZru0l2RaNNdI

https://mega.co.nz/#!gV5FBCJT!S0xAC9A_3Ww2pFGQ8gG0pxP80OO5hIpltclIuZL8XBA

https://mega.co.nz/#!8MgmARyT!zvhwcWf-rJrghQcIwPhWBim2brpSBzbE5lfqqmJHV1Q

(The missing videos are errors in recording, they were only a few seconds long.)

[N_las]

A kid can stop the movement of the swing by shifting its center of mass, the same way it can increase the swing movement. But maybe thats not what you mean. I take a look at the videos. Can you do a quick overview which video contains what?