Disc Physics

[dgdave]

Sure we do. Its a slow, very understable disc thrown with lotsa hyzer and height.

 

[dgdave]

There's a video of Barry S throwing this shot on the 05 or 06 MS DVD. It wasn't with height. It was down hill. Weird looking. Steve Dodge says "I wish I had a disc that did that. "

 

[JoshEpoo]

I don't think curve 3 really represents the first shot you mentioned. For that the disc is slowing down at the apex and spin is decaying (JHern's model is for a constant spin rate). If there is enough spin and the disc is nose down enough at it's peak, it may come out of a slight fade and start to turn again but that is from gravity accelerating it. That would give the line we're talking about but not follow the model.

The shot from Barry I don't understand unless it is the same principle as the first shot (i.e. it slowed down or stopped spinning and then sped up again). I've never seen an understable disc that, if you threw it hard enough, would turn, then fade, then turn, then fade again.

 

[Moseley]

I don't think curve 3 really represents the first shot you mentioned. For that the disc is slowing down at the apex and spin is decaying (JHern's model is for a constant spin rate). If there is enough spin and the disc is nose down enough at it's peak, it may come out of a slight fade and start to turn again but that is from gravity accelerating it. ...
I've never seen an understable disc that, if you threw it hard enough, would turn, then fade, then turn, then fade again.

Yeah, I don't think curve 3 makes sense either... curve 2 would be a really hard-to-control disc, meathook until you throw it a tiny bit too hard, then turn and burn. I have a hard time seeing any disc go much beyond curve 1, maybe JHern put this in there as a test for the rest of us? :?

Also, these curves represent different discs' behaviour with a flat throw, as I understand it?
What you can accomplish with oat, hyzer angles, help from the wind, elevation etc is another story

 

[JHern]

Oh i should have been more precise i wasn't speaking in the terms of physics i was thinking of practical throws in real life where an inch or a half degree of movement isn't usually gonna be the deciding factor to the disc getting to birdie putt range.

Right, I'm sure you understood, but wanted to use that moment to be explicitly clear in the discussion, for posterity's sake.

As long as we are talking theoretical issues that may not be that visible in the accuracy of a throw: How about the magnus force vs gravity or should i say lift vector change with varying steepnesses of the throw? Ain't gravity pulling the disc one way, lift pushing the disc in another and magnus force possibly introducing a third force momentum going in a different direction as the other two forces possibly in another direction than the spin is going in real life applications if there is wind? And the resultant effects on turn however minute that may be. Even if it is just for one second. And that time certainly varies in practical throws from disc to disc if OAT is present. Putters won't stabilize as quickly as hog drivers. In those cases accuracy in a real throw may go down enough to be significant to scoring.

Nobody believes that the Magnus force is important for discs, the force is already relatively weak, and it acts on the side of a disc, so there isn't much surface area on which to exert such a force. It is more important on spinning balls in flight, but you can see how the force scales with the size of the ball: baseballs don't move nearly as much when spinning as a spinning soccer ball will do (aka "bending" the ball). Compare the cross-section of a base ball with the cross section of a disc (viewed from the side), and you'll see that there isn't much going on. Blake told me once he thinks he saw a Magnus influenced flight, but it involved hyper-spin and even then the effect was very very slight (just a tad lateral motion before the hyper-spin dissipated). Also, the Magnus force will not cause the disc to turn/precess in flight.

OAT is a very important subject, indeed. I will return to that later.

As long as we are talking theoretical issues that may not be that visible in the accuracy of a throw:

How about the Coriolis effect? =)

Actually, I can easily include the Coriolis force, which will introduce a latitude dependence to the throw, also a difference if you throw E,N,S,W. However, it shouldn't be important for real disc throws, since it doesn't traverse a significant arc length of the Earth's surface.

Do you think the curves will be symmetrical like that for discs in practice?

The curves aren't actually symmetric in detail, most cases I've run have a smaller slope on the upward part of the flight and a steeper slope on the downward part. This asymmetry increases as the throw velocity and distance increase. I think this part is quite realistic.

I suppose there will be inconsistencies during the inertial phase you're talking about. The most relevant part would be the portion where precession approaches 0.

Actually, the inertial part of the flight should work very well, since it depends only on the disequilibrium between lift and gravity. Lift and gravity are probably the most reliable forcing terms in the governing equations, and I'm solving them to 4th order with a small time step, so there should be no problems at all. I've simulated air bounces, pushing through hyzer, etc., and it all works just as I thought it would do. I was previously deriving all this stuff by pencil and paper, and it is fun to see all of the phenomena pop out so clearly in the numerical model.

Curve 3 confuses me some since we don't see discs flying like that in real life.

Yeah, I don't think curve 3 makes sense either... curve 2 would be a really hard-to-control disc, meathook until you throw it a tiny bit too hard, then turn and burn. I have a hard time seeing any disc go much beyond curve 1, maybe JHern put this in there as a test for the rest of us? :?

Curve 3 is a counter-example, a bizarre case to show just how crazy a disc would have to behave in order for the precession to have multiple zero-turn crossings as the velocity increases, while still fading left in the limit of zero speed, and still eventually turning over at the very highest speeds (these limits should be robust). You can see that it would have to be like this disc, and it probably would not fly well even if you could get a mold to re-produce these kinds of torques (I can simulate that, too).

Also, these curves represent different discs' behaviour with a flat throw, as I understand it?
What you can accomplish with oat, hyzer angles, help from the wind, elevation etc is another story

The code accounts for hyzer angles, the wind, and elevation. I do plan to study OAT in more detail, as well as how a disc breaks in as it collect multiple impacts with objects...this is something that will come up later. For now, I'm enjoying all the knobs I can turn on this model.

Here is a booming hyzer flip throw (67 MPH release velocity). It starts from 6 degrees hyzer and subsequently turns over. (A flat release, keeping everything else the same, results in a turn and burn, just as it does in reality.)

 

[JHern]

There's a video of Barry S throwing this shot on the 05 or 06 MS DVD. It wasn't with height. It was down hill. Weird looking. Steve Dodge says "I wish I had a disc that did that. "

Any links?

I went to the field the other day with Avery Jenkins and Nate Doss to discuss disc physics, and throw plastic around. Derek Hastings was filming it. Soon we were discussing extreme flight behavior. Avery threw a flippy disc high and fast, and it flipped over faster than it came down, enough that it landed upside down (This could actually be a useful throw in disc golf, if you could get it going reliably). It should be fun to get a simulation of this kind of throw working.

 

[dgdave]

I'll try and find it. At least I'll video the video on my phone and post it up. I watched some, but haven't found it yet.

 

[Moseley]

The code accounts for hyzer angles, the wind, and elevation...

I can see how it would be useful to be able measure a discs properties and then predict how factors like initial hyzer angle, nose angle etc will influence precession across speeds, since those factors can be manipulated easier than spin rate. Could you show how the disc in curve 1 would be affected by changes in hyzer and nose angles?

As to changes in wind - I guess you mean changes in wind from the side then, rear/tail wind would just act like change in speed, right?

 

[JHern]

The code accounts for hyzer angles, the wind, and elevation...

I can see how it would be useful to be able measure a discs properties and then predict how factors like initial hyzer angle, nose angle etc will influence precession across speeds, since those factors can be manipulated easier than spin rate. Could you show how the disc in curve 1 would be affected by changes in hyzer and nose angles?

As to changes in wind - I guess you mean changes in wind from the side then, rear/tail wind would just act like change in speed, right?

Yeah, I'll get around to showing how the flight changes from a standard throw as different parameters are varied, such as hyzer angle, wind, spin rate, velocity, nose angle, etc.. Also how it varies for different values of the drag and lift coefficients, and the parameterization of the pitching moment. That amounts to a very large variation of parameters. However, if I could compare a simulation with an accurately measured and timed throw (e.g., using video from multiple angles), then I could narrow in on the parameters that best match the flight.

Note that everything I just described above was already done for an ultimate lid by a former graduate student at UC Davis, Sarah Hummel, for a masters thesis, and is posted online. It was a good proof of concept, although the parameterization of pitching moment was probably not correct.

With wind tunnel data, I could remove all these uncertainties because I could just measure the drag, lift, and pitching moment as a function of velocity and angle of attack, and thereby be able to predict exactly how the disc will fly in the calculations. That is to say, the models, backed by experimentally verified parameterizations, have predictive power for the actual flight of a disc.

 

[JHern]

OK, I've been moving forward again on this project. I have rigged up a web-based CGI script on my computer (running Apache web server). Basically, the user can select parameters from a form on a web page (speed, spin, hyzer, nose, launch, height, etc.), click a button "simulate flight," and the simulated flight path pops up on a results page. This is great fun to play with, actually, which is the whole point of making it in the first place. What I'd also like to do is establish several sets of flight parameters belonging to several generic molds (under-stable putter, stable putter, over-stable putter, then mids, then drivers, etc.), which can also be selected in the simulation.

Either way, I think this will be an interesting educational tool, and I look forward to getting it up on the web some place where it can go live. I've already been learning a lot from it, myself. One of the things that still amazes me is how sensitive the flight of a disc is with respect to small changes in the release orientation and speed of the throw. Throwing disc is difficult!!! One degree hyzer change can make a very big difference with some of the parameter sets. I think this just illustrates why throwing a disc with distance and accuracy is such a precision skill, perhaps more than the thrower even realizes.

 

[Apothecary]

dang, jhern...youz edumacated.

 

[JR]

Big time.

 

[jaboc83]

OK, I've been moving forward again on this project. I have rigged up a web-based CGI script on my computer (running Apache web server). Basically, the user can select parameters from a form on a web page (speed, spin, hyzer, nose, launch, height, etc.), click a button "simulate flight," and the simulated flight path pops up on a results page. This is great fun to play with, actually, which is the whole point of making it in the first place. What I'd also like to do is establish several sets of flight parameters belonging to several generic molds (under-stable putter, stable putter, over-stable putter, then mids, then drivers, etc.), which can also be selected in the simulation.

Either way, I think this will be an interesting educational tool, and I look forward to getting it up on the web some place where it can go live. I've already been learning a lot from it, myself. One of the things that still amazes me is how sensitive the flight of a disc is with respect to small changes in the release orientation and speed of the throw. Throwing disc is difficult!!! One degree hyzer change can make a very big difference with some of the parameter sets. I think this just illustrates why throwing a disc with distance and accuracy is such a precision skill, perhaps more than the thrower even realizes.

I would love to see this both as a disc fanatic, and as a software engineer. Don't suppose you want to open up that source eh?

 

[JHern]

I would love to see this both as a disc fanatic, and as a software engineer. Don't suppose you want to open up that source eh?

Not yet, though it is pretty straightforward, 4th-order Runge-Kutta integration of the equations of motion.

Playing around with the parameters, here is an example output from a big turn over flight, apexing at around 40 ft...

I'm going to work on getting this up on a site soon. I'll let you guys know when I do...

 

[Muddyboots]

Yeah, that would be a big turnover. What kind of parameters led to this flight? 275' of lateral movement at a decreasing speed and spin? Was there a tornado nearby? Just kidding, and curious.
KP

 

[JHern]

Yeah, that would be a big turnover. What kind of parameters led to this flight? 275' of lateral movement at a decreasing speed and spin? Was there a tornado nearby? Just kidding, and curious.
KP

No wind here, just a big arm. I don't recall the parameters.

Its more like 150' of right-left-right, when measured off the line from release to landing.

 

[JHern]

OK, so I've got this code to simulate disc flights. I'll put it online where everyone can play with it, sooner or later (when I find a server that I can get all the necessary stuff installed on...). It would be fun to put it on discgolfreview.com if that were possible.

In any case, I need some data, for anyone who is willing to pitch in on this effort. I need to calibrate the model for some standard molds, to get the correct lift and drag force coefficients. This could be done with a video camera, a tape measure, and a stop watch. I need to know how the speed at release, total distance travelled over flat ground until touch down (no wind), and total duration of flight in the air, are related for several stable molds. I would propose using Innova molds: Aviar, Roc, Teebird, Wraith. It would be best if the flights were as straight as possible, but a statistical sampling of a bunch of throws that are almost straight is also OK. Speed at release can be measured by filming from a direction orthogonal to the release point and direction of flight, putting a grid (on a painted board, or something) on the other side of the thrower (the board should be parallel to the throw direction) with measured increments clearly displayed. Then the speed can be calculated from the video (this requires good resolution at high speeds).

If anyone else has better ideas to obtain this data, I'm definitely all ears...

 

[JR]

And didn't the grid need height information as well and shouldn't it be spread out to a large distance? That has intimidated me from trying this and the fact, that the field i was planning to try this on is uneven and the surface changes. It is artificial turf with ground rubber pieces between the plastic fake grass slivers. The field is brushed with a tractor pulling the brush and the surface is always changing and uneven. So i can' get accurate elevation data there. After thinking of this and checking out several sand fields in the town i've noticed, that all of them have so large diameter stones in the mix, that the surface is uneven. I'm not taking cameras out in below freezing conditions on ice and ice is uneven too on many occasions. So i too await info on how this test could be performed with real world constraints.

My high speed camera does not have good resolution, 446x384 or something like that at 210 FPS.

 

[JHern]

And didn't the grid need height information as well and shouldn't it be spread out to a large distance?

No, this is a simpler scheme than before, the grid needs only be beside the thrower, and using this only to get the speed at release. If you have a radar gun, that could also work.

My high speed camera does not have good resolution, 446x384 or something like that at 210 FPS.

The high frame rate is more important than pixels, as it allows a small grid to be used. The only thing that is really important, is to film as far away as possible, and zoom in to the throw as much as possible, to compress depth and avoid geometric parallax or optical distortion. Ideally, the poster board will be exactly orthogonal to the direction you point the camera, which will in turn be orthogonal to the throw direction.

If you could also figure out a way to get the height at apex, and the nose angle (relative to level ground), it would be even better. Maybe even rough estimates would work.

 

[JR]

Apex height could be quesstimated by an object of known height. If one were to trust the bubble scale on the tripod the camera would be level. I would not trust my system. Exit speed can be calculated from the disc diameter, because my camera has fast enough shutter speed to eliminate visible blurr from the disc. Of course a 21.2 cm disc leaves measurement inaccuracies. Could you write a program to detect the nose angle of the disc if such image recognitions software isn't available? I would think that image editing softwares would have an angle meter. Paging graphics guys help please for suggestions on software, that would do this.

Could you give specs for the grid? I can't easily haul around large objects. The larger the grid, the larger the cost. There is an accuracy throwing goal on one course around here. Too bad it's aligned so that one would have to throw across two fairways. And the ground is a bog after all the rain. The ground won't carry and will be less slippery only after it freezes. That means that indoors throwing is needed and lighting is gonna induce flicker to the pictures and force too long shutter speeds, that can easily lead to blurred disc images. Indoors halls may very well have indoor soccer goals so measuring one could help in creating the grid.