Short-Link physics bugs

[Calastigro]

One thing i dont think you're factoring in... Cable is heavy!  we're talking huge steel cable here!  after a certain point, cable should snap under its own weight.  i'm gonna go find that point out now....

[edit]  AHA!  that point doesn't exist in Pontifex!  (either that or the maps simply aren't big enough...)

I've been trying to weigh the difference between a 4-length of heavy steel and 4 1-lengths, but my studies are hindered by the brittleness of 1-lengths.  go figure.

[/edit]

(Edited by Calastigro at 2:11 pm on Oct. 23, 2001)

[baggio]

See http://pontifex.mendelsohn.de/forum/sthang1.pxb and sthang2.pxb.

[Calastigro]

AHA!  I was right!  they scale up (or down) the model members in all directions!  hahaha!  

[OverClok]

This has been mentioned before, but not resolved (to my knowledge) - why is it that the strength of a beam/cable is proportional to its length? Such that a 1-unit beam/cable will support virtually no additional stress, but longer structural elements will be much much stronger...

For example, here are three Heavy Steel pillars (just to give a common footing) supporting the highest tower that won't collapse with 1-unit, 2-unit and 4-unit light steel beams:

As you can see, the 4-unit beams can have 9 beams without collapse (10 collapses after several seconds), 2-unit beams can only go up 5 beams, and 1-unit beams, wow... it can't even support anything beyond the weight of itself! This has got to be a physics bug of some sort (mat-c may have mentioned the problem here?) The same issue is apparent with cables - short ones can't support much at all

This is high on my priority list of things that should be fixed, I think...

[JohnK]

Maybe the rigidity (is that a word?) of the joints has something to do with it? The boxes are very strong & heavy.

[mendel]

Kudos to OverClok for actually measuring this effect! Great idea!

I used to think this was physically ok, because the short box diagonals are at more of an angle (45° for 1x1 as opposed to 14° for a 4x1 box). However, setting 100% as the bar breaking load, a 4-box can take 1176% against 1-box 965%, and even with the added relative weight of short boxes (15.3 vs. 12.2 per unit length), this does not account for the strength loss you've observed.

In real life, the longer a (solid) bar gets, the more likely it is to fail by buckling outward, so the compresive strength actually decreases with length. (If you pull, you have no buckling, so strength against tension remains constant).

[mendel]

I did more precise testing, using only one "short" box and stacking big boxes on top for height, adjusting the top box to the limit. http://pontifex.mendelsohn.de/forum/stest4.pxb / http://pontifex.mendelsohn.de/forum/stest4.gif (90kB) full size shot of the image at the top.
The two large sizes (7/8 HD units) crack because the link cube breaks, the boxes stay intact. The small (2 HD) box at the top of the 6 unit tower breaks, too - this is the opposite of a small object breaking at the bottom of a cable...
Looking at the heights of the towers for box sizes 2-6, the effect seems non-linear, possibly quadratic.
 
To remove this "feature" from Pontifex would probably mean breaking a lot of working bridge designs.

[falkon2]

This might be the result of some formula manipulation to make the game more balanced and/or more fun to play..

I like it the way it is, though. As it is now, you have to strike a balance between strength and budget for lateral supports. If the members for the short bars were as strong as the long ones, you can make lateral supports all 2 units and this would make it really easy to arrive at the cheapest possible solution (as opposed to constant tweaking currently needed to get the right balance)

edit: By lateral supports, I mean bars that connect two arches or pillars together so that their tension/compression is distributed. I.e. cables between a heavy steel arch and the deck

If the short pieces could withstand the same load as the long pieces, 4 2-unit long members instead of a 8-unit member would give you free joints (because the joints cost nothing). Thus, there wouldn't be any reason to build your members longer than 2 or 3 units.

(Edited by falkon2 at 5:44 am on Oct. 23, 2001)

[OverClok]

Switching over to cables:

I just put a bunch of cable lengths and as many heavy steel beams as it could support. Any cable 6-units or longer could support 6 beams (the longer cables were stretchier, but supported the load no better and no worse). But once you get below 6 units, things get much worse:
6-units = 6 Heavy Steel beams
5-units = 5 Heavy Steel beams
4-units = 4 Heavy Steel beams
3-units = 2 Heavy Steel beams
2-units = 1 Heavy Steel beam

The discussion of internal structural-elment angles and such shouldn't apply here, should it? A cable is, well, just that, there's no joints, angles, etc (or at least there shouldn't be). So what's the deal here?

In the case of cables, they don't snap in half, they break at the joint with the steel - the steel isn't any different no matter how long the cable, so it can't be the steel breaking, but why should the strength of the cable be dependent on its length?

[OverClok]

I just tried Light Steel tension ability - similar story:

There's no problems with the 4-unit beams, the 3-unit beams do a bit worse, 2-units a lot worse, and once again the 1-unit beam can only just about support its own weight! No news here, just showing it's the same for tension & compression.

Something unusual I did stumble across. Remember that objects-fall-at-equal-rates theory? Well, doesn't seem to work in Pontifex:

You can see in the first shot that the longer beams broke first and yet just before they hit the ground, you can see that the 1-unit beams are falling much faster than the 2-unit ones... (ignore the 3-unit beams, they've already partially crashed into the ground in the 2nd shot). Why is this? Shouldn't be air resistance (if that's even factored into the engine) since they all have the same horizontal surface area, and other than air resistance, all objects should fall at the same rate, right?

[mendel]

My tower level (see above) provides anchors at the top for cable experiments... :-)
Hanging light steel (tension) has the same strength as making towers (compression), I could hang as much weight from it as I could stack with the tower config.

The length of cable seems to affect the strength of the link box it attaches to.
If you're using a cable which has two joint = 3 section, and only the middle section is short, it will hold much more weight than it would if you attached a box to the short piece (or switched the short section down).
Of course, with an anchor at top, you don't need a top section really.
You can also use a cable joint to attach more cables, like an inverted Y. Even if they're the same length, the lower cables will tear first (from the box).

See http://pontifex.mendelsohn.de/forum/sthang1.pxb and sthang2.pxb.

[corrected URL]

(Edited by mendel at 9:08 pm on Oct. 23, 2001)