Algodoo

Well. I've always liked this kind of slow engine with a lot of piston travel, but at the same time... I've dynoed your engine, and this is probably the closest you can get to an electric engine The max torque of it would me MUCH higher (i'd bet around 35000) if you could get it to speeds lower than 300 RPM without it popping half of its pistons. Which gets me to my point: It'd be better if you used scripts to limit low speed torque and make a realistic, non-exploding curve (which is really easy if you know how to manipulate functions - i've baked that into my 5k engines). or made it fast, but that takes a lot of work.
in any case, you can use the new dyno3 if you want to tune it. i guess it'd be nice because you would be able to stall it somehow...

Beside that the scene is a 10. i'll rate r n

- What dyno you use and how you read the values?
- How you "force" widget to draw curve you want?
- This is my second (first I4 OHV in previous scene) combustion engine that have real size and weight and still can run good enough to be used in truck, but thanks for response.
- about curve...something like this or this?
- Some tips how can i get MOAR POWAH from it?
Theoretically:
- Longer rods > lower side pressure > less friction + better conversion of oscilating motion to circular motion
- Bigger flywheel > bigger lever > more torque
- More piston travel > more time to convert energy from particle to piston
- Bigger bore > bigger particles (but let´s keep it in real size...)
Correct me if i am wrong

About the dyno, it's literally the last scene that i uploaded, pretty straightforward. In this case you would have to program the RPM range from about 300 to 1000 so that the engine doesn't blow up...
You can tune the torque curve by making density of the spawned balls a function of RPMs. For example, since the spawner RPMs will be the same as the crank:
spawnCircle{(
randomStuff:=x
density:= clamp(3 * Scene.my.throttle* e.other.angVel, 0, 100)

for example in that case your torque would be limited until (3 * e.other.angVel) reaches 100, which would be about 33 rad/s. That would be a linear limit, so you might find the curve weird, if that's the case just try curves for example quadratic or root functions, maybe also log()...

I was thinking a curve like this for a normal, slow car engine, and for this kind of engine you should put the peak torque earlier, about less than 1/3 of the max RPMs, but that's as simple as using a higher multiplier on the script above (instead of 3 use 4 for example)

About constructive details:Well, that's wrong in Algodoo. Longer piston travel means more torque, but it also means EXPONENTIALLY less RPMs. Therefore the power generated by the engine is pretty low. but it's usually all about finding the exact point. usually it's square, maybe slightly less, but never undersquare.
Rod length isn't too important but you might as well keep it like that.
The size of the spawned balls should be no more than the cylinder's size if your engine is square or oversquare. Of course you may need bigger ones for undersquare like here, but for oversquare it just makes no sense and breaks the engine more often than not.

Also two tricks:
I see that you use braked axles for the pistons (i'd recommend constant 1e-30 but i guess 1 does the job), but there is also something more to basically make pistons popping to the sides impossible. Put adhesion -0.1 on the pistons, but nowhere else. This will allow you to use incredibly stupid ball densities, at least until the engine starts to levitate.

Also negative adhesion on the balls. It's insane. I've been making a 120 Hz engine and could get 250 HP out of a 20 kg V4. But I guess you should try it for other kind of engines instead. the trick is adding many V-shapes and horizontal cylinder lids overlapping (about 10 each) but making them very light, less than 10 density. also the balls should have max radius and low density, and very high (low) negative adhesion, like -7000. you control throttle with the adhesion, idle would be maybe 0

And another idea for the torque limiter would be this
https://www.desmos.com/ca … 3gfbrcfwlm
in this case you just have to put the function in clamp, and it will give 0 below 16 rad/s (eliminating the need for the speed limiter). also, just replace the X's with "(- e.other.angVel)" and you won't have to use any scene.my.

Thanks for response Linkage:
- great dyno, already figured it out
- scene.my.rpm works better tha independent e.other.angVel for each "spark" isn´t it?
Because when i use something that "load" value from another object scene starts to lag
- You forget to save your graph, you just made links to my graph again
(login with google/YT account)
- Rod length is now same as flywheel radius, will increasing help? Because now it look that in lowest point piston have great de-/accelartion
- What do you think about hinge using graph that adjust torque (=green output of graph) from rpm (=input X of graph), thus adjusting power of hinge (=red output of graph)
- Undersquare (Z/D<1) mean wide piston than move only a bit up and down, right?
- Particles 2-times bigger gives 5-times more torque, and engine was made for this, work pretty stable for me, pistons next to each other have different collision layer
(expect low RPM, but i will fix that)
- Braked axle with 1e-30 will freeze whole thing at higher frequency of sim (don´t know why), I found out that 1e-39 is max that will NOT freeze at 120Hz but i´m used to use 1e-10, for timing i use bendTarget to adjust ignition (actually neutral = 0)
- Ok, i will try that negative adhesion trick, but shouldn´t POSITIVE values stick keep piston inside?
- Layered pistons, negative adhesion, many V-shapes, everything light, got it...maybe!

-scene.my.rpm doesn't necessarily work better. but you might as well use it. The problem with it is that you are writing to it in every step. If you read from the other object you will only read from it for 1 step on each engine revolution.
-Oops. Well anyways I was expectiog something like the graph that my demo engine on the dyno scene outputs, but with more torque at lower RPM.
-Maybe a little bit more is better. But not excessively
-i don't get it, but if you mean braking the engine, then absolutely not. Pistons will pop everywhere... Just put that script that i told you up there on the other comment and dyno the engine
-Not only a bit, but just less than z/d. Try with different ratios starting from square.
-Hahahaha sure. Pretty stable... Try to dyno the engine with range [100,1000]
If you fix the ratio you will be able to use higher density (or lower density and negative adhesion) balls and you won't need higher radius circles
-whatever
-no. negative. Positive will stick the piston, you want the opposite. Negative adhesion makes the hitbox slightly bigger though, so be careful
Also, I think that you are lacking an incredibly important part. The cylinder lid. i'm not sure but it seems that you only used V shapes... If that's the case then you'll have to put about 5 one above the other that span the entire engine length. This is what gives adhesion engines the boatload of power
-Everything light up to a point. Keep the pistons around density 100-150. Don't layer the pistons, layer the CYLINDER LIDS AND V SHAPES. And those two can be about density 5 or 10

Do you have FB account? YT account? Or just mail? I think chat will be better for this...
Okay, let´s start typing numbers, because i´m a bit lost there
- 1. I use scene.my.rpm as output for display so updating every tick is good, but if you are loading onCollide event it still only happen once per revolution, scene.my.rpm or e.other.angVel so not make change in this, i think, that loading one number (variabile) is better than load 4 different numbers (on each spark), even without using any variabiles
- 2. Do you know you can use sliders to change that graph i made?
- 3. Not braking engine, look at my "math + hinge" scene
- 4. I just wasn´t if "slighty less" mean "a bit of under-square, but only a bit"
- 5. Stop laughing at my engine! I raised it like my own...algodoo child?
Yeah, i will use negative, of course. It just make no logical sense to me because adhesion is "sticky" so if cylinder walls are "sticky" then it should hold piston inside.
But if adhesion is negative, then it is like "restitution", but in both cases forces work agains each other, so it will cancel that side forces....i think
- 5.1 So top "cover" is actually part of engine? I didn´t know ybout that, i though that "house" shape is enough. Thanks from now on i will install 5-layered covers on top
- 6. But if i want more toqrue instead of rpm i can use layered pistons?
- 6.1 how much density of part will affect engine? Sometimes heavier is more stable, sometimes it can ruin engine...So keep only flywheel and rods (or use springs?) at high density?

Alright, that's a lot to talk about. If you have Hangouts we can add you to the algodoo chat (it's pnvv, faytree, fra32 and me) though I must warn you it's a total mess.

Just one thing - Don't ever try to use logic on adhesion engines. I believe that negative adhesion keeps the piston on the cylinder because the hitbox is bigger but doesn't bug everything, so the forces made to keep it inside may be huge. But positive adhesion doesn't do crap.

Also the piston works much better if it is light. You can use my Dyno2 (not the 3 yet) with an engine with ignition off and it'll tell you how much wasted power there is because of reciprocation and friction. It's usually huge for 5kg pistons but they are needed on non-adhesion anyways

hangouts.google.com ?
I never heard about it
Add me

You have to tell me your google+ username for that

Lol no, your google+ username. You need to make a google account and send me the google+ username, not an email

Jakub Pollák

how do you make it move???

Move what? Truck? Dashboard? Scales?
Contact me via FB, hangouts, discord my name is "Jakub Kubo" everywhere