Jump to content

[Plugin/Parts] Kerbal Foundries - Continuation [Latest: 1.9g]


Recommended Posts

I'm not sure this will work; I'm running w3m on my steam-powered 2001 laptop in a pub in Whitby. I was quite surprised I could _read_ the new forums, let alone post.

I am not so sure I believe in regenerative braking. Joules at least fall out of the sky onto our solar panels; the extra mass needed for regenerative braking equipment has to arrive on rockets. Mind you, many of the KF parts don't exactly look like they were designed with gram-shaving in mind.

A part specified as "max torque | max power" wouldn't have the same problem with feeble reverse torque at high speed. At present some of the parts do because the floatcurve has dropped torque to near zero; but a part specified as I suggest would always deliver full power at high speeds. (The torque would drop off proportionately to speed but never wither away to nothing.)

On the issue of maximum speeds, I think partly that small wheels _should_ have low speeds. Railway wheels can be small because they travel on perfectly smooth surfaces, and even motor car wheels travel on relatively smooth ones. Off-road vehicles tend to have large wheels for a reason (and likewise the ISO 622 mountain bike wheel has largely supplanted the ISO 559 wheel simply because with a larger wheel less needs to be done with suspension).

That so,

Well, it almost worked.

That said, it's basically down to gearing. Electric vehicles don't have conventional gearboxes because they naturally have good performance over a wide range of speeds (rather like a steam locomotive where at higher speeds there are more valve events & the exhaust is a draught on the fire so more steam is generated when more steam is being used.

Part of the gearing is "baked in" by the wheel size. If I take an electric motor and double the size of the wheel it is attached to, it produces half as much shove at ground level (so half as much acceleration) but can go twice as fast (ground speed) before hitting that rev limit.

However, the rev limits of electric motors are very high (eg on a Tesla the motor goes up to 16000 rpm) and there's usually a fixed gearing stage to reduce the revs and increase the torque. That might be configurable in the VAB.

The point of this rambling story is that - with sensible defaults and maxima - the user should be able to adjust torque and rev limit in the VAB in a way that (torque * rev limit) would remain constant. Across the range of parts, (torque * rev limit) would probably be proportional to mass of the part. The maximum value for rev limit would be about the same for all parts (maybe higher for parts whose text says they are "fast" like the long tracks, probably higher in general for wheels); you can make the small parts go reasonably quickly, but not as fast as the big parts, and by doing so you put up with lower torque. (A half-track could then be powered by large tracks and steered by small wheels without the small wheels spending all their time revved out).

MechJeb implements a feature where, with Stability Control, the brakes are disabled if less than configurable% of the wheels are on the ground. This largely eliminates nose-wheelies (although beware; if you have just driven a long rover halfway over a cliff it may catch you out). Likewise, although I think this is a thing to do in MechJeb not KF, it would be possible to disable drive torque if a configurable proportion of the wheels aren't on the ground, which would reduce the tendency to do wheelies when starting.

(I also think Stability Control's main feature of keeping the rover body level to the ground could use a configurable option to only do that when the wheels aren't _on_ the ground - on high-g worlds Stability Control otherwise burns a lot of EC for no reason).

Link to post
Share on other sites

You wouldn't need any special equipment in the tracks, though, to at least take advantage of the effect to brake. All you need is a speed controller that shunts the motor contacts from the EC source to resistor networks and you get the braking side. You waste the energy, admittedly, but still stop no prob getting the braking effect I'm calling for. There's more than enough room in the long tracks, at least, to hide said resistor networks. Could also be done with a seperate part, I s'pose. Have the plugin scan the craft for the resistor network part on unpack, if detected, enable the electric braking. If not detected, no electric braking. Or, we could have resistive networks the default state and have a seperate part that, if detected, would dump power back into the batteries first rather than just waste it as heat?

 

It'd be nice to, rather than waste it, dump the EC back into the batteries. But, I 'spose we'd need those resistor networks anyway, for realism's sake. Batteries have a fixed amount of charge they can take and it's not exactly realistic to exploit how KSP doesn't seem to factor that in when applying EC to a fully charged battery. You wouldn't want to dump what, in all likelihood, is several hundred kilowatts per second(You do the math on a ~15 ton flatbed truck trying to whoa down from about 50MPH I don't understand those sorts of equations :P) back into an already fully charged battery after all.

Edited by Kenobi McCormick
Link to post
Share on other sites

At the moment, all of this is a moot point. Well, while I *can* get tracks to (sort of) work, they're about as stable as a stack of ping pong balls. Launch..... fine. Maybe one side will collapse and the craft will go flying sideways. Maybe it'll flip over! Maybe it'll drive fine, the suspension will collapse, get driven sideways by some phantom force, THEN flip over. I've no idea what Squad have managed, but things that work absolutely fine in vanilla U5 simply don't in KSP. Writing this one down to "not gonna happen" for the moment. Other options are... drastic, and probably not practical, but have yet to be explored.... Wheels are a maybe, but can probably work with the stock module fine anyway.

What you want is a wing, not a repulsor, MokStar :wink: Been discussed many times, and not possible I'm afraid.

Link to post
Share on other sites

In real units that's a 15 tonne [1] vehicle going at 22 m/s, and let us say it's descending a 15% slope and just trying not to speed up. In earth gravity it's hence getting 500kW from gravity. If by a misfortune it's on an airless body with earthlike gravity (I'm looking at you, Tylo) air resistance won't help, so all of that (whether consumed by rolling resistance or intentional braking) will become heat. This is, well, a lot, and McCormick's figure above was spot on enough that I suspect it was calculated. :-)

I'm all in favour of being able to brake by reverse motor torque; I just think (I think, in agreement with the above) that the extra equipment to squeeze the energy back in the batteries would be difficult to supply. Perhaps as a happy medium reverse torque (reverse to the direction of travel, that is) should consume much less charge than forward.

That said, 500kW is a lot. I tend to feel that parts that use EC should in general generate heat, but that's a giant can of worms (and a job for another plugin which would ModuleManager itself into everything)...

[1] ish, depending on which ton you meant.

lo-fi: would it be possible to use the stock Unity wheel module and bypass the KSP one altogether? I'm afraid from here I can't really be much help.

Link to post
Share on other sites
1 hour ago, lo-fi said:

At the moment, all of this is a moot point. Well, while I *can* get tracks to (sort of) work, they're about as stable as a stack of ping pong balls. Launch..... fine. Maybe one side will collapse and the craft will go flying sideways. Maybe it'll flip over! Maybe it'll drive fine, the suspension will collapse, get driven sideways by some phantom force, THEN flip over. I've no idea what Squad have managed, but things that work absolutely fine in vanilla U5 simply don't in KSP. Writing this one down to "not gonna happen" for the moment. Other options are... drastic, and probably not practical, but have yet to be explored.... Wheels are a maybe, but can probably work with the stock module fine anyway.

What you want is a wing, not a repulsor, MokStar :wink: Been discussed many times, and not possible I'm afraid.

Tracks still not working? I think I'm goong to cry. ;.;

Link to post
Share on other sites
1 hour ago, damerell said:

lo-fi: would it be possible to use the stock Unity wheel module and bypass the KSP one altogether? I'm afraid from here I can't really be much help.

That's what I've been doing :/ My other option is to try messing with the EVP classes which are all rolled into KSP now, but without paying for EVP, I've got no documentation and left looking at it thinking "What The.....?"

8 minutes ago, Eskandare said:

Tracks still not working? I think I'm goong to cry. ;.;

Believe me, it's almost brought me to tears too....

Link to post
Share on other sites
2 hours ago, lo-fi said:

What you want is a wing, not a repulsor, MokStar :wink: Been discussed many times, and not possible I'm afraid.

AH. Bummer.

What I was hoping to achieve was a floating base/city/refueling station. Not actually in orbit. Just hovering out over the ocean within the atmosphere.

Well, back to the drawing board. :P

Thanks for the quick reply.

Link to post
Share on other sites
14 minutes ago, MokStar said:

AH. Bummer.

What I was hoping to achieve was a floating base/city/refueling station. Not actually in orbit. Just hovering out over the ocean within the atmosphere.

Well, back to the drawing board. :P

Thanks for the quick reply.

TCS "hover hold", engines and unlimited fuel? Not easy but totally possible.

Link to post
Share on other sites
6 hours ago, lo-fi said:

At the moment, all of this is a moot point. Well, while I *can* get tracks to (sort of) work, they're about as stable as a stack of ping pong balls. Launch..... fine. Maybe one side will collapse and the craft will go flying sideways. Maybe it'll flip over! Maybe it'll drive fine, the suspension will collapse, get driven sideways by some phantom force, THEN flip over. I've no idea what Squad have managed, but things that work absolutely fine in vanilla U5 simply don't in KSP. Writing this one down to "not gonna happen" for the moment. Other options are... drastic, and probably not practical, but have yet to be explored.... Wheels are a maybe, but can probably work with the stock module fine anyway.

What you want is a wing, not a repulsor, MokStar :wink: Been discussed many times, and not possible I'm afraid.

I feel your pain lo-fi.

Try taking your U5 test cases, and adding a jointed rigid-body to the main RB that your wheels are attached to (all in the Unity editor).  Watch the fun/bouncing/jumping/random stuff that happens.  Something about the vanilla wheel colliders does not like having jointed rigidbodies.  Or at least that was my experience.  I could get them to work fine in the U5 editor with no other RB's jointed to it.  I can also get that same simple test part to work fine in KSP... as long as no other parts are attached (e.g. only has the single rigidbody).

I'm actually (very slowly) working on custom c# wheel-collider solution using simple raycasts and rigidbody.addForceXXX... but with my lack of experience on physics programming is likely going to take awhile to get implemented properly.  Works fairly well in the u5 editor so far, and the initial testing in KSP was promising; no flipping, spazzing out, or general weirdness; but that was only a test of the very basic suspension functions.... haven't even begun working on any kind of friction, torque, etc.  I suppose shoot me a PM or something if you are interested in any collaboration on this front (or even just interested using what I have when it is in a more finished state).

 

Link to post
Share on other sites

I do hope Lo-Fi is working on this, Squad's wheels suck even worse than they did BEFORE 1.1.  And no, this isn't a request or pestering for an update, I'm just saying that Lo-Fi is the only person who's ever created usable, functional wheels for KSP.  Ever.  Period, full stop.  Well, that's not quite true, there was another wheel mod way back pre-0.9, RollKage, that had some pretty good wheels at the time, but just as 1.1 has utterly borked Kerbal Foundries, so did whatever update it was that eventually killed the RollKage wheels.  So the worm turns, I guess.

But, in any event, hopefully Lo-Fi and his partners in crime can get 1.1 figured out and we'll eventually have some decent, functional wheels again. :)

Later!  :D 

Link to post
Share on other sites
9 hours ago, lo-fi said:

That's what I've been doing :/ My other option is to try messing with the EVP classes which are all rolled into KSP now, but without paying for EVP, I've got no documentation and left looking at it thinking "What The.....?"

Shoot, I'll buy you EVP if I have to...

Link to post
Share on other sites

Appreciate the kind words and support :)

I'm afraid there's no quick fix here, but rest assured that we're working on it, and quite literally going to reinvent the wheel again...

Link to post
Share on other sites

I'm so happy that you're back on this Lo-Fi, you and yours made one of the best mods imo. Stock KSP wheels just stink, but the tracks and wheels you and yours made were some of the most fun parts I've played with.

 

Thanks for the hard work and I wish you luck on figuring out the rest... even signed up just to let you know how thankful I am you're working on this :).

Link to post
Share on other sites
11 hours ago, Neutrinovore said:

 I'm just saying that Lo-Fi is the only person who's ever created usable, functional wheels for KSP.  Ever.  Period, full stop.  Well, that's not quite true, there was another wheel mod way back pre-0.9, RollKage, that had some pretty good wheels at the time, but just as 1.1 has utterly borked Kerbal Foundries, so did whatever update it was that eventually killed the RollKage wheels.

Think you forgot this blast from the past:

And, of course, the predecessor of KerbalFoundries:

 

But yeah, same as everybody else here im really hoping there is a workable fix for tracks in 1.1, i highly prefer them to wheels. But even old KerbalFoundries wheels were a huge improvement over stock.

Good luck and thank you for all your work, lo-fi and contributors.

 

 

Edited by Vrana
Link to post
Share on other sites

Thank you all, on behalf of all our many contributors! Kinda humbling knowing how many people are waiting eagerly on this...

The current plan involves creating our own wheel collider - one that works properly in KSP. Whether this is practical is open to debate, but @Shadowmage has made a start, and I have some ideas. Plenty of people have gone this route before in other games, having found that wheel colliders are simply not up to it. My only concern at this point is how we achieve it without overloading an already stressed physics engine. I'm afraid it won't be quick, but we'll have something quite special if we can pull off such an ambitious project... @damerell, I could probably use some help figuring out how best to apply forces, making a grip model and how best to calculate a few things, if you don't mind me picking your brain?

Link to post
Share on other sites

there is a placeholder physics material part module in KSP 1.1; interesting times ahead for wheels I expect. well beyond 1.1

To be fair, 1.1 wheels are not bad at all, aside from the clipping issue, which is more of Unity issue. just the stock settings aren't very good. 

Link to post
Share on other sites
1 hour ago, lo-fi said:

Thank you all, on behalf of all our many contributors! Kinda humbling knowing how many people are waiting eagerly on this...

The current plan involves creating our own wheel collider - one that works properly in KSP. Whether this is practical is open to debate, but @Shadowmage has made a start, and I have some ideas. Plenty of people have gone this route before in other games, having found that wheel colliders are simply not up to it. My only concern at this point is how we achieve it without overloading an already stressed physics engine. I'm afraid it won't be quick, but we'll have something quite special if we can pull off such an ambitious project... @damerell, I could probably use some help figuring out how best to apply forces, making a grip model and how best to calculate a few things, if you don't mind me picking your brain?

I have tested this physics engine. I put a full scale O'Neill Cylinder in orbit, but as a static. The engine seems robust enough even against a brief kraken attack from placing a masive colider in game. I've been lovingly pushing the engine and it seems to be playing well.y next approach is to make it a part and see what happens.

Link to post
Share on other sites
2 hours ago, Eskandare said:

I have tested this physics engine. I put a full scale O'Neill Cylinder in orbit, but as a static. The engine seems robust enough even against a brief kraken attack from placing a masive colider in game. I've been lovingly pushing the engine and it seems to be playing well.y next approach is to make it a part and see what happens.

Well that's... comforting!

Link to post
Share on other sites
14 hours ago, Vrana said:

Think you forgot this blast from the past:

And, of course, the predecessor of KerbalFoundries:

 

But yeah, same as everybody else here im really hoping there is a workable fix for tracks in 1.1, i highly prefer them to wheels. But even old KerbalFoundries wheels were a huge improvement over stock.

Good luck and thank you for all your work, lo-fi and contributors.

 

 

Nope, I didn't FORGET those, I include those in the list of wheels that never worked properly for me.  Of course, I guess that when they were NEW, TT's Multiwheels and the Caterpillar track mod would have been hot stuff, considering that before them there were basically NO wheels in KSP whatsoever, but I wasn't playing the game quite that early, lol.  I think I came in just about when version 0.20 was released, so by then KSP already had stock wheels.  Like I said above, they SUCKED, but at the time we didn't really have anything to compare them to, better or otherwise.  Later on, however, when I personally discovered the mods you mentioned, I didn't really feel like they worked as well as all of the hype surrounding them indicated, so I always found them pretty disappointing.

Anyway, yeah... I guess that's all I have to say about that.  :) 

Edited by Neutrinovore
Link to post
Share on other sites
Guest
This topic is now closed to further replies.
×
×
  • Create New...