How do I get rover wheels to act like wheels?

[Krazy_Kerbal 2]

Seriously, I thought 1.2 was supposed to make wheels work, but it feels like a big step backwards. I figured out how to overcome the torque and SAS issues in previous releases, but now I am having these darn issues again! And with a very simple 4-wheel rover on the Mun! If I can operate a rover on Minmus without flipping it, why the heck can I not get it to work on the Mun before it just spins out of control and flip.

My basic design disables steering and power in the rear wheels, and leaves the front two wheels as the control point for that. I again disabled torque and SAS, but that doesn't matter. The rover wheels just spin out and flip the probe. So what am I doing wrong? They're acting like frickin' sleds.

Is there a video or tutorial which teaches you how to adjust traction and when to override vs leave on auto? So frustrating.

Thank you for any pointers and videos you might know of. I watched a lot of Scott Manley's stuff, so if he covers this specific topic, please point me there.

[bewing 3,053]

.craft files are just plaintext files -- so you can just upload them to pastebin and that works fine.

Traction control applies the brakes to individual wheels when they start spinning with no resistance. Thus, traction control drastically lowers your torque, because it is applying the brakes all the time.

Friction effects are 98% side-to-side, and 2% along your direction of motion. So cranking up your friction to 5 has very little effect on torque. If you have a lot of friction ahead of your CoM, you will spin out. If you have low friction in front of the CoM, and lots of friction behind the CoM, then your craft will be very stable.

 

[bewing 3,053]

We're going to need a craft file, in order to give advice.

[Krazy_Kerbal 2]

A newb question first, and then a follow up as I continued to experiment and investigate the rotation matter. First, where can I upload my craft file? And then, to update this query, I discovered I attached my wheels incorrectly in radial positions because mirror was not working right. I knew the wheels were going to spin against each other, so I reversed the opposing side, etc. Interestingly, that still didn't work out, so I attached the wheels individually, and reversed the motor control on one of the front wheels, which suddenly made the thing work really well. Except it still tips over forward when braking.

The main things I seek clarification on though are traction control and friction control and how (if at all) they influence torque when it is running on SAS. If this has already been discussed, what are better keywords to use since I missed it. Thank you for your interest in this matter.

Edited October 22, 2016 by Krazy_Kerbal
rephrased question

[swjr-swis 2,937]

11 minutes ago, Krazy_Kerbal said:

First, where can I upload my craft file?

https://kerbalx.com/ is a pretty good site, and it just recently comes complete with a mod you can install that allows uploading/downloading/updating craft files right inside the game.

 

Not much help on your actual question... still trying to figure that out for myself.

[Ty Tan Tu 389]

Low center of gravity. I put the Probodobodyne RoveMate on the bottom of a M-2x2 Structural Panel and it seems to work pretty well for me. That is the only generic advise I have.

[Ty Tan Tu 389]

I should start this by admitting that I am still new at this and even then the OP is probably asking a more nuanced question than I am answering. But in case there is another newbie like me - here is a very simple rover design that workers for me on Mun.

 

The Probodobodyne RoveMate is on the bottom of a M-2x2 Structural Panel. A tiny docking port is added to the front so that the NAV ball will be horizontal. This makes it wide and with a low center of gravity.

There is a bit more weight in the front with the battery to keep it from doing wheelies, but I try to balance the weight a bit with my Surface Scanning Module on the back. It only works in light and even than only for a short time, but it does not tip over on Mun.

Edited: Actually it works for as long as you would want in light. Hills put a strain on the battery, but it chargers up very fast.

 

Edited October 22, 2016 by Ty Tan Tu

[Clipperride 327]

For larger vechiles, where the CoM is higher than you would like it to be, I've found some Place Anywhere 7 Linear RCS, or Vernier thrusters on the upper side can help. If your vechile starts bouncing, or looks like it's about to tip over in low gravity, you can use them to push the thing back down onto the ground.

[natsirt721 134]

I recently landed a 6t 4-wheel rover on the Mun with a SAS module for stability.  As far as I can tell, you want to set the traction control wayyyy down, if not zero and override the friction control lower to avoid flipping.  I used a traction control of 0 and a friction control of 0.6, but these will vary with the vehicle.  Lowering the friction control will allow the rear wheels to slide a bit before catching, which is what really causes the roll.  You're going to want to quicksave and play around with the tweaks until you find something that works for you.  

In my experience, using reaction wheels to mitigate rolling is the best way to design a rover.  Keep the CoM low and build your wheelbase wide, and you should be fine.  Don't try to turn faster that 25 m/s or so (on the Mun anyway) unless your SAS is crazy powerful. If your pilot can lock prograde, driving with SAS locked to prograde will help stability during turns.  You also get some air control when you launch off a cliff or something.

Something else to keep in mind is that you can set the friction control to 0 (I think this is a bug?), and the wheels become 100% frictionless.  Your rover is now a sled.  Have fun!

[Sharpy 1,977]

On 22.10.2016 at 6:34 PM, Ty Tan Tu said:

The Probodobodyne RoveMate is on the bottom of a M-2x2 Structural Panel. A tiny docking port is added to the front so that the NAV ball will be horizontal. This makes it wide and with a low center of gravity.

 

A VERY bad idea.

RoveMate is very fragile, and the rover WILL be hitting the ground after jumps pretty hard. You'll lose the probe core and the docking port in no time. Generally RoveMate is a terrible choice for pretty much anything. Flip that HECS prograde using the Rotate tool and you're much better off.

1. You really don't need the heavy structural panel. If the rover hits so hard that the body touches the ground, it's game over, the panel won't help - even if the bodywork survives, all the important soft pieces will be gone. So you may just as well go with wing segments or other lightweight structures. Use the spare weight for way more payload.

2. The front-facing port is a decent idea, but if it's not the root part (the rover is not attached by it to the rocket the moment it separates) it will drive you mad. Make sure you have the right front-facing orientation by default.

3. The lousy four tiny panels won't keep up with your battery. And if you run out of power while going downhill, you won't even be able to activate brakes. Give it more solar power, e.g. the large static panels from Asteroid Day.

4. Reaction wheel. Reaction wheel is 10 times more important than low center of mass. The poor HECS is not enough. Give it the small reaction wheel.

OTOH praiseworthy: almost all of the rover's systems are within the frame of the wheels. That would be a pretty bulletproof setup if not that low-hanging RoveMate.

5. Rebind drive controls to arrow keys (unbind camera controls from these; nobody uses them anyway) Set SAS to prograde. Use simultaneously A/S and arrow left/right for (gentle!) turns and you can easily go 40-50m/s. If you do tumble, let SAS point you prograde and use Q/W to set yourself horizontal, while in flight. The rover can take quite a bit of abuse if it lands on the wheels and has a good clearance below.

6. A tiny rocket engine and a tank (Ant, Vernor, monoprop RCS is fine too) will be immensely helpful to save you in case you make a mistake and, say, drive off a crater's edge at 40m/s.

[EpicSpaceTroll139 2,079]

13 hours ago, Sharpy said:

unbind camera controls from these; nobody uses them anyway

Ummm...  Where do you get that from?

[Sharpy 1,977]

8 hours ago, EpicSpaceTroll139 said:

Ummm...  Where do you get that from?

seriously, using the arrow keys vs dragging with RMB?

[EpicSpaceTroll139 2,079]

1 hour ago, Sharpy said:

seriously, using the arrow keys vs dragging with RMB?

Idk about other people, but I find it nice to have my cursor available (even while my camera is being panned) to use part menus and whatnot. Also the arrow keys can move the camera continuously, unlike the cursor, which eventually reaches the edge of the screen, and thus must be moved back, and meanwhile I might be missing my plane impacting the control tower. Doesn't mean I never use RMB drag method though.

Whatever, to each his own

[Plusck 476]

24 minutes ago, EpicSpaceTroll139 said:

Idk about other people, but I find it nice to have my cursor available (even while my camera is being panned) to use part menus and whatnot. Also the arrow keys can move the camera continuously, unlike the cursor, which eventually reaches the edge of the screen, and thus must be moved back, and meanwhile I might be missing my plane impacting the control tower. Doesn't mean I never use RMB drag method though.

Whatever, to each his own

I too use arrow keys almost exclusively...