My Adventures with Direct Steer w/ Torque

[passerBy]

The handlebars will assume proper position because GPB will dictate the position and a control loop will enforce the position.
You will put a torque on the bars to make them move and your torque will be fed to GPB that in turn will "allow" them to move.
The control loop in this setup is the analog of the control loop you have in DSA (where you input the angle and GPB computes the torque to apply to the virtual bars).

If you put it like that, that might work... provided, the controller has means to enforce the position. From what I can tell, most of the wheels (all of them?) are reactive. This is something not exactly suitable for a torque-based controller.
How about using a linear motor(s) pulling/pushing on an inverted "T" (the horizontal part is the bars), while the pivotal main axis is equipped with some friction braking mechanism also commanded by the FFB?

But if you insist on saying that all this is not about the torque I can't keep on discussing.

I think I said it's both?

You didn't get my point: when you use DST or DSA, I think the Direct Lean setting has absolutely no influence, it's bypassed.

I thought so too... Until I actually tried both 0 and 100% DL and started turning the wheel while standstill. The virtual bars were lagging behind the wheel more noticeably at DL 100.

It would be nice if PiBoSo could confirm, but in the meantime: http://forum.piboso.com/index.php?topic=379.msg26565;topicseen#msg26565

It's dated by 2015. Something might have changed since then. But I will double check tomorrow, just in case.

[passerBy]

heheh Garry's mod is not very realistic but it does follow basic principles to an extent but it is not the realism that is rewarding but to see your own creation in action which is why i love Garry's mod

Fair enough Still, give that sim demo a try. I think you'll be pleasantly surprised

[Become dust]

heheh Garry's mod is not very realistic but it does follow basic principles to an extent but it is not the realism that is rewarding but to see your own creation in action which is why i love Garry's mod

Fair enough Still, give that sim demo a try. I think you'll be pleasantly surprised

will do

[HornetMaX]

The handlebars will assume proper position because GPB will dictate the position and a control loop will enforce the position.
You will put a torque on the bars to make them move and your torque will be fed to GPB that in turn will "allow" them to move.
The control loop in this setup is the analog of the control loop you have in DSA (where you input the angle and GPB computes the torque to apply to the virtual bars).

If you put it like that, that might work... provided, the controller has means to enforce the position. From what I can tell, most of the wheels (all of them?) are reactive. This is something not exactly suitable for a torque-based controller.

That's why I said we have no suitable input device for DST. A usual driving wheel won't work fine with DST, but may eventually work with DSA.

[passerBy]

That's why I said we have no suitable input device for DST. A usual driving wheel won't work fine with DST, but may eventually work with DSA.

Yet you ignored my proposal Again, what we are currently lacking in FFB controllers in general is the simulated resistance to displacement done right. What does a G27 do to enforce soft-lock, for example? It checks whether the current displacement goes past the limits, then "opens up" the motor against your effort until the wheel is back within the limits, which, among other things, can lead to some overshooting past the mark. An awful design, if you ask me... Now imagine the same wheel having a disk on the axis... and some brake pads/calipers, ready to hug the disk. Now once the wheel is turned to the limits, the brakes engage, and we have a problem, Houston But at least now the wheel is fixed in place. If we also have torque sensors on the same wheel, they could help unlock it once the sensors are positive you are trying to turn the wheel in the opposite direction. This is the kind of resistance control I was talking about. Except, for the bars it should probably be metal on metal and engaging the shaft itself rather than a disk with pads.

Why linear motors for the actual feedback? I'd assume this would make for a cheaper "direct drive wheel". For handlebars we don't need a lot of rotational span, but we do need quicker response. Therefore, I feel this kind of a drive is more suitable. Maybe not as much a motor, as a couple of solenoids or something would do.

[Become dust]

I was wondering..... how advanced are the tire mechanics in this sim? I know type and pressure can be changed but what else is there to tires?

[HornetMaX]

That's why I said we have no suitable input device for DST. A usual driving wheel won't work fine with DST, but may eventually work with DSA.

Yet you ignored my proposal Again, what we are currently lacking in FFB controllers in general is the simulated resistance to displacement done right. What does a G27 do to enforce soft-lock, for example? It checks whether the current displacement goes past the limits, then "opens up" the motor against your effort until the wheel is back within the limits, which, among other things, can lead to some overshooting past the mark. An awful design, if you ask me... Now imagine the same wheel having a disk on the axis... and some brake pads/calipers, ready to hug the disk. Now once the wheel is turned to the limits, the brakes engage, and we have a problem, Houston But at least now the wheel is fixed in place. If we also have torque sensors on the same wheel, they could help unlock it once the sensors are positive you are trying to turn the wheel in the opposite direction. This is the kind of resistance control I was talking about. Except, for the bars it should probably be metal on metal and engaging the shaft itself rather than a disk with pads.

Why linear motors for the actual feedback? I'd assume this would make for a cheaper "direct drive wheel". For handlebars we don't need a lot of rotational span, but we do need quicker response. Therefore, I feel this kind of a drive is more suitable. Maybe not as much a motor, as a couple of solenoids or something would do.

I ignored it for a reason    but ad you asked for it, here it is (just my opinion of course): the idea of using linear motors for a naturally rotating device sounds bad to me.
Just like the one of a brake behind the wheel (good luck in calibrating the actual torque such a brake would give in any situation, including wear of the disc/pads).
Obviously you would still need a torque actuator (a motor) as the brake cannot generate a torque by itself.

Lock/soft-lock is not the issue here: I'm firstly concerned with what happens when you stay well inside the lock-to-lock range (which for race track bikes is essentially always the case).

I was wondering..... how advanced are the tire mechanics in this sim? I know type and pressure can be changed but what else is there to tires?

Tyres are described with a fairly accurate model used extensively in the industry.
There are a lot of parameters but they are only needed to model a tyre: in the garage you can only choose a tyre compound and set the pressure, just like in real life (OK, if we want to be picky, in real life you can pick the temperature of the tyre warmers, minor detail for me).

[Become dust]

That's why I said we have no suitable input device for DST. A usual driving wheel won't work fine with DST, but may eventually work with DSA.

Yet you ignored my proposal Again, what we are currently lacking in FFB controllers in general is the simulated resistance to displacement done right. What does a G27 do to enforce soft-lock, for example? It checks whether the current displacement goes past the limits, then "opens up" the motor against your effort until the wheel is back within the limits, which, among other things, can lead to some overshooting past the mark. An awful design, if you ask me... Now imagine the same wheel having a disk on the axis... and some brake pads/calipers, ready to hug the disk. Now once the wheel is turned to the limits, the brakes engage, and we have a problem, Houston But at least now the wheel is fixed in place. If we also have torque sensors on the same wheel, they could help unlock it once the sensors are positive you are trying to turn the wheel in the opposite direction. This is the kind of resistance control I was talking about. Except, for the bars it should probably be metal on metal and engaging the shaft itself rather than a disk with pads.

Why linear motors for the actual feedback? I'd assume this would make for a cheaper "direct drive wheel". For handlebars we don't need a lot of rotational span, but we do need quicker response. Therefore, I feel this kind of a drive is more suitable. Maybe not as much a motor, as a couple of solenoids or something would do.

I ignored it for a reason    but ad you asked for it, here it is (just my opinion of course): the idea of using linear motors for a naturally rotating device sounds bad to me.
Just like the one of a brake behind the wheel (good luck in calibrating the actual torque such a brake would give in any situation, including wear of the disc/pads).
Obviously you would still need a torque actuator (a motor) as the brake cannot generate a torque by itself.

Lock/soft-lock is not the issue here: I'm firstly concerned with what happens when you stay well inside the lock-to-lock range (which for race track bikes is essentially always the case).

I was wondering..... how advanced are the tire mechanics in this sim? I know type and pressure can be changed but what else is there to tires?

Tyres are described with a fairly accurate model used extensively in the industry.
There are a lot of parameters but they are only needed to model a tyre: in the garage you can only choose a tyre compound and set the pressure, just like in real life (OK, if we want to be picky, in real life you can pick the temperature of the tyre warmers, minor detail for me).

i see you mentioned temperature. what about temperature change on the track? also. i got a licence :3 YEY

[passerBy]

I ignored it for a reason    but ad you asked for it, here it is (just my opinion of course): the idea of using linear motors for a naturally rotating device sounds bad to me.
Just like the one of a brake behind the wheel (good luck in calibrating the actual torque such a brake would give in any situation, including wear of the disc/pads).
Obviously you would still need a torque actuator (a motor) as the brake cannot generate a torque by itself.

I just think it would make it much easier to achieve high torque and high turn rates (yes, they are the highest near the 0 angle position, but nothing that cannot be solved with a good controller also, the connecting rods don't have to be put on the same line), while retaining good angular resolution (translated to the armatures linear position).

Well, if it's two stainless steel tube halves engaging the shaft of the controller, using yet another electrical magnet, I don't think both the calibration and the wear present that much of an issue.

The bottom line is, I'm sure such a device can be manufactured more easily and much cheaper than something based on an industrial servo-motor.

Lock/soft-lock is not the issue here: I'm firstly concerned with what happens when you stay well inside the lock-to-lock range (which for race track bikes is essentially always the case).

Nevermind soft-lock Since this kind of a controller will be useful almost exclusively for bikes, we can make do with only hard stops. Sturdy enough hard stops, that is

By the way, Max. If you haven't tried to use your headtracker for rider lean yet, give it a try With the angular direct steering it's simply a game changer, but I expect it to help a lot even with the default steering.

i got a licence :3 YEY

Congrats! So, how do you like the difference between the 125 and some proper bikes?
Honestly, I have no idea how Piboso is expecting to sell much with that 125... It's just bad. I gave it a go for a bit, and I hated every second of it. The 250 in the alphas was much better.

[HornetMaX]

By the way, Max. If you haven't tried to use your headtracker for rider lean yet, give it a try With the angular direct steering it's simply a game changer, but I expect it to help a lot even with the default steering.

Lol, I just posted in the other thread that I can give this a try.

But to be honest, personally I have no major trouble with auto (or manual via joypad) rider lean: the auto fwd/back is pretty bad (but I think it can be improved) and the left/right doesn't bring a lot, at least with the standard steering. Once again, it's not manual rider lean that will make DST/DSA any easier.

[passerBy]

Lol, I just posted in the other thread that I can give this a try.

I just noticed that and answered

But to be honest, personally I have no major trouble with auto (or manual via joypad) rider lean: the auto fwd/back is pretty bad (but I think it can be improved) and the left/right doesn't bring a lot, at least with the standard steering. Once again, it's not manual rider lean that will make DST/DSA any easier.

Wrong The effect is huge. I'm not sure why, but it helps tons with the angular direct steering. It feels like I'm actually balancing the bike out myself. And that's even without any calibration or other adjustments.
I don't think it will help you that much though, since you have EDTracker... But I already shared a couple of ideas regarding this in the other thread

After trying the headleaning, I'm sure that having a separate thumbstick for controlling the rider lean helped Klax with his direct steering a lot.

Another thing. Direct Lean seem to not have any effect on the bars in the direct steering mode, after all. I seem to have some input lag in both cases.

[HornetMaX]

But to be honest, personally I have no major trouble with auto (or manual via joypad) rider lean: the auto fwd/back is pretty bad (but I think it can be improved) and the left/right doesn't bring a lot, at least with the standard steering. Once again, it's not manual rider lean that will make DST/DSA any easier.

Wrong The effect is huge. I'm not sure why, but it helps tons with the angular direct steering. It feels like I'm actually balancing the bike out myself. And that's even without any calibration or other adjustments.
I don't think it will help you that much though, since you have EDTracker... But I already shared a couple of ideas regarding this in the other thread

After trying the headleaning, I'm sure that having a separate thumbstick for controlling the rider lean helped Klax with his direct steering a lot.

But the point is: it shouldn't.
On a real bike you can easily run a lap on any track at tourist pace without any rider lean. I don't see why in GPB rider lean should be such a huge help.
It should make a difference only when you're really really pushing hard, while you seem to say that it makes a huge differenbce in terms of how easy it is top ride the bike at all.

Another thing. Direct Lean seem to not have any effect on the bars in the direct steering mode, after all. I seem to have some input lag in both cases.

Good, as expected
If I recall correctly, we used to have some weird behaviour with DST/DSA and steering with the bike standing still.

[passerBy]

DST is quite difficult but once you get the hand of it its pretty fun! My best laptime without crashing is 3min 2 seconds

Today I decided to give the demo configuration a try while using the headtracker to control the rider lean. Surprisingly enough, I did a better lap on the 125 than on the BES WSBK Yamaha: 2 minutes 3 seconds compared to 2:05.
I gave each around 4 or so laps and wasn't trying too hard, since a lap record wasn't a priority for me (and I'm not the biggest of fans of Phillip Island), but all in all I'd say the angular steering mode is more than adequate for riding.

That 125 cc bike is the worst thing I've even ridden, to be honest. Almost no feedback, no feel... it tries to capsize on every opportunity too. But with DS2 and headtracking rider lean it's still manageable. The biggest problem is that I don't see much of the bike in the 1st cam.

[passerBy]

But the point is: it shouldn't.
On a real bike you can easily run a lap on any track at tourist pace without any rider lean. I don't see why in GPB rider lean should be such a huge help.

At tourist pace (as I perceive it) I can ride on a typical track without crashing using DS2 (who knows what else there's to it besides the angle ) with no rider lean input pretty easily. The only things that can complicate this are sloped tight turns. "Victoria" has two particular corners that are pretty hard to negotiate though, but at a tourist pace they are not that hard.

It should make a difference only when you're really really pushing hard, while you seem to say that it makes a huge differenbce in terms of how easy it is top ride the bike at all.

Of course, a bit over 2 minutes is not what one could call "pushing hard" at Phillip Island, but leaning into corners still made a lot of difference for me. I attribute this mostly to that it feels more natural to ride this way. I hope you'll get a wheel soon, Max, and see for yourself. Except you don't have a translational tracker yet, but I guess you'll think of something.

If I recall correctly, we used to have some weird behaviour with DST/DSA and steering with the bike standing still.

Not sure. I only tried DS1 prior to the second mode was introduced, then picked GPB again after it hit beta 5 or so, I think. Maybe beta 4... So, maybe I missed something.
The old DS1 implementation used to have the bars drop to a side and it was impossible to keep them straight at a standstill (because you can do only as much using solely the torque input), which was extremely annoying and I asked Piboso if it was possible to make direct steering and default steering switching between one another with a press of a button, but if I remember correctly, he told me that it wasn't.
I heard that now this mode at least allows to have the bars lined up straight before getting going, which is definitely a plus If I would use DS1, which I don't

[Become dust]

I loove the bigger bikes more mass means i can really push the bike to its limits soon enough i will beat my laptime record! Funny enough i prefer the 500cc or the 1000cc but that might just be the fact thst im so used to a "slow bike" like the 125 and yes, it is practically crap, with dst or dsa i struggle just to keep the bike upright, i feel like it would be cool if we could create a system that self rights when controlls are in neutrual much like with the default steering, but that issue is not apparent with the bigger bikes, luckily. 125 as a demo bike? Its like if you attrmpted to create a forza 6 demo and the only car you can drive in is a fucking honda civic. One thing is for sure this sim/game has potential, the question is just how we can uterlize it. It does reminds me of this great game called Rawbots, but the devs stopped working on it and closed the website, ive bought a licence but because they've closed the website i can't download the game rawbots is a robot contruction game where you are in a sandbox and you can build anything you want, every mechanicle part is programmable and the way you program things is very easy to learn!!! Much like Garry's mod i guess, i once built a B2 stealth bomber in that game, so laggy xD