Discussion in 'General Discussion' started by jual, Oct 15, 2018.
BeamNG Drive already has a destroyable gearbox. Maybe Not great implemented but already great
Ronaldo likes this
Ah, sorry, I think we were talking about different things. You seem to be describing something with FFB in it, ie it has variable resistance based on some sort of parameters? When you said load cell I just meant the method of reading input force, usually put on a pedal with a combination of spring and rubber resistance.
A clutch pedal that can simulate those failure modes would be cool, though probably on the lower end of the priority list. There are various pedals around with mechanical setups to mimic clutch pedal feel which should be enough for most people. The shifter itself is the big one, need FFB to stop you (dis)engaging gears when it shouldn't be possible. The old tranny F1 plugin (and Seven Smiles' new version, I assume, I haven't tried it personally) get halfway there but obviously can't stop you physically engaging the shifter which is a bit unnatural.
No matter what it would be nice if this got updated at some point. But hey, i tested the Radical the other day and the AI is still too fast, so...
All you need to simulate clutch failure is a solenoid that pulls out the return spring somehow I can think of many much more frequently-occurring failures that come a lot higher on my wish-list though!
Both tranny plugins prevent you selecting a gear, as you say that's the best you can do unless you have an FFB shifter.
Yeah, pretty much the only thing that impressed me in iRacing is how dogbox is simulated on Skippy. Really hope to see gearbox simulation a high pri in rF2.
rFactor already provides all but one variable needed for FFB Shifter. The added variable is the angular displacement of engagement between the engagement dogs between the side of the gear and the sleeve. That can be set externally.
The shifter would need an embedded processor to query ratios, speed, RPM and engine inertia along with a few other datapoints to calculate lever resistance.
Careful design of a feedback actuation could use a permanent magnet (PM) with a lower power requirement than a reluctance motor with the caveat that the PM would always place some drag from the neutral position, and stator laminations would need to be mighty thin to respond at the needed frequencies (saturation). With a PM one could get away with 50-75 watts of power. A reluctance (look up switched reluctance motor for more detail) would take about 200 watts without the saturation issue.
I've done a good bit of electric motor design, including true cogless BLDC drivers for precision positioning, and thats what my past learnings would consider in a system design.
As for functional design, when the gear and sleeve are out of synch it applies resistance as a return to neutral. What you'll feel on a mis shift is pretty much the same as on a real car, except that with enough force you can overcome the force and push the lever into the gate. However it won't allow the gear to engage and letting up it returns to neutral. True you could up the power to a point where it won't allow the lever reach the gate, but that would probably lead to breaking a few wrists.
Expect a production cost of about $1000.
Just my nickel's worth.
@John R Denman I wouldn't think processing in the shifter would be the right approach. The game should literally just provide a 'return to neutral' strength, and the shifter rotates to centre with that strength. (from above, the shaft the shifter is attached to would run laterally) So you just need a motor attached to that shaft able to rotate it one way or the other, a lot like a wheel. Some switches (or, in the future, progressive sensors of some type) to sense engagement in each gear.
The simplest and cheapest version of this would feel less like a real shifter because of the short lever and alternate arrangement for reverse (perhaps a simple 'collar'), but if it stops you engaging gears I'd consider it an improvement. More powerful motors could then be used for a longer lever and better feel. Cost wise it would have to be on a par with a mainstream wheel at most.
*Post-dead-battery-edit: At its most basic level I think the game would just have to provide that single output, again analogous to wheel output (-100 to +100 %), for argument's sake -100% would be a full return to neutral while +100% would be to stay in gear. Some fairly simple electronics on the shifter could then work out which way to power the motor from those (with a deadzone in the middle so it's not oscillating), the +100% would certainly be scaled down or it'll burn itself out, but it only really needs to be enough to hold the shifter in gear plus a little more for feel. Then you have the potential to apply a little pressure and feel when the clutch disengages.
A higher end version would need a bit more info from the game, and have a better mechanical setup, but I think even a basic version would add a lot.
Would it not be easier for a FFB shifter to block the gear until the game says the shift is successful rather than trying to bounce the shifter out? I can imagine it might not feel as authentic, but $1000 price tags are ridiculous for an H-shifter.
The prices are so relative...
If someone had told me five years ago that i would have now an OSW, Simulaje F1S and Heusinkveld pedals, i would have called him crazy.
I think it's up to the game to refuse the gear selection, and tell the shifter not to engage the gear (physical). Then it's up to the shifter whether it just keeps pushing the knob back to neutral, or it physically blocks engagement until it gets the all clear. But I think there's scope for a couple of different modes here based on shifter capability, even cheap FFB could vibrate when you don't properly engage, and if you force it into gear it can soon pop it back out when you let go. In fact in that case you want the shifter to say "I'm in 2nd!" so the game can say "whoaaa, that wasn't done right, I'm keeping this idiot in neutral, here's some vibration to try and wake him up". The game should be doing all the thinking.
A better class of shifter might be able to report partial engagement (think analog vs digital) so the game can try and give better feel (again, depending on what's being driven) as you engage and disengage gears, depending on what you're doing with the pedals and how the game says the knob should react.
I can see users mis-shifting and the plastic gearbox not opening up the slot for the shift lever to lodge into. Said user then exerts more pressure to force the shift lever into it's home and POOOF! PC Gearbox broken!
...and then they sell you a new shifter! You’re a marketing genius!
I have actually had my Microsoft Flight Stick, yes the old big clunker, hooked up & working as a gear shifter in rFactor 1, but without a gate to hold it in position . . . & no CNC machine in my garage to make one . . . outta luck, again, but it did work! I set it up as a 4 speed & only had to hold it in the corners of the joystick. the ffb would vibrate back to neutral to keep it out of gear when I didn't use the clutch properly. associated sound files made it pretty cool, but without the gate I was not able to race with it. spent the next couple of months trying to prototype something together out of coat hanger wire, wood strips, lots of duct tape, & little bits of Super Glue, without real success. technology makes us lazy . . . needs to be laser scanned & CNC made.
It comes down to how close to realistic does one want?
Personally I'm fine with paddle shifters, as anything less than what I described wouldn't be any better than a paddle shift. Others will feel differently and thats OK. My only goal was to describe how a realistic feeling shifter could be implemented.
Driving in R3E with a proper implemented clutch/H-shifter and then moving to rF2 without it, is a bit of a downer though.
HE Ultimate pedals / Fanatec CSW shifter (waiting for the HE H-shifter).
I do hope it gets added soon.
@Fanapryde In the meantime, have you tried https://forum.studio-397.com/index.php?threads/realistic-gearshift.58578/ ?
I'm just going to try this now myself.
I just wrestled through the topic...
Seems to be difficult to get it working as expected.
I can understand from reading the thread, but actually the hardest part for me was finding the right files to download on the GitHub site. Most of those posts you're seeing are from earlier versions, the 'Configurer' let me select the devices directly and assign the controls, no file editing necessary. Anyway, further discussion of it should probably be in there, but give it a go or ask questions in that thread if you're not sure, I think I've sorted it out
Separate names with a comma.