Powertrain Feedback Collection

I'd settle for some kind of manual mode or even sport mode on the Gavril 6-speed. PRND21 shifting on a 6-speed makes the automatic Roamer RS nearly undriveable at race speeds; once you're above redline in second you have no control whatsoever over what the transmission does, which means that for any corner where the ideal speed falls between top of second and flat out, the stupid thing is going to rush straight for top gear on entry and then take its time downshifting on exit. (Like pretty much every 4-speed, but it has two more gears to be annoying with.)

 

The topic/issue to bring up is...
Turbo spooling in neutral in 0.28

This happens when...
Flooring the gas pedal in neutral in any turbo vehicle.

Instead it should...
It should do what it did in 0.27, initially spool in response to load but then drop off. Turbo's shouldn't build boost with no load. Every car with a turbo acts as if it has a trans brake, which is wrong.
Two step is a different story however. Some turbo cars retard timing and enrich the fuel in order to build exhaust pressure and spool the turbo.

 

Ok, I haven't played on the new update yet, so I am not sure how it actually physically handles this, but in a real car, you can totally build boost when just free reving on some engine setups. The difference is how the BOV (blow off valve) handles what is done with that boost once you hit redline and how the car handles that redline limit. Even if the engine is just bouncing off rev limiter, there is more than enough exhaust flow to build decent rpm's on the turbo if you have full airflow and minimal intake resistance. The BOV will then vent the extra boost to atmosphere to remove a lot of the resistance for the intake side of the turbo allowing the turbines to really get spinning.

So while true, its not "building boost", once the car takes off (if the car handles it correctly), you can have a lot of turbine RPM built up and the BOV will slam closed under load and allow a very quick full boost situation to happen. It won't be instant, but it will be quicker than just starting from a dead stop.

One thing I will add though is that the TYPE of rev limiter makes a big difference. Like you stated with two step, even a stock cars rev limiter will make a big difference in how much the turbo will spool. Obviously Two Step will get you the best bang for your buck as you are putting maximum heat energy into the exhaust to really get that turbine spooling.

Fuel cut rev limit will reduce the size of explosions in the engine leading to a fairly smooth rpm limit and can cause lean conditions. This massively reduces the energy going to the hot side of the turbo, and if the BOV remains shut (or it just doesn't have one... every car is different in this regard) then too much pressure will build up on the cold side and the RPM will not be sustained. If the BOV remains open, the flow through the hot side will be enough to keep it spinning if it already was (full load acceleration and then hit rev limiter) but it won't be enough to spin it up from a standstill to full RPM due to the lack of energy.

Ignition Cut rev limit isn't really super common in factory cars but I do think it exists on some. Regardless, when the ignition is not cut you have a full throttle full fuel cycle being dumped into the exhaust. Once ignition is cut you have a full throttle full fuel cycle, but no ignition... this can lead to the ignition taking place somewhere in the hot exhaust system, usually the cat or the turbo. Not great for an exhaust system and thus why factory cars don't usually use this method. Race cars are a different story but I will skip over that bit. Anyways, due to the added energy of the cycles that do ignite and the extra energy of the ones that don't but got lit by the turbo anyways, this can lead to full boost being made even without Two Step... but its much more unpredictable and has the potential to cause some real damage if the exhaust isn't up for the job.

Enter Ignition Retard rev limit. This is basically Two Step. It was already explained, so the simple is it ignites the fuel so late in the combustion cycle that the engine makes very little power (limiting the RPM) and the combustion keeps happening even after the exhaust valve is already open. This added energy is then given to the spooli bois and lots of boost is made. Pretty simple.

Throttle rev limit. This is really only a new thing on cars new enough to get an electric throttle body... which I guess isn't really that new... but regardless. You hit your rev limit, the computer automagically reduces the throttle to just hold an RPM... pretty simple. This is also how soft rev limits are done in cars now adays. They just change the number, and the computer cuts throttle when you hit that number. Super fine control over the RPM and you can attain a dead flat RPM at the rev limit. In terms of boost, while we aren't cutting anything else, technically you are off throttle and thus "off boost". The BOV will release the pressure in the intake and the reduction of throttle (and the reduction of air through the engine) is massively reduced to a grand total of 6 to 10hp worth so barely any throttle at all. This is nowhere near enough exhaust to properly spin a turbo, and thus no where near enough to build any real boost... unless its a very tiny turbo on a much larger engine... but then I have to ask... why you have that?

No rev limit... so... this would really only come into contention on engines that are old enough to not have a rev limiter, and probably have a custom home made turbo system... or maybe a Maserati Bi-Turbo might act like this... hard to say. But if you keep your foot planted in neutral, the engine will rev up. The faster it spins, the more air it pumps. The more air it pumps, the faster the turbos will spin. The faster the turbos spin, the more boost they will make. The more boost they make, the more air is shoved into the intake. The more air shoved into the intake, the more fuel the carb will add. The more fuel the carb adds, the more power the engine makes. The more power the engine makes, the faster it spins... you can see where this is going. End result? You either float the valves or you send a piston into orbit... but boost is definitely made unless they have an RPM BOV limiter that I am unaware of... but if there is no BOV... well... yeah...

But yeah... its not quite as cut and dry as one solution or another... there are definitely factors that need to be considered.

 

Yeah what happens currently is it will rev to a soft limiter at like 4K RPM and somehow manages to build 15 lbs of boost and hold it. What it seems like is that it’s a form of two step but the devs forgot to add anything that appears or sounds like two step, it’s just a soft rev limiter.

Edit: this is my experience with the new performance ETKs, I was unaware that this persisted between multiple vehicles.

 

Problem: Torque converter Automatics slip between gear changes when accelerating at a moderate pace in D. Happens with any motor that has a lot of torque.
Instead it should just go into the next gear without the rpm jumping

 

I believe this was addressed in the Updated Shiftlogic Transmissions mod back when it was still maintained. If I recall correctly, default_player's explanation of how it actually worked was that it varied based on when the transmission was manufactured.

Early automatic transmissions with only hydraulic control circuits would often leave the torque converter clutch disengaged by default. It would only come on once a certain gear was reached, usually overdrive. On occasion, some transmissions would leave the clutch engaged once it came on in overdrive even when downshifting to lower gears, only disengaging it once more when it reached second or first gear.

Modern computer-controlled transmissions like on the ETK 800 or Bruckell Bastion have different logic, usually controlled based on either the wheel speed, engine RPM, or both. Some transmissions engaged the clutch as soon as second gear, leaving it partially engaged and slipping until a higher gear. Most notably, and most relevantly to your post, the clutch would be left engaged during shifts, since automatic transmissions don't need to worry about their epicyclic geartrains grinding from shifts under load. This contrasts with the behavior you're observing, where the game is disengaging the clutch between shifts. Some heavy vehicles even lock it in first gear once the torque converter has reached the coupling phase to minimize power loss.

BeamNG in general has always modeled automatics inaccurately at low to medium speeds, though it's very accurate to real life when shifting aggressively at high speeds. I'd love this to be implemented some time, and I'm hopeful since the devs recently implemented proper regenerative braking and one-pedal driving in the winter update.

 

Just a little sidenote: Default Player is back on updating this awesome essential mod!
He announced it yesterday in the mods thread.

 

Very interesting that the powertrain team seem to be focusing on overhauling the manual transmissions when they're already excellent, whereas the automatic transmissions need the most work but haven't had attention for a long time.

 

Cant agree more: Automatic transmissions are downright unrealistic in game. no matter if its a torque converter auto or a DCT. They are all bad and in need of HEAVY rework.

 

AT in this game is not bad at all, keep in mind that most driving/racing games don't even have proper torque converter simulation or adaptive shift RPM(lower throttle, earlier upshift)

 

Honestly i dont share your opinion in that regard.
Torque converter lockup holding much too hard, most low shift rpm is not appropriate when minimal throttle input is given while driving, DCTs do not even have idle creep. These are all not really hard to fix points since your mod handles it with ease and you said for yourself that you are an amateur modder so i dont really see the reason to not include stuff like this vanilla.

 

Kinda wish the old clutch physics could be legacy supported as on controller it was more usable. On my sim-rig yeah the new physics are more realistic. But now clutch on controller requires so much thoughtful input that it has become far more difficult to performance drive on controller.-

 

Old autos hold low gears when you are braking, this is probably not realistic given that they are controlled using hydraulic valve bodies irrespective of brakes. Modern autos do hold the gears however.
This makes older cars 'OP' as you can just add 5% of brake through a corner to make sure the gearbox doesn't 'upset the car'.. in the real world this would likely be different.

 

It should be noted that idle creep exists purely because of the way a torque converter works with an always spinning combustion engine.

As such there is no technical reason for an DCT to have idle creep nor is there for EVs.

Some EV simulate idle creep because unlike a combustion engine, electric motors have basically no minimum RPM and as such can be operated at very low speeds without issues.

For a DCT - and in fact any clutch based transmission - to have idle creep in combination with a combustion engines is only possible with clutch slip - which will wear the clutch because that is what clutches do... transmit torque via friction and wear.

So while some real life DCTs might do that to placate people used to classic automatic transmissions, it really is a poor design choice because substantial usage will wear the clutch many times faster than one would otherwise expect.

 

I might consider making my own thermal system.. But the current one in BeamNG does not accurately simulate wax thermostats.
Wax thermostats gradually open, they do not instantly open. In BeamNG they open and close too fast, sometimes only over the span of about 2c.

If my car has an '87c thermostat' (normal for a german car) it might start to slowly open at around 70c (extremely slightly, like 0.1). Then at 87c be open at like 0.2-0.3. Then at 90c perhaps be open 0.6 - 0.7. Then at 94c be open 1.0.

ECT - Thermostat Co-efficient
<70c = 0.0
70c = 0.1
80c = 0.2
87c = 0.3
90c = 0.6
94c = 1.0

This gradual opening causes the coolant temp to fluctuate a bit whilst driving.. This is entirely normal. Most modern cars have gauges which lie to you (just stick at "90c" displayed even across a range of like 80-95c) Or some older cars have a gauge which shows a normal range.

Having gradually opening thermostats would be nice on old cars where we could simulate stuck thermostats (not fully opening / only fully opening too late) causing realistic overheating issues.
Hopefully this is a fairly easy thing to code/improve

This is by no means a recent issue, it's been an issue since the engine thermals were added many years ago lol, but thought i'd give some feedback on it.
Not sure if this is the right place to talk about it either, if it doesn't get traction here i'll maybe make a thread in Ideas&Suggestions.

 

then we could finally simulate my car irl (it simply acknowledges the fact it is overheating and goes "damn, someone should do something about that")

 

noticed theres provisions to slowly feed in nitrous to the desired start RPM, tried it in jbeam, shows working as intended on the torquecurve app but doesn't behave any differently to default, im guessing this isn't intended behaviour?

attached image of modified to suit jbeam & ingame driving clip



this is admittedly minor but would be pretty nice for making modified street cars smoother to drive and less blowwy-uppy

 

Having more than one configurable system of nitrous would be great for drag racing, instead of a 500 shot out of the hole it can start with 100 and be able to turn on another 100 after a half a second and so on.

 

All automatic transmissions will refuse to upshift when 1% brake or more is applied, will only upshift when you hit rev limiter so you can stay near redline indefinitely on a downwards slope for example.