LE torque curve seems strange now.

[zimluura]

when i did the LE.car file at first i tried to get really accurate data on the torque curve. it was a very slow car then. recently it got changed to have much more low-end torque. which makes it accelerate much faster. probably true-er to the real car. but it would be nice to be able to put accurate values in for the torque curve and have it drive well too.

with the slower curve i was able to get it to accelerate ok if i did neutral to get the revs up, and then shifted into first. but it still wasn't as good as it is now. does all the blame lie with the autoclutch? or is the rotational inertia too high as well? or was the previous torque curve just not at all correct (maybe i messed up some metric conversions)?

[NaN]

The torque curve is from fudej's settings a 1ZZ-FE. Your values are from a K-Series I guess? Maybe we should offer both. It looks like there are half dozen engine options for the S2.

[zimluura]

i thought i got them from a 2zz 7th gen celica and added a little at the horsepower peak. maybe i copied the hp curve instead of the torque one.

does vdrift use Newton * meters for torque?

[NaN]

Newton * meters for torque?

Yes.

I have to admit I haven't checked the numbers. For the 1ZZ-FE torque starts at about 130Nm, 1000pm and will max out at 165Nm, 4400rpm. www.spyderchat.com/1zzfe.pdf

Maybe it's time to create some engine config files.

[fudje]

The original values in there had a progression that didn't make much sense for a torque curve, and also continued beyond any – A little checking shows that if you had a power curve instead of a torque curve you might have got your figures from the QE2 (1050Nm at 500RPM!), in which case the interesting parts of the curve are missing

That said, the numbers I put in are far too optimistic on the low end. I only had figures at 2200, 4200, and 6400rpm to go on, the tech data NaN linked to gives a much different picture of what the torque curve should look like.

In any case, if you calculate Nm figures below about 80 or so with 300 <= rpm <= 6000 you should probably double check your maths.

[zimluura]

cool, i'll try to do a re-check soon, just 'cause i want to know what i did wrong. i was pondering writing a little gl app that would let me draw a torque curve and then do unit conversion automatically, then spill out a vdrift text block.

for engine config files there isn't any necessary code programming left to do, correct?
all we need to do is put an engine directory in the carparts folder, write a bunch of engine config files, and then include which-ever one we want in whatever car?

[NaN]

The only problem is engine position/rotation. We should separate them from engine description. Any ideas?

[zimluura]

hadn't thought of that.

maybe we should do

[engine]
load engine/2zz-fe.txt
position = x,y,z


worried about directory getting really sloppy (hundreds of engine models * a few different generations * a couple variations). maybe by region

load engine/jpn/2zz-fe.txt

[NaN]

Maybe more like:
[engine]
type = engine/jpn/2zz-fe
position = x,y,z

Or (override the default parameters):
[engine.2zz-fe]
position = x,y,z

[fudje]

Surely then you're have something like:

engine/jpn/* # 5,639,812 engines
engine/ita/* # one engine
engine/fra/* # half an engine
engine/usa/* # 4 engines, but they're all derived from the first one

You'd have to start splitting japan up into industrial complexes....

Anyway, can't you already do

include ../../carparts/engine/model

[engine]
position = x, y, z

?

And is rotation in any way interesting?

[zimluura]

maybe: if z = 90 then it assumes a transverse layout. so the mesh models of all engines will be longitudinal.

[fudje]

But the number of cars where you'll ever see the engine is so low that it's really quite pointless.

On a related (and amusing) note, did anyone yet notice that the texture slapped onto the back of T73 is a photo of the engine bay (specifically the carb and generator) of a post-1963 VW Beetle?

[zimluura]

i don't know about that. now we have the ability to model the hoods as separate meshes (which can fly off if they someday take damage) and could be rotated open just for bling.

i think i'll be trying to model doors this way as well. with the shadows as they are it will probably look pretty good.