The online racing simulator
How to: Making Accurate Dyno Graphs for LFS!
1
(44 posts, started )
How to: Making Accurate Dyno Graphs for LFS!
Well guys, i have figured out how to accurately graph the power of the cars in LFS. Not sure if anyone else has done this yet, but here we go.

Here are the first two "Dyno" Graphs i have made for the XRG and XFG.

Theres an equation i found in one of my physics books that relates gforce to torque. I attempted to join this with another equation that related gear ratio, final drive, and weight to feet/second. Combining these equations, and with a bit of reading, i was able to come up with this equation....

(A) (Gforce)
(Gear Ratio x Final Drive Ratio x 0.85)

(B) (Driving Weight x Rolling Radius(ft))

Torque= A x B

Now lets apply that to making Dyno graphs for the cars in LFS....

Gforce: Press F9 in-game and you will see Lateral and Longitudenal G-force information displays under the steering meter. This is a key tool.
Gear Ratio x Final Drive: You can do these "dyno runs" in any gear, however it is easier to test in a higher gear so that the rpms do not jump up very quick. Take the Gear ratio of the gear you are in and multiply it by the final drive ratio.
0.85: Just a mathematical constant used in the equation.

Driving Weight: This is the weight of the car + driver + whatever fuel you start with. Make sure when you factor in this weight in the calculations, that you click on the Driver and Fuel tabs on the left in the garage to add those items to your race weight.
Rolling Radius: Knowing wheel specifications of each car is important. The rolling radius is half the diameter of the whole wheel (center of wheel to end of treds). (NOTE: Visit this website http://www.carbibles.com/tyre_bible.html and scroll half way down. There is a Rolling Radius calculater on that page, giving you the Rolling Radius in Millimeters. You then will need to convert this to inches, and then to feet. An easy way to do it: Take that RR number and put it in the "mm" box on this website: http://mg-jewelry.com/mmtoinches.html . This will give you the RR in FEET).

Now that you know the contents of the equation, lets go through the procedure.

First off, take out a peice of paper (unless you want to apply the data directly to excel), and make a table; 3 columns wide, left side RPM, middle is Torque, Right side is Horsepower. In the RPM column, start at 1000rpm and list down the line in 500rpm intervals until you reach the car's redline.

Secondly, for those of you without a clutch pedal, make sure you DISABLE autoclutch so that you can retreive torque readings below 2200rpm.

Lets start with the XFG. The XFG weighs 2231lbs with 1 driver and 5% fuel. The Rolling Radius of the wheels is equal to about 1 foot. As a control, i use the (Hard track) settings for every test. This means the gearing is pretty much standard. Open up Drag Strip.

The first test point will be 1000rpms. Accelerate up easy. 3rd gear is good for this car for this test. My 3rd gear ratio was 2.180 with a final drive ratio of 3.20. Get in to 3rd gear and slow down until the engine is revving at about 800rpm. Mash the gas and watch what the geforce meter reads as your engine hits 1000rpms. Do this a few times in order to get an average accurate reading. In my tests with the (hard track) gearing, 3rd gear at 1000rpm full throttle yeilded about 0.155g's. Lets plug this into our equation.

(A) (0.155)
(2.180 x 3.20 x 0.85)

(B) ( 2231 x 1 )

A x B = 59.59 ftlbs torque @ 1000rpm... illepall

Now, in order to convert to horsepower, all you do is the following....

HP = (Torque x RPM) / 5252.... So (59.59 x 1000)/ 5252 = 11.35 HP

So , as you begin cussing at your XFG for its pitty of an automobile, you mark this down on the table. Then you will do the same exact thing, but test at 1500rpm. Then 2000. and so on....

After you get your full table of data with torque and horsepower, open up Microsoft Excel. In the left column, starting at 1000, list the RPM's in 500rpm intervals going down until you reach the redline rpm (just like you did for your data table on a peice of paper). Then the next column list the Torque numbers corresponding. Then the Horsepower numbers in the next column. Then, click on the Chart Wizard icon. Set series 1 as horsepower and series 2 as torque. Assign the horsepower series with an X value of RPM and the Y value of your horsepower numbers. Do the same for the Torque Series and torque numbers. Then you will select a Line graph, the first one. It will plot your dyno graph. Add titles and labels as you wish.

And thats it guys!!! Keep in mind, these numbers are actually power to the drive wheels (so they will not necessarily match the power information in the Info tab in the garage). I think someone should be able to make a program that automatically calculates everything using the information i provided. Unfortunetly, i dont have those skills!!!

Enjoy!!!

-HK-TheStig
http://www.hkracers.com
youknow.... with this someone could make an app using outguage and outsim to do it automaticly
You didn't really need to explain all the maths, and you're going about this the really hard way.

I already did this back in S1 days using RAF outputs and 5 minutes of Excel work, gives a much smoother curve and takes 10 minutes per car to get ~1200 points for the curve. I was meaning to do this again for S2 (to include all cars and to check what has changed for the old cars) for my own purposes anyway.

This file was attached on RSC but seeing as that forum won't even load to let me link to it I'll post them again here.

Note: these were for S1, see further down this thread for S2 graphs.
Attached files
thanks for the info
#5
I could put this into practice for my idea...
Quote from Bob Smith :You didn't really need to explain all the maths, and you're going about this the really hard way.

I already did this back in S1 days using RAF outputs and 5 minutes of Excel work, gives a much smoother curve and takes 10 minutes per car to get ~1200 points for the curve. I was meaning to do this again for S2 (to include all cars and to check what has changed for the old cars) for my own purposes anyway.

This file was attached on RSC but seeing as that forum won't even load to let me link to it I'll post them again here:

Pretty cool. However......

First off i had to explain the math because not everyone knows what all of it means. Its makes it much easier to come out with the correct number using the equation rather than having the end user go searching for how to use the equation.

Secondly, It seems as though your graphs display engine output, and not torque/horsepower to the ground. This is alright, as some dynos strickly measure just the engine output leaving out transmission and drivetrain loss. However, my method allows to find out what each car is putting down the the ground because Geforce numbers are brought in to the equation, and geforce cannot be calculated without change in motion. What makes the car go? The engine makes the power which turns gears in the transmission which transmit that power in a ratio to the axles and then spins the wheels. After all that, it is the wheels that move your car... right?

The XFG does NOT put 115hp to the ground.... its more like 98-99hp. This is logical because you have a car that puts out 120hp, and then you factor in drivetrain loss; in the XFG's case, it is losing about 18-20% through the drivetrain and wheels. Just because car manufacturers market a car making 200hp does not mean it is putting 200hp to the ground.

And how is it difficult? Im sure you had to start out some way similar, making data tables and coming up with equations. I do not have the knowledge to create any problems to make this easier, as stated at the end of my initial post, so i post my initial experiment to allow for others to see the core of it. I then encouraged someone to make a program using these equations.

... I wasnt trying to steal your fame or anything, i was just making a post about what i found out. You came in here riding on a high horse and it doesnt come off to me as proper. Jump off, accept whats happened, and then maybe your posts will come off as proper.

-HK-TheStig
Yep, those graphs display engine output (in theory, they're not perfect either).

Since creating those graphs I've found out all (non AWD) cars in LFS have a fixed 85% drivetrain efficiency (or 15% loss), so it's easy to jump between power at the crank and power at the wheel. Crucially, it allows you to prove the graph is right by comparing to the peak power/torque figures in the LFS garage, and the shape of the graph remains unaffected.

Sorry if I seem to be on a high horse, I'm sure you enjoy doing stuff like this as much as I do. I just don't want you to waste your time and end up a bit disappointed, the whole thing does end up quite time consuming. You're way isn't really that difficult, but using RAF outputs is a breeze so it's just the relative difference.

Speaking of which I updated the graphs for S2, but I've ommited turbocharged cars this time round because it's such a pain in the ass working with that turbo lag. Also the graphs don't match up perfectly due to rolling resistance and angular momentum. Rolling resistance can be minimised by using maximum tyre pressures, angular moment by using a tall gear for low acceleration.
Attached files
-
(nisskid) DELETED by nisskid
Quote from nisskid :sorry, but since when was torque and power directly related? they are seperate measurements, i really dont see how u can equate horsepower from torque.

power = torque * revs

plus there is no way i can think of to measure power directly (any type of power)
Quote from nisskid :sorry, but since when was torque and power directly related? they are seperate measurements, i really dont see how u can equate horsepower from torque.

Oh My God.

Someone who likes cars enough to play car related games like LFS to the point of posting about these 'technical' things on a forum actually thinks that?

You are going to struggle a lot.

Power is the rate of doing work. Torque is the work. So power is how quickly you do torque, and thus are directly related.

Ever wondered why every single hp/lbf.ft dyno charts cross at 5252rpm? It's not because engine tuners make them like that, but because they HAVE to as a consequence of physics.
Quote from tristancliffe :Power is the rate of doing work. Torque is the work. So power is how quickly you do torque, and thus are directly related.

not quite ... torque isnt work ... it has the same unit as work but its not work
For the purposes of my analogy it's good enough. I know torque isn't work, which is why power /= torque * time, but torque * rpm.
Quote from tristancliffe :Oh My God.

Power is the rate of doing work. Torque is the work. So power is how quickly you do torque... and other geeky stuff

If talk is the work and power is the rate of talking then how much power do you have

Sry for off topic - I'm bored, tired and should go to bed...
Quote from tristancliffe :Ever wondered why every single hp/lbf.ft dyno charts cross at 5252rpm? It's not because engine tuners make them like that, but because they HAVE to as a consequence of physics.

No i haven't ever wondered that but now i do.
[noob question] I'm an engine noob really but all the different factors that effect hp and torque how is it always a constant? [/noob question]
#15 - JTbo
Curves in first post do look much more like real engine curves than Bob's even both surely are right, I feel that Bob is smoothing curves out too much so they do look artificial, like there would be just few datapoints and formula used to draw curve, bit like what you get from car commercials/brochures.

Real power curve

So I do prefer style from first post, it does give more information what rpm to use if it is truly accurate, imo.
you gotta hate how the Torque Just Dies after 6000rpms
Quote from Greboth :No i haven't ever wondered that but now i do.
[noob question] I'm an engine noob really but all the different factors that effect hp and torque how is it always a constant? [/noob question]

It's because the equation that links power and torque (using lbf.ft and hp) is

Power (hp) = [torque (lbf.ft) x rpm] / [5252]

5252 is derived from various pis, constants and conversions which aren't necessary to explain.

Power (kW) = [torque (Nm) x pi x rpm] / [30000]

Look, an extra pi that is incorporated into the 5252 of the previous example

Quote from lalathegreat :you gotta hate how the Torque Just Dies after 6000rpms

Looks like a diesel torque curve to me, so the engine will be limited to lower rpms, and thus have a broad flat torque curve with a harsh drop off.
#18 - JTbo
Quote from tristancliffe :
Looks like a diesel torque curve to me, so the engine will be limited to lower rpms, and thus have a broad flat torque curve with a harsh drop off.

Those LFS curves looks bit funny, like engine would been revved well past peak power, what is rarely seen in real dyno charts, anyway LFS XRG's engine is completely out of steam at 6k, small valves and mild cam, not much more sport in it than in my trackt... ehm.. car.

From cam in my engine boys say that there is no low end torque at all, well I think different.
Engine should rev to 7000rpm, but had little problem with rpm signal that wen't crazy after 5000.

That is 2ltr 8valve NA engine, here is curve from almost stock engine, no point to rev over 4k. This again is with original ignition that was pretty much shot, spark was weak and also timing was jumping all over, also it had limiter at bit over 6k.

I can get very smooth curves out too, but those would be more off, now I don't know if Bob's curves are more accurate or less, LFS can be very well generating just simple curves too and then Bob's curves naturally are more accurate.
The difference between "my" curves and the curves theycallmeebryan extracted, aside from greater accuracy due to method, is that Bryan manually took about 15 data points to create the curve, whereas I just used RAF outputs from driving and got around 2000. The LFS torque curves are just a polynomial equation so will be perfectly smooth. The graphs are created in Excel as scattergraphs and joined up with straight lines, there is no smooth what so ever.
#20 - JTbo
Quote from Bob Smith :The difference between "my" curves and the curves theycallmeebryan extracted, aside from greater accuracy due to method, is that Bryan manually took about 15 data points to create the curve, whereas I just used RAF outputs from driving and got around 2000. The LFS torque curves are just a polynomial equation so will be perfectly smooth. The graphs are created in Excel as scattergraphs and joined up with straight lines, there is no smooth what so ever.

Yes, then LFS curves are still far from being realistic. Perhaps we see improvements to this in future as that is quite important for feel of engine.
Improvements to the engine simulation inside LFS have been mention by Scawen as being on the cards, usually when referring to sounds, as the more variables there are for the engine, the more detail and varied the sounds can be. Since patch V was compatible I am assuming the new sound system uses all the current variables to the max and that the next big step in sounds won't happen until more detailed engine simulation is performed.
Bob, what exactly are RAF values. Can you explain that a bit more?

I like my method. It can also be applied to real life if you have actual weight and know your gear ratios. Ofcourse, since we know that LFS has 85% efficiency rate to the wheels (15% drivetrain loss) in all cars, we can make a direct and accurate conclusion. However, there are many factors in real life that can change that. Using my method in real life would generally give you a good assumption. Oh, and there are many ways to make homemade g-force meters, for example using a tube with water and a marble inside....just search the internet.

I think it would be awesome if in the future of lfs they made it possible to be able to actually tune the engine. Adding different size exhaust manifolds/headers, selecting different camshaft profiles (Duration, Lobe seperation, overlap, etc), actually being able to adjust ignition and fuel injection timing. We all know that one engine setup isnt the best for every track, just like one camber and tire pressure setting isnt the best for every track. With alot of work, im sure it could be done. Ofcourse it all involves alot of very complicated mathematical formulas, but if the Devs really want to prove that they arent lazy, they could pull it off.

If I created this game and was part of the DEV team, i would see it as an oppurtunity to step over the mountain that has seemed to hault many racing sims of this day.
The main issue with your way is that you're measuring g-force, which comes from the resultant force of engine output vs drag. So all aero drag, tyre resistance and moments of inertia all reduce the g-reading, and vary with speed, thus greatly reducing the accuracy of your graphs. With RAF outputs you can get the forces direct on the tyres, which excludes aero drag, which really helps high speed testing (which is somewhat inevitable at high revs if you want to reduce the effect of angular momentum).
#25
Quote from JTbo :Yes, then LFS curves are still far from being realistic. Perhaps we see improvements to this in future as that is quite important for feel of engine.

I've seen much smoother curves even in full race engines than any of the RL ones posted here, it will of course vary between engines.

As for the low power after peak power on the RL curves remember that the drop off is caused by the lifting of the throttle, if you watch a graph on a rolling road it actually goes round in circles, but for the sake of neatness they chop that bit off the finished graph. In some cars there is a good amount of power after the peak but normally there's enough of torque available in the next gear to make it not worth the risk.
1

How to: Making Accurate Dyno Graphs for LFS!
(44 posts, started )

© 2002-2019 Live for Speed - Scawen Roberts, Eric Bailey, Victor van Vlaardingen
All timestamps on this website are displayed in your local timezone
FGED GREDG RDFGDR GSFDG