Maximum Tractive Effort Curve

[Ed Sclater]

Hi Guys,

I would like some help modeling our car for the straight line acceleration test please.

I would like to create a maximum tractive effort curve + other data with it, that makes it useful. This is so the FD ratio can be optimised.

Has anyone done this? Would anyone be willing to post an excel spreadsheet so i could get an understanding and produce my own?

Any help or advice is much appreciated.

Many Thanks

Ed Sclater

[Ed Sclater]

Hi Guys,

I would like some help modeling our car for the straight line acceleration test please.

I would like to create a maximum tractive effort curve + other data with it, that makes it useful. This is so the FD ratio can be optimised.

Has anyone done this? Would anyone be willing to post an excel spreadsheet so i could get an understanding and produce my own?

Any help or advice is much appreciated.

Many Thanks

Ed Sclater

[murpia]

I've done this, but as I'm a Design Judge I don't want to release my spreadsheet. The output graphs should appear in the presentation 'Develop to Win' on http://www.formulastudent.com soon, to give you an idea.

I will say that you will learn a lot more by producing you own spreadsheet than using someone elses. It's a straightforward problem. I will hint that you need the following data for your car as a minimum:

Torque curve, gear ratios, final drive ratio, mass, wheelbase, centre of gravity position and some form of tyre data.

Also see:
http://fsae.com/eve/forums/a/tpc/f/125607348/m/71810543821

Regards, Ian

[Denny Trimble]

Agreed, I made one in Excel with the vlookup function, and learned a lot.

[magicweed]

Would it also be effective if you used a weight distribution percentage?

Say you had a car that was balanced 47/53, use the 53% as a basis to develop a tractable torque curve. This is if my understanding is the usable amount of torque in each gear before wheel spin. Otherwise, could you two elaborate on what youre developing and have developed.

My spreadsheet only uses wheel torque outputs at certain rpms, and I compare them to a max amount of torque needed to produce f_staticmax (friction).

[BeaverGuy]

It sounds like you are on the right track there. Start with an FBD of the car and use the corresponding equations to develop your equations for propulsive force. To get an idea of actual accelerations you will need to compare the maximum force available to the maximum force the tires can sustain.

I'd post the spread sheet I've created but like the others I think it is something that is relatively easy to do and you will learn quite a bit by doing it.

[murpia]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by magicweed:
Would it also be effective if you used a weight distribution percentage? </div></BLOCKQUOTE>

By CoG position above I meant both height and longitudinal location. Lateral can be assumed symmetric. That way you can calculate weight transfer to the rear wheels.

The only Excel functions I used are 'LOOKUP' and 'IF'. Using a simple sum to integrate timesteps worked well enough for me, just make the timesteps small...

Regards, Ian

[Ed Sclater]

Mainly for Ian, but if anyone else can help I would be greatful.

On your graph for modelling acceleration what governs the rate at which the RPM trace increases and the rate at which it decreases when changing gear?

Also when changing gear is the time lost about 0.2s?

Many Thanks

Ed

[Marc Jaxa-Rozen]

You'd need to factor in engine inertia and a more detailed model for clutch engagement, but it probably won't change the results much. I just use a fixed drivetrain loss percentage and a simple on-off clutch which picks up the engine torque instantly after the shift delay, seems to work OK.

For the others who have done this, do you see a significant change in ET with the final ratio? I get a difference of about a tenth of a second between 3.5:1 and 5:1 on a 4.1 ET, which is consistent with results from a Performance Trends drag racing program, but I would've expected a bigger change. Probably depends a lot on the way you handle launch.

Marc Jaxa-Rozen
École Polytechnique de Montréal

[BeaverGuy]

The way you handle the launch and also the shift times affect how the final ratio changes the ET. If you set the final drive low enough you can end up with an extra shift or two and depending on how long you have set for shifts this can negate advantages from the greater acceleration.

[murpia]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Marc Jaxa-Rozen:
I just use a fixed drivetrain loss percentage and a simple on-off clutch which picks up the engine torque instantly after the shift delay, seems to work OK.

For the others who have done this, do you see a significant change in ET with the final ratio? I get a difference of about a tenth of a second between 3.5:1 and 5:1 on a 4.1 ET, </div></BLOCKQUOTE>

Ditto for the shift modelling. ET varied by less than 0.1s between final drives 2.5:1 -&gt; 4.0:1 for 1 -&gt; 4 shifts respectively. This is somewhat intuitive, as in general the most critical portion of launch is grip not torque limited, and the improved tractive effort curve from a short final drive is offset by the extra shifts required.

Ian

[Underthefloor]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by murpia:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Marc Jaxa-Rozen:
I just use a fixed drivetrain loss percentage and a simple on-off clutch which picks up the engine torque instantly after the shift delay, seems to work OK.

For the others who have done this, do you see a significant change in ET with the final ratio? I get a difference of about a tenth of a second between 3.5:1 and 5:1 on a 4.1 ET, </div></BLOCKQUOTE>

Ditto for the shift modelling. ET varied by less than 0.1s between final drives 2.5:1 -&gt; 4.0:1 for 1 -&gt; 4 shifts respectively. This is somewhat intuitive, as in general the most critical portion of launch is grip not torque limited, and the improved tractive effort curve from a short final drive is offset by the extra shifts required.

Ian </div></BLOCKQUOTE>

Have you guys been able to test these different ratios and verify your spreadsheet results?

How usefull do you guys think a detailed spreadsheet is vs a simple one based on torque curves, gear ratios and final drive? Most of the times in Detroit varried by more than .1 seconds for the different accel runs.

[Charlie]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by murpia:
<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Marc Jaxa-Rozen:
I just use a fixed drivetrain loss percentage and a simple on-off clutch which picks up the engine torque instantly after the shift delay, seems to work OK.

For the others who have done this, do you see a significant change in ET with the final ratio? I get a difference of about a tenth of a second between 3.5:1 and 5:1 on a 4.1 ET, </div></BLOCKQUOTE>

Ditto for the shift modelling. ET varied by less than 0.1s between final drives 2.5:1 -&gt; 4.0:1 for 1 -&gt; 4 shifts respectively. This is somewhat intuitive, as in general the most critical portion of launch is grip not torque limited, and the improved tractive effort curve from a short final drive is offset by the extra shifts required.

Ian </div></BLOCKQUOTE>

Depends on your shift time. The longer the shift time the less sensitive the car will be to final drive because of the shift time offset. If you have faster shifts the final drive becomes more influential.

Obviously, unless the car is already completely grip limited at the end of the accel event, it will benefit from more tq to the wheels from a lower final drive ratio. But the shifting comes into play.

[Jersey Tom]

Here's a dumb question.

What are you plotting on a maximum tractive effort curve? What, as a function of what?

[terra_dactile]

hi,
i could be mistaken but it would probably be engine rpm on the y axis and actual vehicle speed on the x axis.
correct me if im wrong,

Jude Berthault
ETS FSAE

[terra_dactile]

sorry for my last post it was obviously not correct, i read the question wrong,

I believe that you compare the force created by the engine on the tire in logitudinal direction with respect to the amount of force the tire can resist with out slipping.

the latter is very difficult to know for sure since their are so many parameter s that ifluence the force the tire can resist,
but the dat we have from the tire testing consortium should guide us in the right direction,

Jude Berthault
ETS Formula Sae 2003-Current

[murpia]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by Jersey Tom:
What are you plotting on a maximum tractive effort curve? What, as a function of what? </div></BLOCKQUOTE>

Tractive 'effort' would be better named Tractive Force. Calculate the longitudinal force available at the contact patch and plot against vehicle speed. Usually transmission losses are included, but not any road loads. Tractive force can be converted into vehicle perfomance only by including road loads and tyre grip.

Ian

[Jeff The Pyro]

one thing I was wondering how you get around is longitudinal weight transfer. In order to know how much tractive force is available you have to know how much weight is in the rear tires, and to know that you have to know how much you are accelerating, but to know this you need to know the tractive force, etc. etc.

I suppose this could be accomplished with a simple iterative method... get one parameter on the left of the equation as well as buried somewhere in the right side, plug in a guess on the right side and see how close it comes out to the left side, and then plug the new value back in again and see if it converges. sorta like what you would use to calculate a four bar linkage.

[Underthefloor]

In my opinion, the most important factor in choosing a final drive ratio is how easy it is to launch. We were able to drop our times 4/10ths of a second from 05 to 06 with similar power curves by choosing a final drive ratio that was easier to have clean launchs without excessive wheel spin.

Things like shift time don't matter all that much.

[Marshall Grice]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Things like shift time don't matter all that much. </div></BLOCKQUOTE>

I watched many teams do poorly in accel due directly to shift times. Any shift that has an audible delay is costing you big time.

We also made improvements in our accel times by going taller on our gearing.

[awhittle]

All the calculations in the world will make you feel better but the less shifting that you do on the autocross course will give the fastest lap times. The problem will always come back to the drivers ability to process information. The less things the driver has to do the better. Most FAST autocross cars have two pedels and only a temp gauge. If the gear shift lever is there at all it only has 1st , Neutral and 2ed.

AW

[Marshall Grice]

shift lever...psh.
video here (http://www.csupomona.edu/%7Efsae/video/dyno_vid06.wmv)

[wozniak8]

So what exactly is going into a spreadsheet here? The graph of the maximum tractive effort is vehicle speed vs longitudinal force correct? Also, how would using a CVT change things here?

[murpia]

<BLOCKQUOTE class="ip-ubbcode-quote"><div class="ip-ubbcode-quote-title">quote:</div><div class="ip-ubbcode-quote-content">Originally posted by wozniak8:
So what exactly is going into a spreadsheet here? </div></BLOCKQUOTE>

Usually you would want to include basic car data: Mass & CoG position, wheelbase, torque curve, gear ratios, final drive ratio and tyre size. Then construct a basic model to predict acceleration performance, probably including a very simple longitudinal tyre model. That will allow you to analyse trade-offs during the design process and optimise various factors (most usefully torque curve and final drive ratio). Note that the same model works for the acceleration event and corner exit analysis. Of course there are many ways to model the car with more sophisticated tools, but this is a good way to get started.

Regards, Ian