Hi all,

I've been playing with the MRA nondimensional model and comparing it to the raw data. I've found that for the Avon (tested in round 2), the camber effect in the raw data differs quite a lot from the camber effect produced by the MRA model (taking into account the various sign conventions etc).

Is there a way to account for this by varying any of the parameters in the MRA model? Ie, be changing anything in the tyre .mat file? I'm keeping everything within the range of the test parameters.

Thanks,

Dean McGeary
Redback Racing - UNSW FSAE

This message has been edited. Last edited by: nos,

Not sure how kosher it is discussing actually data trends in public forum.. but..

1) Do you need an accurate model? For a lap simulation tool? I would imagine at the FSAE level most benefit can be reaped by characterizing the raw data

2) Have you looked at the Pac96 / MF5.2 (whichever it is) coefficients provided by Stack? How does that fit overlay with the raw data, and with the MRA model?

I am not that familiar with the MRA model, but I do know you can go in and manually tweak coefficients for the Pac96 model to change curve shape, camber sensitivity, blah blah blah. Stack provides the underlying equations so you can at least see how everything works. Isn't encrypted into a P-file.

3) Are you staying within the range of test data? Ie are you trying to extrapolate trends up to say 8-9 deg of camber when the test and fit data only go to 4?

Hello to fase.com,

Dr Kasprzak is very busy with other work these days and recently he asked me if I could take on some of his TTC volunteer work, including watching the fsae.com forums. So I've just joined and will monitor this thread, and maybe some other TTC related threads if I have time.

Looking back through some recent posts, the topic I'd like to pursue is a secure forum for TTC members to post data and modeling questions/results. If it was really secure, maybe it could also be a distribution channel for future data--sending out 150+ DVD's after the next test is not something we volunteers are looking forward to doing! Has anyone given this more thought?

Re: MRA model basics -- anyone interested in the MRA-provided tire model can get a general idea from RCVD Chapter 14.

-- Doug Milliken for the TTC

FWIW: I have some experience with the primitive tire model used Bosch/Lapsim. It's Pacejka with just 4 coefficients for each tire and well suited to FSAE car lap time optimization. Folks in the Racing group use it (full blown version) for many of their own vehicles and have a high regard for it (for reasons already mentioned: operating at the limit. While there is no Lapsim input for Mz or Mx, I can show you how to produce Lapsim tire model coefficients for Fy, Mx, and Mz, slip, camber, LR, etc from the Calspan datafiles using Matlab if anyone is interested. (not sure how to post code, though, unless its with a Forum blast. I personally have not yet looked at the TTC data but am very familiar with Calspan .tir and .dat files. I suspect some extra=points seeking teams might want to run lap time sims with manageable tire model elements. I'm not dissing MRA or Pacejka'xx, or MFT, just recommending a KISS approach. The Lapsim model is high value and has a lot of low hanging fruit to be picked.

Bill,

Nice to see you posting here. If you would like, your Matlab code could be distributed along with the FSAE TTC data. Your code sounds like it would be a "roll your own model" alternative to the two different "canned" tire models that are currently included. Support for the free Bosch/Lapsim is a great idea.
The next TTC DVD will be assembled after the fourth round of testing and will go out to all current members (over 150 universities). This might happen this summer, depends on workload of the TTC organizers. Email me at:
doug . milliken on gmail . com
or we can continue the discussion here online.

-- Doug Milliken for the TTC

Hi all,
What’s about the MRA update? Has anyone solved the problem with the MZ data? (Matlab Error: ‘Error using rdivide’).
Currently I try to plot the data using the SES Pacejka coefficients to compare the plots with the MRA output. Has anyone of you already done this?
Cheers, Haymo

Haymo,
Unfortunately for you, our paying work comes before volunteer work for FSAE/FStudent. Dr Kasprzak is very busy and I can't predict when he will have time to look at the model we provide.
Plotting the MRA model against the SES model is an interesting idea--maybe you can post some comments on your general results?
-- Doug Milliken

Hi all,

This is a bit early...But,

Any thoughts to new compounds for round 4? We feel some of the new Hoosier B compounds should be on the list:

43161R25B 20.0x8.0-13 road racing slicks R25B compound

43162R25B 20.5x7.0-13 road racing slicks R25B compound

Also, On behalf of Rensselaer I would like to thank you Dr. Kasprzak and Doug Milliken for your work. We understand you are busy and have lives as well.

Respectfully,

-Ian

Can RPI get help from Alumni ?

Bill Cobb
'71-'72

I wouldn't be opposed to it.

Respectfully,

-Ian

Dear all,

At the moment I'm trying to implement the TTC data into an ADAMS model I'm making of our FSAE car.
Doing this I ran into the following problem: TTC delivers a Pacejka 96 model, whilst ADAMS comes with 89, 94 and 2002. After investigating all models by reading the TTC/ADAMS manuals and browsing the internet I came up with the following partial solution:

Take the 2002 model and put the 96 coefficients in. Take all unknowns 0, and disable transient behaviour. This effectively creates a 96 model in the 2002 one. However a number of things cannot be ignored, but are also not present. The TTC data does not include (or I have not been able to find) the vertical damping, peak value and high/low load variation of the rolling radius. These last three might be extractable from the TTC data. But I have my doubts about the damping. ADAMS does need these however.

Also the PKy2 < 0 (Hoosier 20.0x7.5-13 tires) is not accepted by ADAMS.

It amazed me a little that this problem has not popped up on any of the fora before. Is there a solution to this ("known?") problem (converting '96 to '02)?
As well as why PKy2 has the wrong sign by ADAMS liking?

Thanks in advance for any help / pointing me in the right direction.

Gimme a tire brand/size/construction that you are wanting to run an experiment on. If you have Matlab, I can maybe provide you with a MF5.2 pure slip fitting routine. A Roll Your Own, so to speak. It that works for you, you will be able to quickly figure out how to do the rest, given 'good' quality tire data. Keep this phrase in mind: "Good Tires Test Well". I suppose that also could mean "Good Tyres Test Well", too !

Interested?

I beleive the answer to your question is as simple as "No ones doing it".

I still think it'd be cool for Round 4 (whenever that is) to do some wet testing. Inevitably it tends to rain on at least one day of driving in Detroit.. and from the Oz comp apparently even they are not exempt from bein rained on. Good rain tire would give a pretty big advantage.

Would also just be cool to see what the difference in grip levels is, breakaway traction effects, etc.

I've been trying to get in touch with Dr. Kasprzak to update our contact information, but have never gotten a reply. We never received a DVD for round 3, but we did get the update ADAMS model via an alum of the team that was kind enough to forward the data to our adviser. Does anyone know who I should contact to make sure that our info is correct?

Matthew,

I haven't seen an e-mail from you...don't know why. I'll send you a personal correspondence.

We can check the mailing address and arrange for another Round 3 DVD to be sent. Be sure to look for it when it's expected to arrive. University mail systems are, in general, above average at making things disappear.

It's true I've been very busy this year, but I'm still around and putting time into the FSAE TTC as I can.

Regards,
Edward

Hi Dr. Kasprzak

Is there an expected date for round 4 testing?

Also can we still expect an update to correct problems found with round 3 tire models?

Thanks!

GooDLuC / Université Laval --

No date set yet for Round 4 tests, we will post here (or start a new thread) when we know more.
Likewise for our MRA-donated models.

-- Doug Milliken

Dr. Kasprzak,
E-mail sent! Thanks again.

quote:

Looking back through some recent posts, the topic I'd like to pursue is a secure forum for TTC members to post data and modeling questions/results. If it was really secure, maybe it could also be a distribution channel for future data--sending out 150+ DVD's after the next test is not something we volunteers are looking forward to doing! Has anyone given this more thought?

It's a bit disappointing that nobody else has put their hand up for this. Now is a really bad time for me to go through the quite time consuming task of setting up a good facility for this, with a lot of work ahead of me in the next few months, but if nobody has taken on the job by the end of the year then I'll do it.

Malcolm,
Thanks for your offer of help with a secure server or other system. This problem has not gone away, and the task of data distribution obviously gets a bit larger every time the TTC gets a new member.
-- Doug Milliken