Sign in to follow this  
Followers 0
markop

Wheelcheck

66 prispevkov v tej temi

Evo, sem sprobal na 20% in je občutek veliko boljši. Več stvari se čuti, recimo kakšna grbina, ki se je pred tem ni, pa na volanu se čuti več mase avtomobila. Skratka pozitivne stvari. 

2

Deli ta prispevek


Link to post

Sinoči me je BMW prehitel. G27 Koderman

Kaj naj naštelam :blink:...

0

Deli ta prispevek


Link to post

Napisano (urejeno)

Kot si sam povedal Marjan ti WheelCheck ne bo pomagal. [owje]

G27 ? Min force daj na ene 10% do 15%. Bom updejtal grafe v kratkem. Tudi s tvojim.

Uredil - markop
0

Deli ta prispevek


Link to post
On 6. 1. 2017 at 10:24, markop je napisal:

Pojdite konzervativno z min. force. Če graf kaže 20% probajte z 10% in pol počasi navzgor, dokler volana ne začne tresti samo po sebi, ko avto stoji.

Tale metoda ni ravno 100% točna. Točna metoda je preko neke komplicirane analize v Motecu, ki je ne mislim izvajat ;) To kar mi delamo je "good enough" po moje.

Kako bo graf drugačen, če praviš, da naj min. force v Ac spremenim, test bom pa potem spet delal brez Ac?

 

0

Deli ta prispevek


Link to post

@markop

Daj še moje gor. :D 

CSPV2

0

Deli ta prispevek


Link to post

No, jaz sm mel na 25% tale minimum force... sm dal zdaj na 15% je bl realno glede na wheelcheck. :P

0

Deli ta prispevek


Link to post

A to v igri daste. Jaz imam 0 v igri min. force.

0

Deli ta prispevek


Link to post

Tako Marjan. V igri daš to.

0

Deli ta prispevek


Link to post

Dodal sem še preostale volane. Ni mi jasno pri nekaterih kaj se odgaja. Neki z nastavitvami mora bit.

0

Deli ta prispevek


Link to post

Če misliš pri tistih ki imajo "hribe in doline"...

Jaz sem imel isto takšne rezultate nad nekim % moči FFB, ko sem imel na nastavitvah volana (na originalnem fanatecovem volan displayčku) prižgan SPR in DRI na neki stopnji od 1-5 in ne na 0 oz. OFF. 

Seveda takoj po testu dam nazaj na nastavitve 5 oziroma -5, ker drugače se mi zdi FFB odziv volana na "driftanje" čuden. 

0

Deli ta prispevek


Link to post

Vidim da je moje gor dol :D 

Imam DRI na 3 in to je baje najbolj linearno.

0

Deli ta prispevek


Link to post

@mare87: Vprašanje glede tvojega g27 volana. Imaš originalen obroč gor ali imaš kaj drugega ?

 

0

Deli ta prispevek


Link to post

Original, star 2 leti + 1 mesec :). Sem zdaj še enkrat naredil bench z nastavitvami za GT3, in 100%, če bo kaj drugače..

 

 

 

 

Test.zip

0

Deli ta prispevek


Link to post

Glede retultatov sem se obrnil na IR in tistega ki je napisal WheelCheck in FFB za IR nasploh. Bom kar copy/paste prilepil:

---------------------

Hi, I would need some help. I've asked the members of the local sim-racing community to perform the step2 test for their wheels. What I've got is somehow strange. The G27 wheels were pretty much in line, so were the T500. 

But, others were showing a zig-zag behaviour. Also the deltaX degrees was greater for the g27 wheel (red line) when compared to a T500. 

The OSW (black line) was totally off, even showing negative deltaX degrees. 

Can anybody explain where I'm doing it wrong ? 

---------------------

David Tucker:

The step test is really only good for comparing wheels that are very similar to each other. To get a proper view you want to switch over to physically measuring the wheel force, rather than the motors ability to overcome the wheel rims inertia and friction. Also from the post right above yours, you may want to set kd back to 0.25 to help smooth out the results. Finally you can't really get anything usefull out of an osw or accuforce wheel, there motors are so strong they accelerate to max velocity long before the test is complete, and often they end up slamming into the bump stops and bouncing back, etc. 

You can pick up a used USB postal scale for about $15 on ebay and with a bit of ingenuity you can turn that into a test rig for wheels. If you want to be super classy you can pick up a load cell and make your own strain guage for about $60. I can fill you in on the details of that if you are interested. 

http://members.iracing.com/jforum/posts/list/3482238.page

---------------------

Thanks David. 

Before turning to you/forums I've read a whole lot of this stuff here and in the wheel profiler forum. My goal was to have a relative comparison and not in the absolute torque measurement. Further, I wasn't interested in the comparison of the wheel base solely, but with the rim attached to it as this is what the user feels at the end. 

I'll subdivide what I don't understand (and asking for help) in three parts: 

1) Given all this it's my understanding is that the log2 test should be proportional to torque if you take the X position reading at the same time from feeding the input to the wheel. Indeed, all the G27 and T500 wheel looks nearly the same. Then there is that g27 (mare87_g27 red) graph that stands out in regards to other g27 wheels. 

2) Then, there are other wheels beside the OSW exhibiting that zig-zag behaviour for some unknown reason. Those are not DD wheels but for instance the Fanatec V2 base does the same - see the "alkoro CSPv2" graph. I doubt it hits the bumpstop in 300ms. I think it has more to do with the Fanatec DRI setting. 

3) For the OSW wheel I even tried to lower the "step time" to 25ms for the OSW, but the result wasn't the expected one either. 

Is my understanding of the log2 test fundamentally wrong ? Thanks. 

---------------------

David Tucker:

No you have it right, but you aren't seeing all the variables. In an ideal world, if there was no resistance on the drive train and the wheel rims all had a uniform weight, then the step2 test would always allow you to compare each wheel to the other directly. However different wheels have different weight rims (some with offset weights even) and they all have different levels of drag in the drive train. 

The belt driven wheels have a lot more drag than the gear drive or direct drive wheels. This can be a good or bad thing but it definitely changes the results. That is why you can see the G27 matching the T500 even though the T500 has 2x the torque. This is also why an older broken in G25 wheel will often beat out a new G29 wheel, the gear train has less drag the more it wears. 

I use a PID loop to center the wheels between samples in the step2 test, and without the kd value being set to 0.25 the wheel (apparently) is overshooting and oscillating instead of just returning to center. This is affecting the results, causing a wild oscillation on some wheels so be sure to run the test again with kd set to 0.25. That should make the Fanatec work better at least. 

OSW wheels are just way too strong to ever work with this test. They have large motors with little drag and they have intelligent controllers that limit there max acceleration and velocity. The net result is that after only a few milliseconds they are already spinning at there full speed rather than accelerating up to speed over time like all the other wheels. This has two effects, it messes with the numbers (of course) and it allows them to spin into there bumpstops causing them to bounce back off of them really hard. If you watch the wheel running the test it looks like a jack hammer that is flopping all over the place. You can tweak the settings to make it work (sort of) but the end result is still not worth much. Think of it like trying to use an RC airplane engine dyno on a corvette, the engine would be so powerful that the measurements are meaningless. If you really want to make the OSW work then slap a bunch of weight on the rim. I think 2-3 sand bags would do the trick. 

There is still something to be gained from running this test and comparing wheels, you do get an idea of how much resistance each wheel produces and that in turn gives you a feel for how responsive the wheels will be. As it turns out the belt drive wheels with there extra mechanical damping feel smoother (or sluggish depending on who you ask) while the gear driven wheels feel snappier (or more rattly)

Really a step response test should only last for 30 ms or so, but I'm pushing it up to 300 ms because our update rate from the wheels just is not fast enough. 300 ms is way too much time and lets the stronger wheels accelerate too much, but lowering the number just makes a mess of the response plot (as you found out). You can compromise and use 150 ms but that is not low enough to fix the problems so it does not seem useful to me. 

Also you really should be looking at the derivatives of the plot to work out velocity and acceleration but again we lack the ability to measure the wheels position with good precision so the derivatives are a real mess even with lots of filtering. 

In short this is the best you can do with what little we have, but it is not very good compared to using a load cell or some other setup.

Oh, and drift mode on the Fanatec wheels is a feedback loop so it will make the test more unstable. Your best bet is to turn it off when running these tests.

---------------------

 

2

Deli ta prispevek


Link to post
pred 8 minutami, markop je napisal:

Oh, and drift mode on the Fanatec wheels is a feedback loop so it will make the test more unstable. Your best bet is to turn it off when running these tests.

 

this! 

0

Deli ta prispevek


Link to post

Zdej pa primerjajte ta odgovor s tistim od Štefana ...

2

Deli ta prispevek


Link to post

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Registriraj nov račun v naši skupnosti. Zelo je preprosto!


Register a new account

Prijavite se

Že imate račun? Prijavite se tukaj.


Prijavite se takoj
Sign in to follow this  
Followers 0
  • Temo prebirajo   0 uporabnikov

    Te strani si ne ogleduje noben registriran uporabnik