Formula-Retro drivability

Yep, no problem. I simply won’t touch it until it’s fixed.

 

Nevermind.

 

Guys I have what is probably really good news. Going back and forth with the preeminent cringelord, CrimsonEminence, I had the idea to do some testing with the fiddle brake.

Hypothesis: Clutched diff is not functioning properly in any car, this is a systemic problem with the game engine.

Test: Use the fiddle brake on the FV10 Mclaren to test to ability of Mclaren MP4/12's differential to unlock.

Procedure: Similar to before. We will begin by setting all diff options to full loose. Zero preload, 2 clutches, 89 degree ramp angles. Then park car half on pavement, half on grass/dirt. Dump clutch in first gear and observe the behavior of the car and the behavior of the tires.

Pass/Fail criteria:

Pass: Tires spin at different rates at some point. Tire with less traction should get all the power. At Jerez, the fiddle brake is on the right side of the car, so we will park that side in grass. We will then test with and without fiddle brake applied. This should result in massive tirespin on right side of the car in Test 1, and then it should result in ONLY tirespin on the left side of the car in Test 2, as the fiddlebrake will lock the right side tire, and then force all the power into the left side tire.

Fail: Tires remain locked despite ideal circumstances in both tests. Fiddle brake cannot overcome differential locking force, nor can the traction disparity.

Results:

Test 1: Both tires spin initially, but it is VERY obvious that the right side wheel begins to spin MUCH faster than the left side tire and the left side tire slows down to essentially just roll along with the car with little to no power being sent to it.

Test 2: Fiddle brake locks right side tire completely, ALL power is now transferred to the left side tire, car pivots around stationary tire and turns aggressively.

Hypothesis disproven in both tests. Clutch diff works correctly in this case.

I also noticed in some of my tests, that the car was EXTREMELY nervous going through the hairpin at Jerez here to get to the main straight. It just wanted to spin and behaved "strangely". This is intuitive, obvious confirmation in "feel" of the car that the diff IS behaving as a nearly open differential.

This confirms that the clutched differential CAN work correctly within the madness engine. The only caveat I have for htis is that the Viscous diff appears to also be activated on this car. I am not certain if that is correct or not, but it may play a role in it behaving predictably.

I repeated the test with the diff settings at max, but viscous still set to zero. There will be a video of that coming as well, but it will likely be tomorrow. What's interesting about this one is the diff acts more or less like a spool, as expected, but in slow motion actually can see that the fiddle brake does force some amount of slip into the differential. I was lucky enough to have the tire logos lined up nearly perfectly at the beginning, and you can see they do not stay synchronized as they do with a spool arrangement. This is with MAX LOCKING settings, so it's actually quite remarkable that it was able to get any separation at all from side to side, and you can see the rear fiddle brake temperature go almost immediately to 900C!

This proves the Madness engine and clutched diffs can operate correctly without having to trash the engine and overhaul the game entirely. This should be solvable. In the meantime, we can confirm that the Mclaren MP4/12 diff works correctly! It is VERY apparent in the feel of the car if you compare both extreme settings of the diff. You guys will enjoy the videos when they get added. I hope youtube doesnt compress them to potato quality!

I recommend you all go and try it.

 

I gave several cars a try with this basic test, to jsut do a burnout with half the car on the grass...

F-Retros seem to be the only car to be significantly affected. The F-Vintage actually appear to be allowing for diff separation. I think the issue is just that they're too tight. Some cars have much more obvious diff separation than others, and I think the F-Vintage tire grip level is probably masking any separation that is happening from the diiff working properly.

I dont know if this video will tell us much, these things are a lot clearer in the full screen replay than they will be compressed on youtube. But it's interesting anyway. Note the rubber tire marks will end once car speed is high enough, you have to look closely at the tire rotation speeds. If the car is veering to the right, then the left side wheel is still spinning with traction. Note wheel inputs to see how hard I am fighting the pivot in each case. The cars with lettering on the tires make it much easier to see.

I think there is still an issue with the minimum locking force just being too high across all cars, but it isnt as bad as I had thought originally, and it seems especially bad on the F-Retros. Again, with the diff set up as loose as is possible, these things should act like open differentials. They really should probably do one tire burnouts. But for all the other cars the behavior is more or less accurate, just perhaps too much baseline locking force.

 

Thanks a lot for the testing!
Let’s clear ground here with some fact checking.

Baseline implementation of the differential ‘works’ in Madness engine.

Since some builds ago the full open diff cars like the Fuscas (or the stock 86) had very high inner wheel spin making them almost stop accelerating while cornering.

Some cars (like the most recent Puma P5) are still doing it.

Some combinations of diff parameters had (have) more believable results than others.

If it would be possible to export telemetry data to any accessible numeric format without wasting hours in picking numbers one by one from the graphs I could also have a quite clear estimation of WHAT is not working.

The issue is so big that Renato & C. will easily identify and solve it.

Don’t be concerned...

 

Thought yall might enjoy this:

 

All versions of the track in ams2 would be good, esp Historics no chichane

 

I've given these cars a real go again, and they're absolutely undriveable. It's gobs of understeer until you get to violent sudden snap oversteer. I miss the old diffs a lot, they masked a lot of this behavior somehow or another.

 

I decided to buy PC2 and see how the 72E felt in that game... it's an odd comparison. There is none of the instability, the sudden snap oversteer, or the insane understeer, so whatever those problems are, I don't think you can claim they're part of the engine and are unavoidable. And even though the PC2 72D drives much more intuitively... it drives much more "normal"... it drives the way you'd expect, and it feels polished... It's missing some things that AMS2 have. In AMS2 you can really feel the car bouncing on its tires and on the suspension, and you can manage four wheel drift a lot better.

In summary, I recommend you all try it, but...
PC2 Lotus 72D = 5/10
AMS2 Lotus 72E = 3/10
AMS2 Lotus 72E with geared diff = 8/10. Drives beautifully... just needed the crest instability dialed out.

And for the record, I tried the PC2 72D at Bathurst, you can take that dip at Skyline as aggressively as you want, the car does not spontaneously want to become a space ship, or have some other sort of panic attack and want to go anywhere except the direction it's going. It behaves as you would expect, slightly less grip, but not neurotic. So it can be fixed. I hope it will be.

Another interesting note ,there are a couple of odd graphics anomalies in the AMS2 Bathurst circuit that I noticed early on. Happened to notice today that they are present in the PC2 version as well. (You can see them at the inside wall of The Cutting and The Dipper) Since the same flaws exist in both tracks, AMS2 Bathurst was obviously ported over from PC2. That means a lot of stuff from PC2 is probably going to make it to AMS2 eventually. I also want to make a point that the AMS2 Bathurst, even though it's clearly ported over, is FAR superior to the PC2 Bathurst from what I can tell. Its much more true to life and a lot nicer to look at. They are doing good work with that part of it, but I was surprised to see the same graphical glitches in the same places in both circuits, even if they looked and felt wildly different.

 

I would recommend these ffb custom files , ULTIMATE Custom FFB - NEW update V5 The definitive version try again! also many other cars. Lotus 40,
e type jag, vintage aston.

 

I have actually put AMS2 aside for these -but not limited to- reasons and enjoy the other sims for now (i wont mention them to avoid any flame).

Its very strange that some cars that were already in AMS1 and drove so well, are suddenly pretty bad in AMS2. I also find strange that those who work on these cars/physics dont test them more than 2 laps? No offense but how can someone think this is OK? Not just the retros/vintage but other cars like Camaro, etc.

There is also a lot of QOL/Features that I was used to in ams1 that still havent made it through, then the terrible replay system with lack of "physics simulation" where I cant even skip to specific parts of the race just makes me want to uninstall. AMS2 honestly looks great and FFB is one of the best and I believe Reiza will make it eventually... but for now I'll give it some time. And sadly I expect many others will too looking at online activity.

 

where did you get that information?

Perhaps those cars aren't that bad,
I follow the discussions about these cars and most of the time I try the cars and try to feel what supposed to be wrong, and most of the time it doesn't match my experiences.
Sometimes there is room for improvement, but the cars aren't bad or broken
Same with the FFB, The custom files were never an improvement on the default

 

If you didnt notice the 2 lap point wasnt serious... I exeggarated to make a point.

Anyway, if you read through the threads @Maser V6 posted an interview with a man who drove/drives these cars and clearly stated even at higher locking (and more plates) there is no such understeer or strange behaviour and the cars were relatively controlable.

But my main point was, Reiza had these cars in AMS1, the behaviour is very different (close to rf1 and rf2) in a good way. I dont understand how they thought this was better or as good

 

If I may interrupt here then I think this change is not caused by Reiza "thought this was better or as good" but because they cannot 1:1 control how the cars from AMS1 will behave in AMS2 because of the transfer to Madness physics and TM.
As I understand it then because a car looks like in AMS1 (somewhat ) then the underlying code controlling the behaviour has nothing to do with the code in AMS1.

 

Oh yes I know and agree I also get they wanted to release the cars as soon as possible for AMS2, just think they could have spent more time tuning. The way it was released just looked like it was ported to the new engine and that was it. I mentioned the diff issue after trying to setup the retros when this was released and posted about it here somewhere and bug thread. I also sent a PM to Renato about this 2 months ago but at that time it wasnt a popular opinion I think.

At the same time I do hope they get the grip of madness engine as much as possible to make this a great sim.

 

The Lotus 72 is one of my faves in PC2, but I agree it feels a lot more domesticated and can take much more abuse than in AMS2. Still a lot of fun hotlapping the Nordschleife in that old banger.

I struggled a bit with the F-Retro at first; horrible mid-corner understeer and wicked, wicked corner-exit oversteer. It was literally a shuntfest. I found that tightening up the preload, softening the rear suspension and keeping the slow rebound higher than the slow bump helped a lot.

With a stiffer front and softer rear one can use the throttle to control the understeer and sort of push the car through the corners, much like balancing a broomstick on your finger, if that makes sense.