v 0.10.2 FDE Improvements?
-
@sender46 It is very difficult to quantify, hence the general terminology used. As the PA28s were generally well received, I have not wanted to stray too much from what was there, possibly smooth off a few edges. However, to do that with the core flight model as it is is very difficult - I have always used accurate geometry as a base for my flight models, from when I started developing back in the early days of FSX. Then accurate coefficients on top of that, then minor tweaks based on the perceived accuracy of the calculated behaviours when compared to available documentation and direct input from current (or as recently current as possible) pilots.
With MSFS, the individual aircraft's flight model is supposedly dependant on accurate geometry - my first step. However, this most definitely does not provide accurate behaviours even in the most straightforward aircraft such as the PA28s, let alone anything with a slightly 'different' shape such as the Hawk. Most of the coefficients and tables we use to accurately define behaviour in FSX / P3D have been made non-functional in MSFS and no definitive list has been provided of what works and what doesn't. Plus the core flight model frequently changes with sim updates. So the second step in my process is effectively made incredibly difficult, if not impossible. This is where much of the last few months has been spent - tinkering with coefficients, trying to find a combination that may work, reverting back if it causes a problem in another area. To actually define what has successfully been done is, therefore, not the easiest, and lists of alterations similar to those I supplied for our P3D aircraft are not practical - I have made adjustments based on the Warrior as that seems to have been the best-received (and being the latest had the advantage of my learning from the Arrow and Turbos) but those adjustments have had to be carefully tempered to retain the individuality of each of the PA28 variants. Probably the most noticeable area for you will be the ground handling, which should be a lot less squirrelly.
Needless to say, developing accurate flight dynamics for this sim is not the enjoyable experience it should have been - much of the control we previously had has been taken away and when an aircraft that is accurately defined does not behave correctly there is little we can do about it except try and paper over the cracks which are resultant from the core flight model
I have yet to find an experienced flight dynamics developer who will speak positively about the current situation: the workload is unnecessarily high, there is a mass of confusion and we are having to move away from accuracy into what effectively is guesswork to try and create some semblance of flight characteristics.Hope that gives you some idea of why there has been silence occasionally and the time involved in making these updates.
Paul. -
@delta558 You did a great job ! The ground handling is definitely better, with no squirelly tendency anymore, and the adverse yaw is much improved as well: still noticeable but not to the point of being a handling problem on final as it was before. Did not fly in turbulence, so don’t know about the jerks though…
I hope for you that SU8 will not mess things up. It’s a pity Asobo is making things so difficult for developers, because for the rest, with such an add on as your arrow, this sim is awesome !
Keep on the good (if tiring) work 😉 -
@delta558 OK, ground handling a lot less squirrelly is very welcome. Can't get to fly it 'til the weekend but thanks very much for that.
Thanks also for taking the time to explain the difficulties. I guess every developer (payware and freeware) is having to learn for themselves by trial and error what works for them with MSFS.
-
@delta558
The new flight model is much better in my opinion. I perfectly understand your hesitation to implement some strange, artificial values in parameters, but probably there is no other way. After all, what matters in simulators is the feeling, not numbers. -
@tatar are you a flight model developer or an end user? If the latter, it may well 'feel' better to you but that is completely different to the point I was trying to make. I agree that there is now a fluidity within the environment and that that can make for a more immersive experience, however that does not necessarily mean accuracy. I want a PA28 that I build to fly as closely as possible to a real PA28, I do not want it to fly like a generic C172. Being able to build accurate behaviours within that environment would be great, but we are prevented from doing so - take, for example, the stall behaviour. This has been noted by other devs as needing massively inaccurate figures to create, as accurate figures do not give the desired result. NOTE: it is the SDK itself which states that this sim relies on accurate input of figures, it's not just me wanting to do so! I am not 'hesitant to impliment strange, artificial values' - if you go back to my original comment you will see from stage three that I am always prepared to nudge things sideways a bit based on documented evidence or pilot input. My problem here is that the first stage, accurate figures, does not work anywhere close to as well as it should and the second stage (the input of figures) has pretty much been removed so the only way we can nudge things is with the base geometry. That always creates changes in areas other than the one you want to affect. The core flight model should be a core to interact with the environment. Then we should be able to adjust with clear, well-defined and understood parameters (e.g. The aerodynamic coefficients which are understood with reference to control and motion, rather than made-up terminology that often does not make much sense).
If you are one of the former, then you are indeed a rarity!
-
-
@jmarkows I think continued and constructive feedback directly to MS/Asobo is the only course of action at this stage. In the meantime we will do all we can to simulate our aircraft as realistically as possible within any constraints of the sim, no matter how frustrating that can be at times.
-
@delta558
I'm only an end user but beeing also an real life glider pilot I have some understanding of aerodynamics. Obviously it is completely different perspective since I use MSFS for fun not as a training tool. It's always nice to have plane behaving as real as possible, but I'm not going to fly pa28 in real life.
Maybe I'm just used to the fact, that all PC simulators have been more or less inaccurate in the past - especially when close to stall conditions.
But I understand how much extra work is required to model fde correctly under these circumstances - with no guarantee that the internal flight model will not change in the next MSFS update.
