Download Logo 10E 3D for Reflex

[Neil Walker]

Some screenshots:





The readme:

Logo 10E 3D v1.0 for Reflex XTR
===============================

Model & physics: Neil Walker & Russell Bowman
- with a few bits and pieces shamefully stolen from the Reflex Logo 10

To install: Unzip into your REFLEX\Heli directory and make sure you have "use folder names" on. The model will end up in the "Mikado Model Helicopters" group.

This is a model of the Logo 10 3D on 8s, with the "Extreme" carbon frameset, a stretched boom and 550 blades. The real heli it is modeled after has the following specs:

Mikado Logo 10 3D mechanics
D.F. L10 Extreme frameset & helical gearbox
Neu 1512 3D
Hacker O-77 Heli ESC
Thunder Power 8S2P 4000mah
550mm Carbon Rotortech m/r blades
Stretch Kit 620mm
Logo 14 75mm t/r
L10 3D paddles and flybar
Carbon Sideframe t/r gearbox
t/r thrust bearing kit
Hardened mainshaft
JR DS3421 CCPM servos
Futaba 9253 t/r servo
Futaba 148DP Rx
UBEC
Futaba 401 gyro
1900 rpm on curves
± 12 degs collective
approx 6 degs cyclic

The physics were arrived at with the input of the pilot (Russell Bowman) and are pretty close to how the real bird flies. There is, however, a slight difference in the high-speed flight performance. It just seems to "wobble" a bit, in all axes. This behaviour seems to be present in every heli in Relfex, so it seems that it is a flaw in the physics engine rather than physics parameters of any particular model. Some helis in Reflex, like the stock Raptor, are quite bad in this respect, others are better but they all exhibit this "loose" behaviour when flying fast. The real Logo 10E 3D this was modeled after is rock-solid in FF and FB flight. It's a minor nuisance with the model, though, and many people won't notice it. Needless to say it really annoys Russ

Note that at 100% swash travel, the cyclics are pretty crazy. A more realistic (at least closer to how the real heli flies) is to set the swash travel to around 80-85% in the "Model Parameters" menu of Reflex (F5). I've also found that in general, setting the time expansion factor in the "Simulation Parameters" menu (F6) to 1.05 or 1.1 slows the simulation down just a hair and seems to feel a little more realistic for all helis... but that's just personal preference.

A video of the real heli can be found here: http://www.success-info.com/Xfer/L10E_part_ii.wmv

You are free to do whatever you like with this model and the associated .PAR file. If you do use it as a base for something else, or modify it and redistribute it, it would be nice if you gave us credit and provided a link to the original in your readme file.

For comments or questions you can PM user "Neil Walker" on RCGroups or RunRyder.

Enjoy!


The download: http://130r.net/files/Logo_10E_3D.zip

[CustomPC]

My new favourite reflex heli. It flies just like my Logo 14 and looks very close with that 3D canopy.

[Lost Horizon]

Originally Posted by CustomPC

My new favourite reflex heli. It flies just like my Logo 14 and looks very close with that 3D canopy.

In real life, it will outfly my Logo 14 on 10s2p even... it's a hot machine..

[Neil Walker]

We made a concerted effort *not* to make it totally crazy in Reflex. I've been involved with simulations for quite a number of years (I worked on some commerical driving/racing simulations) and also with the communities of people who make add-on vehicles (cars, or in this case helis), and one constant seems to be that most of them come out with greater performance than they actually have. I think it's a psychological thing - you do a model of your car (or favorite car, if you can't afford an M3) or in this case your heli, and you're obviously proud of it and have a feeling that it is "the best" for some reason or other - in this case because you own it - and the end result is that when you do the physics you end up making it "the ultimate machine".

With this model we really had to tame it quite a bit - even from many of the helis that come with Reflex, which IMO are all a little exaggerated, in particular with top speed. We cranked up the drag on the L10E just to slow it down. Russ feels that the cyclics are fairly accurate, although as you know Dana they're still a little bit sticky when compared to your - I mean *MY* - Logo 10.

One thing we're going to try soon that may result in a physics update is to take an altitude logger I have for my sailplanes and put it in Russ' bird to figure out the max climb rate.. then Reflex has some data displays that we could use to match up the model's power.

[nose-over]

I think it is very important to get the parameters by actual physical measurements. I just recently bought Reflex after owning the "other" sims and am blown away. I have an aeronautics background and after tooling around with all the physics of the big 3 believe that Reflex has the most potential. I'm a member of the same club as one of the W3 guys, and what they have coming out is really impressive. I talked alot with him about how people who were releasing planes out their were doing their physics by "feel", and that just doesn't work. Chris told me that they have spent the last 7 months writing new software to mesasure ALL of the reflex inputs for physics. NO GUESSING. I believe they primarily do planes but it goes to back up your point. Guessing the parameters just doesn't cut it. But it is impossible for most people to measure moments of inertia and so on. I can't wait for their release of all thier new aircraft done with their new software. Great job on the Logo!

Seth

[Neil Walker]

Actually I kind of disagree. I have a significant amount of experience with simulations - both in coding a physics engine itself and in developing the physical parameters for particular models. And have participated in what seems like an infinite number of threads where people argue the merits of Absolute Accuracy (or "the math doesn't lie") vs. Feel ( or "I don't care what you eggheads say, it just doesn't feel right"). And I used to be on the side of absolute accuracy, but over the years that has changed a little bit. (As a side note, if you ever want to start a big flame war, log onto a sim racing forum somewhere, find people's favorite sim and make a post that says something like "I dunno, the tire model's all wrong, I think the slip curves just don't feel right" and sit back and watch the fireworks)

I find it rather interesting that someone is talking about writing software to measure all of the Relflex inputs for physics. Not to put down your friend's efforts or anything (I've seen the W3 models and they are some seriously talented modellers!), but even without actually inspecting the code for the physics engines themselves it is easy to tell that there are quite a number of parameters that make up a Reflex XTR model which are quite clearly "fudge factors" - they have no real-world counterpart. Hovering instability? Rotor head vividness? Rotor head stiffness? Elevator tendency? All expressed by single numbers... I would like to hear a scientific definition of those physical quantities and how to measure them And those are just a few examples.

And I don't mean to dump on Reflex either - it is my favorite sim. But it's interesting to compare it to AFPD. When you look at the phsyics files for AFPD the things you have to specify are much closer to real physical quantities. You're simply specifying masses and moments and positions and joints, and the simulation works out the physical properties of the system you define. Reflex, on the other hand, seems to have taken a mixed approach. There are lots of strict physical quantities you define but there are also a number of "behavioural quantities" that you define. That's a nice way of saying fudge factor. Things like "hovering instability" and "rotor head vividness" are obviously such factors and there are no real-world counterparts for them. The only way to set them is by feel. In fact the only way to figure out what the heck they do in the sim, without having access to the source code, is by feel.

Which brings me back to my original, rambling point. In my early days of involvement with sims I would have said AFPD is the only way to go, hands down. The more accurate the mathematics behind the simulation and the more detailed the measurements of the phsycial parameters, the better it will be - period. But that's not the case, is it? Ask the majority of heli pilots, and the clear winner in terms of best sim is Reflex - the one with all the fudge factors. Somehow the XTR folks have struck just the right balance - pretty decent physical simulation but in the end they've realized the most important thing: The goal of the sim is to convince the user they are using the real thing. However you can best do that is the way to go.

In the end, physical accuracy only counts up to a certain point. I've come to the conclusion that what you lose by using some well chosen "fudge factors" and going by "feel" is dwarfed by the fact that you're sitting in front of a computer screen, and are lacking a whole range of sensory inputs that you'd normally have. I still firmly believe you have to start with a solid foundation in the real world math and numbers.. but if a little fudging and modifying quantities to get a good feel works, then do it.

[nose-over]

I do agree alot with what you said and also believe that in the helis there are more "fudge" inputs than in the airplanes. In the airplanes the only thing that can't really be measured is the damping factors. Which as you know are put in by programmers as fudge inputs. what I got from talking to Chris, is that the more numbers you can muster from correct calculations , the more the damping inputs become a non factor. It seems the damping cures "WAY-OFF" inputs elsewhere. case in point: There was a quiet storm released here on the threads a week or so ago, and I really do appreciate the efforts, but I own a quiet storm and their model flies absolutely nothing like it. and anyone that has any aeronautics knowledge can look at the.par file and see why. it was done by two guys by feel obviously. Didn't work. That model was the topic of my conversation with Chris, He emailed me a Quiet storm that W3 will release in the near future and let me get first look at it. It flies incredibly like my real one and was all done by physical numbers and measurements. He doesn't even own one. So I agree with you on some of those points but it seems I'll take the numbers with just a "tiny" fudging in the case of the planes.

Seth

[[ o ] cap]

Hey Seth, I'm glad you liked the Quiet Storm. I'm sure it can still use a little tweaking but It should be extremely close. Neil is correct on alot of issues with Reflex but not quite all. From alot of expirementing and consulting with the Reflex folks we figured out what all the inputs in the par files actually do. Some of the definitions of these inputs are a little confusing because of "lost in translation" and just how they apply to Reflex and maybe not aerodynamics as by the extreme definition. What we found was that if you put in the correct information (as they apply to Relex) you get out a very close flying model without much, if any, "fudging". What seems to be the problem is calculating many of the inputs. some of those calculations are just over many peoples heads ( including ours at the start of this). For example the most important measurments (besides the physical dimensions of the plane) is "moments of inertia". This is extremely difficult and confusing to measure for a plane and most people wouldn't know where to start. So they just don't. They borrow the inertias from a "similar plane" that someone else borrowed from a "similar plane" and wha-la... aerodynamic mayhem. You start tugging and pulling on different inputs to make up for that input then you have to tug on some other inputs to make up for those and before you know it you have a plane that is so tightly wound up that it wants to explode like a spring. And that is what we call " doing it by feel". I am guilty of it too. So we set out and designed software that could break a plane down to more accurate areas, dimensions, weight distributions and moments (as they apply to Reflex). What we get out of it is a par file that surprisingly needed very little fudging. Turns out if you get all the right numbers in the model, its ...well...very relaxed and not wound up like a spring. Neil was right in the fact that there are still some leftover inputs that need some " feel" added but only a few on the aerodynamic side. more on the drive side. The key was to figure out what Relex's definitions all meant and how real world aerodynamics apply to them. So basically I would spend days on a plane trying to get it to feel right by guessing inputs and still never get it quite right or have a plane that flies like I would "like it" to fly. Now I put in the raw data and get out an extremely close par file all in about an hour with no fudging.
Now, we really have only dabbled in the planes side of it and not the helicopter side, So I wouldn't know about the fudging side over there. Neil may be right about being better off by using more of a mixture of real numbers and the seat of our pants inputs for the heli's and I applaud him for his efforts. I would'nt even know where to start on those yet.

Chris
W3 Group

[Neil Walker]

I have actually in the past (over 5 years ago now!) done exactly what you're talking about - calculated MOE's from components in an effort to achieve accuracy. Somewhere out on the net there's already some software to do it for you. Can't remember what it was called but it was for racing cars, used mostly to design suspension geometries. It did a whole bunch of stuff. It was commercial but you could download a demo that let you do alot.

But anyway, I have been through endless discussions exactly like this (usually on the other side, saying all the things you are saying) and don't care to get into another one. It's just pointless. Keep up the good work!

[guillaume83]

Hi all,

your Logo looks great Neil!! thanks and keep up the good work

Nose-over,

I feel I have to stand up for my friend's work. I was a (very) little bit involved in the development of his Quiet Storm for Reflex XTR,but I hardly did anything compared to what Laurent did. I provided my feelings of the model as you rightly point.I do own a Quiet Storm though,so I should know how it behaves.
I cannot see anything wrong with "designing" a model from your feelings rather than mathemaitcal calculation.

When it comes to Aeronautical knowledge,I might not be the most experienced Aero. Engineer outthere but at least I've got some knowledge about A/C and so does Laurent as we both work in that field.

It really is a shame that you did not post in the relevant thread in order to improve the model and make constructive comments. Laurent would definitely like to have some constructive feedback rather than plain comments.

The set up also changes the behaviour so your QS might be set up differently to ours.

Sorry for the off-topic but one has to stand up for his friends.

Guillaume

[nose-over]

Guillaume,
I appologize if my comments weren't the best choice of words. I really do appretiate your efforts. If I knew how all the inputs worked and what changed what I would know what to change and how much, but I don't have a very good working knowledge of Reflex so I can't. The plane just doesn't feel right to me, Where as the quiet storm I tried from W3 felt right on. I'm not sure how they came up with all there inputs but they obviously are on to something good. I tried changing inputs on your quiet storm to more reflect mine but to no avail, It just made it worse. So I will say your work is better than what I could do myself anyday and you should keep plugging away at it. The more good modelers we have out there the better.

If I have any comments that can help I'll direct them to the quiet stom thread.

Seth

[Rickenbacker]

Originally Posted by Neil Walker

"the math doesn't lie"

This is, of course, entirely true. But ONLY if the sim perfectly matches real life in every respect. I don't think we'll ever have a sim that does that, so some fudging is necessary for the helis to "feel" right.

[Neil Walker]

Originally Posted by Rickenbacker

This is, of course, entirely true. But ONLY if the sim perfectly matches real life in every respect. I don't think we'll ever have a sim that does that, so some fudging is necessary for the helis to "feel" right.

I think that was what my post was trying to say. The way you quote me makes it sound like I really believed "the math doesn't lie", when in context I was being a bit sarcastic when I said that

[Rickenbacker]

Sorry about that, couldn't resist an opportunity to look clever .

[CustomPC]

Chris ([ o ] cap)

Has W3 made any models with flaps. One thing i've noticed about reflex is that any model with flaps balloons dramatically when flaps are deployed. I've been wondering if this was reflex's physics engine or just poor par files causing it.

Any chance of doing an EStar Models OV-10 Bronco?