Very nice breakdown! Luckily the devs have already shared some WIP shots of model improvements on their Discord, chief among them being corrections to the cockpit model.
Thanks! I'll have a closer look when it releases. I don't follow the development in the discord all that closely, especially since I was focused on getting this video out the door before another release dropped and I'd have to redo a bunch of things again.
Personally, I'm extremely appreciative of what the creator has provided for free to all of us rotor heads. Hundreds and hundreds of hours put in to provide us enthusiasts with a favorable, and most welcome addition. I personally have enjoyed tinkering with it and look forward to future, free updates, despite its imperfections! I am confident that other helo simmers share the same sentiments!
I absolutely agree. I'd much rather have it than not. I just felt there were issues that needed to be addressed. I'm sure they will be, and I'm looking forward to the next release. I probably wont be making any more videos on it until it's reached what they consider a "complete" release, though you know these things are never "complete."
Nice presentation. I enjoy this Mod. Good explanation of the mod, the flight controls and the reasons for why they did what they did. Also the tail plane portion reveals this is critical component-I guess so because it has a gauge right in front of the pilot's face...explains the nose-over into heaven. At first I thought this might be a Michelangelo made the David's hands too big ego trip, but then the tenor of the presentation moved to an engineering level discourse but the with real-world textual (Dash-10), and visual explanations-good for single diode operated beings like me (former USMC). I took away much from this presentation and appreciate it. Good job.
I do indeed plan on doing something like this for the Apache. I've already started looking for information, but it's definitely more difficult to find publicly available documents to reference, however.
Good job. I love this mod and if these things are corrected great! If they fix the instrument panel being to long and high and I can see thru the chin I am satisfied
This is a review that should be done for all modules, by those who knows what to look and how to compare. People with experience notice small details that are off, and they can dig up the comparison to public photos etc to explain the differences. These serve as good as excellent bug report, that any developer should take as such. The up coming Gazelle update (maybe on the next update?) needs again a complete revisit for the flight modeling check etc, and meanwhile hopefully these mods gets the corrections done.
Thank you, you spoke out on the point of what I found as well. but I gave up sending a bug report or information. The Community is not acting friendly about this. Well done a solid Quality Report.
I kept looking at the model of this mod and thought it looked slightly off, but couldn't point out what it was, it was driving me nuts! Thanks for showing that I am not completely insane
By far the most comprehensive overview and review of this module. Not sure if the devs are familiar with system control theory though. Most people dont even know what summing junctions are, much less how to derive transfer functions with feedback loops.
Yes, I was careful to not include diagrams with transfer functions and such as that would get too complex for an audience that is broad enough as it is. If you want the full details and all the transfer functions, NASA CR-166309 has everything you need. It also contains a full mathematical model of a Blackhawk that is essentially the component-wise model I suggested at the end of the flight model section.
In my opinion, "doing it for free" has become the most go-to excuse after "tested by real pilots". Speaking of which, it has become almost customary to include the latter line in every civilian flight sim addon disclaimer. People almost unanimously fail to realize just how different "flying" an aircraft (or operating other manned vehicle) in a sim is compared to reality where your sight almost takes backseat to the other senses, not to mention how much less you are getting through vision in a sim. "Subjective" is putting it VERY mildly - if you ask a dozen of real pilots to create "flight models" based on their perception alone, you are getting a dozen of very dissimilar such flight models. Engineers, on the other hand, don't really operate on perception - and that's why videos like this are massively more useful than "real pilot" input. I highly appreciate what you are doing here, Erik (also for free, by the way, and not even allowed to go to the full extent). Only turning to open source and having both the code and FM behavior reviewed by *competent* parties is a good recipe for proper simulation. By the way, have I understood you correctly that for rotary wing aircraft BET-like models are much more preferred to the "whole thing" aerodynamic table array FMs similar to what can be found in CR-3144 or TP-1538 alike? My defense for the tables is that Aerodynamics is way too complex to have any hope to easily break down the airflow at the components and parts level and integrate the results back up - especially using something as primitive as "Blade Element Theory" that is simply unaware of vortices - while the results that go into the table are usually "solved" by nature itself (although wind tunnel testing has its own set of flaws too). But I guess with so many degrees of freedom as helicopters have, populating enough tables to cover for most situations becomes a seriously arduous task.
Thanks for the comments. You make a good point about the benefits of having the behavior of the aircraft rolled up into a table of derivatives. You're absolutely right that this permits you to incorporate various complex aircraft behaviors without actually modeling the physics. While I have never worked on fixed-wing simulations myself, I know that this method is used extensively for high level flight simulators used for fixed-wing flight training (think "Class D" flight simulators). The required level of precision in making the aircraft behave correctly is quite strict, and so they turn to modeling the behavior rather than the physics. To to this, they actually measure the stability and control derivatives using real-life instrumented aircraft in an incredibly large number of flight conditions (airspeeds, weights, CG positions, altitudes, etc). This is a great method if you never plan on flying in conditions significantly outside the range of conditions spanned by the data and you really don't care about the contributions of each part of the aircraft. Additionally, these types of simulators are not typically interested in modeling aircraft damage (though they will model things like engine failures and other emergency situations). Modeling the effect of damage is a fairly crucial element of a combat simulation like DCS, so having a model that can handle that fairly seamlessly is a great benefit. You could have an entirely new set of tables to represent different damaged states, but you will need some way to generate that data in the first place. You are pretty much on the money with that last sentence. There are indeed many more things you have to worry about with a helicopter, and the level of detail can also vary wildly depending on the level of depth you want to go to. For instance, you could get a decent simulation by providing each blade a flapping degree of freedom and assume the blade is otherwise rigid (this is what DCS does, I believe). You could then add a lag degree of freedom for each blade, and you could also start to include blade flexibility, which means a degree of freedom for each elastic mode you wish to include. And then you have to model the inflow. Most real-time simulations will use a 3-state inflow model which approximates the induced inflow across the disk as a uniform plus lateral and longitudinal gradient. For non-realtime analysis, freewake codes have become popular. These model the wake as a series of vortex filaments that are free to move and distort based on the influence of other vortices within the wake. Modeling the rotor inflow and flapping degrees of freedom are probably the most essential aspects of making a helicopter simulation behave correctly. There is a noticeable time delay associated with putting in a cyclic command and the aircraft response due to the time required for the blades to flap. A simple 6DOF linear model will not capture this, and the discrepancies become significant at higher frequency inputs. On the other hand, however, there's no reason why you couldn't include the rotor flapping and inflow states as part of your linearized models. These would likely have to be derived from an analytical non-linear model, however. Lastly, I will say there is absolutely still a place for these linearized models in an engineering context. Control system designers rely on them extensively, and simplified (inverted) linear models may even be used in the control laws themselves to predict the appropriate inputs to make for a given command before waiting for feedback. This is called "model following control."
Excellent critique. I hope the devs take it in the spirit in which it was intended I.e. constructive criticism. Incidentally, l would be keen to hear which particular pc sim helicopters ‘raise the bar’ in terms of flight modelling, in your opinion? (Whether they be on DCS or another simulator). Enjoyed your video, thank you 🙏
I really love this mod. I really enjoyed the video. You covered a lot. And they still have the warning panel on the dash clipping which is all it is it’s clipping so they need to just raise it a touch and it wouldn’t so add .2 maybe to the proxy in the code. So I am hopeful it’ll all get done if it’s to be released because the flight model is very good for what I was hoping a lumpy stable work horse how true that is to real life one can hope but in game at least it’s very good taxi’s well I love flying it personally.
Great breakdown, very informative! Any plans to revisit v1.2 and highlight the changes and improvements, maybe as a regular incremental series, as the mod evolves?
I've bought, but haven't learned/used my heli modules in DCS yet so this was just generally very informative for me in building up to flying any of the whirlybirds. They still seem like stones waiting to fall compared to a fixed wing aircraft to me though 😄
Yeah, I timed the release of this video pretty poorly considering the 1.2 release just came out. I haven't tried the 1.2 release yet, so I can't comment yet. I don't think I'll be doing a video on it, but I'm interested to see what's been changed. I do know they addressed the field of view out the chin window by reducing the scale of the front panel. I don't know if other geometric errors have been fixed yet. I did send their 3d modeling a bunch of my own photo references recently, so I do know things are being addressed. Regarding the torque, I think there may be various kluges and fixes that can be addressed to match the performance in steady cases, but I think it's still going to be very difficult to handle the dynamics properly without moving to a fuller-fidelity non-linear model. I'm not involved at all with the development, however, so that's all just speculation on my part.
Hi Erik, I really appreciate the level of specificity you bring to these videos. I would really like to know what you think of how to Huey is modeled in DCS. It's one of my favorite modules, so I'd really like to know if it is accurate to actual flight data.
So professional and concise! Would be awesome if you also reviewed the top msfs helis, like the r44 and the h145, who are said to be the most realistic in msfs
Would you like a US Army Maintenance Test Pilot flight manual for the UH-1H? I believe I have a copy in a box somewhere with data on how the aircraft should perform, as well as autorotation data.
Had a giggle at 12:50. Anyone who's made even a somewhat serious simulation model of any kind of vehicle will resonate with this. For me in particular it's road and race-car modeling in a semi-serious amateur context and exactly the same applies. Pilot feedback about transient behavior is worth squat most of the time if their hardware isn't up to snuff.
I think a position as SME is more what you could do with such a project. Well, with this video you already did it, kind of at least. I am also a SME for a DCS project I was working on IRL. You should be able to know better than anybody else what you are allowed to tell and what not. Of course I can fully understand that you don’t want to Modell or even code, I wouldn’t do that as well. But it’s exactly what the former pilots do in these DCS projects. The give hints, tell stuff, etc., but they know what they are not allowed to tell, as I am in my project.
I'm really struggling to get this thing to hover - yes I've switched off the TCS (I think it's called) but I'm still having issues. The biggest thing I can't seem to manage is to slow down to land. I try to bring the nose up and decrease collective but the speed dosnt come off very quickly.
Enabling the FCS should help you hover, but you need to hit the trim release to have it capture the aircraft attitude. But still, hovering with just the SAS isn't that difficult (though it seems they've changed the SAS a bit in the 1.2 release). Hovering helicopters generally just takes practice and developing a bit of feel. There is a lag associated with your inputs. The UH-60L captures this somewhat, but in real life there would actually be a slightly longer time delay associated with the blade flapping that really can't be captured with this simple model. The point is, there's a delay between when you deflect the stick and when the aircraft responds, so it can be easy to overcontrol. What I recommend is if you need to make a correction, put in an input and then immediately take half of it out. With practice, you'll end up doing this naturally, and you'll learn to anticipate the aircraft's movements and make corrections seemingly before you even consciously notice the aircraft drift. I also highly recommend focusing on a distant object when hovering while tracking closer references in your peripheral vision. this can really help you control your attitude and prevent you from over-controlling. Focus initially on arresting translational velocity rather than holding a position. once you master hat, then move on to hovering above a chosen spot. Regarding deceleration, you again have to be patient. You can yank the nose way up, of course, and this should increase drag substantially, but doesn't with this model. Regardless, slow decelerations is what you want to start with, and is what you should be doing under normal circumstances. Start by flying down a runway at about 100 knots or so and pitch up a few degrees from trim, hold altitude by managing collective and see how far you drift down the runway before stopping. Do this for a variety of initial speeds and pitch angles and you'll start to get a feel for how it decelerates. Once you start to feel you can control the deceleration, then work on decelerating to a hover over a specific spot. Overall, it takes patience. I did say in the video that this is the easiest helicopter to fly in DCS, and I stand by that. However, it is still a helicopter and is still difficult to fly when compared to fixed-wing aircraft. It will take practice and requires developing a feel for how helicopters behave. It you want to really get started with helicopters in DCS, I recommend the Huey module. It's a bit old, but it's probably the most bare-bones helicopter experience there is in DCS. And it's a great way to learn and develop your skills.
I find a constructive and helpful analysis good! But I also think that a free model should not meet the same requirements as the overpriced paid models! For a real flight sim I agree with you, although I don't think they would listen. I learned to drive in a battle tank sim, and the images of the terrain were extremely bad! And this sim cost millions! take X plane for example how bad are the terrain and buildings there!
If I was the developer of the UH-60 mod, I would value a video like this very highly and see it as an opportunity to make what I'd done already, even better.