f3-res Mad Hatter F3-RES

[martynk]

Sticking my neck out a bit again..

Those following the F3-RES discussion in "Thermal and F3J" may understand my enthusiasm for the class. I have had my thinking head on for a couple of days and this is my contribution to the class (so far and will probably be amended as we move along).

Details are:

2m span  (just under)

AG35 wing section

Carbon boom and vladimir mount from Hyperflight

Carbon caps in the wing - top and bottom - but no Kevlar wrapping

Tailplane is scaled down from Bubble Dancer (again).

Fin and sub fin is scaled down from Cheshire Cat

Root Chord is 235mm straight tapering to 155mm at the wing tip

Target mass is 0.75kg - possibly lighter if I am extra careful - wing loading about 190g/dm^2 with the capability for lots of ballast if needed

a few numbers.

Any thoughts on this?

 

 

[PeterT]

Well Martin, the 3 view looks very nice to me even though I have an aversion to tapered wings with every rib different It's all right for you laser cutters I suppose!

Is the 750g target weight correct, as most of the models appear to be sub 500g? Other than that, the maths is well above my labour grade so i'll take your word for it

Peter

[martynk]

Tricky one. I would like to get it down to sub 500g - this is A/2 (I remember and used to fly those before they got posh) territory but with more wing area and the handicap of radio gear as well - albeit this shouldn't add more that about 50gms if I am careful so it should be achievable. However, I have forgotten how to build light..

Its worthwhile trying for but I am also being a bit realistic. TBH, in UK conditions if I even managed to get down to 500g we would probably end up ballasting up again anyway - even on a 'nice summers day' - whatever that is..

[martynk]

I forgot to mention. Andy Ellison has cast his eye over the design and thought the nose was a little too short (currently 250mm).  I'll probably extend this by another 75-100mm - that should help keep the weight down a bit. Fuselage is actually drawn in error as well - its too wide by a factor of 2.

Regarding tapered wings and ribs. I am totally sold on laser cutting for models like this. I have never managed to successfully cut a set of ribs using the 'sandwich' method and the demanding requirements for accurate wings are such that I could never build to the accuracy required using home cut ribs using a template - as I used to do in the old days (and still do for some sports models).

While I was lying in bed this morning at about 4:30am (insomnia) I started to think about reducing the width of the spar caps (and the associated spar/webs - again to save weight from 10 x 0.5mm in the centre panels to 6 x 0.5mm and from 6 x 0.5mm to 6 x 0.25mm in the tip panels. Don't want to end up with a too fragile model though. I have more than my fair share of self induced clumsiness to contend with.

[Whitmore]

Looking good Martyn.

Just a couple of thoughts, but bear in mind I'm also feeling my way on 2m design!

Your EDA looks low to me, based on the single V dihedral models I've designed (all DLGs) a minimum of 10 degrees or even 11 would give you better turning performance working small lift. It will be controllable as is but may be a bit slack in yaw. I'd suggest looking at the VvB parameter in Curtis Suter's Sailplane Calc (a new version is available but I haven't checked it out yet.)

Airfoils - AG35 is a good root airfoil, but will be thicker than necessary at the tip. If you taper the airfoil from AG35 to AG38 you will get a slightly better sink rate and a real improvement in penetration. With a straight tapered planform, the airfoil blending is very easy to do in XFLR5 - happy to help here. Tapering spars also makes sense.

Your aspect ratio is on the low side (10.25?) but quite ok if you expect a bit of weight. 11-12 is about normal for F3-RES. I've investigated higher aspect ratios but above 12 doesn't perform as well as you might expect - you start to run into Reynolds number limits. It depends on the airfoil of course but high aspect is not always better in a span limited class.

When you get a finalized design, send me the details and I'll run in XFLR5 alongside my Halo and the PuRES etc

Jon

[martynk]

Thanks Jon

EDA - I spent a lot of time trying to sort out the EDA - like you I thought it looked low as well. I use Martin Brungards EDA Calculator spead sheet.  Increasing the EDA (in any way) up to say 10deg results in the Blain Rawdon Stability Parameter increasing to 7.5 - optimally I think this should be just over 5 and its currently sat at 6.. This can be reduced by reducing the Vv value - moving the fin in front of the tailplane or reducing the moment arm or significantly increasing the span (which we cant do). Its a very iterative process and I am not sure that I have got it right but those were the best numbers I could come up with. If you look at the front projection, it doesn't look too bad  (TLAR overrules anything else I do)

The wing section. My original plan was indeed to taper from AG35 > AG37. 

I have also had a close look at this. I use Profili for analysis and I do not profess to be an expert in any way. I analysed AG35 and AG37 between Re 60000 to 120000 and the results were a little surprising. Cd for both sections is very similar with unsurprisingly, the AG35 being more draggy at low alpha but very similar Cd at higher Re. But if the alpha increases above 5, then the AG37 becomes much more draggy at all Re whereas the Cd for AG35 remains fairly flat until alpha 7 then rises more slowly. As I have learnt (the hard way) in the past, thin section stall more easily and the AG37 shows a (tip) stall developing (across all Re) appearing at alpha 8 to 9 depending on Re.

The downside of this is that Cl for the AG35 is marginally worse across all alpha from 0 to 10. However, if you want to fly a model fast (to escape from sink) then I think the AG35 is a better performer. For still air gliding at lower Re then the AG37 wins providing you don't stick the nose up too much

Agreed on the Aspect Ratio. I think you need a big jump in AR to gain any real performance benefits. I have opted for increased area instead and a model that will operate at higher Re where the AG35 performs slightly better.

I am more than happy for anyone to tell me that my analysis is a load of codswollop..  I am trying to learn..

 

 

[Whitmore]

Martyn,

Firstly on EDA, dont worry too much about spiral stability, a lot more than 6 is not necessarily terrible. Small R/E DLGs tend to have high B parameter values - the Elf is 8.3 and my Vela flies happily at 11.4 with no hint of dutch roll, albeit with rather higher Vv (0.43 on the Vela) than RES designs.

These high B values occur when you put a conventional cruciform tail layout on a low aspect wing. Horizontal tail moment arm (lh) can be thought of in relation to mean chord (3-4 times the chord is normal) but Vertical tail moment arm (lv) should be related to the span (something like 0.3-0.5 times span) Like you say, to satisfy all the parameters you need to make the positioning of tailplane and fin independent of each other. A high aspect glider wants an lh rather shorter than lv so the tailplane is put ahead of the fin. As aspect ratio decreases this flips round at some point. In practice you can have a conventional arrangement (more practical?) since the B parameter is fairly flexible.

EDA is directly related to the smallest radius of turn possible. EDA sets the 'non side slip' bank angle for a R/E model, turn any tighter than this and the model will be yawed and just sink rate increases due to the extra drag. More on this at the bottom of this page: http://ocw.mit.edu/courses/aeronautics-and-astronautics/16-01-unified-engineering-i-ii-iii-iv-fall-2005-spring-2006/systems-labs-06/spl8.pdf

Considering this, I'd be inclined to satisfy the EDA requirement (my Halo has 12.5 degrees) rather than the more flexible B requirement.

It's probably also a good idea to give yourself some flexibility on fin size, perhaps a rudder that can be swapped out for a bigger or smaller one. This will let you tune up the handling when flight testing.

Jon

[PeterT]

Quote

A high aspect glider wants an lh rather shorter than lv so the tailplane is put ahead of the fin. As aspect ratio decreases this flips round at some point. In practice you can have a conventional arrangement (more practical?) since the B parameter is fairly flexible.

Jon, I'm a bit puzzled by this. I don't understand the calculations, but I can compare what you are saying with  F1 class free flight designs, where the aspect ratios are much higher and the models pretty much without exception have the fins in front of the tailplanes. Have I misunderstood something? Not being controversial, I'd just like to understand

Peter

[Whitmore]

Peter, thats a good point! And also where aspect ratio is lower, ie FF power models, it's the other way round

All I can say is that free flight trimming probably has its own logic going on here and the requirements for locked down and VIT models will vary when compared to RC. My initial thought was ease of DT and avoiding blanking the rudder in highly banked transition but I'm not sure now.

There was this discussion if you remember: http://www.hippocketaeronautics.com/hpa_forum/index.php?topic=12918.0

One thing I've learned is there are large differences between RC and FF glider design parameters and while they look similar they are doing very different tasks aerodynamically. I do use the same formulas and checks but I've had to work up some quite different numbers from those suggested by Mark Drela for radio control models.

[martynk]

Thanks Jon

That is an excellent explanation. I'll get back to the spreadsheet and do some more tweaking. 

Regarding Freeflight models, the size of the fin is very much smaller than those used on Radio gliders. Directional (yaw) stability is much less and I remember that a good freeflight model should tighten nicely in lift and go looking for lift in a large open wandering circle when the air was not so good. That doesn't really answer the question though

[martynk]

I've tweaked the dihedral, lengthened the fuselage nose and corrected a couple of other minor issues. Also added the spoilers.

How does that look. I have also added a dimensioned PDF version of the plan if you want a closer look

I feel quite comfortable with this at the moment. Any other comments?

Thanks again.

Mad Hatter Wing Plan.pdf

[Jef Ott]

Hi Guys,

I must have read the rules wrongly, but I thought carbon booms were not an option. 

Hope I did get it wrong, because there are a few carbon fuselages in my collection awaiting re-tasking, and I have some built up 6ft wings that would be good on windy days.

Yes the 2m RES looks like a good leveller. 

Best regards,

[martynk]

Carbon or other Glass composite booms are OK Jef. Can't protrude further forward than 50% Wing MAC. 

 

[Jef Ott]

3 hours ago, Whitmore said:

One thing I've learned is there are large differences between RC and FF glider design parameters and while they look similar they are doing very different tasks aerodynamically. I do use the same formulas and checks but I've had to work up some quite different numbers from those suggested by Mark Drela for radio control models.

I don't know diddly-squat about theory, but when an RC DLG model, with a Drela section wing is trimmed perfectly, it will free flight superbly after the initial bunt to flat slow turning.

 

[Whitmore]

Looks good Martyn. The H tail moment arm is on the long side at 4.3 times the mean chord, not excessively so though. It's within bounds and there is some flexibility anyway. It's personal choice to some extent.

Arguments for longer tail moment arm: Going longer improves dynamic stability in pitch and increases B for the same EDA. It also means areas can be reduced for the same Vh and Vv, so less drag, especially in penetration.

Against: Putting the mass of the tail further out requires more nose weight or a longer nose, so means more weight overall and more inertia. This nibbles away at the aforementioned dihedral and stability benefits (inertia is not represented in any of the tail sizing checks but has an effect.)

...so it's a balance.

 

Regarding airfoils, your Profili analysis is worthwhile but you also need to analyze the wing in 3D. Profili runs Xfoil to give you 2D 'wind tunnel' results of the airfoil. XFLR5 allows you to model the 3D wing, select its airfoils and compare sink rates, penetration, glide angles etc. You can also model the tail surfaces to model the whole plane.

This gives you a much more powerful tool to assess airfoils. What you will find it that when you consider lift distribution and downwash across the wing, a thinner section at the tips (AG37 or 38) provides real benefits, as does tweaking wash out for best efficiency.

Your concerns about tip stall are valid, but you need to consider the wing as a whole. The AG35-38 family are designed to work together and the thinner sections are appropriate for the lower Reynolds numbers at the tips.

Tip stall can also be assessed in XFLR5 because it shows you lift distributions and airfoil transition behaviour.

When you are ready I will run the Mad Hatter in XFLR5. It really is a powerful program and worth getting to know.

Jon

[Whitmore]

Jef, yes an RC model can be trimmed out to fly hands off the difference is that a FF model bunts (or transitions) on its own, tightens its turn when it finds lift, and can fly both very fast (climb) and very slow (thermal) with no intervention from the pilot.

This requires some quite different numbers in the parameters we are discussing here. EDA of 21 degrees is normal on a catapult model for example. And Tail volume of 1.0 or more is normal on a locked down model.

My point is just that aerodynamically speaking each discipline has a specific task. Understanding that in aerodynamic terms is the key to optimising designs. That said there are often structural, practical or class rule based reasons why things are done a certain way. Or just fashion

Jon

[Jef Ott]

2 hours ago, Whitmore said:

the difference is that a FF model bunts (or transitions) on its own, tightens its turn when it finds lift,

Now that is clever.

[martynk]

Been playing with the numbers again for the past few evenings. I have also imported the model into XFLR5 but having problems with the calculations at low Re on the tailplane and wing tips.

Anyway following a few suggestions, I have reduced the tail moment and increased the tailplane area, personally, I now think it 'doesn't look right', but the stability numbers are looking OK. I am used to longer tail moments and smaller tailplanes on free flight models I think

 

The nose also looks far too long now but it wont need much nose weight..

Your thoughts please?

Mad Hatter Wing Plan v3.pdf

[Whitmore]

Dont hate me Martyn, but I quite liked the long tail version it had a bit of the DLG look about it

I wouldn't worry about nailing specific numbers too precisely, rather just understanding where you are on the 'spectrum' for each parameter, and then weigh your expectations against reality when you start flying it.

If you lay out some aims ("I want good speed", "I want to work small bubbles", "I want a high launch") and then adjust the design to target those aims. The more clearly you set the task (aerodynamically) the easier it is to settle on a design compromise.

Jon

[PeterT]

Quote

Dont hate me Martyn, but I quite liked the long tail version

Moi aussi. If 'TLAR' counts for anything it's the previous version for me.

Peter