Uhhh - well in front of the leading edge???

The author claims that this works for canards, I think:

http://www.zenithair.com/kit-data/ht-90-4.html

I haven't built a canard, but if I do I'll make a little foam or balsa glider, scaled down from the plans until it's around a 12" wingspan.  I'll keep trimming and adding nose weight until it's stable.  Then I'll figure that because its for CL, the "real" cg is going to be between that point and about 20% of the wing chord forward of that.

Try it -- if I'm wrong you can either tell me here, or you can proceed confident in knowing that at some point I'll mess up exactly the way you did!

There's a way to do this mathematically that jibes with setting the CG on a conventional-wing RC plane.  But I've never tried it out on a canard for anything but gliders.

When in doubt, tether the model at the tip and swing it around your head on a reasonably short string. Move the CG until it tracks true. Reality trumps theory EVERY TIME.  This is how I trim my weird configurations! 

I tried once with a canard configures RC glider.  I did area calculations and at least it glided (more like fell with style) fairly level till it landed safely.  My number was well in front of the main wing.  IIRC I assumed the point of lift in the 20%-25% of wing chord region for both wings, then averaged from a fixed point like the nose.  Bunches and bunches of trapezoids that were all calculated separately too!  I haven't tried another canard.

I credit the overweight brick outhouse nature of my building as the main reason for the failure to preform.  This was 20 years ago, and I was a teenager at the time.  I stripped all the hardware and electronics out of it long ago.  It might look pretty cool hanging as artwork, but I doubt my wife would be impressed.

Phil

I haven't built a canard, but if I do I'll make a little foam or balsa glider, scaled down from the plans until it's around a 12" wingspan.  I'll keep trimming and adding nose weight until it's stable.  

That's what I did.  I never got around to making the actual canard stunter.  If you want to make a stunt plane, try the glider with deflected controls and see if it has enough lift capability to do a loop when scaled up.

There's a way to do this mathematically that jibes with setting the CG on a conventional-wing RC plane.  

Or any airplane, for that matter.  Canards came up from time to time where I worked.  Last time was about 2000, where we spent a year working on an airplane that looked like a duck. Artificial stability makes canards practical for some full-scale applications now, particularly airplanes that don't have to operate with a wide range of CGs.  One of my favorite ideas is the floating canard, which provides lift to balance the wing pitching moment without the usual destabilizing effect.  

The one canard R/C glider that I designed required about 10 deg incidence in the front stab before it would fly.  I think this is typical.

One tid-bit of information that I've heard that applies to canards is that when you locate the CG so that the plane will be stable when trimmed, the front surface, whatever it may be, needs to operate at a higher coefficient of lift than the rear one.  This presents a challenge to the designer of a canardly airplane, because with a conventional tail you don't lose much by setting things up so that the rear surface has zero or negative lift, where with a canard you have to start with the lift coefficient needed in the back, then make more lift in the front.

Bob,

You're killing me!   

I had a tough time finding this math, but I dug it up just for you.

Well, actually I'll need it one day because I expect to complete this model.

My Aviojet jet drawing of the wing and the math to find the correct CG. Hardly readable, but you can get the idea.

Did this a long long time ago in a galaxy far far away.

 Charles

That, Larry, is quite good!

A topic of papers from some decent aerodynamic minds, it's not so difficult to show that without computer enhanced "reflexes", stability demands that the main wing of canard craft operate at less than its maximum lift. This and destructive interference reduce performance to below that of the conventional trailing-tail configuration. Three (or more)-surface planes may be something else again.

SK

I ran across this earlier today in my files  - from Stanford. I may have posted it before.

SK

I ran across this earlier today in my files  - from Stanford. I may have posted it before.

That is a cool plot.  It took me awhile to see that it's all possible combinations of fore plane and aft plane sizes.  There must be more to the story.  What got held constant?

OK...

I had just posted the graph for those who seemed to care, because it was available in my file. 'didn't realize that it would draw interest. I'll try to make the thing more palatable; maybe the attached will help some.

SO...This is a graph of relative maximum lift coefficients for various canard and tailled configurations, ranging from tiny "pure" canards to conventional configurations with tiny tails. Typically, when the graph lines are highest above the bottom horizontal line, the lift coefficient values are highest, and for lower heights you have lower lift coefficients, etc. The lift coefficient is a value that indicates how high the total lift is for some chosen wing area, speed, and atmospheric conditions. Lift for any plane is proportional to its lift coefficient. So you can just assume "equivalent" planes (in areas or drag - not sure which here) and  see how well they lift for different wing/tail (or canard) proportions by seeing how high their graph lines are. They're just comparative. The horizontal scale across the bottom compares forward and rear wing spans as the front wing rises from zero span and the rear wing reduces to zero span, left to right.

I don't have the original full paper at hand, but I wrote down on this graph page that for these data, the static margins were optimized. That means in simplest terms that the c.g. has been placed for best stability vs. performance - the original topic of this thread. Total area of forward and aft wings may have been maintained maintained as equal, but they may have gone with constant induced drag. I don't know.

Anyway, the scale across the bottom records the ratio of forward (canard) wing span to rear (main) wing span as it progresses from 0.0 to 1.0 at the graph's center, where they are the same span (ratio = 1.0). Past that, as the front wing becomes the main wing and the rear one reduces to a tail towards the right, it gives the ratio of of rear (tail) span to Front (main) wing span, until the tail span becomes zero. They switch nomenclature and ratios to avoid dividing by zero and having to show ratios approaching infinity. "b" is almost universally used for "span" and shows up here as span. The three graph lines correspond to three chosen ratios of forward and aft "wing" aspect ratios. On the left, from top to bottom, the ratios of canard to main wing aspect ratios are .5, 1, and 2 respectively. The lines are continuous across the center, but the ratios then become  inverted so that the smaller is always compared to the larger span - tail to "wing" on the right. So on the right, the aft wing (tail) aspect ratio is now in the numerator. That's confusing in words, I know, but you can see that each graph plot (line) represents the same continuing ratio of fore/aft aspect ratio. It's just that the canard has morphed into a main wing, and the original main wing has morphed into a tail.

SO... the graph just shows that as the front surface increases in span relative to the rear, total aircraft's maximum lift increases until it is no longer a canard, but a main wing and the aft wing has been reduced to a tail. Best lift comes torward the right where the plane has the more conventional aft tail, rather than canard configuration. I've tried to illustrate that in the picture below, but it was done with Microsoft Word drawing tools which do not allow great accuracy in sizing or positioning of anything. I hope this helps with understanding the original graph.

SK

   Do you know why they call them "canards" Because..................................................................................They "canardly" believe that the damn thing would fly!! Kind of like a carardly diamond, so smal you canardly see it! 

   Howard already kind of touched on it........ I would call 1-800-RUTAN for the answer!

     BAck in my free flight days, there was a guy named Roy White in our club who held many, many indoor and outdoor records for ornithopters. They design trend for indoor models evolved into a four winged, alternating flapping canard design. It really allowed the time aloft to greatly increase, and had to be seen to be believed. Talk about flying works of art! One strange quirk was they were almost imossible to get trimmed into a predictable flight pattern. He could try every trick in the book to get one to hold a circle in either direction, and it might start out the flight like he wanted, but it was like they had a mind of their own and would often just start circleingthe opposite of when it was launched, then go back again!
   Type at you later,
  Dan McEntee

Howard, and anyone else listening -

This material belongs in the engineering section, but since it is in direct sequence with the above, I'm posting it here. I don't want to hijack this thread further, since my first post was only partially related, but I'll fill in the report info here. The parameter held constant for the CL_Max graph was the total wing area (both surfaces). Method was from vortex type code they developed for this work that was supposed to reduce memory needed for a computed analysis. Rather than trying to explain, I've scanned the relevant parts of the paper and posted them below, with  appropriate graphs - except for the one already posted above. I had mine numbered wrong and had to re-examine them. The one already posted is Fig. 10, which I had mistakenly labeled as #11.

You'll see that the canard can only nearly achieve the low drag and higher lift of the aft-tailed configuration, if stability constraints are removed. Also, it is quite sensitive to aspect ratio, and what is best for low-drag is worst for maximum lift. I still think that the graph previously shown (fig.10) is for optimized static margins. Howard, please correct me, if I have misunderstood that. Below, in order, is the text (followed by graphs), relating to trimmed drag and CL for these two-surfaced computer models. They total 840 Kb; so they'd better fit. - SK

Thanks, Howard. I should have thought of that. I forgot that the "figure number" placeholders in my print-out had actually been links. So you get the whole thing in order. Much better! Perhaps I should delete that post and save the forum some capacity. I still haven't figured why the lift disappears at the extremes, where an entire wing remains. With constant area and A/R, there must be an easily computed finite span and MAC. This still doesn't help with the c.g., although static margins for these are computed in the program. I'll just shut up now.