Sine wave spar

[Konrad]

In this thread I want to go into detail as to what a sine wave web spar is and how it works. Listening to guys on the slope there is a bit of confusion as to what it is, how it works and the construction limitations. I'd like to ask if Aeroic has any drawing or a white paper on this structure?

Let's set a base line the I beam and box beam. The I beam offers a lot of strength for any given amount of material (cross section) against bending. (The further apart one can place the spar caps the stiffer the beam is against bending, as I recall this goes up by the square root of the distance). But the I beam performs badly against any torsional loads it is very easy to twist. The box spar performs much better against torsional loads. But if keeping to the cross section (mass) the same size spar caps would need to be brought closed together to allow for the material of the box webs. This is a huge cost in bending moment if one recalls the advantage of keeping the spar caps apart. So most box spars will take the weight hit in added material weight to keep the spar caps as far apart as possible.

Now the sine wave or serpentine web offers the best of both the I beam and box beam. That is with the vertical web going from one side of the spar caps to the other side of the spar caps the sine wave beam is much stronger against torsional loads than the box spar. The sine wave beam is almost as light as the classic I beam as the number of vertical web and the cross section of the web is the same. There is a bit of weight gain against the classic I beam as the effective length of the vertical web is longer.

This is easy to see and measure if one is making the beams out of an extruded monolithic material. (Maybe use a 3D printer to model the beams, and test to destruction).

Once one is confident that the geometry of the sine wave beam is far superior to the I & box beams. We then need to study the real world limitations of the construction of said beams. It is in the construction limitations of these beams that folks get confused as to which beam is superior in our wings when being used as a spar. Hint; it is the sine wave spar if looking at the same kind of construction methods. The problem is that some methods are prohibitively expensive when it comes to the sine wave spar. But composite molding can offer some cost relief but not much.

I'll wait to see if Aeroic has any drawing or a white paper on this.

 

[Konrad]

There is a bit of weight gain against the classic I beam as the effective length of the vertical web is longer.

That statement is a bit mis leading. If the web is a true sine wave the added length is more like 57% longer for the web vs the straight web of the I beam. So the web part of the spar will be 57% heavier than the web for the I beam. But the torsional resistance of the sine wave beam is far greater than the box beam (spar).

 

[Doc James Hammond]

That statement is a bit mis leading. If the web is a true sine wave the added length is more like 57% longer for the web vs the straight web of the I beam. So the web part of the spar will be 57% heavier than the web for the I beam. But the torsional resistance of the sine wave beam is far greater than the box beam (spar).

Hi Konrad, thanks for the explanation - which is pretty much on the ball.

Normally the box spar (ie TWO I Beam Spars) is used for high stress hollow moulded model flying surfaces, because as you explain above, the single beam spar really is no use in torsion, and would probably soon fail unless it was incredibly strongly made - by using a lot of material and therefore adding a lot of weight.

So the box spar is two beam spars that have to be spaced apart by some sort of substrate, that has to be bonded to the carbon webs so as to give enough distance to cope with at least some of the torsional loads - but like many aeronautical solutions, its a tradeoff. Also, being a twin beam construction its already likely to be heavier than the single beam Hammond Sine Wave Spar (HSWS). In addition, Konrad's 57% weight gain assumes that the HSWS is the same thickness throughout, which it does not need to be, and in fact never is. Of course the higher the aspect ratio - ie: model gliders - the more critical this strength/weight tradeoff becomes, so tapering becomes a very attractive idea.

Of course the frequency of the parabolic wave on the HSWS is also critical, as a very low frequency wave and low amplitude would detract from the torsional performance, whilst a very high frequency arrangement is not needed and would use too much material/add too much weight. Oddly enough for our wings, the frequency wave and amplitude of some kinds of plastic corrugated roofing is ideal at about 50-60mm amplitude, and 100-120 mm peak to peak.

In an ideal situation it is eminently possible to vary the frequency of the HSWS along the length of the wing to follow the elliptical arrangement that is ideal for fast model gliders, ie use a variable construction with higher frequency/less amplitude towards the tips. I have experimented with this but frankly the advantages do not outweigh the difficulty in making them so I have opted for the constant wave type.

Because the HSWS, although monolithic, is actually quite flexible until it is bonded into the wing, the spar does not need to be straight, and the position does not have to be in a straight line, and as long as the spar caps are also suitably placed, can follow the maximum thickness of the wing - another advantage. Added to that, the layers and directions of the plies that make up the spar are extremely easy to vary, making the strength, and indeed the weight distribution "tunable" throughout the length - although the same could be said of a conventional box spar.

Critical is the bonding of the HSWS to the spar caps, and this is best done when the wing skins are green. Fillets at the bottom/top of the HSWS are ideal but not that easy to do - though we are working on a reproducible way to reliably achieve this.

So far it works well, buts with all things there is ALWAYS room for improvement, and thats just what I have been doing these last few months.

I hope that sheds a bit more light on it, but I also understand that for someone with a very mechanically oriented mind the dynamics are easy to visualise, but for those perhaps not so lucky in this respect, it can be rather difficult.

Cheers,

Doc.

 

[Konrad]

Doc, Thanks for joining in! Do you have any drawing or a white paper on this spar?

My 57% weight gain comment was for the vertical web of the HSWS only, not the complete spar. This was assuming that the period was twice the peak to peak amplitude using the same units of measure. I need to find a citation, but the sine wave beam is generally thought to be 2/3 as heavy as a comparable performing I beam. (Maybe someone with a 3D printer can make up some test coupons to show this)

I'm thinking of drawing the classic moldie wing with what is effectively 2 "C" channels facing each other to make the box spar. This allows for very effective fiber wrapping as there is no overhang of the spar caps.

Now I stymied by the last Apple OS upgrade in that all my 16 bit and 32 bit drawing programs are bricked. I know you and I understand what we are talking about. But my aim is to help (show) others what are the advantages and limitations of the spar system in our models. Hopefully most can see the the HSWS is the superior one for most applications.

For now I'll be sticking to a simple constant cross section beam for simplicity.

If I can't get a program working I may have to resort to photo graphing pencil and paper drawings. Yikes!

All the best,
Konrad

P.S.
Hammond has said many times that he did NOT invent this beam. Rather he thinks he is one of the first to use it in the application of our toy wings. This is why we will be using the term HSWS. And if you give any credibility to Boeing's engineering, this beam is used through out the carbon bulkheads of the B-787.

 

[Konrad]

Wow, that was a pain! Re learning just some of the basic drawing skills that I haven't used in 20 years! I'm sure there were some easier ways to do this but this is what I got.
I'm trying to show the basics of the Sine Wave Spar. I took the opportunity to show both the tapered spar caps and the progressive web getting stronger against torsion on the right.

Doc. before I go any further is this about right?

I also included the classic I beam and box beam.

 

[Doc James Hammond]

Wow, that was a pain! Re learning just some of the basic drawing skills that I haven't used in 20 years! I'm sure there were some easier ways to do this but this is what I got.
I'm trying to show the basics of the Sine Wave Spar. I took the opportunity to show both the tapered spar caps and the progressive web getting stronger against torsion on the right.

Doc. before I go any further is this about right?

I also included the classic I beam and box beam.

View attachment 5027

Hi Konrad - yep you have got it down pretty much. I don't have any useful white paper or even much technical test info - bar the physical tests that we made, but I'll try to find some photographs. The HSWS for me was one of those ideas, or more correctly re-engineered applications, that having thought it all out I was completely convinced was going to work.

So what happened?

When tested "against" or maybe its better put as "Compared" to a conventional box spar, the HSWS showed up as about 30% of the total weight and gave a 15 to 20% increase in tension/compression stiffness while improving the torsional stiffness by about 75%. Quite impressive.

Note that these simple (weight hanging) compression/tension tests were done using a Redshift wing compared to another well known F3f wing. Test wings were securely clamped at the wing joiner and testing was done to destruction.

The torsion test wings were securely clamped at the wing joiner and tests were again conducted by hanging weights, but this time on a cantilever beam attached midway along and at a 90 degree angle to the bottom of the wings, and extending 500mm to the weight hook. While this is not a really valid torsion test, (Combines tension/compression) it does give a good quick and dirty indication when directly compared wing to wing.

Frankly all the tests showed such a great improvement and potential for optimisation that I was quite astounded.

Cheers,

Doc.

 

[Doc James Hammond]

Some pics: As you can see, another advantage - if you are going into production - is that the HSWS can be made in a large sheet then cut to the correct thickness.

The rather crudely made hand drawing was the original test layup which has since been modified quite lot.

 

[Konrad]

Here I'm trying to show that the shear web not only addresses the spanwise shear (red) but also offers a lot of resistance to torsional forces cordwise shear (green). (The I beam and box beam have no such feature)! Not only are the webs placed in the proper orientation they also offer some geometric strength, in that they are curved rather than flat as we see in the classic web.

I see that Doc. is moving into some of the practical concern with the spar. Every moldie spar* that I've seen fail (other than impact) has failed in rotation (flutter) first. This flutter fails the bond between the spar caps and the web. When this happens the spar caps are free to buckle.

It looks like in Doc's drawing the added webs are trying to add more contact area for bonding. I don't believe that there needs to be any more material to resist shear. But as the failure mode of most wing is in the bond area more webs means more bonding area.

Not sure I agree or understand the 90° orientation of the fibers in the web make up.

*Don't confuse this with wing joiner failures.

All the best,
Konrad

 

[Doc James Hammond]

Good stuff Konrad - I see you really do understand how this works. For some obscure reason it has proved to be a bit of a mind bender to visualise for some folks.

And...Crikey it looks like we are on the same page!

OK lets look at your comments:

Here I'm trying to show that the shear web not only addresses the spanwise shear (red) but also offers a lot of resistance to torsional forces cordwise shear (green). (The I beam and box beam have no such feature)! Not only are the webs placed in the proper orientation they also offer some geometric strength, in that they are curved rather than flat as we see in the classic web.

Exactly.

I see that Doc. is moving into some of the practical concern with the spar. Every moldie spar*that I've seen fail (other than impact) has failed in rotation (flutter) first. This flutter fails the bond between the spar caps and the web. When this happens the spar caps are free to buckle.

Yes this is true, which is why the torsional resistance is so critical.

But I do find that the wing design or rather the planform of the wing design has a lot to do with flutter. Things are changing - maybe I helped a bit, I don't know, but I have been using high aspect ratio reverse elliptical wings for some years, in fact soon after I quit working with RCRCM. At first my wings were described as "Likely to be as tip stall prone as hell" by a well known designer. But they are not.

I didn't see how the planky wing shapes with all that area out toward the tips helped, so I simply put the lift where I knew it was needed. In my opinion the "me too" planky wing is likely to be MORE tip/high speed stall prone that one with optimised lift distribution, since a lot of the wing is at too high a chord to be working well most of the time. Wrong place, wrong time - most of the time.

It looks like in Doc's drawing the added webs are trying to add more contact area for bonding. I don't believe that there needs to be any more material to resist shear. But as the failure mode of most wing is in the bond area more webs means more bonding area.

Got me! The extra fibre is only really to add to the thickness and also provide a kind of angular transition from 45 degrees to 180 degrees. Its vertical in the hills and valleys of the spar so it's easy to do using UD fibre.

Not sure I agree or understand the 90° orientation of the fibers in the web make up.

As above, and maybe a bit of a hark back to the vertical balsa webs we used to bung into vac bagged foam wing spars.

*Don't confuse this with wing joiner failures.

All the best,
Konrad

 

[Konrad]

LOL,

Doc. I'm not trying to "get" anybody. What I'm trying to do is show how the sine wave beam works in the HSWS, at least as I understand it. I also hope to give the designer a chance to discuss (correct) any misconceptions I might have. What drove this discussion was that in my talks with guys out on the slope many don't see the advantage of this beam.

With the constraints of wing manufacture on a production scale, this spar still has the same constraints as the box spar. That is the failure point is in the bond between the web and the caps. But the fact that there is now webbing to constrain the spar from twisting, this bond is under a lot less stress. Meaning that HSWS for any given weight can handle a lot more load before it fails.

Now if one can add some fiber across the bond zone then we gain a lot of performance! With the box spar this is "easy" with some Kevlar wrapping or a carbon sleeve (sock). The HSWS does not readily lend itself to this solution.

Now I'm trying to limit the discussion to one element (part) of the wing structure the main spar. True, the performance of a wing needs to be looked at as a system. For example the poor performance of the I beam can be addressed by adding a second I beam to the wing. But this has a huge cost in weight and material. Or to my way of thinking isn't an elegant solution. The HSWS is very elegant.

I'm just trying to expand everyone's understanding, particularly mine. I pissed off the powers that be with a similar discussion about the short coming of another spar design. Many of its defenders said; but but...

Dumas spar in their Electric R/C kits - RC Groups

Discussion Dumas spar in their Electric R/C kits Scale Kit/Scratch Built

www.rcgroups.com

All the best,
Konrad

 

[Konrad]

I think we agree that the true limitation in the molded wing spar is the bond area of the web to the spar caps. The addition of fibers across this joint will add greatly to the durability of the wing. With the box spar this is often done with a carbon sleeve where the fibers are biased to the span. Also Kevlar thread is often used as the wrap keeping the spar caps in place against the shear webs.

But most molded wing are configured with the spar caps having an overhang segment. This would limit the effectiveness of the wrap or sleeve.

Now if the web could be laid up as an integral part of the spar cap this issue with the bonding area would go away. I see what looks like a start of this in this thread (post#323). The proud fibers just need to be pressed down against the core and the caps added (manufactured in place).

Show us what you are working on..

Got it! Sounds great. Are you doubling the whole wing (chord) to the half way point or just the spar cap?

forum.alofthobbies.com

True this would add to the cost of manufacture. That is in man hours and also that each wing design would need its own molded spar. The up side is that there wouldn't be as much need for multiple webs and there would be a saving in material and weight. With the use of low grade wooden molds or even sacrificial foam molds this is doable. As these spars are internal structure their appearance wouldn't be an issue.

If anyone is interested, one can see the HSWS in the bore scope photos in this thread.

Redshift; Used Purchase

Its been 45 years since I heard a Tie fighter. I must not be in the correct Cl range as my Redshift was all but silent in today's race.

forum.alofthobbies.com

 

[Doc James Hammond]

Hi Konrad - I have been looking at that. Here are a couple of test pieces with some glass sheet attached to the HSWS in order to use it as a kind of mould to extend the webs with fillets etc. Still playing with it but it shows promise.

 

[Konrad]

ALRIGHT!!!

What folks need to know is that there is often a huge leap between what is possible and what is manufacturable, in the real world and still be financially viable. I recall there where a bunch of idiots on the other forums who didn't grasp the basic concept of the spar nor did they understand the limitation of the manufacturing technology.

In case anybody is reading this and just to be clear, I don't claim to have any manufacturing solutions to the problem I just put forward.

 

[Doc James Hammond]

ALRIGHT!!!

What folks need to know is that there is often a huge leap between what is possible and what is manufacturable, in the real world and still be financially viable. I recall there where a bunch of idiots on the other forums who didn't grasp the basic concept of the spar nor did they understand the limitation of the manufacturing technology.

In case anybody is reading this and just to be clear, I don't claim to have any manufacturing solutions to the problem I just put forward.

Alas 'tis always the way.

Rant begins:
What gets me about other forums (RCGroups)is that there are a lot of people there who simply don't get any fun out of the hobby - which after all should be why we do it. Speaking frankly there are many idiots out there, and I think RCGroups garners by far the biggest share.

But...what I love more than anything are the "sages" - the armchair professors with no background, no training, no actual experience, no academic qualifications, and yes, the most amazing - nay, astounding point of all - they are still completely convinced that they know what they are talking about.

I'm also advised that Trolls are not only tolerated but actually encouraged as they boost the readership from the same kind of people who slow down to take a look at bad road accidents. Is any of that why I fly model gliders? To these people I have some sage advice: Te futueo et caballum tuum.
Rant ends.

But, back to the point. I think a large part of my job is to develop advanced concepts and then to develop special manufacturing processes to bring them to practicality. In my "day jobs" I have done this, hopefully pretty well, all my working career.
I am open to, and actively encourage suggestions by my ever faithful evaluators - yourself included - as anyone can see that more brains = more horsepower - ask SETI. But in the same way it's up to me to evaluate any suggestions for improvements and incorporate the good ones - as long as they do fall into the basic formula that you outline above.

Improvement Value
Cost to develop

But in summary I think that anybody who takes the time to read this thread will now come away from it with the basic knowledge of what is a HSWS and why is it useful, and it don't get better than that, Lads!.

Cheers,

Doc.

 

[Konrad]

Damn!

I see that Boeing has done just what I describe and filed a patent in 1996!

US5919543A - Composite sine wave spar - Google Patents

A method of forming composite sine wave spars from composite material. In accordance with the method, the sine wave spar is formed of web plies, filler plies, separator plies and cap plies. The edges of the web plies are cut to form a plurality of teeth having different geometric configurations...

patents.google.com

Maybe it isn't so bad as the expiration date for this patent is 2016. I wonder if Boeing has revised it any, resetting the clock.

 

[Doc James Hammond]

Damn those pesky Aviation companies!

Doc.

 

[Old_Salt]

I ran across the idea of a sine wave wall many decades ago. Jefferson had one at Monticello. Apparently, it is done to save material and still resist falling over compared to a thicker rectangular, straight wall. I am periodically reminded of the idea when I see a venerable brick wall gradually falling over.

Crinkle crankle wall - Wikipedia

en.wikipedia.org

 

[Konrad]

I ran across the idea of a sine wave wall many decades ago. Jefferson had one at Monticello. Apparently, it is done to save material and still resist falling over compared to a thicker rectangular, straight wall. I am periodically reminded of the idea when I see a venerable brick wall gradually falling over.

Crinkle crankle wall - Wikipedia

en.wikipedia.org

OK?
You lost me, not hard to do. What does this have to do with shear webs and load bearing beams?

 

[Old_Salt]

The idea of a sine wave shaped beam or wall rather than a rectangular one has been around for a long time. A way of looking at the sine wall is one that resists torque and twist, hence it is resistant to folding and falling down. The masons used thin sine walls rather than a thicker rectangular, straight wall with the idea of minimizing the material needed. Those ideas have been around for a long time. The idea of using a sine spar can be viewed as an update of an earlier application or as a new discovery with roots reaching back at least two centuries and probably further. I agree that an I beam has minimal resistance to twisting and the sine I beam is an improvement that resists both bending and twisting but with a weight increase. The same features relate to the sine spar in a wing.

The sine spar has been discussed above in relation to I beams and box structures. I wonder how it compares to a tube or conical spar with a very gradual taper. I would imagine that a tubular spar wound have resistance to both bending and torsion. It would not have the change in cross-section that is pointed out with the sine spar. Tapering and thickness variations are also practical. If made from filaments many more options are available (pultrusion vs. filament winding). Tubular structures would not have the profusion of adhesive joints often used in the sine spar. How does a tubular or conical spar compare to the sine spar?

I would anticipate that the sine spar would show irregular response to both bending and torque depending on the local cross section (see #8 AA and CC vs. BB). If the sine spar were to fail I would suspect that it would occur where the sine support was at the edge of a cap rather than where the cross section was an I beam.

This discussion points to developing a spar that has both bending and torsional resistance. An I beam resists bending and has little torsional resistance. How does a cylinder compare to the sine spar? Would an elliptical cross section be another interesting comparison? The balance between torsion and bending could be varied with an elliptical cross-section. Lame curves alone?

Konrad, did this explain why I made the comment and how it is relevant? I thought the historical use would be interesting.

 

[Doc James Hammond]

Spar cap adhesion.

Doc.