Hello guys,

I am opening this thread regarding the last article I made in RCJI.

I know that it is a little bit technical. So here is an interactive way to understand the concept better.
Just ask any question that you came up with while reading the article and I'll be happy to answer it here ( if I can ).

Also feel free to submit cases and examples here at your convenience...

Sounds Interesting! I will have to rush out and get a copy.. whats the subject?
Andre

I guess I need to renew my JCJI subscription! I let mine lapse [&o]

Good luck with that!

Flight control surfaces simulations, servo requirements, test protocoles...

Sounds good. I think more tech articles, less politics is just the thing!

Bob,
BTW I have had my RCJI lapse many times due to my own forgetfullness. Everytime I renewed I got my mag on time without fail, dont understand why USA should be different. Then again post between UK and SA has always been excelent?
The news stands here stocks 1 issue old (late) at double the price.

Very useful article, thank you Olivier. I have a question for you. Ys and Yc are distances from the "chord line". What is the chord line? If in a wing do you mean literally the wing section's chord line? The diagrams of the linkages suggest that the chord line is a reference line tht passes through the centre of the servo output shaft, and the hinge line of the control surface but if that is the case then Ys must by definition always be zero so why include it as a variable? So what exactly is the "chord line", is it simply any reference line that the user chooses?
Harry

I am hopeful that the Traplet Publications bottleneck in what is otherwise an outstanding mag has been unclogged. Previously I had terrible dealings with them and only put up with the hassle because the contributions made by Oli, Tom, Dave, and others was just too valuable to pass up.

But I recently had such an outstanding customer support experience that I have been won over and have made a point of saying so publicly.

IMO this is one mag you shouldn't do without. Just one hint or tip from Oli and his peers can save you enough money to literally pay for a lifetime subscription MANY times over.




Oli, if this tech thread is even half as informative as your previous one regarding fuel traps, then you will have once again made a very meaningful contribution to RCU, as well as to RCJI, and I am looking forward to reading every post !!


Don Ray.

Many thanks Don,

This thread will only be as informative as you guys want it to be. Ask as many question as you can. I will try to answer them. If I cannot, there will always be someone around who can.

Hi Harry,

The chord line is a line that passes across the airfoil section from the leading edge to the trailing edge ( flight control surface set at zero degrees ). It is the white measurement line in the graph below:



Here are the definitions again:

Xs: servo distance forward of the hinge line
Ys: servo distance above/below chord line
Xc: control horn distance aft of the hinge line
Yc: control horn distance above/below chord line

To compute Ys, I make am approximation here: I measure the wing/surface thickness ( skin to skin ) at the servo axis location and divide it by 2. I assume that the chord line passes by at this distance. I then check the distance of the servo axis from one side of the skin and deduct Xs. There will be a few mm mistake in case of a non symmetrical profile, but this is non critical.

Yc: The program assumes that the chord line and the hinge line are superposing. This is not always the case ( live hinged surface on a thick wing for example ). In that case you need to take the control horn distance above/below the hinge line. Of course, this distance is the measure from the used control horn hole to the hinge line.



In the example above, the distance from the chord line ( white dot ) to the control horn hole differs greatly from the distance from the hinge line ( green dot ) to the control horn hole

The web address for the spreadsheet given in the mag doesn't work for me, all those %20 seem to confuse things as the browser (firefox) kept changing them to blank spaces, I checked again and again that I had not mistyped it. I hunted around Olis website and found it, I was about to copy and paste the working link here (complete with all its genuine %20 !) but stopped in case Oli didn't want it linked to here! So Oli, perhaps you could publish the link here for folks to use rather than have everyone laboriously type it in themselves?
So far I am stunned at how little force the ailerons on my F-100 generate, a tiny servo could drive them! Not that I would try that, I like good chunky gears in any servo on a primary control to help guard against things like flutter.

Looking forward to the next article about all moving tails since most of mine have AMT's as will many scale jets

Harry

Hi Harry,

Here is the link for imperial units:

http://www.geohei.lu/olin/data/model...alc%20imperial

And for metric units:

http://www.geohei.lu/olin/data/model...0metric%20.xls

You're corresct, %20 stands for blank spaces. I use firefox exclusively.

My ftp is here to be browsed...

The next article is much more technical.

Here is a sample of what can be done to simulate all flying stabilators:

http://www.geohei.lu/olin/data/model...03D%2070ms.avi

http://www.geohei.lu/olin/data/model...03D%2015ms.avi

This is useful to confirm the position of the CP depending of the AoA. The excursions of the CP will dictate a position of the pivot point. As usual it is a trade between flying surface stability and servo efforts.

The CP position is shown by the big green arrow and the XCP vlue at the bottom right.
The lift field is shown in green.
The viscous drag field is shown in magenta. Inviscid drag is not represented here for more clarity.

Oli,

I used your servo torque calc spreadsheet to check the servo requirements for my Viperjet. Nice tool you have built there. Really appreciate you making this available publicly as well as offering support.

PaulD

Thank you Paul,

Just as a reminder, the spreadsheet was designed by Craig Tenney. He's the one to thank for the hard job and for giving it to the community.
I just modified and adapted it for our use, and made up the metric version.

If you have any question, please feel free to come again here.

Oli,
Dont have the article yet but I have seen methods like this before. A very interesting subject though. I wonder if anyone has ever verified these servo loads with a direct measurement, ie a force gage. I suspect that most of the digitals we use today are more than enough for our applications. All flying scale stabs maybe an exception, where the hinge line was designed for supersonic conditions maybe?
Even logging precisely the individual servo current throughout a flight together with roll rates, g etc may give a clue?
Andre

Hi Andre,

I have a servo test bench using my laboratory dynamo meter. All the servos I have tested are making the torque claimed by the manufacturers initially.
When the servo warms up, the stall torque drops gradually.

I have not yet published the results because it is a lot of work coming ahead. However similar thing has been done before and is available on line here:
http://www.troybuiltmodels.com/categ...00/Servos.html ( check the "TBM test data" and "servo testing" tabs )

I explain in the article why it is worth checking the torque required for the proper application. In summary if you get the right servo size:

1. You'll optimize the price of the system ( I do not talk about low cost application here )
2. You'll optimize the current consumption of the servos ( http://www.rcuniverse.com/forum/m_92...tm.htm#9257664 )
3. You'll optimize the weight of the Rc system ( servo size, connectors, leads, power bus, batteries )

All flying stabilizers are very specific and I wrote a 10 pages article to come in the June/July RCJI publication about this matter ( it will probably become 4 pages after editing ).

I have also been logging the servos consumption in flight using my Eagle three system ( the torque to current relation is linear for most of the range of use with today's digital servos ).
This enabled me to validate the figures found with the method. They are relevant.

Thanx Oli,
That is good info indeed. I look forward to the articles. I think it is commendable that someone goes to this effort on what is really quite an important subject.
Keep up the good work.
Best Regards
Andre

Oli, can you help me with this please? I am getting bizzare numbers and diagrams playing around with the inner flaps tab - I am using that one to get the assistance with what each cell means. Since flaps only travel downwards I always built the horn at a raked back angle but the spreadsheet seems to assume that the horn must be at right angles to the flap and I can't find a way to apply a rake angle. Even worse comes with the servo, the spreadsheet puts the servo arms and control rods at bizarre angles. Is there a box where I specify the neutral rotation angle of servo? The angle of the yellow line to represent the servo neutral angle, as chosen by the spreadsheet seems very odd. I am sure I must have missed setting a servo neutral value somewhere but I can't find any box that works!
Also can you explain what the black line is showing?
Harry

Hi Harry,

The program is finding the offset required for the intended travel. You can play with the servo neutral position by changing the servo travel values.
Here the Rs+/Rs- values.
Instead of the +45/-45 you inputed, I would suggest you to try +0/-90.
The black line shows the 0% stick position.
The rake angle can be simulated by changing the Xc value to what you need.

Let me know if you get what you want that way...

OLi.

That doesn't seem to work Oli, see screenshot below. The black and red lines overlap. The Rs values behave as travel values either side of the servo centre (the black line) so a value for Rs+ of 0 means no rotation in that direction, a value of 90 for the Rs- means 90 degrees of rotation from centre in the other direction. A typical servo rotates 35 to 45 degrees either side of centre, so values between 35 and 45 will be the normal values. What I can not find is a way to set the angle of the black (servo at centre) line. The spreadsheet is choosing an angle for the black line and it chooses utterly bizarre angles that often lead to the rotation and pushrod going over-centre.

Harry

Bingo! I've got it to work, just can't post a screenshot at the moment.

I realised that the black line, representing servo at centre, could not lie horizontal for flaps. It does for things like elevator, rudder etc which travel either side of a horizontal neutral. But when a flap servo is at centre, the flap is not horizontal, it is (roughly) half way down. When the flap is horizontal, like an elevator, it is not at centre but at one end of its travel. Somehow then we have to tell the program that the Control at centre is not horizontal but angled downwards, and that is done in the Oc box.

If you put in these values, it all works nicely:
Xs 8
Ys -0.5
Xc 1 (to rake the control horn back a bit)
Yc -2.5

Rs+ 45
Rs- 45
Rc+ 22.5
Rc- 22.5 (the angle of 22.5 either side of centre gives a flap travel of 45 degrees)

Oc -22.5 (this says the flap will be half down when the servo is at centre, this is the crucial bit that was missing before, a value of 0 says flap is horizontal when servo is at centre, so flap will travel up and down from horizontal, whereas we want up and down from half-way down!)

And now the diagram is exactly what we expect to see.

Harry

I had been looking back at Oli's other thread ( http://www.rcuniverse.com/forum/m_92...tm.htm#9257664 ) and questions from Tassos about the flap, see post #25 onwards. Both Tassos and Oli are running into the same problems I had with flap, and neither solved it. They were using flap and servo travels of zero in one direction from centre, and massive (impossible) travels in the opposite direction from centre, as per Oli's suggestion a few posts above this one of using servo values of zero and 90. It can't work like that, the Rs+ and Rs- values are degrees of rotation either side of servo centre, so zero and 90 means no servo rotation one way, and 90 degrees the other way from centre.

So, rules of thumb for using the spreadsheet:

The values for Rs+ and Rs- are degrees of rotation either side of centre and should be in the region of 40 for both. Not +40 and -40, just plain 40 for both. Using a minus value in one of the boxes will make the servo rotate the same way no matter which way you move the stick! The actual value will be whatever servo travel you have available and whatever travel % value you have set in your Tx, your servo may travel 45 degrees each side of centre, maybe space limits mean it can be 45 one way and 30 the other etc etc. But something in the region of 40 to 45 for both directions is a good starting point. Neither of them should be a zero, as this means no rotation away from centre. Neither of them can be 90, as 90 degrees either side of centre would give a servo with 180 degree rotation, and I am not aware of any that do that apart from retract servos.

The Rc values behave the same way and should not be a zero. Setting flap Rc+ to zero does not mean it is horizontal (i.e flaps up) at one end of travel, it means it has zero travel away from the centre of travel. If you have flaps on a 3 position switch, switch at centre is where the black line on the diagrams should be (assuming you have not adjusted the centres and travels away from default values in the Tx) and the Rs and Rc values are travels either side of this. Clearly you don't want a servo travel of zero or a control travel of zero when the switch is in the up position.

The value for Oc is where the control surface is sitting when the servo is at centre, 0 being horizontal, minus values being angled downwards. So for example an elevator might expect an Oc value of 0, but a flap connected to a servo that is at its servo centre will be (roughly) half way down. Remember that the servo at centre (black line) is where the servo and flap will be when flap switch on the Tx is at mid position (assuming the tx values for centre and travel are at default, you can of course alter these in the tx to give a different mid switch position).

So, to simulate flaps in the spreadsheet, lets try an example. You want a flap travel of 45 degrees down. In that case the centre of flap travel is roughly half that, 22.5 degrees down. So the centre value for the control, its "Origin" angle called Oc in the spreadsheet, is not zero (horizontal is defined as an Oc of zero), its centre value is minus 22.5 degrees and that is the value that goes in the Oc box. Immediately you will see the black line at the control surface go to flaps half down position. Now put in 22.5 to each of the Rc values, and see how the flap travels 22.5 degrees either side of its origin. This makes the flap travel from horizontal, to 45 degrees down. You should not make the flap control horn simply at right angles to the flap, it should be raked back or forward a bit so that roughly when the flap is half down, the horn is at right angles to the pushrod. Do that in the Xc box. As explained above, the Rs values should ideally both be in the region of 40 to 45. Rc and Rs values of zero or close to zero should not be allowed.

Harry

Here's a screenshot to go with the above

I have also worked out what those little numbers at the trailing edge are - they are the distance of the travel in cm or inches depending which version of the chart you are using. So if you know you want the t/e travel to be 2 cm or inches, just adjust the angle in Rc until you see the diagram give you the travel value that you need - of course you must have already set the correct chord in box Cavg, but in order to find the angle that gives the correct travel set Cavg to the chord at the travel measurement point e.g the root, and not the average chord. Once you see what angle that requires, you can set Cavg back to the average chord in order to do the rest of the calculations.

It's astonishing what values are calculated, for example it says that at 150mph the ailerons on my F-86 need a servo torque of, wait for it, 0.7kgcm !! Even with a 50% uplift to provide a margin for error and not strain the servo, that's tiny electric foam model territory, why did I fit expensive, powerful, coreless servos? Because I didn't have this spreadsheet!! A very standard servo would do the job. It says the rudder needs a mere 0.3kgcm, a housefly could pull that! Having this for future building projects will certainly be changing my servo buying decisions.

Here's an example of how to hunt down the reason for some bizarre results and rectify it.

The screen shot shows what is roughly the values for the elevons of my Grumania Eurofighter, in elevator mode. They have a lot of up travel but not much down travel. (In aileron mode they have very little travel at all) As you can see from the first diagram, the position of the servo arm is bizarre. Can you see what is wrong in the values to cause that? I have put in my default servo travel of 40 degrees either side of centre, and the required elevon travel of 40 degrees up and 15 degrees down. That's the clue to the strange result, if the servo rotates the same amount both ways, how can it make the up elevator more than twice as much as the down elevator? It has to use a huge amount of differential, this can happen at either servo or control but the program is constrained to only adjust the servo, because I have defined the control horn offset in box Xc. And that is why the servo arm has been rotated so far round. In reality we would position the servo arm at right angles to the pushrod, and in the Tx we would turn down the travel % for the down elevator to give less travel. So that is what we have to tell the spreadsheet. In the second screenshot, in the Rc- box I have turned down the servo rotation to 15 degrees (as I would by turning down the travel value in the Tx from 100% to something like 38%). Now the diagram looks normal again, as the servo arm no longer needs to have a huge offset angle to provide the necessary differential.