R/C gyro unit

I have such a device in hands actually. It is dedicated to controlling heli back rotor, but it works pretty well with any R/C servo. It allows to set gain and neutral by resistor trimmers. Did anyone try to attach it to the rudder of stunt model to balance gyro moment of prop? Now I see the whole device can be done in 1 or 2 oz including battery. What you think?

>I have such a device in hands actually. It is dedicated to
>controlling heli back rotor, but it works pretty well with
>any R/C servo. It allows to set gain and neutral by resistor
>trimmers. Did anyone try to attach it to the rudder of stunt
>model to balance gyro moment of prop? Now I see the whole
>device can be done in 1 or 2 oz including battery. What you
>think?
>
>igor

What I think, Igor, is that you ought to try it. We need someone with good flying skills, engineering background, ability to work up calculations and do trial and error testing. You're it!

I know I will regret this, but I believe this is what will begin to happen if we don't get the "only" controls to flying surfaces and engine must come through the control lines or other "natural" forces.

What I mean is that for sport flying, do whatever you want. But competition will ride a very slippery slope if we start building full house servo controlled stunters. What could be next? How about a micro-processor controled flight?

Ok, ok I know I am overreacting!

Alan

ps, these gyros are used all the time on rc-heli's to correct torques due to engine irregularities by feeding into the tail rotor. I certainly use them for that.

>I know I will regret this, but I believe this is what will
>begin to happen if we don't get the "only" controls to
>flying surfaces and engine must come through the control
>lines or other "natural" forces.
>
>What I mean is that for sport flying, do whatever you want.
>But competition will ride a very slippery slope if we start
>building full house servo controlled stunters. What could be
>next? How about a micro-processor controled flight?
>
>Ok, ok I know I am overreacting!
>
>Alan

Alan, I believe this would already be outlawed already in FAI, which is what Igor flies. And, if so, for him it would merely be an experimental device. Now, one could argue that this is a trim control, only, and that the directon of flight is controlled totally by the pilot at the handle, so...?

Now I am in full agreement that all flight controls (including throttle control, if any) should be controlled by mechanical means through the lines (not electronic pulses, even), although I am still not sure of my own feelings towards a timed shut-off (presently illegal in FAI), and would definitely support a mechanical, timed gear retract system (triggered by the helper at launch) where the gear would remain retracted as long as the engine was running and automatically come down when the engine quits (triggered by absence of sound.) Again, this would not be a flight control, so that all flight control inputs would still be handled by the flyer at the, er, handle.

Allan, I see Len already answered what I wanted to write.

For me this challenge is most important point of this hobby. I want play and try untried. Building, training, contests and all that "serious stuff", is what I do as a “side effect”. We have only 5 flyers in Slovakia and if I wouldn’t do it seriously, we will lost 20% of our flyers (now you see how important I am ).

Nice try Len. Yes, I certainly intend to risk my model, I have the unit, 4,5g heavy servo small battery from telephone and one problem – not enough time, I just wanted to know if there is someone who can tell me especially what NOT to do.

In any case I have million of questions which I can probably answer or test myself, but it is always good to know where to start if someone already know something. Especially I am asking myself if to attach it to yaw axle or pitch axle. … But As I see lack of answers here, I will start on green field and it will be long autumn

...hmmmm... don't forget PA planes fly (approximately) on the surface of a hemisphere

Igor, The trick with the gyro is to have a servo that is fast enough to stay up with the gyro. You can not do this with the light/cheap servos. If the servo is slow it is to late to make the correction. You will want something in the nature of .01 rate(seconds). Most electronic gyros can way out race the servos ability to send output to the rudder. This will put you in the $75.00 to $125.00 USD range.
Richard Oliver

Yes Richard, good point, I need more such. You are right, this problem I tried to solve some years ago. My dream was to make “proper” ratio between flaps and elevator. But no servo is usefull till present time - combination of veight, speed and torque is not available. Now I have servo with set time 0.1s/45 deg. It is relatively slow, but shorter distances are much quicker and I plan to have 6V battery it will add some speed also. If not, I am still ready to use only one direction. The gyro unit is adjustable the way that it works only to one side, so I can use it Al with his rudder. It will prevent wrong function if the deflection will come too late. The pHowever I think you mean to put the gyro to yaw axle. I am rathe thinking of pitch axle. I hope little slower servo will not make so many troubles. It will need precise trimming, but should work better in wind. Am I wrong?

igor

>Igor, The trick with the gyro is to have a servo that is
>fast enough to stay up with the gyro. You can not do this
>with the light/cheap servos. If the servo is slow it is to
>late to make the correction. You will want something in the
>nature of .01 rate(seconds). Most electronic gyros can way
>out race the servos ability to send output to the rudder.
>This will put you in the $75.00 to $125.00 USD range.
>Richard Oliver

Funny but the Helis are not using such fast and expensive servo's with their gyro's?

They do not, but few degrees yaw do not play role on heli fuselage.

Ron, I do not know what you mean. I am reading rules, but I do not see any statement telling the lines must be tight and not trawled by the ground

BTW, variable line length on stunter is my next project
I beleive it will solve lot of problems in wind.

what I mean is that our stunters fly in approximately a hemisperical surface: a 3 dimensional circle (half of it normally above ground, hopefully). Gyro flight control system work by integrating rates. in this case yaw rate I assume. Your gyro will see a fairly constant yaw rate (for a 5 second lap time that would be a yaw rate of 72 degrees per second)and try to achieve a zero yaw rate and probably simply run to max control deflection in an attempt to stop that yaw rate. All maneuvers will have that non-zero yaw rate (in fact from the gyro's perspective all maneuvers contain yaw, pitch, and roll rates). If you can modify the feedback amplifiers to detect acceleration rather than rate it might work...

I don't think I understand the goal of using one in the pitch axis...

Ron, I thought it is a joke

>>>Your gyro will see a fairly constant yaw rate (for a 5 second lap time that would be a yaw rate of 72 degrees per second)and try to achieve a zero yaw rate and probably simply run to max control deflection in an attempt to stop that yaw rate.<<<

The gyro unit does not work this way. It does not integrate and does not remember initial point. I wrote it yesterday in message #11. Now I see the equation is wrong so I try it as follows:

output angle on servo = angular speed of device x gain + neutral point

It means the servo angle is linear to angular speed only, does not depend how far you are from its original point or how log you are turning the device. I tried it yesterday evening, I am sure.
>>>I don't think I understand the goal of using one in the pitch axis... <<<
I see on your next message #14 you got it.

Igor,
Also keep in mind that the gyro is not self contained. It must see a referance signal from the receiver or it will never know which direction to drive your rudder servo. You'll need to build a referance signal device that outputs a 4.8vp-p, 1.5msec pulse width signal. Either high-going-low or vice versa. 1.5msec is roughly "neutral" for all radio systems.

Another thing you'll have to take into account is that if you get the gyro response gain too high, the tail will wag. If left in this condition long enough the tail will swing to the point where its uncontrollable. You'll need some way to kill the engine and get the plane down until you come up with the optimal gain setting. Which by the way, can change with outside temperature and the quality of the gyro your using. Typically you want to run the highest gain you can without getting "wag". Basically what is happening is that you can reach a point that the servo cannot keep up with commands from the gyro. One fix is to use the highest quality, fastest, rudder servo (usually a $115+ proposition) you can find but it depends on the gyro you have.

Gyros also drift off center..... If left to their own devices the tail on a helicopter will drift around requiring you to "command" it back to the position you want. The newer heading hold gyros take care of this to some extent but the lower quality samples will still drift. Remember, the gyro was only made to "dampen" unwanted torque changes. It was never meant to totally take control. What I fear will happen is that at the beginning of the CL flight your tail will stay on course but the gyro "centering" may drift the rudder to the point that the aircraft nose is into the circle or to the point that the nose is pointing directly out of the circle. Either situation would not be good. I'm not sure how you would overcome this with out getting overly complicated or limiting rudder control to the point that you negate the purpose of the gyro.

>>>Also keep in mind that the gyro is not self contained. It must see a referance signal from the receiver or it will never know which direction to drive your rudder servo. You'll need to build a referance signal device that outputs a 4.8vp-p, 1.5msec pulse width signal. Either high-going-low or vice versa. 1.5msec is roughly "neutral" for all radio systems.<<<

Right, I got it in shop. It is called servo tester. I replaced the trimmer by fixed resistor and now it is single chip device less than 1”x1”. The gyro itself has trimmer for centering of neutral so it enough to have “any” approximately 1.5msec impulses.

>>>Another thing you'll have to take into account is that if you get the gyro response gain too high, the tail will wag. If left in this condition long enough the tail will swing to the point where its uncontrollable.<<<

This is exactly what I afraid of. But as I wrote to Len – yes, I am ready to risk my model .

>>>Gyros also drift off center..... If left to their own devices the tail on a helicopter will drift around requiring you to "command" it back to the position you want.<<<

I hope that lines will keep the fuselage in its position. The gyro will add some deflection after start, but lines should handle it. I can deflect the rudder and model is still flyable. But you are right I will better start with only small deflections.

>>>Remember, the gyro was only made to "dampen" unwanted torque changes.<<<

That is exactly what I expect it will do. The lines must keep position, while the device only damp impulses from prop.

>>> What I fear will happen is that at the beginning of the CL flight your tail will stay on course but the gyro "centering" may drift the rudder to the point that the aircraft nose is into the circle or to the point that the nose is pointing directly out of the circle.<<<

Now you are surprising me. I understand it as follows (and I wish I am right): The gyro unit function is, that the output servo angle is linear to the angular speed of the unit. It means fixed unit keeps neutral angle, if you turn it lowly it keeps little deflection, if you turn it quickly, it keeps more deflection. If I shall to describe it analytically then:

= x +

The device does not integrate, so it does not “remember” original position. Thus I expect following (I think we are both speaking of gyro on yaw axle). The model on ground will keep the rudder (say) in neutral - that is the initial point. After starting, when angular speed is 360deg/5.3s (means relatively slow), the device will make some, but small “rudder out”. Tight lines will easily handle it. If I hit the corner, of if a whirlpool will kick the fuselage to some direction, the gyro will try to solve it by deflection. So I do not afraid the nose will point out of circle after ¼ of lap. If I afraid of something, then if the gain is too low, the gyro will not kill the impulse, but it will preserve the yaw after. If the gain is too high, and the servo is slow, it can oscillate.

This is why I tend to do it different way – gyro in pitch axle. It will work EXACATLY and ONLY against the gyro moment of prop – just if you pitch down, it yaws the nose right, that is all. May be not as pretty as the first solution, but I think more precise.

igor

>I have such a device in hands actually. It is dedicated to
>controlling heli back rotor, but it works pretty well with
>any R/C servo. It allows to set gain and neutral by resistor
>trimmers. Did anyone try to attach it to the rudder of stunt
>model to balance gyro moment of prop?

I never actually tried it, but I did look at it. The bandwidth tends to be too small to really respond fast enough. Also, you need to go in and put in a coupling capacitor to make sure it doesn't respond to DC, or putting it another way, it knows about the secular rate of simply going around the circle.

I looked at putting in a feed-forward to help the response. But, once you do the math, it begins looking an awful lot like the feed-forward alone is probably good enough. I would note that the parameter you would use for a feed-forward is control deflection. Which, through a bizzarre coincidence, ends up being precisely what the Rabe rudder does!

By the way, as a passive, set-and-forget type of device, it should be perfectly OK as a rudder controller in the proposed FAI rules. In fact, the stuff I sent the committee addressed exactly this device and suggested that the rules should certainly permit it. What you can't do is hook it to the elevator ('devices controlling elevation must be mechanically coupled' or some such words)

Brett, you wrote that it is exactly what Rabe rudder does. So you mean if Rabe rudder acts “too early” because of fuselage moment of inertia, the gyro unit with slow servo does the same just too “late”. Do I understand it well? It can be true, I am little disappointed, but I will try it anyway.

What about use both, it will be perfect … or?

Ah, now that more comments have come in I see what you mean by putting the sense axis in pitch; in effect you would have a "fly by wire" Rabe rudder... with the advantage of being infinately adjustable

interesting

Why not just do it the easy way and use a physical, horizontal pendulum linked to the rudder. It will respond instantly, and definitely would be legal anywhere. The only catch is that the model slows and speeds up both in maneuvers and with the wind.

Larry, just because there is no advantage of the pendulum, and it will come certainly heavier.

However, you are leading me to question I am thinking already longer time. What is difference between pushrod and wire??? Why one is considered legal and another not? Why electronic counter shutting the engine is not legal, while mechanical counter of drops (I mean tank volume) is legal??? Why mechanical device for measuring RPM and limiting of power (I mean pipe) is legal and electronic equivalent is not legal (or even mechanic pendulum on R/C carb is also not legal). I think current rules are very very conservative – means saying “use whatever we use now, do not try anything new”. It is far from I expect of this hobby …

So I did it after all. Here is a report from first session with gyro unit. I used self-adhesive foam tape to attach all parts from inner side of fuselage – to have it on eyes. The gyro is oriented to damp yaw (on vertical axle). The rudder is in neutral at ground, but makes 1 or 2 deg offset out at 5.5s lap time.

The first flight was disappointing. The unit did not work enough; the model was like too much tip weight (I note all units are on inner side of fuselage). Model also needed twist flaps to right. It is quite repetitive - after dismounting I had to give all back – so I think it IS property of gyro functioning. The gyro also did not damp all movements. It is because of function of the unit – if it should act, it NEEDS some movements.

For next flights I readjusted the device for maximal gain (still no tendency to oscillating), I also tried to adjust tip weight and flaps. Model performed much better; also gyro damped all movements much better, but still not at all.

If I compare that model with fixed rudder, I can say following:

1- The effect of gyro unit is only minimal at lap times at or under 5.3. It looks quick model is stable enough. It is probably because of better rudder efectivity and especially more line tension.

2- The effect is very clearly visible at speeds 5.4 to 5.5 or even more at lower speed (which I prefere). Model fly clean corners, all nervous movement or waves of such slow model are well damped. It is clear they do not come from imprecise wing or elevator. They come from imprecise yaw stability.

3- Model goes easier to corners. The rudder helps to pitch, while stability in straight segment is better.

4- There is also one clear disadvantage. If the model already HAS som yaw, the gyro tries to preserve it. It happens in very tight negative corners, when gyro does not fix all yaw. The model stays pointed nose toward the circle and unit tries to keep him there. The result is, that straight segmen after the corner gives less line tension, but the flight is still clean.

Those are mixed results from first day … if nothing else, it is nice toy

Now I try to put the gyro to pitch axle, so second part of report comes in few days.

Here are some pictures. The first shows all units - battery, generator of neutral servo singnal, the gyro unit and the servo. The second shows detail of gyro unit and small 4.5gram servo.

Here is promised second part.

I attached the unit to the top of fuselage (pitch axle), so its work is similar to the Rabe rudder. The first flight was similar to the previous tests (on yaw axle). The model needed remove tip weight again (related to the flight without unit). Now I saved it all and my scale home told me 1.85g. Now I start to believe, it has really what to do with the yaw stability. It looks controlled rudder needs less tip weight. Probably because of sweepback and negative yaw in negative corners without controlled rudder – the AC moves right, so the CG must also and thus more tip weight is necessary (probably – just my “first look” theory). I also got feeling that leadouts needs to be moved back, but I had only limited time, so I did not touch it yet.

Here is my one-day opinion:

1/ Again the same – quick flight does not need it. Slow flight is significantly improved.

2/ I think this version is little better than yaw-oriented device. It can solve almost all prop induced gyro yaws. The yaw-oriented device cannot fix all thanx to limited gain (if it acts, the yaw had to be present), while pitch oriented device can prevent it. And the biggest difference – it does not try to preserve erratic yaw.

3/ Compared to regular Rabe rudder coupled to elevator, the device works little better. I was able to adjust gain to fix almost all visible yaw, while regular Rabe rudder on the same model did not - there was still some yaw – if I damped all yaw in corner, I had parasitic yaw before or after, If I used less deflection the parasitic yaw disappeared, but yaw toward the circle in negative corner was present again to some degree.

4/ The deflection of the rudder was very close to that I got from calculation. I ended up close to +/- 10 deg on relatively large fin. If I did the same with Rabe rudder, it was too much. I never got good result with more that +5 / -0 on Rabe rudder on the same model. The gyro unit needs symmetric deflection and twice more.

5/ The speed of servo is really little trouble. I did not feel it bad in whole flight, but if I tried to do VERY STRONG second corner of hourglass, when the line tension is not very good, I got feeling like the model is somewhat flexible. But model has strong massive construction with balsa covered foam wing. The only answer can be, that the device worked little late and it was reason of some shake or what.

However, it works well. I expected worse result. The question is if it is worth to carry the battery and all that hardware in model, and additionally to maintain the battery. I do not know, I will use it for a while, I will try to trim the model again (just to try to take all of its potential with the unit) and especially I will try all again in wind, it can be I will see it from another point of view

This seems to be proper solution. It kills gyro yaw by rudder linear to pitch rate, while accidental unbalanced yaw (turbulence, slow servo or whatever) is damped also. I would prefer two units (with independently adjustable gain), but this solution costs less.

I got first serious trouble. I forgot to switch the device on and I launched. It was windy and I did not recognize any problem in first (positive) corner of wingover, but I almost crashed in the second (negative). I did usual action by the dandle, but model did not do usual 90 deg turn, it ended up pointed down by almost 20degrees. The rudder did not work, and it was reason, that model was much less sensitive (especially in negative corner).

I note I have up line back – gyro moment makes it less sensitive due to negative feedback.

Ha! Now you know what those RC people feel when they launch without turning on the receiver..

(I guess we're lucky that it's almost impossible to take off without connecting the lines.. though, of course, there's the small matter of picking up the handle upside down..)