FFT Magni rotor Vibration analysis

[Mike G]

Brian and Chuck
I agree there shouldn't be any 2/rev in a vertical descent that's why I chose to balance my rotor in a vert descent. I'm also pretty confused and most of the tests I'm planning are to try to check the instrument I have rather than the rotor. If Magni rotors had a 2/rev problem as bad as my analyser says I have I'm sure we'd have heard about it before now.
Mike G

 

[Mike G]

Chuck
In your undersling calculation do you assume that the blade CofG is at the centre of the blade span?
Mike G

 

[Mike G]

Yesterday I did a zero speed vertical descent and recorded an FFT. And then I unbalanced the rotor by removing the balance weights on one blade and re-did the excercise.
The results are attached and show a clear increase in 1/rev vibration (around 5 hz or 300 rpm) and a slight decrease in 2/rev.
This seem to show that what I'm measuring and recording is true and as Chuck would say "the squirrels are running the nut house".


Any suggestions or ideas about why or how I get so much 2/rev vibration in a vertical descent would be more than welcome.

Mike G

 

[Mike G]

I also tried to look at the effect of forward speed on (what I think is) 2/rev vibration. This seems to confirm, what you'd expect, that 2/rev vibs increase with speed. It also shows that in my case the 1/rev vibes also increases going from 3.9 at 60 mph to 7.3 at 80 and 14.4 at 100. I'd expect a small (maybe 5%) increase due to the increase in rrpm but nothing like what I'm seeing.

Chuck, can the 2/rev excite or amplify the 1/rev vibs??

The graphs aren't as precise as the others because I have to download the data by hand and I got lazy, but the trends and peak values are correct.

Mike G

 

[ckurz7000]

Mike, I can only conclude that something else than variations in aerodynamic rotor blade drag is exciting the 2/rev shakes.

Can you remind me again what the orientation of the accelerometer is you used to pick up the vibes? You might want to measure the 2/revs in all 3 spatial axes at different forward speeds.

I can see how, for example, the turbulence caused by the body might interact with the rotor blades twice per revolution in a vertical descent. This might be sufficient to excite the 2/revs you observe.

You might also want to do the vertical descent with the engine stopped to see if there is any coupling of particular vibration modes from the prop or engine to the rotor or mast.

Keep us posted, Mike, this is fascinating stuff!

-- Chris.

 

[Mike G]

In a further attempt to establish if the analyser I'm using is giving me false data I fixed the accelerometer to the gearbox on the engine.
The prop is out of balance because I stupidly had a heavy landing at the end of my trials doing FFTs on the rotor and the tip of one of the prop blades touched the chassis (it's a problem with Magnis, the prop tip is only 1 inch above the tube supporting the rudder).
So attached is an FFT of the prop/engine/gear at 2600 engine rpm 1071 prop rpm.

You can see that the frequencies (I've converted them to rpm) given by the analyser are spot on which gives me even more confidence in what I'm measuring on the rotor.

I tried to balance the prop but find the balance mode of this analyser very poor. I confirmed this at the manufacturers when I tested my analyser against one of theirs on their test rig. I must find the time to go back and force them to explain why this is so poor a balancer.

Mike G

 

[Mike G]

[/IMG]

Chris
This is my accelerometer installation, you can also see the Magni rev counter and my rev counter (yellow) for the balance mode of the analyser.
I hope inserting this photo works. Mike G

It needs a paint job!!!

 

[Oskar]

I tried to balance the prop but find the balance mode of this analyser very poor. I confirmed this at the manufacturers when I tested my analyser against one of theirs on their test rig. I must find the time to go back and force them to explain why this is so poor a balancer.

Mike G

Mike,

Balancing the prop on a 912/914 is very tricky due to the gearbox ratio. The DSS balancer is great for balancing direct drive props but does not work when trying to balance a prop on a 912/914. So far the only way I have been able to balance props on 912's is to do multiple measurements with an intentional large unbalance weight.

I am busy working on a balancer that will hopefully be able to overcome this issue, it should be ready for the first test on a 912 in a few weeks time.

Oskar

 

[okikuma]

Mike,

The cause of the feeling of 2/rev shake just might be as simple as turbulent air flowing up and around from the fuselage and occupants, then up through the rotor disk that normally will not happen in cruise flight.

Wayne

 

[Mike G]

Oskar
I set the once per rev sensor looking at the prop hub and the accelerometer on top of the gearbox. That way the balancer isn't affected by the gear ratio. I'm not sure what problem DSS balancer gives you unless your once per rev sensor is looking at engine speed.
Wayne
I agree that a possible cause of 2/rev excitation in a vert descent may be the wake coming off the fuselage or the prop wake. I shall do an engine off descent soon to see if it makes a difference. I'm not too optimistic because the prop wake should be small at idle.
Mike G.

 

[[email protected]]

The only shake/vibes (Cabin shake) I experience and experienced ever in my M24 Orion is limited to very narrow margins:

Wenn under thermal conditions and SOLO (both in cruise flight and vertical descent) the rrpm drops suddenly under 351 and revs up then to 362 there is slight yawing shake of the cabin selflimiting to 2-3 seconds.

After reproducing this slight shake exactly it turned out that this shake/vibes can be reduced significantly by not touching the foot pedals to "unlock" the rudder.

The stick on the other hand is rock-solid all the time, not the slightest vibes ..........

 

[Mike G]

Angelo
Thanks, I'm not sure what, if any, significance there is in you experience to mine but it's good to know others are reading and trying to understand. I don't feel so alone.
On my next vertical descent I'll leave the rudder free just to see.
Mike G

 

[bpearson]

Mike. Back when vpm's were first being brought in to the uk some had horrendous stick shake.....it really did bruise your legs. I remember a company coming to our airfield to balance our props and they also did helicopter blades. They had all manner of equipment but could not reduce the shake and put it down to manufacturing intolerances. We tried good (ok'ish) blades on the rally bad aircraft but it still was bad. That's the reason I asked previously about the age of your blades and if your head had been back to the factory! I think it is only in the last few years that they have got the shake sorted.
I hate to be negative but I have seen people spend days trying to get rid of stick shake on magni blades and fail! Any chance of a video of the stick in flight?

 

[Mike G]

Brian
I've been thinking along similar lines myself, I forgot to ask Eric Changeur about the age of these blades when i saw him the other day, I'll try to ring him.
I remember hearing that Magni changed the blade production a few years back and that apparently improved things a lot but I have no details.
Do any of the Magni drivers or distributers know anything about the blade manufacturing change I heard about or was it just a rumour?
Don't forget that the big difference between the M16 and the VPM was the mast. I believe that the VPM had a much stiffer mast which should have made it worse for the VPM.
The problem now is that we (Mrs G and I) are moving to our summer home in Brittany next week for the summer and I'm in the middle of preparing for the great annual summer treck which includes putting the gyro and all my tools on the trailer plus a million and one other things I mustn't forget so that I can work on the house in Brittany. So I doubt if I'll have time to do much more testing for a week or two, but I'll try to take a film of the stick next time.
I'd like to try to organise a trip to Eric Changeur's again to test my rotor with his Vibrex in parallel to my analyser and then do the same on more recent Magnis. However our summer house seems to be the favorite haunt of our children and all our nieces and nephews and friends who take advantage of Mrs Gs cooking and hospitality and my wine cellar so I doubt if I'll get permission to take a few days off.
I suppose the real solution will be to trailer the gyro down to Magni in Italy and try to sort things out there. Perhaps they'd admit some of the things they know are wrong and have corrected on later gyros.
Thanks for joining in, are you still flying you gyro, you were talking about giving up?
Mike G
Perhaps I should just put it on the trailer and drive down to Italy!!

 

[Oskar]

I set the once per rev sensor looking at the prop hub and the accelerometer on top of the gearbox. That way the balancer isn't affected by the gear ratio. I'm not sure what problem DSS balancer gives you unless your once per rev sensor is looking at engine speed.
Mike G.

Mike,

The difficulty arises because of the fourier definition which is based on an infinitely long measuring period. The DSS (and probably the equipment you are using as well) only measures and processes the vibration measured during one cycle, in this case one prop revolution. Using clever mathematical manipulation the various harmonic components can then be calculated. This, however, assumes that the signal is periodic, in other words the vibration for one prop revolution is assumed to be the same as the vibration for the next prop revolution. And it's here where the gearbox ratio messes things up completely.

The vibration sensor measures all vibration at the point where it is mounted. Mounting it on the 912 gearbox means that it will measure the vibration due to a prop unbalance, but also the vibration due to the running engine attached to the gearbox. And with a gear ratio of 2.43 the vibration due to the running engine is not periodic when looking at one prop revolution.

With a direct drive prop the engine vibration is periodic when looking at one prop revolution and the DSS balancer works well.

Oskar

 

[Mike G]

Oskar
I can't comment on your analysis re FFT (you're into a subject out of my scope) but how do you explain the close fit of prop speed, engine speed and blade passing frequency looking at the FFT in post #66?
In fact I'm a bit confused about what you're trying to tell me. We started talking about the difficulty in balancing the prop due to the gear ratio. When balancing I understood that in the balance mode the analyser simply takes the periodic vibration at the frequency of the 1/rev sensor and measures the time between the peak of the (more or less) sinusoidal wave and the 1/rev sensor virtually instantaneous signal. From this time period it's simple to calculate the angle between the 1/rev reflecting tape and the unbalance. I think you may have some knowledge that will help us advance even further so please continue.
I don't see how the problem with the FFT that you explain affects the balancing operation unless the analyser does an FFT each time it does a balance calculation, but as I said at the beginning you're in an area of electronics beyond my experience level.
Mike G

 

[Mike G]

Brian
When you propose sending my rotor head back to Magni, which parts would you include in the "rotor head" and what do you think Magni could do with it to improve it? They can only change the bearings and bolts, otherwise I might as well buy a new one?

Mike

 

[bpearson]

Hi Mike
Got the gyro itch back but currently trying to sort out my single place fuel tank that has been attacked by ethanol....praying the VPM is ok!

Magni replace the ally bearing housing and returned the rest to 'new'. The new head uses two bearings and maybe he re-machines the head to better tolerances.

My feeling is that if the head is not accurate then you will never sort out the shake by messing with the blades. A friend of mine had problems and Magni told him to send the head and blades back to the factory. Both were replaced which makes me think the head can be causing problems. Mine was old and came back like new.....with new blades.

Getting much info out of the factory is hard and their cure involves sending them the cash equivalent of a new small car in return for a couple of lengths of fiberglass.

 

[Mike G]

Brian
My head is the newer two bearing one, I think they did this when they developed the M16 from the VPM.
I've told this story before but here it is again. When I stripped my head down to check the bearings I had the ally bearing housing checked dimensionally to ensure that the four corners were concentric with the bearing bore so that when I tightened the four long bolts holding the ally housing to the pre rotator ring gear I could be sure that when the four corners showed no run out I know that it was parallel and concentric to the Jesus bolt. We discovered that the four corners were perfectly concentric but the teeter bolt hole was off centre.
See attached Magni head error 1 which shows the error we found.
Magni at first refused to accept that there was an error so I sent a photo (DSCF1820) of a set up that showed easily that there was a problem and a series of drawings (Magni head error 2 & 3) to show what we did.
They came back saying they didn't want to argue about it and sent me a new ally bearing housing complete with two new teeter bearings free of charge.
So I know that these are the latest design.
The error I had was probably a CNC machine programming error and would have given a simple 1/rev unbalance effect that would have been difficult to correct without a Vibrex or a similar type of balancer. I wonder when they realised themselves this error and how many old Magnis vibrated because of it, because I understand that the early Magni balance technique was simply to try to get the CofG and mass of each pair of blades as equal as possible.
As for "returning the rest to new" I don't see what they could do except replace the teeter bolt and bearings and include the teeter bolt axial adjustment (teeter adjust.pdf) in the blade carrier. I think that the earlier machines didn't have this which would also have introduced a probable "out of pattern" (CB's expression) 1/rev vibration that would have been difficult to correct without a balancer.
I'm now pretty convinced that the vibrations I see are either due to the older blade design or CB's "in plane natural frequency" problem or both.
As I said earlier it'll be some time before I get flying and testing again so I don't know when I'll be able to move forward.
Also I see that the forum interest in this is falling off and we're probably at the limit of the forum's experience so I'll probably move on to something else.
Mike G

 

[Mike G]

Chuck
In post 2 you say:
"The 2/rev stick shake, fore/aft or lateral depends on how near excitation frequency is to inplane resonance of the rotor. At resonance, displacement lags excitation force by 90º but phase angle changes quickly around resonance.

When you say "phase angle" are you refering to the lag between displacement and excitation or something else?
Mike G