What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Prop vibration analysis

Nathan K. Hammond

I'm New Here
Long time listener, first time caller.

Chasing a vibration that I can’t seam to null out.

3 blade MT on an angle valve AEIO540, mounted on a Chipmunk.

Installed the O/H’d prop on an O/H’d engine, dynamic balanced from .75ips down to .03ips but it still has a noticeable vibration. My running theory (and it’s a SWAG for sure) maybe a blade is out of profile. With the sun at my back, I have 3 distinct reflections at the blade tips. 2 are slightly overlapping, but the third is not tracking with the others. (Aft 4” of the blade tips are painted white for formation work.) That makes me think something internal is off, or the blades are profiled slightly differently.

Trying to eliminate as many possibilities (aerodynamic, resonance, or ___?___) before pulling the prop. Dan Horton mentioned the Resonance App in a previous thread, which I downloaded, but I don’t know how to interpret the data.

So, for the collective knowledge…. What am I looking at here?

2600rpm
200C2583-EFED-43BE-BC0E-CA15C815A0BD.jpeg

2500rpm
FB6B38BF-1368-4BBE-8197-8C1AFF708C5D.jpeg

2400rpm
CE38613A-1C50-4CA6-9F7E-127ABF5FCEB6.jpeg

nkh
 
For the blade tracking, put something under the prop tip like a stool or short table and tape a piece of paper to it. The goal is to have a flat surface as close to the bottom of the prop as possible. Slow turn the propeller backwards by hand and make a mark on the paper where each tip crosses. Most props have a limit of 1/8 to 1/4" max difference in tip position, but you'll have to refer to your specific manual. If you still think one isn't tracking right, find a prop shop that has a vibration analyzer with a strobe light. It's similar to putting the sun at your back but much more accurate. Should be able to identify the out of tolerance blade quickly.

As for the vibe data, it's odd. 20 Hz is equal to 1200 rpm, which is 1/2 the speed of the motor. Whatever is causing most of your vibe is spinning at half the speed of the engine. This would point me towards something in the camshaft or lifter assembly. You can attempt to verify this by removing your prop balance weight and redoing the data run. A run with an unbalanced prop should have the peak at 40hz (rpm / 60 = 1X base frequency in Hz) be the dominant peak by a large margin. I'm not an engine expert, so hopefully someone can chime in on where to look, but most props don't vibrate at 1/2 the base frequency. Subharmonic vibe can also be caused by worn bearings or "looseness".
http://updateinternational.com/Book/VibrationBook6b.htm


Which way was your phone facing for these readings? Is X up and Z forward? Was the phone in the cockpit when you collected this? I've never used this app before, so I don't know how much the data can be trusted. The best way would be to mount an accel to the engine and use a real analyzer with spectrum capability.
 
Last edited:
The following affect half order vibration:
Anything that controls the combustion cycle on the engine pretty wide territory for troubleshooting. It could be compression, mixture, induction losses. valve lift, spark timing, or anything that controls the combustion cycle in the engine!

In other words, one of the cylinders is not carrying it share of the load or is 'mis-firing'.
 
Last edited:
The following affect half order vibration:
Anything that controls the combustion cycle on the engine pretty wide territory for troubleshooting. It could be compression, mixture, induction losses. valve lift, spark timing, or anything that controls the combustion cycle in the engine!

In other words, one of the cylinders is not carrying it share of the load or is 'mis-firing'.

I know almost nothing about harmonics and vibrations. However, I am thinking that a half order harmonic means that it is happening every 180* of rotation and therefore in line with your recommendations here. However, if the issue were with one cylinder, the imbalance would occur every other rotation. Not sure how this would impact a vibration test or if the sensor is smart enough to see the vibrations come and go across alternating revolutions. I am guessing that this may get lost in the smoothing or averaging. However, if the sensor can distinguish a half order vibration that occurs on every rotation vs every other rotation, then the diagnostic value is very high, as you can eliminate individual cyl issues. The number o things that can under perform EVERY rotation is much smaller than those on every other rotation. just adding some thoughts to possibly help the OP and certainly not challenging your expertise here.

One last thought. If a half order vibration does mean every 180 degrees, I am not sure this logic applies to a 6 cyl engine, as it fires once every 120 degrees vs every 180 degrees like a 4 cyl engine. Again, don't really understand what half order means and as anxious to learn as I am sure the OP is.

Larry
 
Last edited:
Half order vibration means it's happening every 720 degrees, not every 180 degrees. If it happened every 180 degrees, it would happen two times per revolution, making it a 2X harmonic.

Another way to think about it is in RPM. 2400 rpm means the crank is coming through top dead center 40 times a second (2400 / 60 = 40 Hz). In your case, the thing causing your vibration is only rotating at 1200 rpm (1200 / 60 = 20Hz. So your crank is coming through top dead center twice for every peak in vibration. Things that operate every other turn of the crank are camshaft or combustion related. Stuck valve, weak cylinder pressure, bad plugs, bad ignition, etc. only fire are all things that only affect operation every other time the cylinder comes to TDC

EDIT: I'd like to thank Walt for getting me back into the combustion engine mindset. Most of my vibe analysis has been on electric motors, so sub-harmonics are extremely rare. Combustion motors have a huge 1/2X harmonic capability since the valvetrain and combustion cycle operate at 1/2 the motor RPM.
 
Last edited:
Half order vibration means it's happening every 720 degrees, not every 180 degrees. If it happened every 180 degrees, it would happen two times per revolution, making it a 2X harmonic.

Another way to think about it is in RPM. 2400 rpm means the crank is coming through top dead center 40 times a second (2400 / 60 = 40 Hz). In your case, the thing causing your vibration is only rotating at 1200 rpm (1200 / 60 = 20Hz. So your crank is coming through top dead center twice for every peak in vibration. Things that operate every other turn of the crank are camshaft or combustion related. Stuck valve, weak cylinder pressure, bad plugs, bad ignition, etc. only fire are all things that only affect operation every other time the cylinder comes to TDC

Thanks for the detailed explanation. Makes complete sense that a single cylinder is misbehaving. Can one deduce from this whether or not this is due to one cylinder OVER powering or UNDER powering? Or does this type of measurement not really get to that level of detail and instead something is just different every 720*?
 
Last edited:
Thanks for the detailed explanation. Makes complete sense that a single cylinder is misbehaving. Can one deduce from this whether or not this is due to one cylinder OVER powering or UNDER powering? Or does this type of measurement not really get to that level of detail and instead something is just different every 720*?

In this case, I think we can just say that "something" is happening every 720 degrees. With an advanced vibration analyzer and a roaming accelerometer you could probably pinpoint which cylinder was the culprit, but by only taking readings with a phone app, it limits how accurate you can be.
 
Amazing! Thanks all.

The blades all tracked less than 1/8” at the tip during install. My theory above was/is the blades are at different pitches in cruise.

Earlier we tried clocking the prop 180° but that actually made it worse.

For this last flight, using the vibration app, Phone was laying on structure square/plumb to the axis.
Z = up/down
X = left/right
Y = fore/aft

Honestly, I never would have thought a cylinder to be the culprit. Engine analyzer looks clean and level through all modes of flight. Adjusting mixture from full rich through LOP doesn’t appear to change the vibration.

Thanks again for the trouble shooting ideas and clear description of the data!

nkh
 
Here's a spectrum analysis from my Microvib II unit just for reference, this one show a bit of a rise in half order vibs, over the alarm limit.

BH%20RV4%20spectrum-XL.png


and here's one that is more "normal":

Steve%20Christo%20IO390%20Vib%20Survey-XL.png


Some more info on vibrations compliments of Microvib:

Normal Vibration Signature

Each combustion pulse acts much like a hammer blow, hitting the engine block with a pulse of energy. The vibration spectrum of such a pulse is a series of vibration spectral lines. These spectral lines will be at integer multiples of the firing rate of each piston. In a four stroke engine the piston fires every other revolution, therefore the fundamental spectral line will be at 1/2 the engine RPM, often called the 1/2 order vibration. The result will be a vibration signature that has spectral lines at the 1/2 order, 1P, 1-1/2P, 2P, 2-1/2P, 3P ... etc. In our experience, we have found that most 4 and 6 cylinder horizontally opposed IC engine produce these spectral lines in varying patterns but the levels are usually in the range of .2-.5 IN/S. For unknown reasons we have found it is not uncommon for an engine to produce a 2P or 2-1/2P vibration in the range of .5-1.0 IN/S. These readings are taken on the ground at typically 2500 RPM.

1/2 Order Vibrations

It turns out that if all of the pistons produce nearly identical combustion pulses, the 1/2 order vibration will be very small, .1-.3 IN/S. When any one cylinder produces less power than the rest, the 1/2 order vibration will increase from .3 IN/S on up to over 1 IN/S for a misfire. Mechanics know very well what can cause one cylinder to be weak. Plug misfire, bad plug, plugged injector, broken ring, leaky valve, low compression, bad magneto, worn cam, collapsed lifter, etc. Any of these things can cause a higher than normal 1/2 order vibration.

1/2 order vibration are especially troublesome as they can be felt in the cabin by the pilot. Low frequency vibrations are not well isolated by most engine mounts, and the vibration will shake the entire aircraft. If not taken care of, a 1/2 order vibration can loosen rivets, hinges, and pivots all over the airframe, as well as causing premature pilot fatigue. This is a serious safety issue.
 
Last edited:
I see the RPM notes now. Duh on me...sorry.

20hz is half order, and as noted, a weak cylinder could do that.

The phone app is not an analyzer. I think it is most useful for elimination, narrowing where to look.

For example, in this case it is unlikely to be a heavy blade or similar, because at 2400 the big spike would be 40 hz, first order, once per rev.

It's not at engine firing frequency, as 2400 with the 6 cyl would 120 hz, third order, three times per rev, thus it's not likely to be something like the vibrating engine contacting the cowl.

It's not an out of balance wheel spinning in the breeze, because that would be some random frequency with no relation to RPM.

There is one caveat. The phone is measuring the vibratory amplitude of the thing you're holding it against, not the source. The component could be resonating at its own natural frequency, excited by an integer multiple of the source vibration. Because it is resonating, it would have a large vibratory amplitude, i.e. a big spike on the screen, and fool you into thinking the spike is the source frequency. To guard against being fooled, just take readings on several different components. It's unlikely they all have the same natural frequency.
 
Vibration on a IO-540

I have test flown an RV10 on behalf of a new owner. A few small snags but one thing that did bother me was the engine vibration that I have not felt in other RV10s including my own.

I have recently downloaded an app that measures vibration. So I have positioned my phone standing on its side on top of the throttle quadrant. I have saved the file and had a look at it. I have recorded while flung at 2400 rpm. I know that a prop out of balance should show up at 40 Hz when flying at 2400 rpm.

I did not see much close to 40 and 20 Hz but did see the highest value at about 28.7 Hz. Anyone who can give some advice or share some insight on this please.
 

Attachments

  • IMG_4378.jpg
    IMG_4378.jpg
    89.6 KB · Views: 22
  • IMG_4381.jpg
    IMG_4381.jpg
    73.8 KB · Views: 9
  • IMG_4379.jpg
    IMG_4379.jpg
    71.3 KB · Views: 6
What prop is on it? A 3 blade prop has a blade passing frequency of roughly 120Hz. Since you're limited to 100Hz sample rate you may be missing the actual spike and getting the 1/4 frequency data spike at 30Hz. If you can, try to record the data again at 500Hz sample rate and lower your FFT time to the 1-2 second range. If it's a 3 blade prop, go balance the prop and check the blade tracking.

If it's a 2 blade, I have no idea what rotates at that freq.
 
Sorry agent, I forgot to say that it is a 3 blade Hartzell (metal) prop. Not sure if the iPhone app can do that frequency, but it is good advice to balance the prop. Thank you.
 
There is one caveat. The phone is measuring the vibratory amplitude of the thing you're holding it against, not the source. The component could be resonating at its own natural frequency, excited by an integer multiple of the source vibration. Because it is resonating, it would have a large vibratory amplitude, i.e. a big spike on the screen, and fool you into thinking the spike is the source frequency. To guard against being fooled, just take readings on several different components. It's unlikely they all have the same natural frequency.

This was my initial thought as well, which could mean that dynamic prop balancing could still be beneficial.
 
It is the MT prop. One of the blades is wrong. Especially if you can see separate reflections. I had a prop shop totally screw up the overhaul on my MT, and this was the cause. Meaning, no matter how well I balanced it, there was an annoying vibration. You could even see it in the stick shake.

Vic
 
Last edited:
It is the MT prop. One of the blades is wrong. Especially if you can see separate reflections. I had a prop shop totally screw up the overhaul on my MT, and this was the cause. Meaning, no matter how well I balanced it, there was an annoying vibration. You could even see it in the stick shake.

Vic
Vic,
How can one know if a blade is incorrect? Any testing, measuring while on the plane?
Johan
 
An aerodynamic imbalance can occur if the blade pitch angles are different.
I have found 3 bladed props on 6 cyl engines to be troublesome for some combinations. Dynamic balance and/or reclocking will often improve things.
 
On a Hartzell its very easy to incorrectly orient the pitch change blocks and have a blade angle mismatch. They are made asymmetrically so the blade angle can be changed by flipping the orientation. I imagine its similar on an MT.

To do a crude blade angle check get a angle app for your phone and pick a spot on the middle of each blade. Get a step ladder and rig up a pointer such that each blade can be repeatedly checked at the same angle. Rest the phone against the spot picked and do it the same way for each blade. Spin the prop and check each blade angle.

Blade angle differences can cause "coning" where one or more blades generate more thrust than the other(s) which creates a vibration.
 
Good advice Vic and Bob - thank you. I do have a digital protractor which I trust more than the app one. I will do an angle check when I go out there again.

We did have an appointment with AMO yesterday to come out and do a dynamic balancing, but they did not pitch up. In our part of the world almost everything comes to a standstill now and they will only come out in January for the balancing when they reopen.

Agent - I went through the settings of the vibration app and it can only go to 100 Hz an not higher.

This RV only has 85 TT hours on it and as far as I can find out, the prop is as it has been received from Hartzell. Could they have the blocks orientated wrong?
 
This RV only has 85 TT hours on it and as far as I can find out, the prop is as it has been received from Hartzell. Could they have the blocks orientated wrong?

I’d flip it 180 deg first, highly unlikely it has an issue from the factory.
 
Back
Top