Need help with a strange cyclic car vibration

[brainz01]

Zeke: Interesting video. I'll have to check runout at the tread.

I drove a friend's 997 GT3 recently - - his was noticeably smoother despite a stiffer suspension.

I'm concerned that two sets of supposedly good wheels are vibrating similarly. Frustrating.

[javadog]

When I would probably do at this point is duplicate the set up that you had where you observed the rear tire jumping rope, as you put it. This time use your buddy’s wheels.

[rusnak]

Can you make a steel plate with holes that line up with your lug bolts, and weld a handle or breaker bar to it? Try to make the rear hubs show flex. Make sure that your wheel hub nut is tightened to spec. and check all upright bolts for torque spec.

Very interesting thread indeed. The guy in Milt's video looked to be shaving the tires.

[brainz01]

Well, here's an update on the situation.

The current plan is to measure the run-out on all the axles and hubs and then run the car on jackstands to see if anything looks off.

I did end up buying a set of wireless Chassis Ears, so I also plan to use to use those to listen for any weird suspension or bearing noises.

I also learned that disconnecting the electronic shocks on PASM equipped 997s should make them go full stiff. I plan to disconnect them one corner at a time and see if the shock feels different - - it should.

Today's findings:

I put the back end up in the air and flipped the orientation of the top dished washer on my engine mounts (the washer dish now faces up). I don't expect that to make a difference with the vibration, but thought I'd mention it as it may increase the compliance of the mounts just a bit.

With the rear tires removed, I went about measuring the runout on the axles. The CV body at the hub measured ~2.5 thou on both the right and left sides. That was a bit more than the hubs and rotors, which were around one thou each, but didn't seem too bad. That would seem to suggest that everything is OK with the hubs.

The axle shafts themselves, however, surprised me with about 10 thou of run-out on both sides as measured in the middle of the shaft. I have no idea whether that's acceptable or not. I find specs online for drive shafts of other makes suggesting up to 30 thou is ok , but I did not find any specs for axle (half) shafts, much less Porsches.

I also measured the inboard CV body on the right side. The runout there was initially about 10 thou, but I was able to loosen and tap the CV to about 5 thou of runout. The high point of the axle runout was the same as the inboard CV high point.

I don't detect any play in the axle. And the shaft "floats" smoothly in the CVs if you push/pull it inboard or outboard. As a reminder, this is a low mile used axle that I regreased and booted. I didn't notice any play or wear while I was rebuilding it.

Interestingly (and thankfully) the runout on the right side transmission flange was very low ~1-2 thou (and that was on an uncleaned surface). So that seems good.

I did not have time to measure the left hand side inner CV and transmission flange. I'll do that another day.

Here's a video of the right side axle runout:

https://youtu.be/TZVYsFfg9jI

As always, I appreciate any thoughts.

[rusnak]

Stupid question from me: When you jack the car up, and observe the axle whipping around, have you taken it out and driven it first? I'd say you need to drive it at least 20 miles to massage out any flat spots in the tires from sitting, even for a day.

Also, I must say check the torque on all of the control arms and strut upper and lower bolts.

[sammyg2]

I'm trained and certified in vibration analysis up to level III (ISO/IEC 17024 and ISO 18436-1 accredited).
I'm also a bit rusty but based on spectral analysis of the limited data, there are some real clues:

The main exciting force is around 18.5 hz and suspected. That is likely the rotational speed of the wheels.
The analysis shows a pronounced 1x, 2x, 3x and 4x and 5x which could be attributed to looseness. I can't determine impacting other than the G amplitude, a waveform would be nice

The cyclic vibration you are feeling is due to vibration at two close frequencies that periodically line up and amplify the amplitude, then separate and cancel each other out, or at least reduce the amplitude.
I notice in a couple of your spectra I see two peaks very close to each other at around 56 hz (3x) and in another at 38 (2x).
Does the tool you are using allow monitoring in live mode or just snapshot?
If it does allow live mode I would expect those two peaks to modulate in amplitude as they line up and then separate.
I wouldn't be surprised to see other peaks do the same.

So why is that?

I suspect you have two very similar exciting forces at very close frequencies. The most logical suspect would be the rotating wheel assemblies where one wheel was turning slightly faster than another but your troubleshooting tends to suggest otherwise.
BTW the tool is set up to display in acceleration so I started integrating the data into displacement and velocity for comparison and flashed back to the vibration institute days, now my head hurts
Haven't had to do that manually since the exam.

Anywho, if the two exciting forces matched in frequency, you would feel the same level of vibration all the time and likely not notice it as much.

There is another clue in the data:
The 7x.
Orphaned, no sidebands visible, not harmonics.

I also see enough non-synchronous noise to be of interest but not enough info to analyze.

I'm betting that 7x peak is important.
ninelevenick suggested rear wheel bearing, that could explain the 7x, can't rule it out.
I think I see a peak that could be inner race frequency but that's a bit of a stretch.

Does anyone know how many balls are in a 911 CV joint? It's been a long time since I messed with one and i've slept since then.

[javadog]

Whatever happened to swapping wheels and tires with a buddy?

[brainz01]

Originally Posted by sammyg2

I'm trained and certified in vibration analysis up to level III (ISO/IEC 17024 and ISO 18436-1 accredited).

I'm also a bit rusty but based on spectral analysis of the limited data, there are some real clues:



The main exciting force is around 18.5 hz and suspected. That is likely the rotational speed of the wheels.

The analysis shows a pronounced 1x, 2x, 3x and 4x and 5x which could be attributed to looseness. I can't determine impacting other than the G amplitude, a waveform would be nice



The cyclic vibration you are feeling is due to vibration at two close frequencies that periodically line up and amplify the amplitude, then separate and cancel each other out, or at least reduce the amplitude.

I notice in a couple of your spectra I see two peaks very close to each other at around 56 hz (3x) and in another at 38 (2x).

Does the tool you are using allow monitoring in live mode or just snapshot?

If it does allow live mode I would expect those two peaks to modulate in amplitude as they line up and then separate.

I wouldn't be surprised to see other peaks do the same.



So why is that?



I suspect you have two very similar exciting forces at very close frequencies. The most logical suspect would be the rotating wheel assemblies where one wheel was turning slightly faster than another but your troubleshooting tends to suggest otherwise.

BTW the tool is set up to display in acceleration so I started integrating the data into displacement and velocity for comparison and flashed back to the vibration institute days, now my head hurts

Haven't had to do that manually since the exam.



Anywho, if the two exciting forces matched in frequency, you would feel the same level of vibration all the time and likely not notice it as much.



There is another clue in the data:

The 7x.

Orphaned, no sidebands visible, not harmonics.



I also see enough non-synchronous noise to be of interest but not enough info to analyze.



I'm betting that 7x peak is important.

ninelevenick suggested rear wheel bearing, that could explain the 7x, can't rule it out.

I think I see a peak that could be inner race frequency but that's a bit of a stretch.



Does anyone know how many balls are in a 911 CV joint? It's been a long time since I messed with one and i've slept since then.

Sincere apologies for the slow reply. I somehow missed your thoughtful post.

I'm using a free phone-based vibration app. It does seem to do something, and is somewhat repeatable, but it's not pro level stuff. Sadly it does not show waveform ot allow tracking of X, Y, and Z axes - - it just shows frequency peaks.

New Info: I discovered I have a failing water pump. It's become noisy as all get out - - has the Mezger "box of rocks" sound very badly. I don't think it's causing my vibration issue, but I do think it's adding to my overall noise signature. I need to fix that first for multiple reasons, but am hopeful that may get rid of some of the higher order vibrations I'm seeing.

Additional new info: Since my last post, I thoroughly checked the LR suspension and swapped in a rebuilt axle (now both rear axles have been swapped). I also flipped the cupped washer on the motor mounts to give them more travel. Net result: No difference in the vibration. Zip. Zero. Nada.

Today I also turned my AWD into a 3WD - - I bought some stub axles from a Carrera and replaced the FR axle with just a stub axle. I also measured the FR runout of the wheel, and it was ~ 5 thou at the edge/face of the wheel. Seemed good enough. Net result: Possibly a small reduction in the vibration (as measured by my app) - - it felt about the same - - still annoying.

New observations: Both front shock mounts appear to be worn out. They're not obnoxious and don't audibly clunk, but the rubber insert appears to have become unbonded. This may be related to my issue.

Interestingly, when I drive 65mph or less, the vibrations are absent and the car is smooth (as confirmed by the phone app). But as I increase speed above 80mph, and especially on the smooth grooved pavement, the vibration becomes very obvious. Maybe slop in the front strut bushings is setting up a chassis vibration? We'll see...

Also, I have experimented a bit with the Chassis Ears. It's a good tool - - transmitted a lot of noise on the RF which appears to be from the strut mount. The LF was less noisy, but then the mount looked better too. It's a fair amount of work to setup the Chassis Ears (especially without a lift), but they do provide interesting information - - lots of noises you'd never hear otherwise, even the sound of the brake pads on the rotors and the electronic servos in the shocks.

Someone asked about trying a different set of wheels. I did that with a known set of new wheels and new (different brand) tires. It made no difference (actually might have been worse as the wheels were heavier and the tires were harder).

My current guess is that I have a tired suspension - - or at least front strut mounts. So after fixing the water pump, my plan is to renew all the suspension arms/links/bushings. I'll hold off on shocks for now as they're relatively expensive and the ones I have seem fine (I did some tests including disconnecting the electronics on one shock, and the whole set appeared to go full stiff as expected - - the car rode like a dump truck while disconnected).

That's all for now. Appreciate any thoughts or theories.

[DanielDudley]

These kind of problems tend to be exacerbated when the body of the car has a resonant frequency similar to the harmonics of the vibration. Early Miatas have this, and so do late model GMC trucks. Some tires can make this worse.

I had a car years ago with one bad roller on an inside wheel bearing. Otherwise the bearing was fine. I take my wheels to a local Chevy dealer now. There are different levels of road force that are considered acceptable. You need more than acceptable. But, you said you had a problem on one side, and replaced a driveshaft. I'd concentrate on that. Wheel bearing, hub, recheck rotor, and alignment.

It may be a PORSCHE, but it is also just a car. You can run up a lot of money in parts if you have the attitude that it must be perfect and it becomes an obsession. So far, you have only found one thing that seems to indicate a problem at one corner. Run that to the ground first. Maybe get it up in the air and spin it up again. Shocks may be a contributor, but they are not the cause. You have a legitimate rotational issue at wheel speed. Stick with that.

Go over to Rennlist, and run this up the flagpole there in the proper forum. I can't believe you are the only person to experience this.

[brainz01]

Originally Posted by DanielDudley

These kind of problems tend to be exacerbated when the body of the car has a resonant frequency similar to the harmonics of the vibration. Early Miatas have this, and so do late model GMC trucks. Some tires can make this worse.

I had a car years ago with one bad roller on an inside wheel bearing. Otherwise the bearing was fine. I take my wheels to a local Chevy dealer now. There are different levels of road force that are considered acceptable. You need more than acceptable. But, you said you had a problem on one side, and replaced a driveshaft. I'd concentrate on that. Wheel bearing, hub, recheck rotor, and alignment.

It may be a PORSCHE, but it is also just a car. You can run up a lot of money in parts if you have the attitude that it must be perfect and it becomes an obsession. So far, you have only found one thing that seems to indicate a problem at one corner. Run that to the ground first. Maybe get it up in the air and spin it up again. Shocks may be a contributor, but they are not the cause. You have a legitimate rotational issue at wheel speed. Stick with that.

Go over to Rennlist, and run this up the flagpole there in the proper forum. I can't believe you are the only person to experience this.

All good advice. I'd tackled this in the past on the 997 turbo forum at RL, but I should have posted to the larger/more active 997 forum, or the track focused GT3 forum.

I have 2 known issues that need to be addressed:

1) water pump - - this is not my primary vibration issue, but interestingly, it does seem to generate a vibration signature that may be the source of my higher order vibrations.

2) front strut mounts are bad - - maybe this is part of the issue - - hopefully it's the issue.

I should also put the car on stands and run it in the air again and take a video of any vibrations. The right rear had looked bad in the past (my axle jump rope), but I've since measured all the runout and swapped the axle to no effect. I need to do the water pump first - - I don't want to chance doing engine damage - - new pump, belt, coolant tank, etc are on the way.

I feel no play or noise in the wheel bearings - - I've now spun them with the axles disconnected and the brake pads pushed back and they feel fine. I have replacement bearings ready if/when I decide to replace the suspension. Might as well "while I'm in there"...

Because I've swapped wheels with a known good set and because I'm not seeing a vibration under 65, I'm increasingly suspicious that there's something loose in the suspension (like the front strut mounts). And I think that the grooved pavement is exciting the vibration. If it was a wheel balance issue, I'd have expected to see a difference changing wheels and/or I should see the wheel vibration at 65 mph and below on the vibration app.

I know it's just a car and it will never be perfect, but I've driven other 997s and they don't have this vibration. Admittedly, it's not a safety/must fix issue, but it's really annoying to me as I drive the grooved stretch of road frequently.

It's always the last place you look, right?

[fanaudical]

Not sure about 997 CV joints, but every other Porsche and Audi CV that I have had apart had six balls in it.

I have some experience in vibration monitoring/analysis for rotating machinery (not nearly as extensive as Sammy and absolutely defer to him as the expert).

First couple of posts - your highest amplitude peak shifted up to around 37 Hz. What were you doing at that time?

You're getting this data from the accelerators in your cell phone, correct? How is the phone mounted in your car? Pictures of mount?

[RKDinOKC]

Where the valve stem caps on when the wheels were balanced?

[brainz01]

Originally Posted by fanaudical

Not sure about 997 CV joints, but every other Porsche and Audi CV that I have had apart had six balls in it.

I have some experience in vibration monitoring/analysis for rotating machinery (not nearly as extensive as Sammy and absolutely defer to him as the expert).

First couple of posts - your highest amplitude peak shifted up to around 37 Hz. What were you doing at that time?

You're getting this data from the accelerators in your cell phone, correct? How is the phone mounted in your car? Pictures of mount?

997 CVs are 6 ball on inners and outers.

The vibration app I'm using in my Android phone is called Vibration Isolator Pro. It's free and was the best/only app I could find without spending $$$. It's not perfect, but it is fairly repeatable, so seems useful for something.

My phone is supported upright in a cradle to the right of the center console. I tried other orientations/mounting locations and they were all about the same in terms of signature.

I don't tend to feel a vibration until there's a force above 0.2Gs. The 37hz spike is a 2nd order wheel speed at ~80mph. That frequency spikes when I'm driving on the grooved pavement. Otherwise, the primary frequency (~18hz) dominates.

I'm doing some additional measurements on my wife's SUV to confirm I'm not chasing red herring data.

[brainz01]

Originally Posted by RKDinOKC

Where the valve stem caps on when the wheels were balanced?

I believe so. I'm running ubiquitous black plastic caps - - nothing heavy or unusual.

[sammyg2]

Originally Posted by fanaudical

Not sure about 997 CV joints, but every other Porsche and Audi CV that I have had apart had six balls in it.

I have some experience in vibration monitoring/analysis for rotating machinery (not nearly as extensive as Sammy and absolutely defer to him as the expert).

First couple of posts - your highest amplitude peak shifted up to around 37 Hz. What were you doing at that time?

You're getting this data from the accelerators in your cell phone, correct? How is the phone mounted in your car? Pictures of mount?

Any guesses on the 7x frequency spike? Sounds like we can rule out a CV joint, it's got me puzzled.
Vibration analysis is not an exact science, at least not in my experience


Strategy can be boiled down to a few things:

1) identify the excitation force and eliminate it., IOW fix the broke stuff, balance it, etc.
2) Change the mass to change the the natural frequency away form a multiple of the excitation force
3) Change the stiffness to raise or lower the natural frequency.
4) dampen the vibration so it is not as noticeable.
5) any combination of the above.

Replacing the suspension bushings/mounts should dampen the noise to get it down to where it's not noticeable. Fingers crossed.

PS my favorite analysis was the Emerson/CSI 2140 data collection unit and their health manager software database, but it's expensive.
'Bout $25k to start and $10k per year subscription. That's why I don't have it no mo'.
Lots more affordable when someone else is paying.

[rusnak]

Originally Posted by brainz01

All good advice. I'd tackled this in the past on the 997 turbo forum at RL, but I should have posted to the larger/more active 997 forum, or the track focused GT3 forum.

I have 2 known issues that need to be addressed:

1) water pump - - this is not my primary vibration issue, but interestingly, it does seem to generate a vibration signature that may be the source of my higher order vibrations.

2) front strut mounts are bad - - maybe this is part of the issue - - hopefully it's the issue.

I should also put the car on stands and run it in the air again and take a video of any vibrations. The right rear had looked bad in the past (my axle jump rope), but I've since measured all the runout and swapped the axle to no effect. I need to do the water pump first - - I don't want to chance doing engine damage - - new pump, belt, coolant tank, etc are on the way.

I feel no play or noise in the wheel bearings - - I've now spun them with the axles disconnected and the brake pads pushed back and they feel fine. I have replacement bearings ready if/when I decide to replace the suspension. Might as well "while I'm in there"...

Because I've swapped wheels with a known good set and because I'm not seeing a vibration under 65, I'm increasingly suspicious that there's something loose in the suspension (like the front strut mounts). And I think that the grooved pavement is exciting the vibration. If it was a wheel balance issue, I'd have expected to see a difference changing wheels and/or I should see the wheel vibration at 65 mph and below on the vibration app.

I know it's just a car and it will never be perfect, but I've driven other 997s and they don't have this vibration. Admittedly, it's not a safety/must fix issue, but it's really annoying to me as I drive the grooved stretch of road frequently.

It's always the last place you look, right?

How many times have you put the car in the air to look at the wheels and spin them, with no improvement? You want to do that again? And now you are saying "replace the suspension"?

I would suggest changing the known bad parts first. And I've suggested many times to check the suspension BUSHINGS AND BOLTS for tightness. Don't forget to check the top mounting nut of each strut. Why would you replace the suspension links rather than check them first? We did a pretty good thread on Rennlist in the late 2010s in the 911 forum (air-cooled). There was a Porsche Mercedes Benz tech named "Whalebird". His advice would be applicable here, I feel.

[john70t]

A daily driver used to make me queasy every time I'd drive it on the highway.
The old tires looked fine and were balanced, but a new set fixed that problem.

[fanaudical]

Originally Posted by sammyg2

Any guesses on the 7x frequency spike? Sounds like we can rule out a CV joint, it's got me puzzled.
Vibration analysis is not an exact science, at least not in my experience


Strategy can be boiled down to a few things:

1) identify the excitation force and eliminate it., IOW fix the broke stuff, balance it, etc.
2) Change the mass to change the the natural frequency away form a multiple of the excitation force
3) Change the stiffness to raise or lower the natural frequency.
4) dampen the vibration so it is not as noticeable.
5) any combination of the above.

Replacing the suspension bushings/mounts should dampen the noise to get it down to where it's not noticeable. Fingers crossed.

PS my favorite analysis was the Emerson/CSI 2140 data collection unit and their health manager software database, but it's expensive.
'Bout $25k to start and $10k per year subscription. That's why I don't have it no mo'.
Lots more affordable when someone else is paying.

Not an exact science in my experience either - far from it.

The 7x spike has me puzzled too. I think it's closer to 8x. I'm wondering if its a structure resonance. Natural rotational frequency of a drive shaft or the center driveline?

Something to try (assuming the app can capture this - can it capture real-time spikes? Not just an average over time?):
- Park the car with the phone in the usual spot and recording.
- Smack each tire right on the tread with a BFH 2-3 times.
- Review the data - see if any of the familiar spikes are there.

(Old-school structural ringing test - maybe it shows something?)

Another thought - What's the signature look like at 40mph? Does everything shift?

What's the signature look like at 80mph both one gear up and down from normal?

Adding mass is always easy. If isolating this as suspension resonance, it might be easiest to put some sandbags in the passenger footwell and storage compartments. Load the thing close to the load limit. Take it for a drive. Any of the spikes moves frequency or amplitude?

Are the front and rear final drive ratios the same on this car? What are the ratios?

Sammy - I never had access to the cool stuff you did. I was stuck with a few hard-mount accelerometers and a borrowed 40MHz LaCroy digital o'scope with FFT and good capture capabilities - had to make do.

[eastbay]

If the 997 suspension parts are the same quality as the 996, I would look at your control arms, ball joints, etc. in the rear hi torque areas. These are garbage components made with nylon on steel balls and bearings with no provision for lubrication. They wear fast and strangley and can cause no end of phantom and fleeting issues.

[brainz01]

Originally Posted by rusnak

How many times have you put the car in the air to look at the wheels and spin them, with no improvement? You want to do that again? And now you are saying "replace the suspension"?

I would suggest changing the known bad parts first. And I've suggested many times to check the suspension BUSHINGS AND BOLTS for tightness. Don't forget to check the top mounting nut of each strut. Why would you replace the suspension links rather than check them first? We did a pretty good thread on Rennlist in the late 2010s in the 911 forum (air-cooled). There was a Porsche Mercedes Benz tech named "Whalebird". His advice would be applicable here, I feel.

I hear you. I'm not looking to spend money on parts I don't need, although I could point to $250 worth of wheel bearings that may be unnecessary...

I started at the back of the car because it feels like the vibration is through the body. But everything seems tight and the rotating parts checked out. There's nothing obviously wrong or worn.

Moving to the front, the strut mounts look very suspicious, but they're not to the point of rattling away. I could hear them occasionally on a hard bump, but they were prominent with the Chassis Ears though. Replacing the strut mounts entails undoing most of the front end (again). So if I'm putting in the time, maybe I also replace the 13 year old control arms and tie rods? It's all gonna need to be realigned anyway.

I don't think I need new shocks, and don't plan on replacing them - - there's no leaks and they definitely firm up when commanded to. I also don't think I need to make any replacements on the back suspension arms either. It all seems tight, though admittedly I've not run the Chassis Ears back there. I suppose the shock mounts could be worn and not making noise... One thing at a time.

My current theory is 2 part:

1). The failed water pump may be the source of the higher order vibrations that might otherwise suggest wheel bearing - - essentially a red herring, but a different failed bearing

2). The failed strut top bushings are allowing a tiny "wheel slapper" oscillation that's felt primarily through the chassis, as the lower control arms and tie rods appear tight? And the grooved pavement helps excite that oscillation. Maybe one bushing is a bit more damped than the other side which gives the cyclical beating?

Also, I ran the vibration app on my wife's Mercedes GL 450 today. There's a different signature for sure - - generally much lower recorded vibration levels - - typically at or below the 0.2Gs I thought was where I could start to feel a vibration. See the attached charts to see what I mean. Note that at freeway speeds (75 mph) her 1st order wheel speed frequency is ~13.5hz. And her diff ratios are 3.45. You can clearly see she's got slight wheel and driveshaft imbalance. Note that you wouldn't know it - - it's a very smooth ride. And no weird high order vibrations. Interesting data - - I'd say the app is working.

That's all I got for now.