Vibration meter

[Bruce Tuncks]

Does anybody have ideas about a panel-mounted vibration meter ?

 

It would be easy enough to do. At the front of the engine, you would mount a sender and then a digital readout on the panel, with some fairly simple electronics to put a displayed number from the sender output.

 

Vibration senders are not difficult to buy. Calibration would be an issue, perhaps you could try it in a known smooth setup to get a baseline. I don't think the actual figures would matter much, but the change would tell you things. Perhaps it would be worth going to the effort of calibrating it in inches per second or whatever they use. You would need another ( calibrated) meter to do this.

 

The reason for wanting one is that I want to know if a subtle change, like a prop coming a bit loose, has occurred. My sense of feel isn't reliable enough . I would like a meter that showed a big increase in numbers if this happened. But would a simple vibration meter pick up useful information or not ? Or would the normal variation in different flight stages swamp the indication you want to notice?

 

 

[mnewbery]

http://www.skybrary.aero/bookshelf/books/2637.pdf

[facthunter]

The Allison turboprop in Hercules and Lockheed Electra had low frequency meters on the props which were very prone to cracking. They are steel sheet fabricated items. The meter would be obviously set tor record specific vibrations which characterised the cracks being experienced. I doubt you could get a meaningful result with out years of analysis, if you install something, without a lot of guidance. You might be better finding some means of measuring any creep on the hub face with the prop contact area.. It apparently gets hot and chars with wood. Nev

 

 

[crashley]

I have also thought about fitting a probe bently nevada make lots of these probes in different ranges typically something like 0 to 1" per second they are generally 24 volt supplied and have a 4-20 ma output

 

does anyone have the typical vibration level of a 912 rotax

 

 

[Bruce Tuncks]

I have seen a Jabiru prop where the bearing face was charred by the rubbing . Yes it was a loose prop, but it didn't feel like this on hand turnover.

 

But I would like more from a vibration meter, such as to pick up detonation events and even "rough running" from too lean a mixture.

 

But if it were this easy, surely it would have been done before?

 

And thanks man baby for the clip, but it didn't fully answer my question about what would show up on one of our planes.

 

 

[mnewbery]

Detonation won't show up unless the sensor is for that. A knock sensor is quite common on pretty much every petrol engine that uses an electronic engine computer (EEC) to control spark advance or retard.

 

Knock Sensor Function Symptoms and Problems

 

http://www.standardbrand.com/media/885972/ST10235KS-FEB14_JTF_Knock.pdf

 

As far as I know the know sensor sends a pulse every time a knock occurs. The internals are piezo electric so you need to integrate that signal into some indicator or recorder. Plus auto engines have one knock sensor. This won't do on an aero engine with widely spaced cylinders. They need one per cylinder.

 

Finally the knock sensors are make/model specific and sensitive to placement so taking one from an auto application may give false positives or not show anything. Either way, they allow really rapid seamless tuning changes at the EEC but often go wrong and cause the check engine light to come on.

 

Integrating this feature into an aero engine may not have the desired effect.

 

Why worry about knock in an aircraft engine?

 

While it may be attractive to tune an engine to the point where it knocks then back "it" off a bit, what use is a sensor when you are climbing out over a corn field/trees on a hot day and it starts indicating?

 

 

[Old Koreelah]

Does anybody have ideas about a panel-mounted vibration meter?...

Yep. I fill in my paper flight plans while the engine is warming up. Look back thru past plans to compare shakiness of writing...or is that Parkinson's ?

 

 

[mnewbery]

A comment on HUMS. Let's say that these heavy and expensive sensors were already installed on a typical recreational aircraft.

 

Firstly they produce an insane amount of data every hour and this needs to be stored on the aircraft until it is possible to offload.

 

Secondly, the data needs to be interpreted for "usage". This is the kind of data that Rolls Royce, MTU, SNECMA and others use to create trend information. A baseline for each individual engine is set up over about 10 cycles and then the variations over time are compared. This is called trend information. Vibration is just one measure. Other interesting stuff is %NG vs torque, ITT vs EGT, fluid temperatures and so-on. As an engine and gearbox wears, these numbers and ratios change. Also, every time the pilot (or mine truck driver) does something a bit extreme like using emergency power settings or a bit of a pickled landing that gets logged too.

 

When a component doesn't make it to book TBO, or it it does but it's stuffed anyway, the manufacturers already have the fingerprints and can look at other examples and the oil if that is involved.

 

Sometimes they have all this and the main rotor or front wheel still comes off, killing everyone.

 

Oil rig helicopter crashes off Norway coast, 13 presumed dead

 

Is instrumenting a light aircraft with a vibration sensor like trying to put an oil temperature gauge on a Briggs and Stratton lawn mower?

 

 

[rhtrudder]

A comment on HUMS. Let's say that these heavy and expensive sensors were already installed on a typical recreational aircraft.Firstly they produce an insane amount of data every hour and this needs to be stored on the aircraft until it is possible to offload.

 

Secondly, the data needs to be interpreted for "usage". This is the kind of data that Rolls Royce, MTU, SNECMA and others use to create trend information. A baseline for each individual engine is set up over about 10 cycles and then the variations over time are compared. This is called trend information. Vibration is just one measure. Other interesting stuff is %NG vs torque, ITT vs EGT, fluid temperatures and so-on. As an engine and gearbox wears, these numbers and ratios change. Also, every time the pilot (or mine truck driver) does something a bit extreme like using emergency power settings or a bit of a pickled landing that gets logged too.

 

When a component doesn't make it to book TBO, or it it does but it's stuffed anyway, the manufacturers already have the fingerprints and can look at other examples and the oil if that is involved.

 

Sometimes they have all this and the main rotor or front wheel still comes off, killing everyone.

 

Oil rig helicopter crashes off Norway coast, 13 presumed dead

 

Is instrumenting a light aircraft with a vibration sensor like trying to put an oil temperature gauge on a Briggs and Stratton lawn mower?

[rhtrudder]

I fitted a knock sensor to my 914, thinking it might show up stale fuel , bolted the Mike somewhere near the rocker cover , series of lights on a display inside cab which light up when engine is high up in the rev range but I don't think it picks up any detonation or maybe the motor has never had a problem

 

 

[Jabiru7252]

perhaps this might be of use.....

 

Vibration Meter - Android Apps on Google Play

 

It's an Android app used to turn the phone into a vibration meter.

 

 

[M61A1]

The vibration meter would not only readout IPS, but also need to give you a frequency, then you have to correlate the frequency at which the vibe is occurring with what ever is turning at that frequency. Helicopter ones run a "spectrum analysis", and maintenance manuals will often list which components (main rotor shaft, input shafts, accessory shafts, tail rotor shafts, intermediate gearboxes and any number of internal engine bits) turn at that speed for fault finding.

 

like this...

 

 

 

[Bruce Tuncks]

Thanks guys... I am getting the message that a cheap vibration meter without frequency output will not be useful.

 

In my imagination, I saw a meter which gave a steady low reading until some event happened like a flywheel bolt coming loose or some bad fuel causing detonation, whereupon the meter would go to a high number. Then I would just have to find the cause and fix it before anything broke.

 

Too easy to be true huh.

 

 

[Kyle Communications]

 

Internet of Things: High Sensitivity Vibration Sensor Using Arduino

 

 

[onetrack]

Thanks guys... I am getting the message that a cheap vibration meter without frequency output will not be useful.In my imagination, I saw a meter which gave a steady low reading until some event happened like a flywheel bolt coming loose or some bad fuel causing detonation, whereupon the meter would go to a high number. Then I would just have to find the cause and fix it before anything broke.

Too easy to be true huh.

Nope. Many many decades ago, when I still had brown hair and lots of it, we utilised a cunning German device called a Kienzle Tachograph. These units were also built by a company called Argo.

The device was designed for trucks and machinery. When fitted to trucks, they contained a speedo - when fitted to machinery, they only contained a clock.

 

The device worked on the vibration of the surface it was attached to - and then recorded the level of vibration the tachograph was subjected to, via a sensitive vibrating pointer, scratching the surface of a rotating paper card.

 

The card was inserted in a pack of 7 tied together by small tabs, for a weeks use - and the cards turned 360 deg each day. Hours were marked in segments totalling 24 segments to the circumference of the card.

 

As one card finished, at the end of each 24 hr period, the tab was split by an inbuilt knife device, and the pointer went onto the next card.

 

Because the pointer was highly sensitive to vibration, rough operation of the truck or machine would show up as spikes in an otherwise fairly steady circular line drawn by the pointer.

 

The truck versions produced a road speed record, idle time record, excessive bouncing of the vehicle, RPM levels (including engine overspeed), and distance travelled.

 

On machinery, they showed operating time, stopped time, and rough operation, indicated by big spikes in the height of the scored line, just like a seismograph shows earthquake events.

 

The tachograph is a particularly reliable instrument - but perhaps this version isn't precisely what you're seeking, because they're pretty heavy - and they only reveal the severe vibration event, as a past, recorded event.

 

Here's a link to the Argo Tachograph. Not sure if simple mechanical tachographs are still made - like everything today, they have been superseded by electronic versions.

 

Perhaps there is an electronic version that can be used for what you have in mind, you need to talk to some precision instrument supplier.

 

http://canwesttach.com/pdf/Argo_Tachograph_Canwest_Tachograph.pdf

 

EDIT - Seems like Argo bought out the Kienzle Company, and I understand that Argo is now a division of VDO.

 

 

[Jabiru7252]

Thanks guys... I am getting the message that a cheap vibration meter without frequency output will not be useful.In my imagination, I saw a meter which gave a steady low reading until some event happened like a flywheel bolt coming loose or some bad fuel causing detonation, whereupon the meter would go to a high number. Then I would just have to find the cause and fix it before anything broke.

Too easy to be true huh.

Let's say you had a steady output from a sensor of around 1 volt RMS when attached to your smooth running engine. Then something goes wrong and the vibration increases. If that causes a sudden change in output level, could that not be a trigger to warn you? I think it's worth some investigation.

 

 

[Bruce Tuncks]

What I had in mind was similar to the pics from Mark Kyle, except it would have a digital display and a different sender, probably a piezo one which is rated up to the frequency of knock sensors.It would only cost about $50 so I may have a go at it for the fun.

 

Maybe there would need to be more than one sensor, a low-frequency one to suit engine rpm and a much higher one .

 

I can imagine the low-frequency figure would be useful in balancing the prop at least.

 

 

[facthunter]

You don't want this stuff in your plane. It's distracting from what you are there for. It's about flying, not having a flying testbed for an engine. They removed the vibration meters from the B727 and that had a full time flight engineer. The early B 747 had so many "failure modes" for rejected take off they had to totally revise the arrangement, down to a small fraction of what it had been as it was creating situations that weren't safe. Nev

 

 

[waraton]

Go for it Bruce, tinker away. You love this sort of thing and are pretty good at it as your temp monitors demonstrated. We all love different aspects of recreational flying and I am already looking forward to watching your progress and results whatever they may be.

 

 

[jetjr]

Smart avionics in UK sell a prop balancer unit which would do what you want. Can be used to watch general vibrations too. Output on android via bluetooth

 

Vibration analysis is common on big plant and machinery with service and replacement done based on the results. Sometimes a spot check other times constant monitoring

 

The gear is out there and not crazy expensive. Making sense of output isnt simple.

 

Baseline readings a are key to getting it to work

 

 

[cscotthendry]

Depending on your electronics ability, I can think of a simple system that wouldn't cost much to make.

 

Using a piezo sensor like in the video posted by Mark, you could amplify that signal through an opamp, then present that amplified signal to a comparator ladder and use the outputs of the comparators to light up an LED bargraph. It might take some trials to get the reference voltages in the comparator ladder set to show something meaningful, but the bargraph display would give a simple readout that would give you trending info.

 

If you went the digital rout with an Arduino or a PIC microcontroller, you could display numeric info, but then you have to READ the meter and interpret the numbers. Of course you could display the numeric info as well as a bargraph on a simple 40character 2 line LCD display, but the project starts to become a career move at that point.

 

PM me if you're interested in more detail of the simple version.