Gear Box Failure @ 43 Hours

Discussion in 'General Auto Conversion Discussion' started by TXFlyGuy, Oct 31, 2019.

Help Support HomeBuiltAirplanes Forum by donating:

  1. Nov 8, 2019 #101

    pfarber

    pfarber

    pfarber

    Well-Known Member

    Joined:
    Feb 21, 2019
    Messages:
    292
    Likes Received:
    15
    Location:
    Pennsylvania
    An optical sensor would be a bad idea.

    From the 'one that I like':

    3. Prism surface must be at least 2˝ from any reflective surfaces.

    6. Inspect and clean lens at regular intervals. Level switch not recommended for use with liquids that leave deposits.

    And this limitation:

    Mounting Orientation: Horizontal or up to 45° from horizontal.

    I like capacitive level sensors. Dead simple, nothing to clean

    https://www.gemssensors.com/level/single-point-level-switches/capacitive/cap-300-level-switch
     
    wsimpso1 likes this.
  2. Nov 8, 2019 #102

    Russell

    Russell

    Russell

    Member

    Joined:
    Nov 30, 2015
    Messages:
    17
    Likes Received:
    23
    Location:
    Ft Worth, Texas
    Chip detectors have been mentioned here several times. Years ago I flew part time for a pipeline patrol company that had six fixed wing and three rotary wing aircraft. The airplanes would usually fly 4 to 7 hours per day. They installed chip detectors in all of the airplanes. Normal wear on perfectly good engines would set off the alarm. False alarms caused maintenance costs and aircraft down time to skyrocket. Within a year they removed all of the chip detectors. As far as I know, none of the alarms ever revealed a real problem.
    Russell Sherwood Glasair/Subaru EG33
     
    proppastie likes this.
  3. Nov 8, 2019 #103

    wsimpso1

    wsimpso1

    wsimpso1

    Super Moderator Staff Member

    Joined:
    Oct 19, 2003
    Messages:
    5,893
    Likes Received:
    3,119
    Location:
    Saline Michigan
    Sounds like a calibration/design issue. We used them in auto trans dyno testing when we were working on gearset and bearing issues, and they picked up incipient failures ahead of other symptoms. We also used them on stationary test equipment to good results.
     
  4. Nov 8, 2019 #104

    wsimpso1

    wsimpso1

    wsimpso1

    Super Moderator Staff Member

    Joined:
    Oct 19, 2003
    Messages:
    5,893
    Likes Received:
    3,119
    Location:
    Saline Michigan
    If internal parts are too close, you can screw in a pipe nipple and attach the sensor...

    If your gearbox oil is varnishing your sensors, you either need an oil that includes a detergent additive or your oil is running too hot or both. Most folks talk about how these gearboxes running cool enough that you can hold your hand on them - 150F is about the max you can do that on metals.

    Yep, something to know when selecting a mounting. Their information has this limitation AND they tell us it can mounted vertically or horizontally AND they tell us it can be mounted in any orientation. Wonder which one is really the case...

    Where was this response when I was looking for low fuel level alarms? I am using capacitance fuel gauges, but also wanted a redundant low level alarm. Got any idea what these guys cost? The 257F max temp is better for gearbox stuff than the 175F max for the Dwyer optical sensors but would still be violated in overtemp conditions... I wonder what the failure modes are in overheat for the Gem? Trigger the alarm? The Dwyer housing will melt in overtemp, and so does look to be a poor choice in the gearbox, but still OK in fuel tanks. My bad on offering it for gearbox.

    Billski
     
  5. Nov 8, 2019 #105

    proppastie

    proppastie

    proppastie

    Well-Known Member

    Joined:
    Feb 19, 2012
    Messages:
    3,797
    Likes Received:
    1,098
    Location:
    NJ
    are they just simple open/closed circuit gap/magnet? would any slight amount of metal close the circuit?
     
  6. Nov 8, 2019 #106

    Russell

    Russell

    Russell

    Member

    Joined:
    Nov 30, 2015
    Messages:
    17
    Likes Received:
    23
    Location:
    Ft Worth, Texas
    The chip detectors installed by the patrol company were a fixed gap that closed the circuit when metal filled the gap.
     
  7. Nov 8, 2019 #107

    pfarber

    pfarber

    pfarber

    Well-Known Member

    Joined:
    Feb 21, 2019
    Messages:
    292
    Likes Received:
    15
    Location:
    Pennsylvania

    My first though was even simpler, just a simple resistance bridge using the oil as a leg. But a capacitor setup seems just as simple. Also the unit I referenced is made for vibration environments.
     
  8. Nov 8, 2019 #108

    narfi

    narfi

    narfi

    Well-Known Member

    Joined:
    Feb 23, 2016
    Messages:
    751
    Likes Received:
    376
    Location:
    Alaska
    I know of a PT6 that had a light some time after takeoff and they returned to base (some 1+hrs in the air) and the engine was ruined, but they made it back safe.

    Yes, that is how the chip detectors work in the pt6
     
    proppastie likes this.
  9. Nov 9, 2019 #109

    Deuelly

    Deuelly

    Deuelly

    Well-Known Member

    Joined:
    Nov 6, 2012
    Messages:
    131
    Likes Received:
    121
    Location:
    Marshall, MN
    Many big radials run chip detectors. Many P-51s also run them for their Merlins. The only times I've ever seen the lights come on are when the engine is on it's way out. When sized and positioned right, chip detectors are very accurate and reliable.

    Brandon
     
  10. Nov 9, 2019 #110

    TFF

    TFF

    TFF

    Well-Known Member

    Joined:
    Apr 28, 2010
    Messages:
    11,639
    Likes Received:
    3,269
    Location:
    Memphis, TN
    Depends. All certified helicopters have chip detectors. Older helicopters may have ones with no light that take manual inspection, but all built since the late 80s requires cockpit indicators. Depends on the type of gearbox. My helicopters have ring and pinion like a car, probably more like a bob truck. R-22s use ford nine inch specifications. They will grind up bits and don’t seem to come apart. Airbus, formally known as Eurocopter, have planetary. Light pops on one one of those, land immediately because in five minutes you ARE doing an autorotation unless it locks it all up, which has happened. A chip detector on an engine that’s not a turbine is a waste. It is fun though when you pull a filter and it all looks like metallic paint in there.
     
  11. Nov 9, 2019 #111

    wsimpso1

    wsimpso1

    wsimpso1

    Super Moderator Staff Member

    Joined:
    Oct 19, 2003
    Messages:
    5,893
    Likes Received:
    3,119
    Location:
    Saline Michigan
    Chip detectors intended for turbines and then used on piston engines are probably not a good scheme. Piston rings and some other engine parts make a lot of tiny iron bits in normal operation.

    In this thread, we are talking about a gearbox with its own oil supply, and it should stay clean until something goes wrong with it. Putting a chip detector on this new model gearbox is what we were talking about while this first flight article is run. Maybe it will be useful... I hope it runs clean and cool.

    Billski
     
    Last edited: Nov 9, 2019
  12. Nov 9, 2019 #112

    Andreas K

    Andreas K

    Andreas K

    New Member

    Joined:
    Aug 19, 2016
    Messages:
    4
    Likes Received:
    5
    Location:
    Beautiful Wisconsin
    I didn't loose the prop, just power. Happened on climbout at 350' AGL. Made it back to the airport. It was according to the FAA not even an incident, so I am not in any statistic.

    It looks like PSRU's seem to work in the lower power range but above 180 hp it gets tough to built something durable. I prefer the LS engine over a Lycoming every day. Just not in an airplane and that is not the fault of the engine.
    Fly Safe
     
    proppastie, Winginitt and dino like this.
  13. Nov 9, 2019 #113

    Winginitt

    Winginitt

    Winginitt

    Well-Known Member

    Joined:
    Apr 5, 2019
    Messages:
    176
    Likes Received:
    47
    Do you have any pictures of the destroyed coupling/shaft ? Do you think something like an ATI or a Fluid damper would have prevented the problem ?
     
  14. Nov 9, 2019 #114

    pfarber

    pfarber

    pfarber

    Well-Known Member

    Joined:
    Feb 21, 2019
    Messages:
    292
    Likes Received:
    15
    Location:
    Pennsylvania
    ICE engines are designed for the main and rod bearings to imbed a certain level of debris. There are much better ways to support the crank than with babbit on steel shells, but most will not tolerate debris.

    I see a lot high revving 2 cylinder engines use ball bearings on the crank... works great.. till the bearings start to see to much debris.

    I don't think a chip detector brings a lot to the table. I would go with a temp probe and a sight glass.

    If you start making chips or lose fluid, oil temp will rise. And if either of former occurs, then your gears as pretty much toast.
     
  15. Nov 10, 2019 at 4:10 AM #115

    Winginitt

    Winginitt

    Winginitt

    Well-Known Member

    Joined:
    Apr 5, 2019
    Messages:
    176
    Likes Received:
    47
    I agree that the two primary things should be a sight glass and a temperature indicator. I think it would also be easy to install an oil level light. As far as metal chips, I would install a simple
    oil dip stick with a magnet on it, or a magnetised dipstick. Check it before and after every flight.
     
  16. Nov 10, 2019 at 5:07 AM #116

    Aviacs

    Aviacs

    Aviacs

    Member

    Joined:
    Oct 21, 2019
    Messages:
    16
    Likes Received:
    5
    Well this is beside the main points of the OP, however:

    You should define "better" then?

    Seems that is a large factor in typical IC engine/vehicle application.

    Again you seem to be arguing against your proposition (there are better ways <multiple ?> to support a crank)?
    Yes/no/maybe: a well designed ballbearing assembly still begins to wear out with the first revolution. For commercial bearings, they are designed to statistically last at design load for 2,000 hours. So some fail more quickly, some live multiples of the predicted interval & changing the loads can yield predictable changes in the life interval . Nonetheless, even with lube, ball bearing sets will gradually fail at a predictable rate. Some preloaded sets will eventually deteriorate from "environmental" stress without being run. OTOH, hydrodynamic bearings need never fail so long as pre-pressurized & run at design rpm range with correct clean enough lubricant. Most power turbines are spinning on plain bearings (steam, hydro, nuke) In some cases ball bearing might not be the best, but like many systems they are easy to monitor & predict the lifespan and (relatively) simple & easy to change in the field, compared to the complexity of good plain bearing systems.

    I like the idea of chip detection in a critical gearbox because even if the wear particles that cause "indication" are small, in a gear set that could be significant. I'd certainly want to know (record) how fast it trips again after cleaning; and at what iteration of tripping events wear on the gear teeth becomes critical. (Also, based on the data, a more or less sensitive indicator could be contrived).

    smt
     
    proppastie likes this.
  17. Nov 10, 2019 at 5:44 AM #117

    AdrianS

    AdrianS

    AdrianS

    Well-Known Member

    Joined:
    Jul 5, 2014
    Messages:
    378
    Likes Received:
    228
    Location:
    Australia
    I doubt that I will still have a photo.

    The solution in that instance was to have a tv damper designed by a specialist company.

    And this is why I bang on about TV : the components that failed saw a loading multiple times the design load.
    There is a naive view that the shaft loads are limited to the engine output torque : unchecked TV can cause a factor of 5? 10? ? multiplier.

    I've also seen soft coupler failure that was traced to TV : the coupling did it's job of absorbing torque peaks, but overheated the "rubber" element doing so, leading to premature failure.

    I am not a transmission engineer - but have seen enough drive failures to have a healthy distrust of "simple" solutions.
     
    proppastie and akwrencher like this.
  18. Nov 10, 2019 at 8:25 AM #118

    mullacharjak

    mullacharjak

    mullacharjak

    Well-Known Member

    Joined:
    Nov 5, 2010
    Messages:
    67
    Likes Received:
    17
    Location:
    Peshawar Pakistan
    The Prop shaft failed so maybe this was a design flaw and needed a remedy.Doesnt mean the rest of the drive was inadequate.I read somewhere the Continental A65 crankshaft failed with the taper fit prop driver.All aircraft specific engines have one piece integral prop drivers for some reason.Vw engine crankshafts also fail at the propdriver/crank interface.All Rotax gear drives have one piece prop shafts at the end where the prop is attached.Did you try to find out where your prop shaft broke?
     
  19. Nov 10, 2019 at 12:57 PM #119

    proppastie

    proppastie

    proppastie

    Well-Known Member

    Joined:
    Feb 19, 2012
    Messages:
    3,797
    Likes Received:
    1,098
    Location:
    NJ
    great idea for homebuilts, not sure how well it would work for aluminum, would be illegal for certified unless you get a STC.
     
  20. Nov 10, 2019 at 2:01 PM #120

    wsimpso1

    wsimpso1

    wsimpso1

    Super Moderator Staff Member

    Joined:
    Oct 19, 2003
    Messages:
    5,893
    Likes Received:
    3,119
    Location:
    Saline Michigan
    We have been exploring several means for having the teeth strip from the gears on the subject gearbox. The teeth appear to have been softer than their spec after failure, but there is also unmistakable evidence of overheating and low oil. We have talked a little about how to catch overheating and low oil levels.

    AdrianS brings up one of my favorite worries in PSRU. Torsional Vibration produces cycling loading that can come to us in a few forms - resonance and impact being particularly bad. Briefly on each:

    Resonance is where some vibratory input lines up with some natural frequency, and can amplify the forcing function hugely. Right at natural frequency there is theoretically no upper bound - in reality things break. Being within a quarter octave of it can be quickly destructive. I have talked about this a lot, and the answer is to drive the natural frequencies away from the vibratory inputs. Straightforward, but can be difficult. Usual approach is to put a suitably soft element in the the right place. Springs, elastomeric elements, quill shafts, that sort of thing, and they all can work if done correctly. Done wrong, they can move resonance off of one condition that drives failure and into another;

    Cyclic loading - Even if you are not in resonance, firing pulses and other prominent inputs from the engine, prop, etc do cause the torque in the system to go up and down. If components are just big enough for the average load repeated forever (steel does have an endurance limit - keep loads below that and they will not fatigue) but the the load actually goes up and down significantly on every turn of the prop or firing pulse or piston cycle, you can see fatigue parts and cause failures. Normally we design these types of systems with a soft element (see the above paragraph), the springiness of the soft element allows resonance to be significantly below the operating range, and the vast majority of vibration from the source is isolated from the rest of the system. If the soft element is not acting as a soft spring, we can have problems. For instance is you have coil spring sets like in a manual clutch disc, the springs can be poorly done and break or bottom. In both cases, the spring rates become much higher and resonance can move from a low and safe frequency to a higher more hazardous frequency or we can just transmit all of the vibration through the system. Gear teeth, splines, shafts, bearings, and other elements of the gearbox must be quite a bit stronger to stand this vibration than if the soft element is healthy and working as intended. If this were a new design system or new engine application for the PSRU, I would be concerned about the soft element's condition. Were the isolator parts overheated, failed in some manner, or indicating bottoming?

    Then there is impact loading. This is cyclic loading where loads go between zero and some dynamic load. Here is your basic estimator - if you put a column on a solid support, and then placed a weight so it just touches but puts zero force on the column, then let go of it so that gravity has its full effect on driving the weight down onto the column, peak load in the column is twice the load from just the weight sitting gently on the column. Yep, no speed to speak of doubles the load. Lift the weight up a little, let go of it so that it gets a running start at the column, and impact increases. So, if torque gets to zero at the gears and the gears completely unload but the lash in the system does not open, we can see double the nominal forces. Open the lash and it gets worse. This can also happen if you are in resonance, have your soft element become stiffer so that you transmit firing torques, and others.

    So, I wonder what the condition was of the isolators between the engine and PSRU of the subject equipment. If the elastomeric elements had been cooked to death, were missing, mashed to pieces, or otherwise AWOL, we might know more about how our parts came to grief.

    None of this insurmountable. ALL of the big round and V engines used in hhuge numbers during WWII and after were geared engines - vibration was figured out.

    These situations are why I tend to prefer that PSRU systems be ENGINEERED to avoid resonance and then TESTED with torsional vibe measurement equipment (and the range of prop MMOI) to confirm that the system is behaving. Even then I am pretty conservative - it should still see a bunch of time running on a stand with the worst prop then in flight for a bunch more hours by someone who knows what is being done. Once we have a good history, I can begin to get behind it. Now this brand new box and prop, well, I would get serious about temps and oil levels and debris detection and isolator monitoring on every flight for quite a while.

    Billski
     
    AdrianS likes this.

Share This Page

arrow_white