LifeofReiley
Well Known Member
What's the use talking here... this Egg guy will be out of business very soon. I feel sorry for anyone who has invested in this.
Van's Air Force
Don't miss anything! Register now for full access to the definitive RV support community.
Egg Gearbox Failure
- Thread starter Yukon
- Start date
Geico266
Well Known Member
My Norwegian is a tad rusty.
It would be interesting to know more about the history of that engine, and how it was maintained. Was there any follow up investigation? On the other hand, one gear box failure certainly is not a trend with as many Rotax 912 / 912S's that are flying. What this tells me is gear boxes can be made reliable for aircraft applications. The reason I even bring Rotax up is to use it as an example so the VAF engineers can look at it and learn.
David-aviator
Well Known Member
Nothing is perfect, not even an analogy.
Some early versions of EGG PSRU's have failed, but generally there is some warning before the failure. If the temperature is running 200+, watch out. I have not been able to get GEN3 near 200 so far. (Last time I flew before this darn ice storm, it ran at 145.)
There have been no Subaru crank shaft failures. If one compares engine crank shafts, this is a non issue with Subaru.
If a Lycoming crank shaft fails there is no warning. It's running fine and a second later it is not running at all except for some noise. It is not a huge problem in than most Lycoming cranks do not fail, but even after some 80 years of development, they have not been able to make them all bullet proof. They change this or that, find a cheaper vendor, and invariably come out with a version that fails and the recall follows. This can not give anyone a warm and fuzzy feeling.
Hey, the first thing that would make this forum more productive is to skip Lycoming/alternative engine comparisons. Lycoming is the standard. Serious work, not BS and not random experimentation, will allow the alternative engines to catch up. Yeah, they both break from time to time, so enough already.
Now, who has even so much as a photograph of Egg Gen III internals? Am I to believe that all Egg owners trade a box full of money for a box full of gears and have never looked in the box?
Operating data: David admits to running his system at an RPM that generates noise from the box. What we don't know is how often, and if it is likely to be common among other operators.
How about it? Can we set aside the flag waving and discuss reality?
Now, who has even so much as a photograph of Egg Gen III internals? Am I to believe that all Egg owners trade a box full of money for a box full of gears and have never looked in the box?
Operating data: David admits to running his system at an RPM that generates noise from the box. What we don't know is how often, and if it is likely to be common among other operators.
How about it? Can we set aside the flag waving and discuss reality?
Jan had a photo of the gearbox guts on his site a few months back but I can no longer find it. Maybe someone has this on their HD somewhere?
Like Rotax, Jan is recommending a 1400 rpm idle now. If you own either package, it's a good idea to follow the recommendation probably.
Like Rotax, Jan is recommending a 1400 rpm idle now. If you own either package, it's a good idea to follow the recommendation probably.
flytoboat
Well Known Member
Egg PSRU
About a third of the way down this page are more pictures:
http://www.eggenfellneraircraft.com/News - Earlier.htm
About a third of the way down this page are more pictures:
http://www.eggenfellneraircraft.com/News - Earlier.htm
The gearbox was sent to Rotax for examination, but no news as of yet. The report doesn't say anything about maintainance and history of the engine, not at this time.My Norwegian is a tad rusty.
It would be interesting to know more about the history of that engine, and how it was maintained. Was there any follow up investigation? On the other hand, one gear box failure certainly is not a trend with as many Rotax 912 / 912S's that are flying. What this tells me is gear boxes can be made reliable for aircraft applications. The reason I even bring Rotax up is to use it as an example so the VAF engineers can look at it and learn.
Thank you Don.
Pictures have limitations of course, but they're a whole lot better than guessing.
Gear A drives gear B, which drives C, which drives D. The D gear is attached to the propshaft. The tail of the propshaft runs in a bearing inside A.
The arrangement of the helical gear angles means torque reversal at a resonant RPM would indeed drive the shafts fore and aft axially (the arrows) at a rate in the hertz range. This would beat the snot out of the bearings. The beating would be much reduced if axial freeplay is carefully shimmed or otherwise kept to a minimum, but offhand I don't think there is any way to eliminate all of it. I can't tell what Jan does to minimize freeplay from just a photo. In this design it needs to be minimized. You can tell the operators absolutely, positively do not run in a resonant range (apparently below 1400 in this case), but that only reduces the total number of hits in a given period of operation. You still must pass through that range at startup, so minimizing axial freeplay to keep the hits small would increase bearing life.
All bets are off if operators routinely ignore the prohibition against low idle speed. I gather there is no mechanical stop or device to prevent low idle, so I'll bet it happens all the time.
Anyway, it's a theory. I could be wrong. Comments?
Pictures have limitations of course, but they're a whole lot better than guessing.
Gear A drives gear B, which drives C, which drives D. The D gear is attached to the propshaft. The tail of the propshaft runs in a bearing inside A.
The arrangement of the helical gear angles means torque reversal at a resonant RPM would indeed drive the shafts fore and aft axially (the arrows) at a rate in the hertz range. This would beat the snot out of the bearings. The beating would be much reduced if axial freeplay is carefully shimmed or otherwise kept to a minimum, but offhand I don't think there is any way to eliminate all of it. I can't tell what Jan does to minimize freeplay from just a photo. In this design it needs to be minimized. You can tell the operators absolutely, positively do not run in a resonant range (apparently below 1400 in this case), but that only reduces the total number of hits in a given period of operation. You still must pass through that range at startup, so minimizing axial freeplay to keep the hits small would increase bearing life.
All bets are off if operators routinely ignore the prohibition against low idle speed. I gather there is no mechanical stop or device to prevent low idle, so I'll bet it happens all the time.
Anyway, it's a theory. I could be wrong. Comments?
gmcjetpilot
Well Known Member
Mic it?
If low idle is a no-no, than the idle should be set high. Right?
I've got generic engineering knowledge of gears. My knowledge of PSRU's is from the few I've seen around the airport and what's on the web. Why helical gears? They develop thrust loads to start with; on the plus side they're stronger than spur gear due to more tooth contact area continuously in contact, making it more suited for high speed.
Still involute spur gears (straight cut gears) are stout and work well. Big plus is no thrust loads. The Lyc accessory case GEAR BOX is an example of spur gears. Spur gears are louder, but so what, they are not weak if designed properly. It's a designer choice I suppose. Spur gears do have impact loads (which is why they are louder). Even if they rattle (which you will hear) they should be overbuilt to take it. At least the bearings will not get the heck kicked out of them.
Double helical is better still, as thrust loads cancel and you get double the shear area. I think the planetary is the way to go, keeping crank and drive all in one axis.
Cars have detonation sensors, which are little microphones. May be a "resonance sensor" on the gear box could warn of gear box rattling? May be the noise is so high it would be impossible to differentiate the noise. May be a vibration sensor. Jet engines have a vibration sensors that displays in the cockpit and recorded by the aircraft. Its used for maintenance.
If low idle is a no-no, than the idle should be set high. Right?
I've got generic engineering knowledge of gears. My knowledge of PSRU's is from the few I've seen around the airport and what's on the web. Why helical gears? They develop thrust loads to start with; on the plus side they're stronger than spur gear due to more tooth contact area continuously in contact, making it more suited for high speed.
Still involute spur gears (straight cut gears) are stout and work well. Big plus is no thrust loads. The Lyc accessory case GEAR BOX is an example of spur gears. Spur gears are louder, but so what, they are not weak if designed properly. It's a designer choice I suppose. Spur gears do have impact loads (which is why they are louder). Even if they rattle (which you will hear) they should be overbuilt to take it. At least the bearings will not get the heck kicked out of them.
Double helical is better still, as thrust loads cancel and you get double the shear area. I think the planetary is the way to go, keeping crank and drive all in one axis.
Last edited:
Trivia: The logo on Citro?n cars, http://www.citroen.com/CWW/fr-FR , comes from double helical gears invented by Andre Citro?n. Citro?n used to produce gears before they started with cars in the early 19 hundreds. The idea was to have a silent gear without the axial thrust forces
lucky333
Well Known Member
I'm learning a lot.
So, Dan, would tapered roller bearings be a better choice? The axial forces you have indicated would not be absorbed by regular ball bearings, yes? If not tapered rollers, what would be a good choice to absorb the axial/radial loads generated by the helical gears? I'm thinking shims would not be a long term choice in an AL housing..
Thanks for your input on this thread. All kool-aid aside, I'm learning a lot!
John
PS for George: the Commander 680 I flew way back when (Lyc IGSO-480s) had geared supercharged (not turbo) engines. The gearing was a planetary affair if I remember right. It had some different operating considerations but when you wound it up to 3300/45" on takeoff, it ran like a scalded cat. Shorter wings than the 500B. A little hot-rod. Don't remember if it had any RPM range limitations. We pretty much shoved the throttles forward and hung on. There were several 680 models, most with bigger (Lyc xxx540) engines.
So, Dan, would tapered roller bearings be a better choice? The axial forces you have indicated would not be absorbed by regular ball bearings, yes? If not tapered rollers, what would be a good choice to absorb the axial/radial loads generated by the helical gears? I'm thinking shims would not be a long term choice in an AL housing..
Thanks for your input on this thread. All kool-aid aside, I'm learning a lot!
John
PS for George: the Commander 680 I flew way back when (Lyc IGSO-480s) had geared supercharged (not turbo) engines. The gearing was a planetary affair if I remember right. It had some different operating considerations but when you wound it up to 3300/45" on takeoff, it ran like a scalded cat. Shorter wings than the 500B. A little hot-rod. Don't remember if it had any RPM range limitations. We pretty much shoved the throttles forward and hung on. There were several 680 models, most with bigger (Lyc xxx540) engines.
Last edited:
I have no experience with these. EPI has some discussion on their site. Robinson uses this on their PSRUs which seems to work fine but beware of TV again if they have not done analysis.
http://www.v8seabee.com/description.asp
http://www.epi-eng.com/GBX-ChainDrv.htm
I think many people believe that TV can't affect chain or belt drives.
As for the Egg drive, yes, off the shelf gears and shafts to reduce costs and lead times and have reliable stress relief and heat treatment. Can't tell if any of the bearings are angular contact ones or not, I hope the twinned up ones on the prop shaft are. The ones I can see just look like standard ball bearings. Transmission countershafts often just use bronze bushings for thrust and they last forever. I guess time will tell if the bearings stand up well.
Idle speed can be just set with throttle or stop as with Rotax as long as pilot understands. Could be ECU controlled too to take dumb pilot out of the loop.
Spur gears have many advantages. The big disadvantage is noise, this can be very high in some designs but like George says, solves a lot of other problems.
Hard to beat tapered rollers bearings to handle all loads. Look at a lathe or an older car front wb setup. With proper lube, these last a lifetime with high thrust and radial loadings.
I built a chip detector for my Marcotte. Gives me a bit of piece of mind to a slow failure at least.
Last edited:
captainron
Well Known Member
Quite right, tapered roller bearings can indeed handle large loads. The concern here is, what holds those bearings? Cases that can flex and move under load cannot support a driveshaft that may otherwise be designed correctly.
The auto transmission components shown in the drive photos look beefy, but keep in mind that these components are expected to deliver final drive torque to the propeller, whereas in a car, these components merely deliver torque to the final drive gearing where the real torque is generated.
On a separate note, the previously mentioned Hyvo chain was used on the early GM front wheel drive cars like the Toronado. If I recall correctly, it was used between the engine and transmission, a low torque environment, rather than after the transmission where torque is multiplied. They also commonly drive engine timing components.
Last edited:
David-aviator
Well Known Member
The engine has to start from zero rpm and there is no way it can idle at 1400 without going through the range that produces some noise. That's reality. I don't sit there at 700 rpm for any reason, I don't like the noise and know it is not good. The fine pitch stop blade angle was recently increased about 4 degrees to have better rpm control on take off. That increased blade angle also loaded the PSRU at lower rpm's which I believe goes a long way toward eliminating gear lash. The original noise has been reduced substantially, in fact I have not heard it all since it has gotten cold. As previously stated, idle has not been adjusted to 1400 because I believe the engine starts easier with the throttle fully closed. As soon as the ECU finds its brain, the throttle is moved up to get out of the undesirable range.
Forward and aft load on the shaft bearings can not be avoided. Eggenfellner even advocates a high rate descent procedure with the prop in fine pitch - the load (a load much higher than with the engine at idle) is all in reverse when doing this. It also occurs when landing, the prop acts as huge speed brake. If the unit can not withstand such loads, we are in deep do-do with it.
I find it difficult to image that the slight gear lash after start can be SO harmful. The loads in flight are much higher both ways. Sure, it is good technique to avoid the noisy range if it exists, but we don't have to go nuts over it. Just don't do it.
To date, I have about 310 hours behind these units. None have failed for me so far.
<<So, Dan, would tapered roller bearings be a better choice?>>
A great many factors go into bearing selection. Those who are really interested in bearing selection and calculation should beg an SKF catalog somewhere. The technical pages are excellent.
Axial thrust, or even reversing axial thrust, is not necessarily a problem. The issue is a possible high-cycle axial impact. The shafts may or may not have some axial freeplay; we don't know enough about the details of this gearbox to be sure. When axial freeplay exists, even if just the internal clearance in the bearings themselves, impact becomes a factor. As freeplay becomes greater, so does the impact. Bearings don't like impact.
Many common applications deal with reversing axial thrust due to helical or hypoid gears. Bearings can be axially preloaded so there is no axial freeplay. You can preload a deep groove ball bearing, an angular contact ball bearing, or an angular contact roller bearing. All will cheerfully accept axial load combined with radial load.
When preload is not possible, the best bearing for dealing with axial impact is an oil lubricated plane bearing. A common example is a crankshaft driving a balance shaft via helical gears. Some Japanese parallel twins privide a nice example; the crank can't be preloaded axially. The axial thrust from the helical gears is limited with half-moon plane bearing inserts at a case web. At idle, such an engine with too much clearance at the thrust bearing inserts sounds like it has a bad rod knock; the crank is clacking back and forth axially with each torque reversal. Get the clearance down to .004" or so with the correct insert and the oil film makes a nice hydraulic shock absorber; it gets quiet. Good way to buy a nice bike cheap if you know what to look for <g>
A great many factors go into bearing selection. Those who are really interested in bearing selection and calculation should beg an SKF catalog somewhere. The technical pages are excellent.
Axial thrust, or even reversing axial thrust, is not necessarily a problem. The issue is a possible high-cycle axial impact. The shafts may or may not have some axial freeplay; we don't know enough about the details of this gearbox to be sure. When axial freeplay exists, even if just the internal clearance in the bearings themselves, impact becomes a factor. As freeplay becomes greater, so does the impact. Bearings don't like impact.
Many common applications deal with reversing axial thrust due to helical or hypoid gears. Bearings can be axially preloaded so there is no axial freeplay. You can preload a deep groove ball bearing, an angular contact ball bearing, or an angular contact roller bearing. All will cheerfully accept axial load combined with radial load.
When preload is not possible, the best bearing for dealing with axial impact is an oil lubricated plane bearing. A common example is a crankshaft driving a balance shaft via helical gears. Some Japanese parallel twins privide a nice example; the crank can't be preloaded axially. The axial thrust from the helical gears is limited with half-moon plane bearing inserts at a case web. At idle, such an engine with too much clearance at the thrust bearing inserts sounds like it has a bad rod knock; the crank is clacking back and forth axially with each torque reversal. Get the clearance down to .004" or so with the correct insert and the oil film makes a nice hydraulic shock absorber; it gets quiet. Good way to buy a nice bike cheap if you know what to look for <g>
Last edited:
<<Forward and aft load on the shaft bearings can not be avoided. Eggenfellner even advocates a high rate descent procedure with the prop in fine pitch ..>>
Correct, but the issue is the number of cycles and the velocity of impact when it changes direction. Going from full throttle to idle and back to full throttle would be one cycle of axial propshaft force. It looks like the
propshaft may be axially fixated anyway; I'm guessing those might be mirror-paired angular contact bearings behind the prop flange, hopefully with preload. If so, there is no bearing impact at all.
Now consider the layshaft. Assume the bearings are not preloaded. If you're at 800 RPM and hearing gearbox clatter, the layshaft is banging fore and aft about 27 times per second.
Correct, but the issue is the number of cycles and the velocity of impact when it changes direction. Going from full throttle to idle and back to full throttle would be one cycle of axial propshaft force. It looks like the
propshaft may be axially fixated anyway; I'm guessing those might be mirror-paired angular contact bearings behind the prop flange, hopefully with preload. If so, there is no bearing impact at all.
Now consider the layshaft. Assume the bearings are not preloaded. If you're at 800 RPM and hearing gearbox clatter, the layshaft is banging fore and aft about 27 times per second.
David-aviator
Well Known Member
The frequency is not that high. I can't find the words to describe it, but the frequency is not that high.
The hex head screw you ask about, I believe it is an alternate port for a temp probe. There is another one like it on top.
Last edited:
This would be my guess as well.
David is right. I start my engine at a low idle and wait about 1 second for oil pressure to come up, then advance to 1000 rpm which is out of my noticeable TV range. You always have to pass through the evil range. Same on the Rotax we tested.
<<The frequency is not that high.>>
Hmmm, that's interesting. Actually 27 hz is based on a wrong input; I was thinking you had a 4-cyl but you have a 6. (800 RPM x 6)/120 is the 3rd order exciting frequency (firing events), or 40 hz. The 6th order (piston inertias) is even higher. Right offhand I can't recall if the 1-1/2 is a major order with a flat 6. Even if it were, it would be 20 hz and still too high by your estimation. One of these has to be matched by some natural frequency to resonate. Lemme think about it.
Can you take a guess? Trivia tip; the average fellow can't verbally count faster than about 4 hz.
Hmmm, that's interesting. Actually 27 hz is based on a wrong input; I was thinking you had a 4-cyl but you have a 6. (800 RPM x 6)/120 is the 3rd order exciting frequency (firing events), or 40 hz. The 6th order (piston inertias) is even higher. Right offhand I can't recall if the 1-1/2 is a major order with a flat 6. Even if it were, it would be 20 hz and still too high by your estimation. One of these has to be matched by some natural frequency to resonate. Lemme think about it.
Can you take a guess? Trivia tip; the average fellow can't verbally count faster than about 4 hz.
David-aviator
Well Known Member
This is the best I can do with a limited vocabulary.
The engine turns the gear box. The prop, with little or no load, turns easily and gets ahead of the gear box. The gears are momentarily unloaded until the prop slows down and the noise occurs when the gear box catches the prop coming down in rpm. The cycle keeps repeating itself if nothing changes. Some of the noise may be the springs in the fly wheel loading and unloading. If rpm is increased until the gear box is under a constant heavier prop load, the prop does not scoot ahead of the gear box and the cycle is broken. The noise stops.
I believe that when the fine pitch stop blade angle is increased, there is an earlier load on the prop in the rpm curve and the "gear lash" (noise) is lessoned because of it.
It just does not seem to me to be a vibration issue or related to engine piston speed.
Last edited:
David-aviator
Well Known Member
It is higher than 4 hz based on counting to 4.
But it probably isn't 10 times 4. Your estimate may be correct.
Looking at Eggs web site it seems that the box does not use engine oil from the engine, but use transmission oil. With all the wheels and bearings in there, this would mean about 2-3 % efficiency loss. 200 HP is 150 kW, and 2-3% of that is 3-4 kW that is going to heat up the gearbox. How does it get rid of all that heat? 3 kW is a lot of heat for that little box.
John,
not quite. Jan has recommended everyone to change their gearbox oil to Mobil One 75W-90 synthetic gear oil, and I believe this was some months ago. For additional cooling, the factory also recommend that cooling air be taken from the 2/3 radiator shrouds and scat-tubed to the gearbox case. From David's experience, I believe that his gearbox temps are a lot lower than the G2 he had.
The efficiency losses and heat generation calculation by SvingenB may be correct for an engine operated at its continuous max rating, but H-6 users normally cruise at 2100 prop rpm, which is a lot lower (and smoother, quieter) than the 200hp shown. I think (hope) the factory has resolved the gearbox issue. Time will tell.
Allan
not quite. Jan has recommended everyone to change their gearbox oil to Mobil One 75W-90 synthetic gear oil, and I believe this was some months ago. For additional cooling, the factory also recommend that cooling air be taken from the 2/3 radiator shrouds and scat-tubed to the gearbox case. From David's experience, I believe that his gearbox temps are a lot lower than the G2 he had.
The efficiency losses and heat generation calculation by SvingenB may be correct for an engine operated at its continuous max rating, but H-6 users normally cruise at 2100 prop rpm, which is a lot lower (and smoother, quieter) than the 200hp shown. I think (hope) the factory has resolved the gearbox issue. Time will tell.
Allan
Last edited:
Mike S
Senior Curmudgeon
Agreed.
No mater if you like traditional, or alt engines, we all benefit when a company in our hobby/sport/obsession succeeds.
For all the perseverance and dogged determination that Egg has shown in the face of development issues, and other impediments to progress, I salute him.
<<It is higher than 4 hz based on counting to 4. But it probably isn't 10 times 4. Your estimate may be correct.>>
Some dynamic event has to drive the cycle. The list of drivers is finite, and being cyclical, frequency will identify the particular driver. My estimates are based on the common drivers, recip and gas pressure forces, 6th and 3rd order events in this case. And frankly, they are not estimates. Given a specific RPM, number of cylinders, and engine type (4 stroke w/ even crankthrow spacing), the frequency is quite precise.
Other drivers are possible. A good example would be a cyclical variation in load. Some twin-engine airplanes have this problem when props tips are place too close to the fuselage side. With a two-blade prop you can get a twice per rev "bump" when the tips pass through disturbed airflow.
Another possibility is a beat frequency. You hear the beat rather than the two underlying frequencies. The twin engine airplane again provides an example. When you run the two engines at slightly different RPMs you get a very annoying wa-wa-wa beat. The beat may be at only 1 or 2 hz, but that's just the frequency at which the two underlying, much higher frequencies go in and out of sync.
Can you think of a driver (an exciting frequency) for your cycle, whatever hz it may be? For sure the prop doesn't just randomly "get ahead of" the engine.
Please don't think I'm picking on you; not my intent at all. I'm really very pleased to see you thinking about system mechanics.
Some dynamic event has to drive the cycle. The list of drivers is finite, and being cyclical, frequency will identify the particular driver. My estimates are based on the common drivers, recip and gas pressure forces, 6th and 3rd order events in this case. And frankly, they are not estimates. Given a specific RPM, number of cylinders, and engine type (4 stroke w/ even crankthrow spacing), the frequency is quite precise.
Other drivers are possible. A good example would be a cyclical variation in load. Some twin-engine airplanes have this problem when props tips are place too close to the fuselage side. With a two-blade prop you can get a twice per rev "bump" when the tips pass through disturbed airflow.
Another possibility is a beat frequency. You hear the beat rather than the two underlying frequencies. The twin engine airplane again provides an example. When you run the two engines at slightly different RPMs you get a very annoying wa-wa-wa beat. The beat may be at only 1 or 2 hz, but that's just the frequency at which the two underlying, much higher frequencies go in and out of sync.
Can you think of a driver (an exciting frequency) for your cycle, whatever hz it may be? For sure the prop doesn't just randomly "get ahead of" the engine.
Please don't think I'm picking on you; not my intent at all. I'm really very pleased to see you thinking about system mechanics.
lucky333
Well Known Member
And the method to caluclate the freq is?
I'm guessing multiples of the power pulse rate for the number of cyls, 2 or 4 stroke, at the RPM but crankthrow spacing-- eh? That and my general ignorance would scotch any real calcs.. So.. what is/are the general formula(e). That is if its, well..., simple. I'm mainly curious.. Not designing a redrive..
Thanx
John
I'm guessing multiples of the power pulse rate for the number of cyls, 2 or 4 stroke, at the RPM but crankthrow spacing-- eh? That and my general ignorance would scotch any real calcs.. So.. what is/are the general formula(e). That is if its, well..., simple. I'm mainly curious.. Not designing a redrive..
Thanx
John
<<This recent discussion centers around idling TV's . Can you address the possibility of cruise power TV's doing the same damage, but at less noticable
intensity, over a longer period of time.>>
It is possible, if F2 falls within the operating range. F3 is probably above the operating range for the typical systems of interest here. We went over most of the theory in the Belted Air Power thread. Have another read there, and look at the sample Campbell diagrams. Basic concept: If a natural frequency is intersected by an exciting frequency, the system resonates.
<<I'm guessing multiples of the power pulse rate for the number of cyls, 2 or 4 stroke..So.. what is/are the general formula(e). That is if its, well., simple>>
Hand calculating the natural frequencies of a system with two or three inertias isn't bad. Penciling four or more inertias is quite complicated (for me anyway), but there is software. Before you get to that stage you also do a lot of calculating or measuring to determine the inertia of all the rotating masses and the stiffness of connecting shafts. You need inertia and stiffness values for the frequency calculations.
In contrast, calculating exciting frequencies is dead easy, and you've already seen the basic equation a few posts back. It is (RPM x #cyls)/120 = gas pressure events per second (hz) for a 4-stroke. The "120" is merely a division by 2 (since each piston fires every other rev) and by 60 (to convert events per minute to events per second). Every other exciting frequency due to engine events is a variation on the same equation.
intensity, over a longer period of time.>>
It is possible, if F2 falls within the operating range. F3 is probably above the operating range for the typical systems of interest here. We went over most of the theory in the Belted Air Power thread. Have another read there, and look at the sample Campbell diagrams. Basic concept: If a natural frequency is intersected by an exciting frequency, the system resonates.
<<I'm guessing multiples of the power pulse rate for the number of cyls, 2 or 4 stroke..So.. what is/are the general formula(e). That is if its, well., simple>>
Hand calculating the natural frequencies of a system with two or three inertias isn't bad. Penciling four or more inertias is quite complicated (for me anyway), but there is software. Before you get to that stage you also do a lot of calculating or measuring to determine the inertia of all the rotating masses and the stiffness of connecting shafts. You need inertia and stiffness values for the frequency calculations.
In contrast, calculating exciting frequencies is dead easy, and you've already seen the basic equation a few posts back. It is (RPM x #cyls)/120 = gas pressure events per second (hz) for a 4-stroke. The "120" is merely a division by 2 (since each piston fires every other rev) and by 60 (to convert events per minute to events per second). Every other exciting frequency due to engine events is a variation on the same equation.
BillL
Well Known Member
Forcing Frequency
Dan, It may not be all the normal engine driving variations. The Sube has fuel injection and A/F control. If it still has the O2 sensor and operates on the original maps, it can pertubate the A/F to the point that it also cycles on engine rpm. (Note: this perturbation done to facilitate NOx removal with the catalytic converter.) This can the in the 3-10hz range. Someone else who is more familiar with automotive may can give more exact numbers. Just another potential variable to consider. If the OBDII is operational it could be tracked.
BTW - On two of my larger projects (at work) with over $1M/yr going into them, both had torsional failures and we looked hard at the beginning. Analytically, the freq analysis was a sorting tool, but we had to do dynamic analysis to get actual stresses on both to solve the problem. It is just not that unusual. Yes, gear boxes are problematic, but issues can be resolved. Good luck to Egg.
Bill
Dan, It may not be all the normal engine driving variations. The Sube has fuel injection and A/F control. If it still has the O2 sensor and operates on the original maps, it can pertubate the A/F to the point that it also cycles on engine rpm. (Note: this perturbation done to facilitate NOx removal with the catalytic converter.) This can the in the 3-10hz range. Someone else who is more familiar with automotive may can give more exact numbers. Just another potential variable to consider. If the OBDII is operational it could be tracked.
BTW - On two of my larger projects (at work) with over $1M/yr going into them, both had torsional failures and we looked hard at the beginning. Analytically, the freq analysis was a sorting tool, but we had to do dynamic analysis to get actual stresses on both to solve the problem. It is just not that unusual. Yes, gear boxes are problematic, but issues can be resolved. Good luck to Egg.
Bill
David-aviator
Well Known Member
I know what you mean by an exciting frequency, a prop failure on a previous airplane caused major vibration with the control surface. But I am out of ideas as to what could be causing this event at low rpm's. If it isn't a gear lash I don't know what it is or what is causing it.
You say you are sure the prop is not getting ahead of the drive unit and coming back into the gear train, if that is so, why is the phenomenon lessened with a greater blade angle?
pierre smith
Well Known Member
Thanks..
To newcomer BillL,
For bringing your expertise to these forums....an incredible amount of talent here! Welcome, and BTW, you can't go wrong with either the -7 or -7A
Regards
To newcomer BillL,
For bringing your expertise to these forums....an incredible amount of talent here! Welcome, and BTW, you can't go wrong with either the -7 or -7A
Regards
Hi Bill,
Thanks for joining in.
<<pertubate the A/F to the point that it also cycles on engine rpm...This can the in the 3-10hz range.>>
Ah ha, an unusual driver! Ross, you're our FI guy, know anything about it?
<<Analytically, the freq analysis was a sorting tool, but we had to do dynamic analysis to get actual stresses on both to solve the problem.>>
Bless you. I keep preaching about the need for live measurement among PSRU suppliers, but sometimes feel like a mad voice in the wilderness. Can you tell us a little about the methods and tools used for the dynamic analysis?
Thanks for joining in.
<<pertubate the A/F to the point that it also cycles on engine rpm...This can the in the 3-10hz range.>>
Ah ha, an unusual driver! Ross, you're our FI guy, know anything about it?
<<Analytically, the freq analysis was a sorting tool, but we had to do dynamic analysis to get actual stresses on both to solve the problem.>>
Bless you. I keep preaching about the need for live measurement among PSRU suppliers, but sometimes feel like a mad voice in the wilderness. Can you tell us a little about the methods and tools used for the dynamic analysis?
David-aviator
Well Known Member
We seem to be getting into the "peer-review process" as emphasized on another forum re climate change data. Some peer-review of the EGG PSRU is a very good idea. This technical discussion is way over my head but clearly we need it.
In aviation use with the OE ECU, the ECU is almost never in closed loop due to the high load and rpm so this would not be a factor on the H6. The non-OE ECU used in the newer Egg engines does not use an O2 sensor at all.
In closed loop, you are correct, the ECU tries to maintain around 14.7 AFR using a narrow band O2. Since this is a feedback control, there is a delay time between sensing and correction so the AFR is rarely dead on 14.7. This delay is dependent on rpm and proximity of the sensor to the port plus software concerns. HP falls if leaner than 14.7 and increases if richer than 14.7.
Usually torque variations in closed loop are quite small but it is true, that this could be a factor in these applications using closed loop AFR control.
This brings up a good point about other factors causing an unknown excitation. Prop resonance, engine mount resonance etc. could be factors as well. My composite prop has some areas of pitch airspeed and rpm that do set up a resonance for instance.
Good discussion and this shows how important actual testing is and flight testing in particular. Theory may uncover something bad before we even run the engine so that can be highly useful as well.
Amazing transformation happening here from "Shoot the Messenger" to "Peer Review Process". No thanks necessary!
lucky333
Well Known Member
Time to hear from Egg?
John, you know better than that.
But I can't help but think that this would be a great time for Egg to chime in and address the engineering issues raised here and how they have been handled in his PSRU.
John
John, you know better than that.
But I can't help but think that this would be a great time for Egg to chime in and address the engineering issues raised here and how they have been handled in his PSRU.
John
David-aviator
Well Known Member
Originally you arrived on the scene with a large John Deere backhoe intent on digging a grave for the EGG PSRU. But some 190 messages later that has not happened so it is ok to pat yourself on the back. None of the information going back and forth would have happened without your first post. Thanks (I think).
Hello David,
<<You say you are sure the prop is not getting ahead of the drive unit and coming back into the gear train, if that is so, why is the phenomenon lessened with a greater blade angle?>>
Actually I said "the prop doesn't just randomly get ahead of the engine." The key word is "randomly"; I was trying to emphasize the concept of a forcing frequency and the need to quantify it in order to know the source.
That aside, let's focus on the "getting ahead" part. You're not right and you're not wrong, but a small conceptual shift would be good.
First, forget that the prop and gearbox shafts are spinning at all. Visualize them as stationary. Torsional vibration is a periodic oscillation in shaft torque. As you know, you can have torque without shaft rotation.
The prop is the largest inertia (by far) in a system of inertias connected by shafts. To illustrate the concept, think of a great big flywheel connected to a little flywheel.
Grab them with your giant imaginary hands and twist the shaft, then release everything and observe the opposing oscillation. You'll see a lot of angular displacement in each oscillation of the small wheel, and very little angular displacement of the big wheel. The big inertia is the anchor against which the smaller inertia oscillates.
Your prop isn't getting ahead. There may be significant resonant torque applied to it's hub, but at most it is mildly oscillating in it's rotation speed. The rest of the system is oscillating against it. The gears (they're somewhere in the shaft in the simple model) clatter with each torque reversal.
There are several possible explanations for the calming effect of prop blade pitch increase. One might be a change in blade bending mode with a change in axis. Another might be an increase of damping in the form of increased aerodynamic drag. If I remember right, it damps the oscillation of the entire system. If the oscillation wasn't very powerful to begin with, some additional damping could be enough to drop the oscillation below the point of torque reversal. I'll see what denHartog has to say when I get home <g>.
<<You say you are sure the prop is not getting ahead of the drive unit and coming back into the gear train, if that is so, why is the phenomenon lessened with a greater blade angle?>>
Actually I said "the prop doesn't just randomly get ahead of the engine." The key word is "randomly"; I was trying to emphasize the concept of a forcing frequency and the need to quantify it in order to know the source.
That aside, let's focus on the "getting ahead" part. You're not right and you're not wrong, but a small conceptual shift would be good.
First, forget that the prop and gearbox shafts are spinning at all. Visualize them as stationary. Torsional vibration is a periodic oscillation in shaft torque. As you know, you can have torque without shaft rotation.
The prop is the largest inertia (by far) in a system of inertias connected by shafts. To illustrate the concept, think of a great big flywheel connected to a little flywheel.
Grab them with your giant imaginary hands and twist the shaft, then release everything and observe the opposing oscillation. You'll see a lot of angular displacement in each oscillation of the small wheel, and very little angular displacement of the big wheel. The big inertia is the anchor against which the smaller inertia oscillates.
Your prop isn't getting ahead. There may be significant resonant torque applied to it's hub, but at most it is mildly oscillating in it's rotation speed. The rest of the system is oscillating against it. The gears (they're somewhere in the shaft in the simple model) clatter with each torque reversal.
There are several possible explanations for the calming effect of prop blade pitch increase. One might be a change in blade bending mode with a change in axis. Another might be an increase of damping in the form of increased aerodynamic drag. If I remember right, it damps the oscillation of the entire system. If the oscillation wasn't very powerful to begin with, some additional damping could be enough to drop the oscillation below the point of torque reversal. I'll see what denHartog has to say when I get home <g>.
<<This brings up a good point about other factors causing an unknown excitation. Prop resonance, engine mount resonance etc. could be factors as well.>>
Whoa bubba, don't get confused. Props and engine mounts have natural frequencies, but have little capacity to bring an exciting frequency to the party. They are usually the victim, excited by something else.
Rare exceptions would be things like the "prop tip too close to the fuselage" example in a previous post. The prop would resonate if it had a natural frequency matched by the frequency at which the tips pass the fuselage. If not matched, it wouldn't resonate but would pass the cyclical variation in load onward to the powertrain.
BTW, "bubba" is a term of endearment here in the Deep South <g>
Whoa bubba, don't get confused. Props and engine mounts have natural frequencies, but have little capacity to bring an exciting frequency to the party. They are usually the victim, excited by something else.
Rare exceptions would be things like the "prop tip too close to the fuselage" example in a previous post. The prop would resonate if it had a natural frequency matched by the frequency at which the tips pass the fuselage. If not matched, it wouldn't resonate but would pass the cyclical variation in load onward to the powertrain.
BTW, "bubba" is a term of endearment here in the Deep South <g>
Last edited:
Well maybe. I sometimes get a quite noticeable vibration on the ground with a gusting tailwind or in flight in turbulence which I attribute to the composite prop blades flexing. This "feels" exactly like my second TV period around 1400 rpm and is worse on descent with lower manifold pressure. Could be just the prop but don't forget those lovely rubber bushings I have on the flywheel/drive plate.
Seems logical to me that any major loading changes in the prop can set up TV on the other end of the system. No? We see this in automotive drivetrains sometimes.
Seems logical to me that any major loading changes in the prop can set up TV on the other end of the system. No? We see this in automotive drivetrains sometimes.
Took a quick look at what DenHartog's "Mechanical Vibrations" has to say about propeller damping. Most of his work in the area involved vibration in ship propulsion (his contribution to the WWII effort), which is a bit different from our case (the arrangement of inertias is swapped, prop inertia being small). Only one paragraph regarding aerodynamic damping of aircraft propellers. In short, there is apparently some aerodynamic damping at low frequencies, but not high frequencies. Given that we're talking about low frequencies in David's case, well, the answer is "maybe". Sorry, best I can do for now. Quantifying the damping would require some data gathering.
Ran across a nice diagram while reading, and thought I'd post it here because it illustrates much from this and other threads. Yukon asked about the dangers of an F2 second mode intersection in the upper RPM range. Kahuna told us about his buddy Subob breaking several crankshafts. In another thread there was discussion of second mode and how to measure or detect it, and I pointed out that the second mode wasn't at all like the first. And yesterday I was discussing relative motion of different sized inertias.
This is a mode shape diagram for a 4-cyl diesel generator set. The inertias are represented by disks. From left to right, they are the generator rotor, the engine flywheel, four crank assemblies, and a damper. A previous illustration listed the inertia values. In this case the generator has 3.28 times the inertia of the entire rotating engine assembly, much like our prop/engine combinations (for which 5x to 10X is ballpark, best I can tell). Our systems may not be exactly like this plot, but they will be similar.
The straight horizontal line is zero angular displacement. Everything above the line is angular displacement (rotation of the inertia) in one direction (call it positive) and everything below the line is angular displacement in the other direction (call it negative). The little circles at the intersection of the zero line and the ordinates are the vibratory nodes (not modes, don't get confused).
The upper plot is F1. It describes system dynamics typical for the idling case David and I are discussing. Note the relative amplitude of the opposing oscillation for various inertias. The dominant inertia doesn't oscillate much compared to the other inertias, as I described in the previous post. First mode so only one node; the big inertia rotates negative and everything else rotates positive.
The lower plot is F2. Note that now the big inertia and the inertias at far right displace positive while inertias near the center displace in the negative direction. Second mode vibration, so you now have two nodes. Notice the node located in the center of the crankshaft, and in particular, the steep slope of the ordinate (do you now understand Subob's broken crankshafts?). Also note the very low angular displacement of the big inertia. Prop telemetry (or vibration perceived in the cockpit) might not tell you much about what is happening with the poor 'ole crank.
This particular example included a true damper at the free end of the crank. DenHartog notes this system would fail without the damper, as resonant torque is "only" about 5 times mean torque, but would be about 50x without the damper.
BTW, this particular little book costs about $15. Given the amount of torsional heartache we see in the alt engine world, it has the potential for a very good return on investment. Perhaps everyone should ask Santa Claus for a copy? <g>
Ran across a nice diagram while reading, and thought I'd post it here because it illustrates much from this and other threads. Yukon asked about the dangers of an F2 second mode intersection in the upper RPM range. Kahuna told us about his buddy Subob breaking several crankshafts. In another thread there was discussion of second mode and how to measure or detect it, and I pointed out that the second mode wasn't at all like the first. And yesterday I was discussing relative motion of different sized inertias.
This is a mode shape diagram for a 4-cyl diesel generator set. The inertias are represented by disks. From left to right, they are the generator rotor, the engine flywheel, four crank assemblies, and a damper. A previous illustration listed the inertia values. In this case the generator has 3.28 times the inertia of the entire rotating engine assembly, much like our prop/engine combinations (for which 5x to 10X is ballpark, best I can tell). Our systems may not be exactly like this plot, but they will be similar.
The straight horizontal line is zero angular displacement. Everything above the line is angular displacement (rotation of the inertia) in one direction (call it positive) and everything below the line is angular displacement in the other direction (call it negative). The little circles at the intersection of the zero line and the ordinates are the vibratory nodes (not modes, don't get confused).
The upper plot is F1. It describes system dynamics typical for the idling case David and I are discussing. Note the relative amplitude of the opposing oscillation for various inertias. The dominant inertia doesn't oscillate much compared to the other inertias, as I described in the previous post. First mode so only one node; the big inertia rotates negative and everything else rotates positive.
The lower plot is F2. Note that now the big inertia and the inertias at far right displace positive while inertias near the center displace in the negative direction. Second mode vibration, so you now have two nodes. Notice the node located in the center of the crankshaft, and in particular, the steep slope of the ordinate (do you now understand Subob's broken crankshafts?). Also note the very low angular displacement of the big inertia. Prop telemetry (or vibration perceived in the cockpit) might not tell you much about what is happening with the poor 'ole crank.
This particular example included a true damper at the free end of the crank. DenHartog notes this system would fail without the damper, as resonant torque is "only" about 5 times mean torque, but would be about 50x without the damper.
BTW, this particular little book costs about $15. Given the amount of torsional heartache we see in the alt engine world, it has the potential for a very good return on investment. Perhaps everyone should ask Santa Claus for a copy? <g>
Last edited:
