Quantcast

BW 350 PSRU Failure, please share the word!

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,771
Location
Calgary, Alberta, Canada
The clutch does nothing to alleviate TV in flight which is the more critical range. If TV breaks the box on startup or idle, you are not going flying.

The small springs in a clutch disc will bottom out at very low torque values compared to what happens at resonance where peak values may exceed the max torque rating of the engine by 10 to 30 times.

What TV studies were done on these boxes?
 

pictsidhe

Well-Known Member
Joined
Jul 15, 2014
Messages
8,065
Location
North Carolina
Yes, those springs will bottom at resonance. What they do is shift the resonant frequency down so you dont have resonance during flight. Running through resonance when starting is indeed a possibility if done quickly enough that the amplitude doesn't become excessive. The lower the resonant frequency, the softer and longer travel the springs need to be, so they're generally designed for resonance not far below the operating range. That's where the clutch comes in. Pedal clutch on car, centrifugal on this box.
 

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,771
Location
Calgary, Alberta, Canada
Yes, those springs will bottom at resonance. What they do is shift the resonant frequency down so you dont have resonance during flight. Running through resonance when starting is indeed a possibility if done quickly enough that the amplitude doesn't become excessive. The lower the resonant frequency, the softer and longer travel the springs need to be, so they're generally designed for resonance not far below the operating range. That's where the clutch comes in. Pedal clutch on car, centrifugal on this box.
Disagree. If the springs bottom at 200 ft. lbs. and peak torque at resonance is 3000 to 10,000 ft. lbs. they do nothing to prevent transmission downstream in the system since they are locked up solid. Thielert had similar issues with their original clutch design as have other PSRUs which have employed similar clutch arrangements. You don't see this stuff being used on RR, Allison, P&W, Wright etc. engines because it's not a valid and reliable solution to TV.
 

pictsidhe

Well-Known Member
Joined
Jul 15, 2014
Messages
8,065
Location
North Carolina
If the springs have enough travel to transmit the full instantaneous torque, they won't bottom, the critical frequency stays below engine speed. They don't need to handle the resonant torque as the system wknt resonate. Hop the engine up, put heavier rods or pistons on, what was safe, may no longer be. Clutch springs tend to be merely adequate as a mfr will change the plate for a more demanding engine.
A separate feature of this psru is the clutch, which will transmit a rev dependant torque. A bit above engine torque at the bottom of the rev range, far more at the top end.
As an electrical analogy, it's an rlc circuit. Most of the inductance (clutch springs) has a voltage clamp (spring travel) A small amount of inductance is the torsional stiffness elsewhere between engine and prop. The prop and engine inertia constitute the capacitance. Spice would be a good way to illustrate TV, but I'm not trying that on my phone!
 
Last edited:

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,771
Location
Calgary, Alberta, Canada
Like I said. The big geared aero engines don't use clutch discs because it doesn't work very well. Professional engineers who've examined TV in detail have better solutions than sticking a sprung clutch disc between engine and gearbox.
 

pictsidhe

Well-Known Member
Joined
Jul 15, 2014
Messages
8,065
Location
North Carolina
Yes, there are more elegant and compact solutions. I believe the big boys have a psru with the critical speed above the operating range, then they may or may not have to deal with harmonics. I IRC, it was the 9th harmonic that was particularly troublesome for the Wright cyclone.
The key to avoiding TV problems is to keep the critical frequency out of the operating range and control any harmonics that may hit the critical frequency. Some ways to skin that cat are more elegant than others. I'm going for direct drive, which puts the critical frequency way above the operating range.
 

Monty

Well-Known Member
Joined
Jul 15, 2010
Messages
1,294
Location
Fayetteville, AR / USA
Disagree. If the springs bottom at 200 ft. lbs. and peak torque at resonance is 3000 to 10,000 ft. lbs. they do nothing to prevent transmission downstream in the system since they are locked up solid. Thielert had similar issues with their original clutch design as have other PSRUs which have employed similar clutch arrangements. You don't see this stuff being used on RR, Allison, P&W, Wright etc. engines because it's not a valid and reliable solution to TV.
The big engines use a torsional spring to accomplish the same thing. The difference is they use a quill shaft arrangement. The quill is the spring, it's also the input shaft. Quill shafts are highly stressed, critical, failure prone components. Material specs, heat treat, and quality control must be top notch, and therefore expensive. The packaging is better though. Especially when you have an engine and gearbox that are integrated. Using a spring hub achieves EXACTLY the same thing. The engine/gearbox will not resonate as long as the operating frequency is above resonance by 2X or so. Placing the spring hub in the system lowers the resonant frequency. As long as there are enough springs to transmit the engine torque they will not bottom. This is a perfectly acceptable solution. Especially when the engine was not designed for a specific application. That is why spring hubs are used in cars, trucks, and boats. If the springs in a clutch hub fail, you can still land the plane. Increased lash also lowers resonance, so the drive would be noisy, but it would still work. You can also inspect this arrangement easily by simply moving the prop during pre-flight. If a quill fails, you have no power delivery to the prop, and there is no good way to inspect it.

What doesn't work is placing a damper between the engine and gearbox, or choosing the wrong springs for your engine/gearbox/prop combination. It is possible for the springs to resonate and fatigue. This can cause problems that are hard to diagnose. It could be some higher harmonic of engine operation that excites the springs in the hub, even though the engine/gearbox/prop system is above resonance. This would be a very difficult thing to diagnose without proper equipment. A quill shaft will also fail if improperly designed, and developing them for those big engines was no small task. Especially with the accessory drive and all its associated problems on the other end of the engine.

But to your other point: In a car the need for the clutch is obvious. I don't see the need for a clutch in an aircraft application. When starting and stopping the engine, moving through torsional resonance should not be an issue. Most aircraft do this anyway. Transients are not the problem, and the clutch can fail.
 

Winginit

Well-Known Member
Joined
Jun 12, 2016
Messages
811
Location
x
Stuart Davis owner of Auto PSRUs said that he felt his product had been wrongly accused in the failure of a gear box built by his predecessor Bud Warren. The gear box is cast as opposed to CNC machined like the current ones and has a 1.75 ratio instead of the current 1.66 ratio. After reading his response on his company website, I did feel he has reason to feel that the drive unit was improperly used, and the complaint was "sour grapes". Now I have to clarify here that my "opinion" is worth absolutely nothing. Not a bonefide expert in the field, but I can read. So here is how it appears to me.

Andraes K Fellow Pilots,

I’d like to share the gearbox failure that I recently experienced in my Vans RV-10.

https://drive.google.com/open?id=0B_...0RTM280TzAyaDg

https://drive.google.com/open?id=0B_...mlmVjJWdEVibUk
Andraes K My setup:

Airframe: Vans RV-10
Engine: GM LS 7 (approx. 400 hp, 800 ft/lbs of torque @ 2700 prop RPM, WOT @ SL)
PSRU: BW 350 (Bud Warren, Geared Drives), Cast Aluminum Case, 1:1.75 reduction ratio
Propeller: Whirlwind Aviation RV-10 composite prop, 80 inch, 46 lbs
First the conversion: 2700 rpms X 1.75 = 4725 rpms...............2700 rpms X 1.66= 4482 Since I can't get these exact rpms on a chart, I'll call them 4750 and 4500 rpms.

According to GM graphs of power output for these engines : LS7= 425 HP (470 lbs ft) @ 4750 rpms (Note: GMs stated 470 lbs ft @4800 calculates to 425 hp and the torq curve on the chart is pretty flat)
The LS1 which Bud Warren originally designed for would be : LS1= 360 HP (405 lbs/ft) @4750 rpms
An LS3 430 HP Corvette engine shows these numbers....... : LS3= 390 HP (420 lbs/ft) @4750 rpms (Don't believe these were available when Bud designed his redrive)

Now if my information is accurate, during any normal takeoff I would expect the LS7 engine at WOT/2700 RPMs was exceeding the maximum rated HP of the gearbox by 25 HP and his torque input was exceeding max torque input (444) by 26 lbs ft.

While speculation about a 3G turn was introduced as the cause of the problem, but no real proof exhibited, my opinion is that the limits of the PSRU were being exceeded during normal takeoff operation. Based on my pure speculation about the problem, it seems that Mr Davis does have reason to feel that his product was mis-represented. Again, thats just my personal opinion and worth nothing, but it does at least put some factual information into play.

As for the clutch/TV component of the discussion, I don't have a current opinion on that part of the equation.

Graphs are LS7 then LS3 then LS1

GM-Dyno-LS7-Charts-73.gif LS3-430HP-Dyno-chart.jpg Slide-27.jpg
 
Last edited:

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,771
Location
Calgary, Alberta, Canada
What doesn't work is placing a damper between the engine and gearbox, or choosing the wrong springs for your engine/gearbox/prop combination.
Yet we have examples placing dampers between the engine and gearbox which have many thousands of trouble free flight hours. Autoflight and SPG to name a couple with hundreds of units flying for over 15 years now...

Engineered elastomer couplers/dampers are extensively used in many industrial, prime mover applications where TV would destroy the machinery. Whole industries exist supplying these. Centaflex and Lovejoy being two big companies in this field. http://www.centa.info/?show=products&c=us&nr=32 http://www.lovejoy-inc.com/products/

The difference between most most quill shafts and a clutch disc is that the first one was actually engineered based on real numbers. The second was likely a common OTS part chosen because it fits with little regard for the numbers. I suspect most quill shafts in engineered PSRUs are MUCH stiffer than the springs found in your average clutch disc. A couple hundred ft./lbs will bottom out most disc sprung centers.
 
Last edited:

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,771
Location
Calgary, Alberta, Canada
You can't say on one hand that you can't exceed say 360hp and then list engines which far exceed that in your display literature on the other hand...

bw350.jpg

Have to make up your mind which one it is.

The AgCat and NXT racer will both require far in excess of 400hp to be useful.
 

Autodidact

Well-Known Member
Joined
Oct 21, 2009
Messages
4,513
Location
Oklahoma
Yet we have examples placing dampers between the engine and gearbox...
A shock absorber is a damper, and with no spring it would do nothing for TV. Elastomeric couplings do have a spring rate in addition to damping properties; are you and Monty talking past each other a little here?

The difference between most most quill shafts and a clutch disc is that the first one was actually engineered based on real numbers. The second was likely chosen because it fits with little regard for the numbers. I suspect most quill shafts in engineered PSRUs are MUCH stiffer than the springs found in your average clutch disc. A couple hundred ft./lbs will bottom out most disc sprung centers.
I get the impression that most of the automotive clutch centers are designed to bottom out; they often have very little angular travel and none were designed with the inertias in mind that exist in a prop drive system. It doesn't mean that they cannot work if selected properly, but even then you may have to run several in series and that would be complicated and maybe even heavy.

An interesting question to me is: Since the quill shaft is designed to flex indefinitely, and Merlins are so old, have there been any problems with their quill shafts fatiguing? I've never heard of it, but that doesn't mean it isn't an issue - I've only read about valve seat problems and the like.
 

BoKu

Pundit
HBA Supporter
Joined
Aug 15, 2013
Messages
2,697
Location
Western US
You can't say on one hand that you can't exceed say 360hp and then list engines which far exceed that in your display literature on the other hand...
I dunno about that. I think that he could make a valid case that the sheet just shows what they have been used on, not necessarily what they were designed or intended for. A "closed course, professional driver, do not attempt at home" sort of thing.
 

Monty

Well-Known Member
Joined
Jul 15, 2010
Messages
1,294
Location
Fayetteville, AR / USA
Yet we have examples placing dampers between the engine and gearbox which have many thousands of trouble free flight hours. Autoflight and SPG to name a couple with hundreds of units flying for over 15 years now...

Engineered elastomer couplers/dampers are extensively used in many industrial, prime mover applications where TV would destroy the machinery. Whole industries exist supplying these. Centaflex and Lovejoy being two big companies in this field. http://www.centa.info/?show=products&c=us&nr=32 http://www.lovejoy-inc.com/products/

The difference between most most quill shafts and a clutch disc is that the first one was actually engineered based on real numbers. The second was likely chosen because it fits with little regard for the numbers. I suspect most quill shafts in engineered PSRUs are MUCH stiffer than the springs found in your average clutch disc. A couple hundred ft./lbs will bottom out most disc sprung centers.
Yes, I understand that. These are also springs, which happen to have internal damping. Properly selected they work fine, but if the spring rate is too stiff, the system will resonate and melt the damper. Damping alone, between the gearbox and engine will not solve the problem if the spring rate is too high and places the system resonance in the operating range.

And I beg to differ on the clutch springs. They all have rated torques, and spring rates. They should be selected based on engineering criteria. Whether they are or not is on the designer of the system.
 

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,771
Location
Calgary, Alberta, Canada
It goes without saying that ALL parts in a successful PSRU unit must be well engineered and made, from gears to bearings to shafts to dampers. If any of them break, bad things will happen.

Eyeball engineering might be successful if you're lucky in this field but the money is on a scientific approach with proper measurement and applied math/ engineering.
 

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,771
Location
Calgary, Alberta, Canada
Yes, I understand that. These are also springs, which happen to have internal damping. Properly selected they work fine, but if the spring rate is too stiff, the system will resonate and melt the damper. Damping alone, between the gearbox and engine will not solve the problem if the spring rate is too high and places the system resonance in the operating range.
Of course, this is understood. Both companies I listed will assist in proper selection through their engineering departments assuming you have some real numbers for them to work with.

I was merely responding that there are many flying examples using these sorts of damper/ couplers. It's simply untrue to say that it doesn't work. Raven uses a friction damper in their drives. Hundreds out there over 20 years I believe along with the others listed.
 

Monty

Well-Known Member
Joined
Jul 15, 2010
Messages
1,294
Location
Fayetteville, AR / USA
Of course, this is understood. Both companies I listed will assist in proper selection through their engineering departments assuming you have some real numbers for them to work with.

I was merely responding that there are many flying examples using these sorts of damper/ couplers. It's simply untrue to say that it doesn't work. Raven uses a friction damper in their drives. Hundreds out there over 20 years I believe along with the others listed.
My comments are directed at those who incorrectly believe damping will solve a resonance issue. It simply will not. The system has to be designed not to be in resonance to start with. Damping WILL help with those higher harmonic issues which is why elastomeric "dampers" are nice. Coupler is a better term, since it is a spring with internal damping, not a true "damper". General usage and engineering terms can cause confusion.
 

Winginit

Well-Known Member
Joined
Jun 12, 2016
Messages
811
Location
x
You can't say on one hand that you can't exceed say 360hp and then list engines which far exceed that in your display literature on the other hand...

View attachment 55108
Have to make up your mind which one it is.


Good points ! I'm going to guess that the one LS7 listed on the brochure ad was the one Mr Andreas had and probably hadn't developed a problem yet. Pure speculation about that. If so, it was the cast one that BW had sold and was operating at above 400 HP because of the higher rpm range needed by the 1.75 reduction ratio. The brochure also shows that the unit being sold now was a 1.66 reduction ratio which would put the LS7 at 4482 rpms where it produces 392 HP. So, it would seem to be plausable that with a new design gearbox, the LS7 might be workable. I will also say that I don't know anything about the AgCat and NXT racer he references. I would further say that expectations would be that if a gearbox were rated at 400 HP it would have some margin of error above that rating, rather than the possibility of failure being right around that rating. My point simply is that the unit is supposedly rated to handle 400 HP and I think that the cause of failure was related to multiple takeoffs at greater than 400 HP rather than a 3G turn. Again, thats just my feeling/SWAG with nothing to really substantiate it. I do feel if 400 HP is the actual point where failure actually becomes a possibilty, then it should be marketed with a lesser rating, but thats just a lot of supposition on my part.


(Note: Remember, I like direct drive when possible, so its odd that I'm saying this stuff. I also believe that whatever is posted should be truthful whether it agrees with my point of view or not, so I felt obligated to point out what I thought was incorrect about Mr Andraes statements. Personally, I hate to see entreprenuers fail in the homebuilt aviation community, especially because of incorrect or misleading information. I also agree with you that the advertisements put out by said entreprenuers needs to be correct also. :)
 

Monty

Well-Known Member
Joined
Jul 15, 2010
Messages
1,294
Location
Fayetteville, AR / USA
The idea of HP ratings is indeed the issue, specifically life at what hp. Assuming there are no torsional resonance issues, and the basic design is sound for flight loads etc...my question about HP is always: "how long?" Static tooth failure is usually not a problem. The problem is usually surface durability or fatigue on the pinion gear. A gearbox designed for 3000hrs @ 400hp would probably easily handle 1000+ hp for short periods of time.

It's all about fatigue life.
 

Winginit

Well-Known Member
Joined
Jun 12, 2016
Messages
811
Location
x
The idea of HP ratings is indeed the issue, specifically life at what hp. Assuming there are no torsional resonance issues, and the basic design is sound for flight loads etc...my question about HP is always: "how long?" Static tooth failure is usually not a problem. The problem is usually surface durability or fatigue on the pinion gear. A gearbox designed for 3000hrs @ 400hp would probably easily handle 1000+ hp for short periods of time.

It's all about fatigue life.
If all the information that has been provided is true and accurate, then it appears there have only been two recorded failures. In both cases the information provided indicates that the pilots were pilots used to aggressive flying, or pushing the envelope so to speak. Also both pilots were using engines that had the ability to exceed the specifications of the redrive. All the redrives being used by other builders appear to be having success. I guess time will tell us the answer.

As for a gearbox designed for 400 hp handling 1,000 hp for a short period of time, that would be a case by case situation, and each successive application of power would build on the previous damage. I know Muncie transmissions didn't last long when moved from small block Chevy to a big block Chevy years ago. If you look at the new aftermarket 5 speed transmissions, torque capacity is increased to either 500 or 600 ft lbs. The relative thing here is that the rating of the trans is changed by changing the first gear, kinda similar to the PSRU being discussed here.
 
Last edited:

rv6ejguy

Well-Known Member
Joined
Jun 26, 2012
Messages
3,771
Location
Calgary, Alberta, Canada
Have to make up your mind which one it is.


Good points ! I'm going to guess that the one LS7 listed on the brochure ad was the one Mr Andreas had and probably hadn't developed a problem yet. Pure speculation about that. If so, it was the cast one that BW had sold and was operating at above 400 HP because of the higher rpm range needed by the 1.75 reduction ratio. The brochure also shows that the unit being sold now was a 1.66 reduction ratio which would put the LS7 at 4482 rpms where it produces 392 HP. So, it would seem to be plausable that with a new design gearbox, the LS7 might be workable. I will also say that I don't know anything about the AgCat and NXT racer he references. I would further say that expectations would be that if a gearbox were rated at 400 HP it would have some margin of error above that rating, rather than the possibility of failure being right around that rating. My point simply is that the unit is supposedly rated to handle 400 HP and I think that the cause of failure was related to multiple takeoffs at greater than 400 HP rather than a 3G turn. Again, thats just my feeling/SWAG with nothing to really substantiate it. I do feel if 400 HP is the actual point where failure actually becomes a possibilty, then it should be marketed with a lesser rating, but thats just a lot of supposition on my part.


(Note: Remember, I like direct drive when possible, so its odd that I'm saying this stuff. I also believe that whatever is posted should be truthful whether it agrees with my point of view or not, so I felt obligated to point out what I thought was incorrect about Mr Andraes statements. Personally, I hate to see entreprenuers fail in the homebuilt aviation community, especially because of incorrect or misleading information. I also agree with you that the advertisements put out by said entreprenuers needs to be correct also. :)

It looks to me that these were never tested to say at least 1.5 times the rated power output so they might have a reasonable lifespan at their rated power levels. Also an NXT needs upwards of 750hp to be competitive at Reno and will also be pulling 4-5 Gs in lots of places on the course. This drive hardly fits that bill with these restrictions.
 
Top