What PSRU Can we Really Trust???

I vote we lock billski in a room with a computer, pencils and paper and not let him out until he has come up with a good PSRU for us. I have come across very few people that seem to know their stuff even half as well.

PTAirco, you guys throw in a bunch of money, access to a machine shop, blank PO's for a gear shop, and the use of a Rotec, and you will have yourself a program! Of course, if you supply just the Rotec and a get a couple people with each of the existing re-drives to fly in for some testing, I can tell you which of the current re-drives is OK vibrationally and perhaps design a better one by monkey-see, monkey-do.

MKIV (I LOVE the Ford GT MKIV! What car that was!) is close to the mark, although I doubt you could convert a good combo into a bad combo with just different crank counterweights. The inertia change is too small to shift things much all by itself.

The big inertias in a typical PSRU are the crank/front pulley/flywheel and the prop. Accessory drives can fall in there too. In between those two are one or two pairs of gears and a couple shafts. Generally the inertias of the gears and shafts are small enough to ignore, but the spring rates of the gears and shafts are really important.

Let's say somebody designs a nice stiff redrive and uses a popular crate small block V-8 engine with a fat torque curve and sets max revs at around 4800 rpm. Firing happens at 320 Hz. They use a popular composite constant speed propellor and the stock automatic transmission flex plate. And the primary resonance of the engine and the prop vibrating opposite each other is ends up at 400 hz. That is 25% higher than the max firing frequency of this engine, and so is barely felt by the system. That primary resonance is out of range high, but only just barely.

Then along comes someone with the a high speed race version with a beefier crank, heavier front pully, and a thicker "heavy duty" flex plate, but detunes it to run the same torque curve as the first engine. And a metal prop. By increasing the inertia by 40%, this newcomer just lowered the primary resonant frequency by about 20%. Now the resonant frequency is only 5% above max rpm, and the formerly "good" PSRU is in a system that is amplifying at max engine speed, and is likely to tear itself apart.

Or someone could walk in with a higher reving version of the same engine, or a V-10 (the firing rate is 25% faster for the same rpm) and get closer to the primary resonance. Boom goes the re-drive!

Here is another possibility. They design a two shaft reducer with a spring isolator (called a clutch damper, but mostly it is just a torsion spring set) between the engine and the gears. With our high inertia engine and prop, it has a resonant frequency that lights off around 600 rpm (40 Hz), so we set min idle on the engine at 750 rpm to put the resonance out of range low. Works fine until somebody puts together a low inertia version we talked about earlier, or worse a V-6. Now the resonant frequency is at 750 rpm, but the owner, not knowing any better sets the idle at the same speed, and the isolator beats itself to death accumulating the damage during idle operations.

Then there are the intermediate masses and the front end accessories that will have vibration modes and interactions with the crank and other modes that can get into operating range too.

Now the difference in stiffness between a system that resonates at 320 Hz and one at 40 Hz is 64 times the stiffness. Not undoable, but that will drive the entire design, including housings, bolted joints, shafts, gears, bearings. Everything will have to be stiff.

Each combination of engine details and prop details matters to these guys. Now I have not done the work. Perhaps a good beefy re-drive is all it takes to drive the primary resonance out of range way high. Then it is just a matter of build quality, gear and bearing lube/cooling, bearing and seal integrity, and a bomb proof setup for running your prop and governor.... Then if somebody comes up with other combinations, it is no big deal, right? That is a rhetorical and sarcastic question. Every change in number of cylinders and inertia of the engine or prop will require validation... Rotec will be an essential part of checking out every new combination that some yahoo wants to put together.

Not what I would call a favorable business model...

Billski

Ok, how about going the opposite way? Instead of a very stiff system, make it very soft and lower the critical RPM to below idle speeds? Is there an inherent disadvantage of doing it that way? Are commonly available couplings capable of this?

These are just some of the amazing variety of couplings out there:
Industrial Couplings Index

Assuming the "soft" were done with the metal parts, wouldn't fatigue become a problem?

Originally Posted by wsimpso1

...Not what I would call a favorable business model...

Billski

Sounds as if the only sensible way to go is to sell engine/PSRU combinations as a unit - no individual sales of the PSRU. Limits your market on one hand, but your liability on the other.

I would like to thank Billski for sticking his toe in the waters of reduction drive design for us. These waters are hot and frothy with lots of contradictory fish. His explanations help to understand why there are not more drives around. The more I read about designing one of these beasties the more reasonable in cost an aircraft engine (used) is and if I remember correctly an aircraft engine is what Billski is using in his own project!
I remember talking to a transmission engineer from one of The Big Three with a masters degree and lots of experience in design and management. He was an EAA member with a Bonanza. I told him he should design, build and sell plans for a reduction drive. His reply was not printable. The short version was "NO!!" He said something about possible failure modes and started getting very excited. And then he mantioned liability.
I work in a shop that builds "one-off" machinery for manufacturers and when they send out P.O.'s sometimes they don't like what they get back as the "completed" product. So there is another issue to deal with and it happens concurrently with design. Last week I was assembling a some parts and I mis-assembled two parts that could not be taken apart and the designer called out the wrong thread for another part. Two hours of assembly turned into a week of waiting and then we will try again.
If I remember correctly the reduction drive that Jim Stewert sold for his S-51 was $20,000 (approx) ten years ago and others that I have seen recently are at least $10,000. Some of these guys are charging $20,000 with the GM crate engine and the crate engine is probably $3,000 to $5,000. I would be curious to see what it would cost to get a timed out aircraft engine and overhaul it yourself. The cost might be the same and in less time.

The issue of using a reduction drive ,I believe, are compounded when you try to figure out how much horsepower you are really getting at the prop. You can dyno the engine but then you have to know the friction of the drive when it is at operating temp. I doubt many folks have actually tested their reduction drive to how horsepower the drive alone needs.

Originally Posted by Topaz

Assuming the "soft" were done with the metal parts, wouldn't fatigue become a problem?

I was thinking more of using one of the many types of rubber/polysomething types of couplings out there.

Originally Posted by PTAirco

I was thinking more of using one of the many types of rubber/polysomething types of couplings out there.

The question has been bubbling around in my mind since we were discussing thin, low-frequency driveshafts. Popped up again here. Seems like if you were to use a long slender 'soft' metal shaft or other component to push the resonant frequency off-range low, it would be getting a lot of deflection in torsion with each start, stop, and throttle change. Maybe I'm missing something, but that seems like a recipe for fatigue failure.

The people who did the P-39/63 drive train put in many joints in their drivetrain that allowed them to control the "whipping" of the shafts. These joints were rigidly mounted to the airframe.

I did mention building a soft unit by including spring package. It has the neat advantages of being tunable and you can put it where it does you the most good too. Change inertias or number of cylinders, and change springs. The single biggest problem with deliberately driving the resonance frequency down, is that you tend to also drive some of the higher order vibration modes down into your operating range... Doing this on a computer, adjusting inertias and spring rates until everything is safe is a task. Imagine doing this by build and test...

Long slender shafts are a soft element. Quill shafts are used for accessory drives on many big piston engines. With any shaft, you must check its max speed against its critical speed. If your operating speed gets close to its critical speed, it can go unstable and flail. Then you have to break it into more than one shaft. You see it in truck drive shafts - with a bearing and a couple Hooke joints midway between trany and axle. And that sounds exactly like the P39 and P63.

Billski

Long slender shafts are a soft element. Quill shafts are used for accessory drives on many big piston engines. With any shaft, you must check its max speed against its critical speed. If your operating speed gets close to its critical speed, it can go unstable and flail. Then you have to break it into more than one shaft. You see it in truck drive shafts - with a bearing and a couple Hooke joints midway between trany and axle. And that sounds exactly like the P39 and P63.

Billski[/quote]

Think I actually watched something along these lines recently with TV coverage of Aussie V8 Touring Cars, they had 'under car' camera focussed on rear of driveshaft & Differential, as car proceeded up thru gear changes & driveshaft RPM approached engine RPM ( 5th/6th gears ) there was a point where the driveshaft tube'''went crazy''' & Im sure the tube appeared to 'snake'. Have also been involved in some discussions re transaxles ( MKIV) recently & one manufacturer was advocating splitting the input shaft into two pieces for similar reasons. Sorry for the sidestep toward car info, but its my background & a lot simpler for me to relate what I have learnt from that field & try to cross reference it to an A/C scenario.

Originally Posted by wsimpso1

...Quill shafts are used for accessory drives on many big piston engines. With any shaft, you must check its max speed against its critical speed. If your operating speed gets close to its critical speed, it can go unstable and flail. ...

Aside from the 'whipping' aspect, is fatigue a factor in slender shafts such as this? Seems like they'd see a lot of torsional deflection in use, between power pulses, throttle changes, and starting and stopping. Or do you design so that the deflections are small enough that the stresses are below the "infinite fatigue life" limit?

MKIV: You just described a shaft going through critical speed. Nothing will last very long with that going on. It sounds like that car is a candidate for a larger diameter driveshaft tube or a two part driveshaft.

Topaz: In this context, soft pertains to spring rate not material condition.

Designing it to live and to give the right spring rate may conflict, but they are not mutually exclusive.

Deflection angle of a torsion bar element is T*L/(G*J) where T is torque, L is length, G is torsional modulus of material, and J is torsional section modulus. Spring rate is thus G*J/L. Make it long (big L) and slender (small J) and it will have a low spring rate. Build it with low G and it will have a lower spring rate too. Oh, and J goes with diameter to 4th power... Now shear stress in a torsion bar element is T*r/J. So the design task is to obtain the right spring rate at a low enough stress. It can be and is done. One thing to remember is that the light way to make a stiff shaft is large diameter and thin wall, while the way to make a soft torsion element is to make it small diameter and long. A lot of attention has to be paid to ends and stress concentrations and heat treat. And then you have one springy shaft. Worked great for isolating pumps and generators and the like from crankshafts.

The problems with using flexible shafts for isolation is several fold. First, if you need a different spring rate (another engine comes along or your design needs some tuning), and you are putting most of your flexibility in that shaft, you need to design and prove out a new shaft. You might not have enough room if you need to make it longer. And the shaft might not be in the right place to do the job, or will require a three piece shaft with two supports...

If instead, you have torsional spring packs, you just design a new coil spring set and have a spring shop make you a few. You can even start out with several in mind and have all of them available for trial with instrumentation...

Billski

I'm curious as to why these geared re-drives are not making use of chip detectors. There are 6 in the helicopter I fly. The main transmission has two, the engine accessory gear box has two, the Free Wheeling Unit (OWC) has one, and the tail rotor gearbox has 1. The gearbox at the T/R also has a temp gauge to detect impending failures.

Even with the money and resources bell helicopter has to design and test these components, they still feel it's necessary to have adequate early warning sensors. I feel good about having them there. Occasionally a mission is aborted when a sensor is tripped by shavings that are later determined to be fair wear and tear by-products. Every once in awhile we do get some serious problems caught in advance though. Just the other day a friend had some good size tooth chunks stuck to his lower transmission chip detector.

Are these practical for PSRUs used in this application?