What's new
Van's Air Force

Don't miss anything! Register now for full access to the definitive RV support community.

Why are PSRUs so hard to design?

Yikes, I seemed to have rattled some nerves.

My point to Freemasm was fuel pumps can be insanely reliable devices. Having only one fuel pump doesn't bother me in the slightest if it's extremely reliable. Lycomings historically have two pumps because the engine driven diaphragm pumps aren't the most reliable. Lycoming engine driven pumps are low pressure. PT-6 high pressure pumps are ~800 PSI. Diesel truck pumps are ~30,000 PSI. Which one is the least reliable? The lowest pressure of them all, the engine driven Lycoming pump.

While DI might or might not be the best for aviation, fuel pump reliability and redundancy isn't the reason why. There's thousands of airliners flying around with only one high pressure fuel pump per engine, and some of them have ETOPS level reliability (speaking to PW100 series engines). Thousands (millions?) of trucks driving around with only one high pressure fuel pump. Not many of them of them seem to be littering the side of highways.

Now I'll sit back down and let the adults talk about how to play nice with others.
 
Most air transport turbines are not even owned, they are leased. The engine owners worry about keeping them going in such a manner as to maintain the needed reliability standards and usually they just get swapped out if there are any signs of a problem. But as I said, those are not high pressure pumps in the way that one has on modern direct injection automotive engines.

Through whatever lobbying / legal connections that exist, the automotive OEMs apparently feel no obligation to provide you with a robust product for on road use. In the overwhelming majority of cases they have never been forced to recall their defective products. You, the customer is stuck with the bill of replacing the unreliable fuel pump with another copy that is no better at enormous expense, usually with collateral damage of the entire fuel system (rail, pressure regulator, injectors, lines etc after their pump starts making metal and feeding it into the fuel system.

If you would even try to use these products in an aviation application, the suppliers would be laughing at you if you had problems, with the added wrinkle of being refused sales of service parts, which you would then be forced to go and find in the junkyard, (if you could).

For gasoline engines, it makes much more sense to either just use a port injection engine (so long as they continue to be produced) or convert them back to port injection if no longer available, which is what Aeromomentum has been doing. Jan at Viking has easy access to the 130hp engine right now, that he can buy cheap out of junked cars, but as soon as they start getting tough to find with low mileage he will kill the engine variant and go with whatever is a comparable production engine again. Just like what happened with the original 110hp Honda Fit engine, try getting any support with one of those today...
 
Yikes, I seemed to have rattled some nerves.

My point to Freemasm was fuel pumps can be insanely reliable devices. Having only one fuel pump doesn't bother me in the slightest if it's extremely reliable....

The apples to oranges simile of the two systems has been nailed down. The challenges of pumping gasoline versus kerosene are known. Probably the most telling data is comparing L-10 bearing life derating factors for the fluids. Extreme contact pressures associated with gears (for those type pumps) would obviously be worse.

Unless things have changed, the (vast?) majority of aircraft turbine fuel pumps were made by a company called Argo-Tech outside of Cleveland. Bought by Eaton IIRC. Short story -> long, they had some stud engineers that I got to work with. They engineered a corrosion resistant flow divider for some sh!t Saudi fuel we had to burn. Still have part of the test article. Bottom line = The reliability of their fuel pumps was much greater than the engines they feed.

Now an interesting question, to me at least. Are the fuel system requirements the same for single engine turbine versus multi; i.e. i wonder what Pilatus approach is? The Vision Jet? Anyone with firsthand knowledge? Am I relegated to the kid's table?

Edit = CR FLow divider pic
 

Attachments

  • Flow divider Agro-Tech.jpg
    Flow divider Agro-Tech.jpg
    655.3 KB · Views: 45
Last edited:
When you only have one pump and it fails, the engine stops.

On PT6 engines, the leading cause of inflight shutdowns by far, is the fuel control unit/ fuel pump. On the CF6, that area ties for the leading cause of shutdowns. This according to a 1993 FAA paper.

Coming back to automotive DI ultra high pressure pumps, there have been multiple mass recalls and class action suits against OEMs for premature failures. They are nowhere near as reliable as typical low pressure electric pumps normally used on port injected installations nor are the injectors used in these systems.
 
Last edited:
Viking

I know the 150 engine that I have has knock sensors and it is recommended to replace them with a different type that is not resonant and then their use has to be programmed on a dyno. If one backs off the timing in response to knock sensors, this failure either would not have happened or else it might have taken much longer to progress giving the pilot adequate time to respond to a dropping oil pressure situation.

Here is the video https://youtu.be/19qpxutW_m4

Jan, for all the things he has done in the past seems to be going to pretty long lengths to protect his products reputation and helping customers stay in the air. Bear in mind the core engine is $1000-$1300.[/QUOTE]


Are you flying behind the 150 Viking? Who designed and who is building their PSRU?

In the video he points the finger at loss of oil due to his "friend" not checking the oil but fails to confirm that this is the root cause. I can't imagine someone routinely failing to check the oil during a preflight. I thought the video was bizzare.
 
Wicked interesting stuff from Ross as usual. I scanned but didn't read for detail. Some screen shots below. The associated granularity of data was a little lacking IMO as they tended to lump some things together. A Pareto (Analysis) graphic representation could have helped but maybe those are buried in the back-up data.

A different perspective that I got some exposure to; I never saw their data but there were a surprising number (% unknown to me) of pump failures associated with fuel quality. Score marks from contamination, evidence of water in fuel, even bulk water exposure, etc. Once the surface finish of bearings, gears, etc. is compromised, the wheels come off pretty quickly.
 

Attachments

  • Screenshot 2022-11-03 102223.png
    Screenshot 2022-11-03 102223.png
    221.6 KB · Views: 47
  • Screenshot 2022-11-03 103240.png
    Screenshot 2022-11-03 103240.png
    241.8 KB · Views: 42
  • Screenshot 2022-11-03 103321.png
    Screenshot 2022-11-03 103321.png
    321.5 KB · Views: 44
The problem with the water in fuel issues is that the consumer has no control whatsoever over it. Usually the problem arises at a particular gas station or with a particular transport company due to deficiencies with their equipment that allows rain to either get into the fuel tanker itself or into the underground storage tank. In some cases one might suspect that it is being done deliberately because selling water at the cost of fuel is more profitable than actually selling fuel.

The vehicles are not set up to be sumped in the same way as an airplane, so verifying the presence of water in the fuel and getting rid of it is much harder. Then there is usually a filter housing somewhere that has a sensor to detect water and the engine controller records any indications that water has been detected and once that has happened, your warranty is already toast.

The OEM treats water in the fuel as if it was the vehicle owner pouring the water in the tank, not as the fault of the gas station or the fuel transport company. I dont think there has ever been an investigation into where the water shows up or some unscrupulous gas station owners would probably be in jail.
 
Im building my own version of the Honda 150. Viking will no longer sell any conversion parts to DIYers so I will not be using the Viking reduction drive. Its a pity, its certainly not a perfect reduction drive but if some of the install issues are taken care of, they appear to last a long time. Certainly the quality of the gears, material, heat treatment etc seems to be stable. He has been making these drives for many years now. There have been some issues with the rubber coupling but that is all I know of.

If you look at car sales websites, they are routinely selling written off cars that cost $50k or more because the owners didnt check the oil level and blew the engines up. So for a Viking 130 that costs about $13k, it would not surprise me at all that the owner didnt check the oil level. If he doesnt do it with his own car, why is that habit going to change with his car based airplane engine ?

The only glaring critique about the root cause is that apparently the knock sensor was unable to save the engine. That should have been caught and the engine should have been able to continue operating at reduced power given retarded timing. It may then ultimately have failed due to the low oil level and subsequent overheating (it did not appear that the pilot significantly reduced the power level prior to landing) so if the pilot just kept flying the outcome might have been the same or worse. So something to think about is whether it may be feasible to get an oil level indication in flight instead of relying on the dipstick.

I know the 150 engine that I have has knock sensors and it is recommended to replace them with a different type that is not resonant and then their use has to be programmed on a dyno. If one backs off the timing in response to knock sensors, this failure either would not have happened or else it might have taken much longer to progress giving the pilot adequate time to respond to a dropping oil pressure situation.

Here is the video https://youtu.be/19qpxutW_m4

Jan, for all the things he has done in the past seems to be going to pretty long lengths to protect his products reputation and helping customers stay in the air. Bear in mind the core engine is $1000-$1300.


Are you flying behind the 150 Viking? Who designed and who is building their PSRU?

In the video he points the finger at loss of oil due to his "friend" not checking the oil but fails to confirm that this is the root cause. I can't imagine someone routinely failing to check the oil during a preflight. I thought the video was bizzare.[/QUOTE]
 
Even the PRSUs designed for aviation use tend to have issues; early GTSIO contis come to mind IIRC.

Boy do I remember those engines. Had to be "gentle" during significant changes in power settings. I want to say we did no more than 1" MAP per second as a "rule of thumb".
 
I know the 150 engine that I have has knock sensors and it is recommended to replace them with a different type that is not resonant and then their use has to be programmed on a dyno. If one backs off the timing in response to knock sensors, this failure either would not have happened or else it might have taken much longer to progress giving the pilot adequate time to respond to a dropping oil pressure situation.

Here is the video https://youtu.be/19qpxutW_m4

Jan, for all the things he has done in the past seems to be going to pretty long lengths to protect his products reputation and helping customers stay in the air. Bear in mind the core engine is $1000-$1300.


Are you flying behind the 150 Viking? Who designed and who is building their PSRU?

In the video he points the finger at loss of oil due to his "friend" not checking the oil but fails to confirm that this is the root cause. I can't imagine someone routinely failing to check the oil during a preflight. I thought the video was bizzare.

I thought the video was helpful and most likely gets the root cause correct. For whatever reason the pilot was never directly asked if he checked the oil level before takeoff or when the last oil change was. The video has a good shot of the cowling and there does not appear to any oil dipstick access door where the dip stick is located on the engine (orange handle on the top of the engine). So the only way to check the oil level and add oil is to remove the upper cowling, which is most likely a PIA on this airplane. The pilot's ADM was surprising, deciding to "press on" with white smoke and an initial oil pressure drop. Also, interesting that he was trying to talk to Jax Center in airspace owned by Tyndall Approach (F95 is in the middle of a MOA) or maybe Tallahassee Approach as he was headed toward their Class C.

I suspect this is a case of someone who does not routinely check the oil level and treats it more like a car in that regard. At least he and the airplane came out in one piece and lived to fly again.

John Salak
RV-12 N896HS
 
In NorCal, presence of water in fuel is due to the delivery from refinery to distributor. For example, there's a pipeline from the East Bay to Fresno. They ship all grades through a single pipeline. They separate grades by water slugs (iirc, 200+ gallons) They have a switching manifold to direct 87 to one tank, 92 to another, etc. The water settles into the bottom of the large tanks and is siphoned off from time to time. Sometimes water gets into the delivery tanker truck.

I found out the hard way. Scrud water into my VW diesel, overwhelmed the filter and gummed up the internals - required pumping out the tank, replacing the fuel pickup and a $1100 injection pump overhaul. Overall, bill went to $1500.

FWIW
 
Last edited:
I had a 2009 Jetta Tdi, always bought fuel from the same gas station. The fueling system failed at about 37k miles and was replaced in totality, under warranty. The bill would have been about $8k had it not been covered under warranty, on a car that 1 year before sold brand new for $28k. If it had failed on me again, when it was no longer covered under warranty it would probably have totalled the car. The Bosch Piezo injectors were about $1300 each the way I recall, then there was the high pressure pump, rail, sensors, pressure regulator etc.

My ex wife got the car in our divorce and she also got the "extra" $12k that they offered over and above KBB during "diesel gate". Some people just have a better crystal ball than others...
 
On the "HondaFit708huECU" group (a group for people who own the Viking 110 and developed their own engine controller to resolve longstanding issues with the Viking engine controller). There was recently a posting by Mark Hubelbank who developed the engine controller, to say that his gearbox input flange had thrown off 1 of its 3 "ears" which had then taken out the coolant system before finally exiting the cowling and falling to the ground.

The attached picture shows the fatigue crack, which ultimately penetrated the entire cross section of the failed drive lug. This is at the very least the second known incident of this failure. The previous example (different builder) failed at 130 hours. Mark did not mention how many hours he had on his engine before it failed. In the BMW application of the flexible element it fits to the gearbox output flange where it is similarly held on cantilevered fasteners, but obviously the drive flange itself is strong enough to withstand the bending forces, even at much higher torque values. On the other hand, since in the Viking application it is turning at crankshaft RPM, the number of fatigue cycles will be higher than on the car.

Bear in mind that Viking has pretty much abandoned the 110 engine. There are very few parts available for it. It was the original "low profile" engine with all of its special CNC parts. The products that came after have mostly been vertical engines with large offset gearboxes. This is much less complicated to build, eliminating the need for special parts except the gearbox itself.
 

Attachments

  • Fatigue fracture.JPG
    Fatigue fracture.JPG
    123.1 KB · Views: 152
Great pic. From it you could observe/would assume:

The fatigue cracking has existed for a while
It occurred in “stages” probably related to operating profile
The final rupture surface is relatively small compared to the micro-cracking

Great example of when “Don’t know all the design conditions, I’ll just oversize it” doesn’t cut it.
 
The parts look a bit different from this photo. Was there more than one design?
 

Attachments

  • viking.jpg
    viking.jpg
    161.7 KB · Views: 137
The parts look a bit different from this photo. Was there more than one design?

Its the input flange into the gearbox. It is hidden in your picture, all you can see is the rubber coupling and the aluminum flywheel. The aluminum flywheel had a habit of cracking too so there was a mandatory requirement to change it to a newer steel flex plate style flywheel with a welded on ring gear.

No you definitely have the right engine there. The pic I posted is taken from a perspective behind the gearbox looking forward and 1/3 of the rubber coupling is gone, together with its drive lug. Its definitely a brittle fracture.

The worst loading on the system would be from idling or taxi but if misalignment between engine and gearbox is a problem one would have a rotating load with the number of load cycles being directly connected to engine RPM. These engines run close to 5000rpm on take off just like a Rotax and that was when both (final) failures occurred.

See attached the Eric Miller failure, he had the newer flywheel and 130 hours and similarly had his cooling system taken out when the failure happened.
 

Attachments

  • Eric Miller failure.jpg
    Eric Miller failure.jpg
    80 KB · Views: 126
Last edited:
There are several configurations of the Viking 110 engine-gearbox coupling. The original was the machined aluminum flywheel that used a bolted SKF flex coupling design (Ross's photo). A "mandatory" upgrade to the steel flywheel/starter ring was issued and new, stronger drive adapters called 'spiders' were made available. One spider bolts to the crankshaft and the other is pressed on the splined shaft of the input gear for the gearbox. Spiders have three pins each and provide a floating attachment for the SKF flex coupler (VKJF98003 if you are interested). The crankshaft side pins pull the flex coupler and gearbox side spider pins. The Flex coupling is also cut into three segments. Flywheel is also a misnomer, it's basically a disk that holds the starter ring gear.

There are different versions of the newer spiders, the latest are thicker and have welded pins. Easy to spot as the welded pins leaving bluing from the heat and any non-welded units should have been recalled.

The two failures in this thread are one of the gearbox spider ears/pins that broke off. The flex coupler damage appears to be from flailing against engine parts after the pin/ear departed and the engine continued to operate. The unknown is which version of the spiders were installed. Also unknown is if the remaining two ears/pins exhibit any signs of cracking. There is a lot of speculation about the cause and very little engineering analysis to back it up. SKF does have allowance for some misalignment in the specifications (it a flex coupler), so having a perfectly aligned engine-gearbox may not be as critical as some speculate. There is a single use UHMW bushing used to align the GB input shaft with the crankshaft center when mounting the gearbox.

The gearbox input shaft has a helical cut gear that has a slight fore/aft movement as the prop loads and unloads power. There is a potential wear issue that might allow the pressed-on spider to contact the gearbox housing. A newer input shaft is available that reduces the fore/aft movement by increasing the gear width between the two support bearings in the gearbox. That movement may or may not have contributed to the failure as it certainly induces some bending forces on the ear/pin.

Almost all of this is documented on the Viking web site under the Service Bulletins tab. The gearbox maintenance videos are helpful if you want to understand the GB internals and how it works.

This photo is what the current configuration should look like. I have 350 hours on mine and will likely pull the gearbox during the CI in December to check for any cracks as a precaution.

John Salak
RV-12 N896HS
 
John, the gearbox owned by Mark had the original aluminum flywheel and the early input shaft spider with bolts. Marks system would have been over 500 hours (I dont have the exact number) but considerably higher hours than the second one I reported owned by Eric Miller.

The Eric Miller system was of the most recent type, having been updated with the new flywheel and input shaft flange. His system failed with only 130 hours, thus suggesting that it is not more robust than the original configuration on Mark Hubelbanks airplane.

I have asked Mark if he would be willing to send me the input flange, I have a metallurgy department at work that may be willing to donate some time to analysis it, do micro hardness etc. At the very least we could take some nice pictures of the fracture, what is visible on the other 2 lugs etc for the sake of documentation.
 
Having been involved with several types of PSRUs over many years and performing a theoretical TV studiy on mine, the lack of flywheel inertia on the Viking setups stands out.

I haven't seen any design yet which consistently lasts very long (500+ hours) with low flywheel inertia on a four cylinder engine. The range below 1500 rpm is almost always pretty challenging in this regard.

Notably, even the professionally developed Rotax 9 series gearbox doesn't solve the TV issue at low rpm, they just tell you to idle above that range. They have almost zero flywheel inertia, just the tiny amount offered by the dynamo ring, which is very small diameter.
 
Last edited:
John, the gearbox owned by Mark had the original aluminum flywheel and the early input shaft spider with bolts. Marks system would have been over 500 hours (I dont have the exact number) but considerably higher hours than the second one I reported owned by Eric Miller.

The Eric Miller system was of the most recent type, having been updated with the new flywheel and input shaft flange. His system failed with only 130 hours, thus suggesting that it is not more robust than the original configuration on Mark Hubelbanks airplane.

I have asked Mark if he would be willing to send me the input flange, I have a metallurgy department at work that may be willing to donate some time to analysis it, do micro hardness etc. At the very least we could take some nice pictures of the fracture, what is visible on the other 2 lugs etc for the sake of documentation.

Keith

The flywheel in the photo of Mark's engine appears to be the steel version. The gearbox side appears to be the original design that used a hex bolt to attach the flex coupler. As I said, there appears to be several configurations in use. I had about 15 hours on my aluminum flywheel before replacing it and it had no signs of cracks but did exhibit some impact marks where the gearbox side bolt nuts hit the flywheel.

Glad to see you are willing to get some scientific failure data that may be useful in determining the origin of the cracks. Hopefully Eric will contribute his failed part as well for analysis.

John Salak
RV-12 N896HS
 
Having been involved with several types of PSRUs over many years and performing a theoretical TV studiy on mine, the lack of flywheel inertia on the Viking setups stands out.

How would one go about doing a TV analysis? I don't even know where to begin.
 
How would one go about doing a TV analysis? I don't even know where to begin.

You can search the forums here. Dan Horton knows a lot about this and helped me out a lot with mine. You must measure or calculate the stiffness and MMOI of each element to do a mathematical analysis.

Alternately, you'd instrument the engine/ drive and run a direct measurement of the frequencies and amplitudes present from idle through to max rpm with the prop in place.

Both are rather involved and the latter would require specialized gear and engineering knowledge to complete.
 
Ross, at the moment the problems with this specific application are related to the type of flexible coupling used and its interface components. In addition to the way the gearbox is mounted which neither provides adequate alignment between gearbox and engine nor is capable of holding it in position adequately when it is under power.

It sounds like the axial clearance on the input gear is also not adequately controlled, thus under power, considering these are helical gears, the input shaft moves in the axial direction and can cause metal to metal contact between parts that should not have any. At the very least some measurement needs to be carried out and 1 or more spacers added to limit axial movement of the shaft.

Finally, the way that the gearbox mounts to the engine needs to be revised to incorporate something better than the pieces of bar stock currently used as standoffs. As one can see from the pictures there is very little support near the position of the flexible coupling, ideally there should be some supports below the centerline of the crankshaft. It even sounds like some mounting locations on the block were cut off by Viking but I will have to form my own opinion when I look at Erics example which had been flying in an RV-12.
 
Given the fundamental design issues described here, TV worries might be down the list of things to correct (but still very important).

Again it looks like another case of TLAR engineering (which can work fine in some instances) coupled with Jan's typical lack of adequate testing and validation before release to customers.

Customer Beta testing uncovers faults and design revision, A, B, C, D etc. follow to correct said faults.

With engines and gearboxes, a vendor should be dyno and flight testing the wares for at least 1000 hours before sales release. 10-20 hours doesn't prove much.
 
Last edited:
See attached the flywheel side drive adapter (shown on the 130 engine service bulletin, but the comment says it will directly fit the 110 engine also)
 

Attachments

  • Viking 110 flywheel side drive adapter.JPG
    Viking 110 flywheel side drive adapter.JPG
    107.7 KB · Views: 84
Ross, the attached was clipped from a youtube video but shows the engine side of the Viking 110 gearbox. Look at the position of the mounts to the engine relative to where the input shaft is at. See how the input shaft is totally cantilevered out into space ? No support of any kind anywhere near it.
 

Attachments

  • Back of Viking 110 gearbox.JPG
    Back of Viking 110 gearbox.JPG
    172.9 KB · Views: 85
A design like this shows a fundamental lack of understanding in how to design a PSRU. This is TLAW (that looks about wrong). Which is much worse than TLAR even.

Wow, just wow...
 
Mounting

Ross, the attached was clipped from a youtube video but shows the engine side of the Viking 110 gearbox. Look at the position of the mounts to the engine relative to where the input shaft is at. See how the input shaft is totally cantilevered out into space ? No support of any kind anywhere near it.

The way that unit is mounted is, frankly, pretty scary. Even the prop thrust center line is cantilevered slightly relative to the PSRU mounting pattern. The long mounting bolts and what appear to be pressed in spacers is another failure point waiting to happen. Imagine the loads that radial imbalance forces from the propeller put on the mounting bolts, as well as the loads on the housing itself and engine block mounting points that planar (tracking) imbalance puts on them. Both of those types of loads are cyclic. Ross is right about TLAW!

Skylor
 
Last edited:
The later vertical engines, which can utilize all the OEM car parts and have no need for custom made parts also have larger offset gearboxes and therefore naturally tend to have a more robust gearbox mounting arrangement, although they are still of the tubular cylindrical spacer type, which is not optimal.

I am buying one of the failed engines and will be making a new back plate for the gearbox in conjunction with a new engine-side adapter plate to provide 3 extra mounting locations below the crankshaft centerline. Which is what should have been done in the first place. I will be replacing the guibo with a flex plate incorporating PU bushings which should offer flexibility in tuning the damping through different shore hardness. Where the current one is filled with cordage and under tension since it has been cut on the bandsaw. Tracy Crook used PU bushings with his Mazda 13b reduction drives and he never had any problems with it.
 
Last edited:
The later vertical engines, which can utilize all the OEM car parts and have no need for custom made parts also have larger offset gearboxes and therefore naturally tend to have a more robust gearbox mounting arrangement, although they are still of the tubular cylindrical spacer type, which is not optimal.

I am buying one of the failed engines and will be making a new back plate for the gearbox in conjunction with a new engine-side adapter plate to provide 3 extra mounting locations below the crankshaft centerline. Which is what should have been done in the first place. I will be replacing the guibo with a flex plate incorporating PU bushings which should offer flexibility in tuning the damping through different shore hardness. Where the current one is filled with cordage and under tension since it has been cut on the bandsaw. Tracy Crook used PU bushings with his Mazda 13b reduction drives and he never had any problems with it.

While your approach may result in a more robust gearbox attachment and drive coupling, the last thing you need on an RV-12 installation of this engine/gearbox combination is increasing the engine/gearbox weight and moving the engine CG forward. I ran the numbers on the 130 hp version, which has been installed on a couple of RV-12s, and the CG and weight were unacceptable. Single pilot and low fuel showed a CG forward of the RV-12 limits. You can counter that by adding weights in the tail, which in turn kills your payload weight.

One thing to keep in mind about the two-gear gearbox is the left turning prop. Adding the third gear makes it a right turning prop. IIRC, one of the original design goals was to have an engine-gearbox combination with a low overall vertical height, which resulted in the tilted engine orientation and the two-gear gearbox to keep the prop shaft line somewhere near the installed Rotax 912's, which was the target for the 110 engines. Keep in mind, the same parts being questioned here are running on the 130, 150, and 190 HP versions of the gearbox.

Perhaps something like the Centaflex A-50 type flex coupler in the photo might be a better approach than the SKF guibo design for this particular instance. The input shaft of the gearbox already has a splined shaft, so you only need to fix the engine drive plate attachment.

John Salak
RV-12 N896HS
 

Attachments

  • Centaflex A-50 flex coupler.jpg
    Centaflex A-50 flex coupler.jpg
    171.6 KB · Views: 46
Tracy used increased gear backlash to attenuate the worst TV periods as well. Not sure about flywheel mass. Two big factors which may not be applicable here.

Every engine, gearbox, prop combo will have different TV characteristics.

A properly selected Centaflex coupler is probably the most elegant solution but you'll need the numbers to pick the right one for this application.
 
And yet, any WW-II fighter with either a radial or a big V12 used a PSRU. And I don't believe I've ever seen a flywheel in one of these. They DID use heavy metal props though, which on our applications typically is a no-no.

Why does that work in those cases, and not in ours? The 9/12/18 cylinders on those engines making for a smoother engine to start with? Brutal force? More clever engineering? Or are these engines flawed too, with red areas in the rpm band and all that?
 
And yet, any WW-II fighter with either a radial or a big V12 used a PSRU. And I don't believe I've ever seen a flywheel in one of these. They DID use heavy metal props though, which on our applications typically is a no-no.

Why does that work in those cases, and not in ours? The 9/12/18 cylinders on those engines making for a smoother engine to start with? Brutal force? More clever engineering? Or are these engines flawed too, with red areas in the rpm band and all that?

Multiple cylinders with no torque reversals (with much lower mean torque variation), clever engineering (quill shaft in the Merlin for instance), rolling pendulum tech on radials etc.

A 4 cylinder engine is a more challenging subject.
 
Multiple cylinders with no torque reversals (with much lower mean torque variation), clever engineering (quill shaft in the Merlin for instance), rolling pendulum tech on radials etc.

A 4 cylinder engine is a more challenging subject.

Speaking of the Merlin, here is a PDF on its TV: An Examination of the Torsional Vibration Characteristics of the Allison V-1710 and Rolls-Royce Merlin Aircraft Engines

By 1940, most of the high output radial aircraft engines were utilizing tuned pendulum dampers to minimize torsional vibrations in the crankshaft- propeller system. The situation in the in-line high output configuration was less consistent. While a number of earlier engines had adopted various types of dampers (mainly of the friction type) by 1940 all but one engine, the Allison V-1710, were without dampers of any kind. Rolls-Royce, Daimler-Benz and Junkers, all with liquid cooled V-12s and operating at comparably high outputs, were damper free. The Rolls-Royce Griffon, under development but not yet in service, never employed a damper.
This paper is an attempt to explain why the Allison engine was unique in this respect.
 
Excellent find. I'll pore over that when I have some time. Thanks for posting the link.
 
Last edited:
Back
Top