PDA

View Full Version : CVT-PRSU what do you think?



Pat_Panzera
10-17-2011, 03:52 PM
Continuously variable transmission propeller speed reduction units – what do you think?

In the Oct 2011 (http://www.eaa.org/experimenter/issues/1110.html) issue of EAA's Experimenter eNewsletter, EAA member Vince Homer submitted an article that asks the above question.

The article is merely a discussion on the potential of utilizing the technology found in the continuously variable transmission driving some all-terrain vehicles, snowmobiles, and a small number of automobiles. The idea is similar to a constant speed propeller except the engine speed (and horsepower) is increased, or otherwise variable, with the prop speed staying relatively constant. Some may already see a flaw in this concept. However, please read the article and share your thoughts.
http://www.eaa.org/experimenter/articles/2011-10_cvt.asp

P (http://www.eaa.org/experimenter/articles/2011-10_cvt.asp)at Panzera
Experimenter eNewsletter Editor

Pat_Panzera
10-17-2011, 05:33 PM
Please note that the article in the Oct issue has not been launched at this writing. Look for it in your email inbox tomorrow afternoon, (October 18th). If you are not a subscriber to Experimenter, click here: http://www.eaa.org/experimenter/

(http://www.eaa.org/experimenter/)

Mike Switzer
10-17-2011, 05:35 PM
I have been considering this, but I haven't messed with one since we used one on our Formula SAE car back in the late 80s. My initial thought when I was pondering it was how would you go about deciding how to pitch the prop, as I believe the prop pitch is based on available HP at a certain rpm. (I would guess pitch it for cruise power & pick a prop RPM, say 2500? but how would it behave on takeoff?)

Mike Switzer
10-17-2011, 05:41 PM
Taking my thoughts a bit further, i am planning on running a supercharged 302 Ford, which I would probably cruise somewhere between 3500 - 4000 engine rpm, with the prop at say 2500 rpm. On takeoff you could take the engine RPM quite a bit higher, but you would not want the prop RPM going more than say 2600 - 2800.

Using a fixed pitch prop with a variable transmission, if the prop was pitched for cruise you would not be getting peak efficiency at takeoff power settings.

(At least those were my thoughts at the time)

Dana
10-17-2011, 06:52 PM
We've had discussions on this over on homebuiltairplanes.com. The usual conclusion is that a CVT doesn't make nearly as much sense as a controllable pitch prop (constant speed or otherwise).

Chris
10-19-2011, 04:26 PM
This just seems all wrong to me. Constant speed propellers turn at a near constant speed, but the pitch changes with the speed of the aircraft. Take off=fine pitch due to low airspeed. cruise=high pitch due to higher airspeed. A fixed pitch propeller is at best a compromise. Most efficient at one airspeed at a given rpm. Varying the rpm and torque of the engine does nothing to increase the efficiency of a fixed pitch propeller turning at a constant speed. Am I missing something here?

Mike Switzer
10-19-2011, 04:54 PM
Am I missing something here?

I don't think so, I was thinking about this one night when I couldn't sleep & that was pretty much what I came up with.

This is one of the challenges we have to deal with if we want to run a non-aviation engine (or any engine with a solid crankshaft). I haven't seen a good solution yet, the electric props aren't really true constant speed props, I liked some of the things about Bud Warren's Geared Drives design but after the latest crash that is probably dead (and from what I have heard the problems were with the GM engine management system, not the redrive)

Vince
10-19-2011, 08:03 PM
As the author of the article I admit bias toward this concept. If you look at a plot of propeller efficiency vs advance ratio one thing pops out. At higher pitch angles the efficiency curve is rather flat for a wide advance ratio range. Conversly, the curves are very peaky for shallower prop angles. My thought was to test at high prop angles at speeds from fero to whatever I can get with my test stand and to do this you need to be do this you need a variable ration PSRU.

I am just starting to build a rolling engine/CVT/propeller test stand and will follow up this article when I have some data.

I am also still looking for a propeller calculation program that will accept RPM, Prop dia & pitch and HP as inputs to calculate thrust and speed.

Mike Switzer
10-19-2011, 10:24 PM
The one advantage I can see, is whether you were running a constant speed prop or a fixed, using a variable transmission you could run an auto engine up to 6 - 8000 rpm on takeoff & get a lot more power, then back it down for cruise. I wouldn't want to run an auto engine over 4000 at cruise, but I have run a 289 at 8000 rpm peak for short durations (for road races basically "cruising" at 6000), and they will do it without flying apart

Dana
10-20-2011, 04:51 AM
...My thought was to test at high prop angles at speeds from fero to whatever I can get with my test stand and to do this you need to be do this you need a variable ration PSRU.

I am just starting to build a rolling engine/CVT/propeller test stand and will follow up this article when I have some data.

I am also still looking for a propeller calculation program that will accept RPM, Prop dia & pitch and HP as inputs to calculate thrust and speed.

A CVT makes a lot more sense on a test stand, when you're not only designing a propeller but also trying to determine the best gear ratio for it.

For a prop design program I suggest taking a look at Javaprop (http://www.mh-aerotools.de/airfoils/javaprop.htm). But you're not going to get thrust and speed out of it; speed is an input and it tells you thrust. Also blade area is a factor.

jmccreight
10-20-2011, 06:44 AM
Hello all,
I'm not sure I see the advantage either, just based on the fact that these are typically used to keep the engine right in the powerband for road use through the whole range of road speeds. Doing this reverse stuff is probably just going to take away some performance, right?

That said, I'll play along: how about Honda's HFT CVT for the job? https://secure.wikimedia.org/wikipedia/en/wiki/Human_Friendly_Transmission

It's hydraulic, so no belt failures here! It should also work great through its entire range of motion, while not requiring anything special at the extreme ends.

Edit: some cutaway shots of it in this video: http://www.youtube.com/watch?v=5b4_MIe5byI

Vince
10-20-2011, 07:01 AM
Anybody have any idea how much one of these units weighs or what their operating efficiency is?

The CVT belt/chain type units appear to be rather heavy. Anyone know what they weigh?

jmccreight
10-20-2011, 08:53 AM
They claim in the video that the hydraulic method is more efficient, though I don't know what the weight or exact gains are from such a system. Unfortunately, I also don't have a DN-01 to disassemble, and even if I did, I can't imagine I'd want to. :rollseyes:

David J. Gall
10-21-2011, 12:15 AM
Vince, you're looking at the wrong curve. Don't look at the efficiency curve, look at the power absorption curve. It goes as the cube of the RPM with minor variations for airspeed and propeller efficiency. (Reference: Fraas' "Aircraft Power Plants", McGraw-Hill 1943, page 88.) So, at full power (not full throttle like on the Lycoming, but at full power) the prop rpm will be very close to its design ("max cruise") rpm even at zero airspeed. (The prop rpm will be a little less than cruise rpm because the prop efficiency is so bad at zero/low airspeeds.)

Then the CVT will allow the engine to spin up to its rated rpm in order to deliver its full power to that prop from brake release all the way up to full-power design airspeed for the prop (and airframe!). However, be careful in your choice of CVT so that you get a range of ratios that will limit the propeller's rpm. You'll want a CVT that is at minimum reduction ratio when you get to cruise airspeed so that your CVT will not allow the prop to overspeed when you glide. Also be sure to use both an engine tachometer and a propeller tachometer, since propellers have operating limits, too.

I think you'll have success with this if you choose components that are matched in terms of horsepower and rpm and put them on an airframe that is matched to the design airspeed of the prop and the horsepower. (I shouldn't have to say that but so many good projects have gone astray on those "details"....) Might I suggest a Cub or Champ as a good place to start when you get tired of the "sled." And try to get an engine/CVT combo that will allow you to use the Cub's same prop for good comparison testing. However, I think that, given the same horsepower as the engine you're replacing, the only advantages you'll see are shorter takeoffs and modestly improved climb. (OH! And a cheaper engine package, too?)

I can't wait to hear a progress report!

Vince
10-21-2011, 08:46 AM
Thanks for the reply.

In the article I show the propeller power absorbed curve and compared that to the engine power curve. I was referring to the efficiency vs advance ratio curves to try to convey my thoughts about operating a fixed pitch prop at varying airspeeds.

The main message of the article was that with a fixed ratio PSRU the propeller power required and engine output would only be matched at one RPM and one airspeed. Assuming that you are comparing a fixed ratio PSRU/cruise prop then I agree that the cruise speed would be the same with a CVT. However, it is my contention that, as you say, the takeoff and climb will be improved with the CVT.

I need to get out to the shop where the test stand needs finishing.

Vince Homer

jmccreight
10-21-2011, 11:42 AM
I did find another link fully explaining the Honda HFT (on their own site, of course!) http://world.honda.com/motorcycle-picturebook/HFT/

On page two, there's a nice "how it works" section. However, I looked in the parts catalog and the whole assembly is about 4 grand as a "replacement part." For reference, the whole bike new runs about 9 grand now, so that's a good chunk of what's there. The included engine is only 60 horse, so it might or might not be usable in some kind of Frankenstein conversion. Hopefully this isn't too much of a diversion from your project, but I kind of think something belt free would be less failure prone. (Or maybe more so? Hard to say.)

geob
10-22-2011, 02:46 AM
I have thought about this topic for a long time. We chatted about it some on Yahoo! group small4-strokeengines, but not as deeply as we have here. My original idea wasn't to functionally replace a constant speed prop. I was attracted to the CVT PSRU because of these factors:
- I was involved with small aircraft with engines in the 30-50 hp range.
- the CVT is already manufactured with parts robust enough for the engine hp, torque and rpm, assuming you use the same engine.
- Depending on how much you have to modify to get it to work (if even possible) it may be lighter than extant PSRUs.
- if you are working from a donor vehicle for the engine then the CVT probably comes along free, perhaps giving a $$ savings.
- I was looking for, hoping for SOME benefit from the 'variable' feature. It was never my expectation that this would become the major focus of such an install.
- you have the benefit of thousands of times more in-service hours for a typical CVT than a typical after-market limited production PSRU. Compared to scooters and quads, any aircraft PSRU is limited production.
- you benefit from an existing pool of trained mechanics (limited value here), and an expensive parts distribution system in place.
- because of the years of racing CVTs there exists a mature body of knowledge on tuning these CVTs, assuming the knowledge transfers somewhat intact into aircraft apps.

Thus I think the diverging discussion to the Honda HFT misses the point. Oh, it's interesting, but it is essentially a different discussion. Yes some cars and heavier utility vehicles use CVTs but I was focused on lighter/cheaper quads and motorscooters with CVTs.

On the reliability of the CVT PSRU... I have no clue. I *do* know that Honda was sued because a drive belt failure on a Honda Reflex 250cc scooter on a busy freeway resulted in the rider being struck by a car and loosing a leg. The drive belt failed because of excessive wear. It was not a premature failure because of some engineering weakness. Honda won but it was a squeaker. They recommend replacement at 12k miles now though extrapolation of wear indicates that the belt should be functional out to 18-24k miles. Depends on the rider of course because some have already exceeded these distances with their belt. My point is that MANY scooters use these CVTs. Many man-miles are delivered by them. People successfully depend on them with stakes nearly as high as in aircraft apps. For the most part they are trouble-free given normal suggested maintenance. I expect the same level of reliability in aircraft.

geob
10-22-2011, 03:06 AM
Also the CVT response to rpm and load can be modified extensively. There are a number of approaches to this some of which can even be combined for even more control over response.

If weight was a huge factore the 'variable' feature can be sacrificed in order to save weight. It is easy to lock the sheaves in place. Some mods to save weight:

- remove the fly-weights.
- typically there are a few other parts associated with the moving fly-weights that could be tossed if the fly-weights were removed, such as any retaining parts and the cover. The cover is a stressed part as the weights exert presure on it.
- remove the centrifugal clutch
- remove the large spring on the driven side of the CVT.

After removing all of these extraneous parts you basically have just a PSRU, depending on how you have to lock the remaining clutch parts in order to still transfer power. There are few parts here that are not on the average PSRU. You would still have the down-side of lower efficiency which might be a deal-breaker. You would have to address the cooling issue no matter which direction you went with the CVT- you could keep the stock housing to use the stock cooling system or maybe a free-air system of some sort.

Vince
10-23-2011, 06:24 AM
Removing the variable feature kinda defeats the purpose. As far as weight reduction, the flyweights actually weigh only a few onces. It would appear that without the springs and flyweights the unit would require some sort of "shims" to set the ratio. In any case, I'm sure that a cog belt PSRU would be more efficient than using a locked CVT.

Along this line. I just read an article in the Nov. 1987 "Kitplanes" about Holcomb's Perigee aircraft testing. They went to a variable pitch prop for several reasons, one was to keep the engine from overspeeding with the fixed pitch prop they initially used. A CVT would eliminate this possibility is the prop was choosen correctly. Even if the engine was a little underpropped, the ability of a CVT to "overdrive" would still keep the rpm in range.

Vince Homer

David Darnell
10-23-2011, 09:16 PM
I might be wrong in my thinking, but suspect the CVT in aircraft is a answer looking for a question

Vince
10-23-2011, 10:47 PM
What is needed is either test data or a good mathematical analysis. Anything else is opinion. Since I lean toward hardware, I will try to develop some data.

Vince Homer

geob
10-24-2011, 01:28 AM
Removing the variable feature kinda defeats the purpose

Did you read my msg, above? There are multiple possible valid 'purposes', each one has its adherents.


the flyweights actually weigh only a few onces.

Ounces make pounds. And I listed a few more parts than that. You don't know how much weight we could lose, I don't know either.


I'm sure that a cog belt PSRU would be more efficient than using a locked CVT.

Yes, and I am sure that is true also. The whole point here is lost if you keep picking at small things. The point is that there is more than one reason to use a CVT PSRU. Some reasons may not be important to you but they might be to the next guy.

d-omber
10-25-2014, 02:40 PM
I like the idea of using a CVT for a PSRU for the falling reasons. The hp curve for an engine is convex, where as the hp required for a propeller is concave. This means that they can only match at one speed. When you back off the engine to let’s say 80% rpm the engine is still capable of slightly over 80% hp. But the prop only can use 50% power. This leaves the engine very lightly loaded. 4 stroke engines don’t like this and 2 stroke engines positively hate it.
The CVT will keep the engine properly loaded at all times.
So in by using a CVT for your PSRU you would need a propeller tack and back the propeller to 80% rpm which would bring the engine down to near 50% rpm.
http://eaaforums.org/attachment.php?attachmentid=4265&stc=1

Vince
10-26-2014, 07:45 AM
I have a rolling test stand that does exactly what d-omber suggests. It has 16 HP engine driving a go cart CVT which drives the prop. The engine runs at 3600 rpm no matter what prop or prop pitch is hung on it. For example; I have a 54 inch prop that turns at 3200 rpm in fine pitch and 1600 when in coarse pitch. Essentially what the VCT does is to lay the engine torque curve onto the curve of propeller torque over the operating rpm range of the engine.

I have a PP presentation on this, but can't seem to figure out how to attach it to a post. Is it possible to make files (Power Point, Word, Excel) available on this forum? If anyone is interested I can send the file through email: vhhomer@hotmail.com

Vince Homer