Tuesday, June 19, 2007

AV - Prop Hubs

It started with this.

At the end of World War II Type 82 Kubelwagens littered the European landscape. In many cases the vehicle was undamaged, abandoned when it ran out of gas.

The Type 82 came in two flavors, those with the original 985cc (70x64mm) engine producing 23hp and the later version fitted with the E-type 1131cc (75x64mm) engine developed in 1943. Given a choice between an early or late Type 82 most folks opted for the later version, whose larger engine produced a neck-snapping 25hp instead of a paltry 23. (Actually, both performed about the same.) It was the 1131cc engine that went into the post-war Volkswagen, remaining in service until the 1956 model year, when the 75mm barrels were bored-out to 77mm, upping the displacement to 1192cc; the '1200' engine. ( Volkswagen engines are designated by a number approximating their displacement whereas chassis are identified by type number. Referring to an upright VW engine as a ‘Type I’ is something of a joke since the Type I (meaning the sedan or bug) used six different engines over the years.)

With the small, light-weight 985cc engine free to anyone willing to pull it out of a defunct Type 82, it wasn’t long before someone decided to bolt it to an airplane, which they could do without using an engine mount thanks to the transmission flange cast onto the front of the crankcase. (With Volkswagens, orientation is always relative to the vehicle.)

Of course, if you bolted the airplane to the flywheel-end of the engine that meant you’d have to attach the propeller to the pulley hub, violating a basic tenet of using a car’s engine in an airplane, which was to put the prop on the beefier flywheel end of the crankshaft. But since they were only looking at 23hp they figured it was no big deal. And it wasn’t.

Before getting into the details it may help if you know a bit of history, such as the fact the Kubelwagen started out as the KDF Wagen, which started out (in 1933) as the NSU Type 32. After Hitler came to power in 1933 the Nazi Party took a strong interest in Professor Ferdinand Porsche’s dream of a People's Car. Over the next five years the Nazi party subsidized it’s design to the point where they largely took over the project, which finally came to fruition in 1938 as the KDF (Strength-thru-Joy) Wagen. Even as the cornerstone of the new Volkswagen factory was being laid the Wehrmacht issued orders to develop a military version of the KDF Wagen, which became the Type 62, the precursor of the Type 82. (As a point of interest, after the basic design was accepted by the Nazis, the factory test drivers were replaced by 200 army personnel who drove the fleet of prototypes day and night until each had accumulated over 50,000 miles.)

The point of all this is that the Volkswagen didn’t just suddenly appear. It was the product of a long, expensive R&D program in which every aspect of the vehicle was carefully studied and engineered, not only as a civilian vehicle but as a war machine as well.

Figure 1 shows the nose of the Volkswagen’s crankshaft. The pulley-hub was designed to transmit about 7hp via a vee belt to the dynamo and cooling fan. To ensure it would never fail at that level of output, it was designed about 5x stronger than necessary, a fairly common practice when dealing with castings. Because that’s what the original crankshaft was. The choice of a cast-iron crankshaft was driven by the mandated design-goal of keeping the price of the KDF Wagen at or below 1000 Marks. Unfortunately, during extensive road testing the cast crankshafts proved too fragile for the task and late in 1936 they were replaced by forgings, although the dimensions remained the same in order to accommodate existing tooling.

The forged crankshaft was mild steel, the DIN equivalent of SAE 1045. Although significantly better than the cast-iron originals it was designed for economy of production, lacking counterweighted flanges and other features commonly found on crankshafts even then. Also note the internal M20x1.5 threads. Metric threads have a sharp 60 degree peak & valley whereas NC, NF and Whitworth threads are rounded. During development the design proved a bit too flexible. Even the fairly light load of the belt-driven blower and dynamo produced a bending-moment sufficient to precipitate the formation of cracks between the sharp threads and the keyways. Volkswagen resolved the problem by installing a 4th Main Bearing immediately adjacent to the pulley-hub. The design has remained substantially the same to this day. (The air cooled VW engine is still being manufactured in Mexico.)

Figure 2 shows the location of the new #4 bearing needed to off-set the bending moment imposed by the asymmetric load of the belt-driven blower & dynamo. The small size of this bearing, only 40mm as opposed to the 55mm diameter of the three real main bearings, makes its ancillary role evident. Now let’s go fly one... and see what happens :-)

I don’t know who put the first prop-hub on the fan pulley but I know how they did it. It was made from a piece of mild steel and probably looked pretty much like Figure 3. There are several variations on this theme. For example, many of the very first prop hubs did not include the reverse thread that serves as an Archimedes’ Pump to keep the oil from inside the engine. Instead, they installed the oil seal from the inner bearing of a Kubalwagen’s front wheel, which happens to have the same diameter as the pulley hub. Some hubs extended farther than others; some where threaded for prop bolts but most expected the prop to be secured with nuts & bolts. Some had a propeller guide-ring as deep as half an inch; others had no guide-ring at all.

Doesn’t really matter. Tapered and bolted to the little 985cc engine, they all flew, after a fashion. Which wasn’t anything new since Volkswagen engines had already flown. In 1937 the Horton brothers were allowed to install a VW engine in one of their flying wings. But in doing so they followed the accepted convention, taking power from the clutch-end of the crankshaft.

The tapered hub worked well enough with the 23hp engine but problems arose when that type of hub was used on larger engines. The bigger the engine, the more the torque and the more critical became the fit of the tapered hub to the crankshaft. Tapered hubs remain available and in use today (2007) but are seldom installed on engines larger than the 1600. Even then, they have a history of breaking off.

You had to be a pretty good machinist to make the tapered pulley-hub precisely match the taper of your prop hub, which was accomplished by lapping the two together. A lot of folks thought there should be an easier way. Which lead to the Shrink-fit Hub, as shown in Figure 4.

To achieve a shrink-fit the interior diameter (ID) of the hub is made smaller than the outside diameter (OD) of the crankshaft. The hub is then heated until its ID has expanded enough to fit over the crankshaft. As the hub cools, it locks itself in place.

According to Machinery's Handbook (an accepted standard), “The intensity of the grip and its resistance to slippage depends mainly upon the thickness of the hub.” (14th Edition, pg 1055). Formulae are provided to calculate the required difference between OD and ID. Unfortunately, the presence of the keyway and the threads of the Archimedes Screw (if present) limits the amount of grip we can produce before causing the metal to crack along the keyway. I’ll get back to this down below but for now you should know that some shrink-fits are not as strong as others. For example, with the hub shown in Figure 4 the strength of the grip will be determined by the depth of metal between the lowest portion of the Woodruff keyway and the lowest portion of the oil-slinger thread, which is .124" – a scant eighth of an inch. That is, when you subtract the depth of the keyway and the threads the effective wall-thickness becomes a mere eighth of an inch.

To get a stronger grip you need a thicker wall, such as the one shown in Figure 5. But a thicker wall on your prop hub dictates the need to machine a larger opening in the nose of the crankcase, a daunting chore for anyone lacking an engine lathe.

Unfortunately, even when you do all of that, you still run the risk of having the prop fall off. A propeller generates some rather massive bending stresses in the nose of the crankshaft. This is because the gyroscopic effect of propeller resists any change in the attitude of the aircraft. And while we as pilots are aware of such changes when they are large, small changes occur constantly during flight. Accumulate enough small stresses, you won’t have to worry about the big ones... because your prop will already have broken off :-)

Figure 7 shows how the crack propagates. It typically begins in the root of the internal threads and connects to the lower corner of the Woodruff keyway. It then follows the corner of the keyway until it intersects the machined groove at the base of the #4 bearing’s journal. In the 1950's I had the unique experience of suffering two such fractures within a 24 month period. The first was a minor event; more of an inconvenience than an accident. The second was a bit more exciting and put an end to my flying for a couple of years. When I finally returned to the air it was behind an engine wearing the propeller on the clutch-end of the crankshaft, a method I’ll describe in closing.

Back then, the experts insisted a tapered hub was the only way to go; if the prop broke off then I must of done something wrong. And the same experts said exactly the same thing the second time it happened. But other folks had also suffered broken cranks. And some of them were wizard machinists. While poor workmanship may have been a factor in some cases it was clearly not true for all.

One very popular solution to the breakage problem is shown in Figure 8. The Woodruff keyway is welded closed, the annular groove is welded full and the three-degree taper is continued right across the journal of the #4 bearing. The internal threads are usually honed away so as to eliminate them as a stress-riser and the bore is re-threaded more deeply into the thicker section of the crankshaft under the cam gear. To index the hub a shallow keyway, similar to that found on the early Continental crankshafts is machined to accept a section of square key-stock. A matching groove is broached in the tapered ID of the hub.

This procedure also applies to the heavy-walled shrink-fit hub.

This does not do away with the gyroscopic loads nor bending moments induced in the crankshaft but the resulting stress is now distributed over a significantly greater area so that the per-unit stress are typically below the level needed to initiate cracking. And if the crack can’t get started, it can’t do any harm. But by eliminating the #4 bearing we have effectively made the crankcase into a bearing and the outer surface of the prop hub into a journal, a role neither was designed to fulfill. This is illustrated in Figure 9.

Running in the parent metal of the crankcase, our longer prop hub will very quickly oval it out due to a lack of lubrication. And without the bearing-support of the crankcase we discover our longer prop hub now serves to amplify the bending stresses that are appearing in the machined groove at the base of the #3 Main Bearing. Our cam’s gear is also developing a very weird wear pattern. It is only a matter of time before we eat the cam gear or suffer another broken crankshaft, this time adjacent to the #3 bearing instead of #4.

The fix for the fix is shown in Figure 10. The prop hub – long-tapered or thick-wall shrunk – is fitted with a sleeve-type bearing. The nose of the crankcase is opened up to support the sleeve and the OD of the prop hub is polished to serve as a journal. There are several variations to this method, most involving how the sleeve receives it’s lubrication. Functionally, all are pretty much the same although I’m a bit surprised that no one has adopted the lubrication arrangement developed by Bob Huggins which was the best of the bunch, in my opinion. But then, my opinion doesn’t count for much. For more than thirty years I’ve been putting the propeller on the wrong end of the crankshaft, filtering my oil and doing any number of things that are all wrong, according to the experts :-)

Figure 11 shows how I make a crankshaft flange out of an old flywheel. The crankcase requires no modification nor is there much machining to speak of. The flange requires a spool to position the prop far enough forward to clear the #3 exhaust stack but a spool is a simple turning; easy to make.

Several years ago while cleaning out a cabinet in the shop I came across samples of all the prop hubs I’ve made and flown behind over the years. I laid them out on a bench and took some snap-shots. When the pictures came back I found somebody with a scanner and eventually posted them to the internet. I think I sent to the FlyVW Group, which used to be on eScribe back then. Later, Yahoo bought eScibe and despite promises to preserve the existing archives, erased them. Since that time several people have asked me to re-post the pictures but I haven’t any idea in the world where they’ve gotten to. Which also goes for the collection of hubs & stuff. I know I’ve given some of the items away, and I used one of the stainless steel hubs on wind turbine. The following photos will give you some idea of what I’ve been talking about. Some are pretty tatty, discovered outside under a bench or forgotten in the back of a drawer.

Image C is a thick-wall, shrink-type hub. This was one of five I made about 1970. The barrel of the hub is 4130, the flange is mild steel. The two are pressed together then welded on both surfaces, heat-treated, then machined. They never came loose but they had a habit of causing the nose of the crankshaft to snap off. Image D shows you the other side. The thing in the background is a basic spool, anodized some silly color. (I was just a kid back then :-)

Image E will give you some idea of the difference between a regular shrink-type hub and the thick-walled variety. The regular one is turned from a billet of 4130. (Expensive!) Image F offers a comparison between a regular fan-pulley hub and the regular (ie, thin-walled) shrink-type hub. Although they don’t look much alike, Image G shows that they have exactly the same hub diameter, which is a tad less than two inches and fits neatly into the nose of a stock crankcase. Installed on a large-displacement engine the thin-walled shrink-fit hubs tended to loose their grip over time. Machine them for a tighter fit (ie, more shrinkage differential) and they would crack along the keyway.

Image H shows what a flywheel looks like from the front when it’s cut-down to serve as a flange. Image J is a view of the back. The holes are threaded because the spool is usually bolted to the flange and pretty much stays there for the life of the engine, whereas the propeller uses regular nuts and prop-bolts to attach to the business-end of the spool. This is one of the first I ever made, probably about 1965. The lack of the O-ring groove sez it’s for a forty-horse crank. Image K shows how it fits the spool.

A lot of folks think there’s a streamlining problem when you put the fan on the clutch-end of the engine. In fact, the VW’s tranny flange is barely thirteen inches in diameter. With a four inch spool and a spinner ten inches in diameter, the tranny flange is completely submerged in the streamline between the spinner and the firewall. (But of course, that can’t be right :-)

With that as preamble lemme give you a glimpse of the future. Image L is a flywheel flange & extension spool from Great Plains Aircraft Supply Company. For folks who can’t get along without a starter, the flex-plate & ring-gear attaches to those extra holes.

The last image is kinda sad. I made my stuff on whatever tooling I had available. I was in the Navy back then, often had to go begging to get some lathe-time. As you can see from the stuff I made, I’m not a very good machinist. (Still learning, though :-) Steve’s stuff is flat-out beautiful. Produced on state of the art CNC machines, marvelously accurate and painfully precise, they are as much a work of art as a piece of machinery. And to own one all you gotta do is give him money. (Seems like cheating :-)

But the real question you gotta ask yourself is why people are still putting the prop on the pulley hub. We're no longer salvaging free parts from Kubalwagens, we're spending thousands of dollars to build 140cid engines based on after-market VW components. All the converted auto engines you can think of - Model A, Corvair, Subaru - bolt the prop to the clutch-end of the crankshaft and go flying. So how did we get trapped inside this box that says Volkswagens have to mount the prop on the pulley hub?

(June 2007)

1 comment:

Anonymous said...

This is the best article on this subject I have ever seen. I want to thank you for sharing this. You've saved me untold hours of experimentation and confirmed a lot of my own suspicions about the pulley end prop drive. As you point out, the only reason these work in the first place is because we're pulling much less power than we think we are out of these little engines.
Excellent point about the precessional or gyroscopic loads on a prop. If you compare an aircraft engine you would see most of the bearing and hub area is designed to take these loads, with surprisingly little attention paid to thrust forces!
Again, Thanks