Saturday, November 25, 2006

AV - Propellers

Before you become too deeply involved in discussions of propellers I suggest you make a few. Traditional designs using Clark or RAF airfoils (which have a flat bottom) are quite easy to carve and with only a modest amount of care (and an accurately made set of templates) the results of your work will closely match or even exceed the performance of most of the props being offered as suitable for VW engines. (Hint: Lay-out your templates with a DeltaCAD. Glue the layouts to the metal then simply cut & file to the line.)

Such props are far from ideal, no matter who makes them, but they WILL fly the plane and do so economically, the major investment being your time.

Once you've carved a few props -- and flown behind them -- not only will the work of Paul Lipps make more sense but you'll be able to actually apply some of his advice to a real propeller.

But don't expect miracles. Most homebuilts have a lot of drag and drag increases as the square of velocity. Even the most efficient prop in the world can't change the laws of aerodynamics.

What you should expect to see is improved efficiency -- making better use of whatever amount of horsepower you have available, which in the case of converted VW engines is never very much. An optimized prop will often allow you to maintain your accustomed cruise speed on less power, meaning your fuel consumption goes down, your effective range is increased and the number of required fuel-stops is reduced.


As someone pointed out, citing only pitch and diameter when selecting a prop is little more than a joke despite the fact that's how most homebuilders do it -- and the fact it WILL fly the plane. But if you've never carved a prop you've got to start somewhere and the truth is, the carving is the easy part. The tricky bits are selecting the wood, gluing-up the blank, doing the lay-out, making your templates and learning how to chunk your way down to NEAR the final shape without ruining it.

Break the task down into a logical sequence of chores and you'll find that making a prop is lot easier than you think. And besides, you only have to carve ONE blade. (Then make the other match :-)


AV - Aluminum Jugs

The use of aluminum barrels in VW engines converted for flight is a nice example of how to separate technologically naive homebuilders from their money.

On average, the cylinder barrels of an air cooled engine contribute only 17% of the waste-heat budget. This figure is true for everything from a Weedeater to a P&W R-2800, the 'on average' waffle factor reflecting how the barrels interface with the heads for the purpose of heat-transfer. In a mono-bloc casting, where the head and barrel are intergral, the figure can be as high as 25%. In the jugs of a real aircraft engine, where the barrel is thread and heat-shrunk to the head, the 17% figure is dead-on. For a compression interface, where the barrel is simply squeezed against the head -- as is the case with the VW -- the figure is actually lower. But since the factory manuals for Lycoming, Pratt-Whitney and so forth cite 17%, so do I since it allows people to look it up for themselves.

Remember the earlier thread about the sealing surface between the head and the cylinder? That's why VW (and Porsche) jugs see so little heat. There is certainly plenty of heat in the heads but there simply isn't enough surface-area between the heads and the jugs to allow an appreciable amount of heat to flow from the heads into the barrels.

So why did Porsche use them?

Because it was running out of cooling air.

To produce more power Porsche had to get more air to the HEADS but they'd already hit the limit on the size of the blower and shrouding they could fit in that particular chassie. Their solution was to send LESS air to the barrels so they could send MORE air to the heads. But to keep the barrels from overheating they had to go to a material having a higher rate of thermal transfer... such as aluminum.


If you can afford to spend $1800 to save 9 pounds there's no reason not to use nickies (say hello to Shad when you order them :-) ...but in a flying Volkswagen that is their ONLY benefit. If someone tries to sell you a set because they'll automatically make your engine "run cooler," just smile and walk on to the next booth.


AV - Lapping Jugs, Sealing Rings and Why the Sky is Blue

Recent posts to this Group (Ed.Note: CX4) caused a number of messages to appear in my mailbox. Unfortunately they could not be answered with a simple yes or no and one was specific to the Type IV, on which I'm not qualified to answer. When a fourth message arrived on (generally) the same subject it seemed that a general answer, publicly posted, would be the best solution. So here it is.

To begin at the beginning, Volkswagen did not lap the barrels to the heads but they DID install sealing rings between the barrels and the heads of the Type IV engine.

Why didn't they lap-in their barrels? Because there was no need to do so. The top of the barrel and the sealing surface inside the combustion chamber were both perfectly flat, hence there was no need to lap them in.

So why does everyone think it's a good idea? Well.... `everyone' DOESN'T think so :-)

Volkswagen was the first major auto manufacturer to use gasket-less assembly. That is, surfaces were machined so accurately that it took only a thin wipe of sealing compound, typically Permatex Type 3, to produce a leak-free fit of the crankcase and transmission halves. (Note: Early VW trannys were split down the middle, just like the crankcase.)

To achieve a leak-free fit between the aluminum heads and the cast iron barrels Volkswagen used some very sophisticated engineering. First, they made sure the sealing surface inside the combustion chamber was perfectly flat and that the depth of the sealing surface was PERFECTLY EQUAL in both chambers. Then they used cast iron barrels having a WIDE sealing surface that was also perfectly flat. Smooth, too. (The sealing surface of new 77mm jugs looked like mirrors.) Finally, they provided approximately 170 FT/LBS OF TORQUE to the head-stays.

Which of course is impossible.

The head-stays are merely hand-threaded into the magnesium crankcase for less than an inch. Even with the coarse pitched M10x1.5 thread you'd need nearly twice that depth to withstand 170 ft/lb of torque.

What VW provided was the amount of TENSION approximately equal to that produced by torquing the head-stays to 170 ft/lbs. How they did this is perhaps the trickiest bit of engineering in the whole engine because the assembly-torque was only 23 ft/lb (18ft/lb for the later model 8mm dia studs).

This seemingly impossible bit of magic was accomplished by taking into account the radically different co-efficients of thermal expansion between the magnesium crankcase (in which the head-stays are screwed), the cast iron barrels, and the cast aluminum heads (which are secured to the head-stays with nuts & washers). Here comes the tricky bit: As the cast aluminum heads heat up, they try to expand AWAY from the cooler cast-iron barrels, which have a much lower coefficient of thermal expansion. But the head-stays prevent any motion between the cylinder head and the barrels. This causes the expansion to appear IN the head-stays as TENSION and it is this tension that clamps the heads to the barrels with sufficient force to ensure a leak-free fit even when subjected to the pressure of combustion. And that's why it wasn't necessary to lap-in the barrels. (Notice the past tense? :-)

So why does `everybody' think it's a good idea? The main reason is because THEIR surfaces are NOT flat. Or they may be flat but of unequal depth. Here's why: Parts heat up when they are machined. Volkswagen machined both combustion chambers simultaneously on a superbly rigid machine, taking the thermal growth resulting from the machining operation into account. The end result is heads that are virtually identical, especially with regard to the flatness and depth of the sealing surface.

By comparison, a shade-tree mechanic opens up the combustion chambers ONE AT A TIME using a spindle-type tool bolted to the head and driven by a drill press or even a half-inch drill-motor. After cutting one chamber, the tool is dismounted and re-assembled over the second combustion chamber and the process is repeated. But after cutting the first chamber, unless you wait at least an hour for the head to cool down, the depth of the second chamber is going to vary by a significant amount due to thermal expansion. Which is only part of the problem...

As for the surface finish, the typical spindle-type cutter has only one cutting edge, which must be at least 3/4" long. When opening up the heads to accept 92mm barrels you're looking at a hole nearly four inches in diameter (~3.978", givertake... ideally, the cut should match the diameter of your set of barrels plus about half a thou per inch of diameter [remember, cast iron expands less than aluminum - - the heads are going to expand more than the barrels, hence the relatively tight fit, which guarantees better alignment during a cold start]). The recommended tool-speed for cast aluminum is about 100 surface-feet per minute, which is also about 1200 inches per minute. Since see equals pie dee that means our cutting tool should be rotating at NO MORE than about 100 rpm and in this case slower would be better.

This cutting speed is easy to achieve with a milling machine but impossible with the typical drill press which usually can't go below 300 rpm. (What's the lowest speed on yours? Many drill presses can't go below 500 and the typical half-inch drill-motor spins between 800 and 1200 rpm.)

Wanna know what happens when you try to open up a set of heads with the cutter spinning at 300 rpm? You get a lot of `ripple' - - the cut surface is NOT FLAT, it's sorta wavy. And the faster you go, the worse it gets.

And that's why `everybody' laps in their jugs... because they HAVE to.

On the other side of the coin are guys who use a real milling machine running at mebbe 50 rpm. The head is rigidly secured in a fixture that supports the over-hanging portion of the combustion chamber. The mass of the machine, which is bolted to the concrete floor of the shop, guarantees there is no vibration, whilst the slow spindle speed - - typically 50 to 80 rpm - - reduces the chance of any harmonics to below the level where they can effect the flatness of the finished surface. In addition, the cutting tool is either flooded with coolant or the head is allowed to cool between cuts so that the finished depths will be identical. End result: Perfect flat sealing surfaces of identical depth... that do not need any lapping.

Now back up about a thousand words and note the third reason Volkswagen didn't lap-in their jugs: The jugs had a nice wide sealing surface. Or at least, they did have, up until the 1500 engines :-) That's when VW bored out the stone-reliable 77mm jugs used on the 1200 and 1300 engines to come up with the 83mm jugs used on the 1500. And over-bored the 83's to 85.5mm for the 1600. Which tended to leak like a bitch no matter what you did.

The reason here was pretty simple: They had increased the bore of the cylinder at the same time they'd reduced its sealing surface. (Hang on to this fact. It plays a major role in most flying Volkswagens.)

Volkswagen knew they had a problem with leaky cylinders. As early as 1965 there were plans to replace the Type I engine's 69mm crank with one of 74mm, and go to an 88mm jug having thicker walls. This would have given them an 1800cc `Type I' engine with about the same cylinder sealing surface of the ultra-reliable 1300. Initially it was to be installed in the Type III's but the odds are overwhelming that it would have found its way into all other chassies. Then Heinz Nordhoff died (April of `68), bean-counters gained control of the company and virtually all R&D was abandoned in favor of short-term gains.

`Machine-in' 88's remained available from after-market sources and once their value was realized they were quickly displaced by `slip-in' 88's aimed directly at technologically naive VW owners who didn't know the difference between `slip-ins' and `machine-ins,' which was profound. Slip-in 88's are merely over-bored 85.5's, resulting in a sealing surface so narrow you were liable to cut yourself. Slip-in 88's quickly became known as the most unreliable jugs ever made for the VW. They are wildly popular of course.

Which brings us to 92mm jugs. These happen to be thick-walled `machine-in' 88 barrels bored out to 92mm. And yes, they leak like a bitch. 94mm jugs, which are based on even thicker barrels (and can only be used on later-model crankcases because of it) actually have MORE sealing surface than 92's.

The quality of after-market VW parts has always been spotty at best. Right out of the box, upon blueprinting a set of barrels - - one step of which is to check their sealing surfaces for flatness - - many jugs were simply unacceptable. In most cases the UPPER sealing surface could be made acceptable by lapping the barrel on a surface plate upon #600 wet & dry paper flooded with kerosene. Flatting the lower sealing surface was more difficult and usually required machining. But the fact professional engine builders often lapped the cylinder's upper sealing surface gave rise to the Conventional Wisdom that EVERY KIND of lapping was a good idea. As you can see from the above, it's not. But all those instant experts who say it is have never paid much attention to reality.

- - - - - - - - - - - - - - - - - - - - - - - - -

As for sealing rings, the proven alternative to gasket-less sealing surfaces is to use a gasket. (duh :-)

When Volkswagen introduced the Type IV engine with its 90mm jugs they finally bit the bullet and installed sealing rings. To keep the cost down the rings were STEEL, coated with pure aluminum. Alas, Volkswagen quickly learned that they could not be re-used, issuing a Service Bulletin to that effect. Properly annealed pure copper rings of the same thickness were an acceptable (but more expensive) substitute.

Copper sealing rings are now available for all commonly available cylinder diameters and are found in most professionally built four-stud VW engines. Their thickness effects the compression ratio and must be included in your calculations. Their use for this purpose isn't anything new, especially among air cooled engines, having been used on the Continental A40 (among others). Properly installed, especially with regard to annealing, copper sealing rings provide a reliable method of sealing a combustion chamber when the wall thickness of the barrel is less than optimum width.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

A converted VW is not a certified engine. Even if purchased ready to run, YOU are the Mechanic-in-Charge. For those who elect to assemble their own, without a good background in Volkswagen engines there's a tendency to do things without knowing why, other than `everybody' sez it's a good idea. Unfortunately, engineering is not subject to the democratic process; Robert's Rules of Order simply do not apply and the fact `everybody' does a particular thing is no guarantee it is the proper thing to do.

Even if you assemble the engine yourself most homebuilders will acquire only one engines-worth of experience in their lifetime. To ensure that lifetime is as long as it should be, you need to THINK FOR YOURSELF. It's important to know not only what others have done but WHY they have done so. If the best answer you can get is, `Because `everybody' does it that way,' I suggest you keep looking.

- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -

This post is meant to be a general answer to a number of specific questions. Rather than answer each in detail I've offered some background that should allow you to answer those questions yourself.


PS - I just threw in that part about why the sky is blue :-)

AV - Urethane Glue

(The following is NOT correct with regard to the 'manufacturer's instructions,' at least with regard to 'Gorilla Glue.' See the footnote as to why.)

When applied according to the manufacturer's instructions, all modern-day adhesives, including one-part urethanes, provide a bond that is stronger than the softwoods typically used for airplane construction.

Homebuilders often compare urethane unfavorably to epoxies but on examination most of those unfavorable comparisons are due to the builder's failure to follow the manufacturers recommendations with regard to clamping pressure.

Urethane glues expand as they cure. The expansion of urethane adhesives is by design, providing the mechanism by which the glue is forced into the micro-structure of the wood. The extent of that infusion -- and the ultimate strength of the bond -- depends upon a significant amount of clamping pressure to ensure the expansion forces the adhesive into the wood rather than simply pushing the joint apart.

To ensure the best bond (ie, typically >80% wood-shear failure in the standard FPL test [spec is 75%]) I've found urethanes need about the same amount of clamping pressure as adhesives containing water, such as Plastic Resin or resorcinol, (Ed.Note: ie, 70 to 100 psi) for about the same amount of time (ie, typically 24 hours). If urethane is applied in the same manner as epoxies, most of which do not require significant clamping pressure, you'll end up comparing apples to oranges.

Although urethanes have been used in Europe since the 1970's they are relatively new in the United States. For an American source of quantified data on the strength of urethane glues, see:

For homebuilders, especially those who have developed a sensitivity to epoxies, urethane glue is Good Stuff but some brands do better than others (as shown by the FPL tests, even though all exceeded miniminum shearing requirements). If you haven't tried urethane glues you can learn a lot from a few experiments.

On the practical side, once you've opened a container of urethane glue it is going to set-up, sooner or later, due to the air introduced into the container. (On this particular planet all air contains some amount of moisture, and since the water acts only as a catalyst, given enough time the moisture in a teaspoon of air will harden a whole pint of glue.)

The manufacturers tell you to squeeze all the air out of the container before capping, which is okay after the job is done but impractical during the course of a big lay-up (I've used it on built-up spars). The container is an excellent air pump, swallowing a fresh gulp of air each time you squeeze out a little glue.

Bottom line is that using urethane adhesive calls for some subtle changes in your gluing tactics and strategy, such as buying mostly small containers, reserving larger ones for bigger jobs. Also, store any container which has been opened upside down. You can't prevent some air from getting inside but when stored upside down the hardening will take place on the bottom of the glue, allowing you to get the maximum usable amount from any container. Another tactic is to store your upside-down container (simply stand it in a tin can) in a refrigerator between use.

As with any adhesive, be sure to spread it properly. Urethane needs only a very thin layer (on both surfaces). Allowing a few minutes of open-time before mating the parts ensures enough moisture for the cure.

I still use resorcinol for props, Plastic Resin when cost (or 'Certified Repair') is a factor and epoxies when I can't provide adequate clamping pressure, but I find myself using urethanes for about half my gluing chores.


PS -- Urethane is a near-perfect adhesive, more than willing to bond to you, your clothes and anything you happen to touch. Lacquer thinner will remove it if it hasn't cured. Accidental drops of glue onto concrete, metal or wood can be chipped or sanded away but if it gets onto fabric, it's there to stay... which lead to some interesting experiments using urethane glue and urethane varnish(!) on fiberglas... or even an old bed sheet. Very handy for making a quick fairing.

PPS -- Urethane glue and a pneumatic brad-driver has become my favorite method of assembling mock-ups, jigs and fixtures.


November 2006

After posting the above a fellow builder who's opinion I respect wrote to tell me that with 'Gorilla Glue' he'd followed the manufacturer's instructions religiously and the results were unacceptable even for household repair work. I've since completed a series of tests using 'Gorilla Glue' (brand name), a product of Denmark available from a variety of retailers including Harbor Freight.

I did 122 tests, all tolled. Test coupons were 1-1/2" wide to 3/4" wide. Woods used were Sitka spruce, Western Hemlock and Douglas Fir. Moisture content of the wood varied from 10% to 15%, as measured by a contact meter.

The instructions on the bottle of Gorilla Glue (and on their web site: ) say to apply glue to only one surface, as opposed to the Forest Products Laboratory who applied it to both. Gorilla Glue says the work need be clamped for only three to four hours to achieve 90% strength, and that the open-time could be up to twenty minutes.

I used two clamping pressures; approximately 45 psi and approximately 75 psi. 'Approximately' because I used spring clamps and did not calibrate them individually.

The tests were done over a period of about two weeks, determined by spare time and the temperature of the area where I did the work, which was when the temp was between 75 and 80 degrees Fahrenheit. Relative humidity varied between 35% and 80%.



All of the test samples in which glue was applied to only a single surface failed to meet the minimum shearing test, typically showing <10% wood-shear. Those clamped with more pressure did slightly better but the difference was not significant.

All of the test samples in which glue was applied to both surfaces passed the FPL shear-test. Indeed, the shear line in a couple of 1.5" DF coupons clamped with 75psi was well away from the glue line.

When clamped for only 3-1/2 hours (ie, splitting the difference in the recommended "3 to 4" hours, none of the coupons passed the shear test. Samples in which glue was applied to both surfaces and subjected to heavy clamping did best but even then the wood-shear was never more than about 25% (the accepted minimum for airplane joints is 75%).

The Open Working Time of 'up to 20 minutes' appears highly optimistic , wildly so with regard to how much glue is needed to make a good joint ("Min. Coverage 1/2oz. per square foot")


The tests were done for my own education. Although I tried to do things in an orderly fashion, keeping notes and taking a few photos, the tests should not be taken as definitive. Except by me :-)

My general impression is that when applied according to the FPL tests (ie, applied to both surfaces then clamped for 24 hours) Gorilla Glue is no worse than other urethane glues when used to join softwoods common to aircraft construction. That includes Excel One, PL and Elmer's Ultimate Glue. But neither is it any better. In my opinion, the instructions that come with the glue do not reflect reality and should be ignored in favor of the method described by the Forest Products Laboratory.


AV - VW Reliability - Compression leaks

A key factor in the long term reliability of the Volkswagen engine, stock or modified, is maintaining a perfect seal between the combustion chamber and the barrel of the cylinder. Although compression ratio and cylinder head temperature can effect this seal the most critical aspect the amount of sealing surface relative to the pressure it must contain.

Here are some typical cylinder wall widths:

Stock 85.5 = 0.165"
90.5 = 0.150" (requires machining the case & heads)
92 = 0.121" ditto
94 = 0.139" ditto

The so-called 'slip-in big-bore' cylinders are merely over-bored stock cylinders. 87mm slip-in's have a wall thickness of about 0.105"

In the Photo archive I've created an album titled JUGS in which you'll find a few photos illustrating the relative diameters & wall thickness of the cylinders above, with the exception of the 'slip-in' type. (*)

(Ed. Note: The photos were posted to the CX4 Group. If I can find the originals, I'll attach them to this file.)

One reason for posting this information was to back-up my comments in recent posts regarding the suitability of various parts. You'll notice that the 94mm jugs actually have a thicker wall than the more common 92's.

As a general rule, compression leaks (ie, failure of the seal between the combustion chamber and the cylinder wall) are not a problem when the engine is properly assembled & maintained, the compression ratio is 7.5:1 or less and the cylinder wall is at least 0.120" in width.

You'll note that 92mm jugs are right on the line. When 92's are used on engines running higher CR's the joint between the barrel and the cylinder heads is usually fitted with a copper gasket, carefully annealed at the time of assembly.


(*) The jugs in the photos were what happened to be in the shop at the moment. The stock jug and the 92 came out of the junk box, the 90.5 and the 94 from engines under construction. I've no examples of the slip-in types because I don't use them.

AV - VW Reliability


When the Volkswagen first arrived on American shores it was a mechanical curiosity, having a suspension system far in advance of anything we'd ever seen, thanks to Ferdinand Porsche, and a cute little engine designed by Xavier Reimspiess that was the cutting edge of 1930's technology.

With a VW dealer in nearly every town and a superbly planned propaganda comapaign run by the ad agency of Doyle, Dane & Bernbach, it wasn't long before the bug enjoyed an enviable reputation for reliability. That wasn't true but in the modern world perception is reality.

The truth is, the Volkswagen is a high-maintenance vehicle, as were all vehicles from the 1930's. It required approximately ten times the amount of skilled periodic maintenance as a modern vehicle. Which was okay when the dealers had a waiting list of factory-trained German mechanics willing to come to America and work for seventy cents an hour.

To the owner of a bug or bus all of this was invisible. They merely took their bug into the dealer and got it back four to six hours later, ready to deliver another three thousand miles of trouble-free operation. They were unaware of the 128 items on the check-off list, nor the 12 different lubricants, nor the average seven years of factory-trained experience in each of the three mechanics and one inspector who performed their periodic maintenance, which the owner thought was a simple oil change.

(Ever own a VW? Back then there was a block in the lower-left corner of the job-ticket labled 'OTHER MAINTENANCE (AS REQUIRED).' Often times when you'd pick up your bug or bus there would be a part number in the box along with a charge; never very much. But the part number might be for a rebuilt head, fuel pump, carburetor or other major component, installed at the discretion of the service-manager. Many VW owners liked to brag their vehicle had never needed any repair when in fact it had been virtually rebuilt, one component at a time. )

Fortunately, the engine was such a simple, robust design that it wasn't too difficult to retro-fit modern-day technology, which is what professional mechanics did in order to develop a powerplant that could not only survive a thousand-mile race over unpaved desert trails but do so at speeds as high as 120mph.

The Myth continued, of course. Indeed, it has even grown through the years: Volkswagens were stone-reliable. Such claims are hilarious bullshit to any professional mechanic but modern-day America has never had a very firm grip on reality.

The HVX mods are a compilation of some of the durability-related retro-fits. Nothing very exotic; every modern engine incorporates the same features. The tricky bit -- and the reason for all the drawings -- is showing how those modern-day features can be retro-fitted to the early VW air cooled engines, which even Volkswagen did, starting with the aluminum-cased 1700. Although occasionally mentioned in the literature most of these mods have never appeared in any of the VW-specific magazines because they are hard-ball engineering, things that must be built-in rather than bolted on. Most shade-tree types have never even heard of them and if they have, pass them off as being unnecessary. Their dune buggy runs just fine, until it doesn't. Even then, it's not too hard to fix. So long as it is loud enough and has enough chrome, they're happy; long-term durability isn't even in the equation.

Unfortunately, a lot of those shade-tree types put a fan on the pulley hub and call it an aircraft engine. And that's when the fun begins. It doesn't do any good to point out that the piece of shit they're flying behind wouldn't make a pimple on a real engine's ass since every successful flight says otherwise. I know I wasn't impressed with the opinion of professional mechanics back when I was in my teens. (And my dad was a card-carrying A&E.) After all, I built an engine and the thing actually flew! What could be better proof than that? An' besides, I'd done everything all the 'experts' said I should do.

It took two lost props and six off-field landings to convince me most VW experts of that era didn't know their ass from their elbow. Indeed, in researching the literature I discovered that none of those experts had actually converted a VW engine for flight and only one had ever flown behind one! (I'm talking the late 1950's here, guys.)

That was the beginning of a long and often difficult education. Which is still going on. But it has turned into a largely personal journey. According to the current crop of experts I put the prop on the wrong end of the crankshaft and do all manner of other things deemed unnecessary, according to the conventional wisdom of dune buggies.

So be it. After leading the horse to water the rest of the job is up to the horse.


AV - Valve Train Geometry

* Date: Fri, 23 Apr 2004 19:23:39

Valve Train Geometry

The basic principle is quite simple: The rocker-arm, which serves as a lever, must act thru an arc. To convert the maximum amount of the arc-motion of the rocker into the maximum amount of linear-motion at the valve, the mid-point of the linear travel must fall exactly upon the tangent of the arc.

You will find the above endlessly repeated in various ways in the hot-rod magazines and that would be jus’ swell... if we could apply the procedure to the Volkswagen. Or Corvair. Or Lycoming, et al. But we can’t, unless we are looking at a bone-stock engine. That’s because the method outlined above addresses only the output side of the geometry equation. The input is not addressed because it doesn’t need to be, so long as we are talking mono-bloc engines, in which the distance & angle between the cam and the axis of the rocker-arm is fixed, or virtually so.

Unfortunately, most flying Volkswagens are big-bore strokers and a properly built stroker is wider than the stock engine. Making the engine wider not only increases the distance between the cam and the rocker-arm axis, it changes the angle between them. To achieve optimum valve-train geometry we must address two arc/lever systems, one for the input of motion to the rocker-arm as well as the output of motion from the rocker-arm to the valve. When dealing with the input side of the equation the same rule for maximum transfer applies, in that the half-point of the push-rod's linear travel must fall exactly upon the tangent of the rocker-arm's arc.

The tricky bit is the fact any change to one side of the equation will be reflected in the other, since the points of maximum transfer of motion must precisely coincide.. Most don't. Indeed, unless you're looking at a professionally built engine it isn't uncommon to see VW valve trains so mal-adjusted as to give away 25% of their potential lift.

The Conventional Wisdom fix to such geometrical disasters is to install larger valves and a cam having more lift. Of course, the larger valves will require heavier springs and the combination of higher lift and greater valve spring compression must be paid for with energy and wear. However, having arrived at this point because the person building the engine doesn’t understand the basic problem, there’s no guarantee they’ll get it right the second time around.

Indeed, across the range of rpm most suitable for slinging a propeller even the largest big-bore stroker has a very modest flow-rate, easily satisfied with single-port heads fitted with stock valves actuated by the stock cam. Assuming of course that the valve train’s geometry is properly set.

In setting-up the valve train's geometry the variables are the length of the push-rod, the length of the valve stem and the height of the pivot-point. Rocker-arm ratio (ie, the length of the input arm to the output arm) has relatively little effect since the length of the output arm remains unchanged and the point of tangency for the input of even the wildest ratio-rockers will still fall within the available limit of vertical travel for the push-rod (ie, in traversing the chord of the arc there is always some component of movement perpendicular to that axis).

Determination of proper valve train geometry begins with the basic blueprinting of the engine, when you measure the actual lift of your particular cam. This data is used in setting up the rocker shaft height relative to the valve stem height and may be done in a simple jig before the heads are installed on the engine.

Another necessary tool is a stock adjusting screw, modified by accurately grinding it to a fine point. The tops of the valve stems are coated with soot, lipstick or Dykem and an optical comparitor is used to determine where the point falls upon the face of the valve stem, the position of which is used to make any required adjustments. The adjustment procedure and a few drawings may be found in the HVX files covering engine assembly but will be included on this blog if time permits.

Push-rod length is best determined for each valve during trial assembly.


Optimizing your valve train's geometry will improve the engine's volumetric efficiency which translates into more torque at low rpm and reduced fuel consumption for the same power output right across the band. Proper valve train geometry also guarantees the system is absorbing the smallest amount of energy, which translates into reduced wear and better output.

Getting the geometry correct isn't especially difficult but it takes a bit of time, calls for precision instruments such as dial indicators, and a simple jig that allows convenient manipulation of the valves & rocker arm.

Most VW 'experts' lump valve train geometry with dynamic balancing and a host of other 'unimportant' details. Rather than address the basic issue they tend to shovel money at the problem in the form of after-market heads having valves the size of dinner plates, hot-rod cams with Himalayan lifts and valve springs more suitable for a punch-press than a light aircraft engine. The fact their engines run and the plane flies is taken as proof that proper valve train geometry is just another of those 'unimportant' details :-)

It's up to you. You're the Mechanic in Charge.