Thursday, May 31, 2007
That's me. Taken at the keyboard by my $5 web-cam as I search GIMP-2 for the Beautify key.
Actually, I'm probably answering mail, a never-ending chore. But someone else noticed the picture of me leaning on the Arctic Circle was taken in 1996 and wondered if I was still alive :-)
Tuesday, May 29, 2007
The air-cooled Volkswagen engine doesn’t have a very good lubrication system. (Which isn’t surprising, seeing as the design dates back to Xavier Reimspiess’ original 1931 ‘boxer’ engine.) Its inadequacies became evident in the late 1950's when me and a few other fools started hot-rodding the things.
The Ford Motor Company had recently published a study of the effects of oil filtration systems on engine wear and the results were so impressive that by the time you could say ‘Jack Robinson’ we’d retro-fitted our bugs with oil filters. But the main problem wasn't dirty oil but not enough oil, especially at high rpm. The inadequate amount of lubrication reaching the heads resulted in excessive friction, leading to high temps and failed valve-train components, which put you out of the race.
Auditing the engine’s lubrication system we found that all of the oil for both heads came through a single 5mm drilling. In theory, a hole that size should have provided more than enough oil. And it did, but only for the left-hand head. And then, only at low rpm. Which was fine for a stock 1200cc 36hp engine, but we’d already bored & stroked that puppy to nearly 1700cc and were running them at over 5000 rpm. But not for very long.
Volkswagen was aware of the valve-train lubrication problems and added a drilled oil channel to the rocker arms and a larger main oil gallery on later engines but the basic problem was that not enough oil was reaching the heads, especially the one on the right-hand side of the engine.
(NOTE: VW orientation is always relative to the driver. The right-hand head is the 1 & 2 cylinder bank, left-hand is the 4 & 3 bank.)
To get oil to the right-hand head, VW cuts a square groove in the bearing saddle for the #2 cam bearing. All of the oil to the right-hand side of the engine gets there through that channel. (And still does, if you haven’t modified the crankcase.)
To make matters worse, the oil to the heads gets there via the cam-followers... but only when that particular valve is actuated. In effect, the VW cam-followers act as a valve, shutting off the oil to the heads for approximately 92% of the time.
To get an engine that could run flat-out for 24 hours we had to get more oil to the heads. To do that we tried opening up the oil channel in the cam-bearing web. That worked but we still weren’t getting enough juice. So we modified the cam-followers to allow oil to reach the heads 100% of the time instead of only when the valve was actuated. Major improvement, but we were still seeing galling on the right-hand rocker-arm shaft, plus an occasional hair-pin fracture. What we needed was still more oil... especially to the right-hand side of the engine.
To do that, on the right-hand case half we extended the oil gallery for the cam-followers to intersect with a new oil channel we drilled into the #3 cam bearing web. Big, BIG change. And lower head-temps, too.
With that as a clue, we grooved the rocker-arm shafts to match the oil channels drilled into the rockers. Fitted with the Ford/Subie type swivel-foot adjusters, which have a matching oil channel, we could now provide the heads with approximately eight times as much oil as before. Heads ran much cooler... which meant the oil coming from the heads was hotter, so we had to come up with a better oil-cooling scheme. Which we did, moving the oil cooler outside of the blower housing.
All of which is pretty old news to anyone hanging out at the finish line. But a total blank to just about everyone else. The magazines were only interested in the mods when there was something to sell, such as an oil filter bracket or an oil cooler core. All of the fiddley bits that made the system work, such as drilling the new oil channels or grooving the rocker shafts, were seen as just more of those ‘unimportant’ details professional engine-builders are always messing with. (Most magazine 'technical' articles are nothing more than infomercials, intended to sell whatever product is being touted. )
A key point here is that a high-output engine needs all of the modifications described above: 100% filtered oil, increased oil volume to the right-hand side of the crankcase, 100% oil flow through the lifters, grooved rockers & rocker-arm shafts, and the Ford/Subie-type swivel-foot adjusters, which act as spray-bars. Some guys would modify the rockers and say they didn’t see any improvement. Others would modify the lifters and say the same. But not one in a thousand incorporated all of the modifications. And still don’t. But it’s interesting to note that Volkswagen included all of the modifications to the Type IV engine. In fact, you can find them in every modern-day engine. Which is just another of those ‘unimportant’ details.
Back in the Day, send a VW crankcase to Jack Riddle’s shop (aka Riddle Machine Company or ‘RIMCO’) for an align-bore or other machine work and it would come back with the some of the gallery plugs pulled & threaded for pipe-plugs, which were included with the returned case.
Why? According to Jack, machining caused swarf to get into the oil galleries and it was impossible to clean them out unless you pulled the plugs.
Which sounds perfectly logical, especially to any experienced mechanic or automotive machinist because seeing oil galleries sealed with pipe plugs is a familiar sight to anyone who has worked on aircraft engines or big V8's. And removing those plugs is a normal procedure in order to clean the oil galleries during overhaul. So pulling the plugs becomes a standard part of building a high-performance engine based on VW after-market components. Your engines run sweet, your customers are happy and those mysterious bearing failures become a thing of the past.
Not so with the shade-tree types, for whom pulling the plugs is another of those ‘unimportant’ details. Lotsa folks still don’t pull the plugs, especially on a new crankcase. Their logic goes something like this: New crankcase has never had any oil in the galleries so there’s nothing for the swarf to cling to; that a blast of compressed air is enough to clean everything up. And about here it might be a good idea to go read...
Starting about 1997 professional engine-builders here in southern California began seeing Brazilian crankcases in which the oil gallery for the #4 main bearing was blocked by the factory-installed plug. Which wasn’t a problem because we routinely pulled the plugs... which was how the problem came to light. Nor was it much of a problem to the dune-buggy crowd, folks who routinely did not pull the plugs. After all, the #4 main wasn’t a real main bearing - - it was added when VW found the asymmetric load of the blower caused the pulley to oval-out the nose of the crankcase. In most cases the blockage wasn’t 100% and the #4 usually got enough oil for passenger-car service. But complaints were heard now & then from the dune-buggy set who came up with an Idiot-Fix: running a 7/32" drill down the oil gallery for #4. Sometimes it even worked :-)
But it was a problem for flying Volkswagens, especially those who put the prop on the wrong end of the engine, which back then was virtually everyone. Here’s what Steve Bennett has to say about the problem...
Read both of the above articles and you’ll note significant differences in our methods; Steve drills-out the offending plug whereas I pull it out - - along with three others of that size. Steve threads the bore to accept a 1/8-NPT socket-head pipe-plug whereas I used whatever is available, my preference being 1/16-NPT pipe plugs.
The VW engine uses four 5mm plugs; one for the #4 bearing, two for the oil galleries feeding the lifters and one for the oil gallery going to the reservoir(s) behind #2 cam bearing shells. I’ve never found a long plug anywhere except on the #4 gallery but other engine builders have said they’ve seen them installed at the other three locations, sometimes with catastrophic results. A long plug won’t cause a problem with the lifter galleries but even a partial blockage of oil to the reservoirs behind the #2 cam bearing shells guarantees the engine will have a short, unhappy life, since that single 5mm gallery is how oil gets to all eight of the lifters and through them, to the heads.
Sunday, May 27, 2007
I was dismayed to learn that some folks having no Volkswagen engine experience have been buying components, mostly from ads in car magazines, expecting to simply bolt things together, hang a propeller on one end, an airframe on the other and go flying. In at least one case a fellow thought he could buy a bunch of parts, haul them up to my shop, wave a lot of money at me and drive off with an assembled engine.
It simply doesn’t work that way. Here’s why:
When you buy a new crankcase what you’re actually purchasing is a ‘universal REPLACEMENT crankcase.’ These were originally provided only to VW dealers, where they were used for the repair of an existing engine whose crankcase has cracked due to age-hardening or collision damage. As received, your new crankcase can’t be used to build an engine from scratch because it is not complete. What’s missing are the things that make the crankcase specific to the vehicle Type and the model year. There is no sump-plate or oil screen, no studs for the oil pump nor fuel pump, no head stays (ie, studs) and no nuts & washers for the studs that are there. You’re expected to remove all that stuff from the original engine, the one with a cracked crankcase.
You can buy all the missing bits either in kits or per-each but if you’re building a flying Volkswagen you’ll be pissing away a lot of money because the parts in the kits are specific to automobiles. For example, in the standard ‘case kit’ (about $20) you get the oil control pistons, springs and slotted cap-screws. But for a flying Volkswagen you need a cap-screw you can safety-wire and an oil pressure relief spring that pops-off at 45psi instead of 27. You’ll also get a mild steel cover-plate for the big hole in the lower right corner of the sump where the dip-stick attaches on the Type III vehicles. Which goes straight into the trash because it weighs three ounces and one made of aluminum weighs barely half an ounce. Ditto for most of the studs since you’ll be using drilled-head bolts which you’ll have to procure and drill yourself. Bottom line is that it generally costs less to ignore the kit and buy the parts onsey-twosey.
A head-stud kit consists of the sixteen stays (in three lengths) that secure the heads to the crankcase, along with the required washers & nuts. Unfortunately, oft times one of the studs or nuts won’t have any threads and you end up having to beat the bushes for a replacement, since any effort to have the retailer replace the kit is like pissing into the wind. Indeed, you’ll often receive a head-stay kit clearly marked as being for a single-port engine that turns out to have the four short stays for a dual-port.
Even when you receive the proper head-stud kit, the things are bare metal. Before you can use them on any properly built engine they need to be plated, painted or coated - - and done well enough to withstand twenty years of exposure. (I usta have all my head-stays cadmium plated but when the tree-huggers forced the local plating shop out of business I went to two-part epoxy paint. Most recently I’ve been using powder coating.)
Finally, you will need nuts and washers and bolts to fasten the case studs and parting-line. Here again, there are kits available but most are the shoddiest stuff imaginable and price is no guarantee of quality. The nuts and washers may have a wash of zinc plating, good for at least a week’s exposure to the weather. Or they may not. And you can toss the ‘exhaust nuts.’ They are copper plated steel. (The good stuff is bronze.) Before you can use any of this crap on an engine you must provide it with some form of corrosion protection. If you don’t, not only with the nuts rust to the studs, you’ll see galvanic corrosion between the washers and the crankcase that will eventually cause the fastener to loosen.
But the biggest problem is that your new crankcase is for a stock ‘1600' engine. Flying Volkswagens tend to be larger, which means the crankcase must be machined to accept bigger jugs and, in some cases, a crankshaft having a longer throw. Plus it needs a critical bit of welding .
In the stock crankcase the spigot bore for the #3 cylinder is sort of hanging out in space. Even on the stock engines this area is prone to cracking. Indeed, a ‘cracked #3' is one of the most common reasons for the existence of Universal Replacement Crankcases. Machine the case to accept bigger jugs and you’ve made the situation worse by an order of magnitude. It’s no longer a question of IF #3 will crack but simply ‘when.’ To deal with that you preheat the new crankcase case and weld in a reenforcing plate using TIG.
A 94mm barrel will hit the threaded steel inserts that are standard on all new crankcases. Not only must you open up the spigot-bores to accept the larger barrels, you must deck the case to provide a uniform sealing surface for the bigger barrels. Since decking the case moves the heads closer to the centerline of the engine, it upsets both your valve-train geometry and your compression ratio. Because of the normal variation in the size of after-market parts, resetting both the CR and geometry is best done by inspection, meaning you’ll need to devote a couple of pre-assemblies to each of those procedures.
Opening up the interior of the crankcase to accept a bigger crankshaft is called clearancing and while most shops use a humongous cutter to do the job at one go it leaves a lot of feather edges that are guaranteed to precipitate cracks, so you have to dress the edges smooth by hand, using a flapper wheel, files and #600 grit sand paper.
If you’re doing the HVX mods you need to pull the plug from the oil gallery on the right-hand side of the crankcase, extend the existing oil gallery and connect it to the #3 cam bearing saddle. This is when you also open up the oil channel behind the #2 & #3 cam bearings (which is how all of the oil gets to that side of the engine.) If you’re installing anything in the distributor hole other than a plug you must also do the grub screw mod.
If you’re going to install the oil temp sensor in the location used by Volkswagen you need to pull the 3/4" plug to the lower-right of the oil pump and thread it to accept a 1/2"-NPT x 1/8-NPT adapter. The oil temp sensor then threads into the adapter.
If you’re running a full-flow oil filtration system (and you should) you tap the main oil gallery to accept a 3/8-NPT to AN8 (flare) adapter. Some engine-builders also thread the oil gallery leading from the oil pump to accept a 1/4-NPT pipe-plug.
If you’re going to run an external oil cooler you thread the oil cooler ports to accept pipe plugs.
And having done all that, it’s time to clean the crankcase.
No, you can’t just blow it out with compressed air. There are a couple of blind corners in the oil galleries that act as swarf-traps. To clean them out you must pull all of the soft aluminum plugs (except the two small ones associated with the oil pressure valve... you can check for contamination by using a mirror down the bore of the valve) . After pulling the plugs you tap the oil galleries to accept socket-head aluminum pipe plugs of the appropriate size: 1/16, 1/8, 1/4 and 3/8. Now you can scrub the bores and visually inspect them.
As with the ‘case kit’ You can buy a ‘plug kit’ but they don’t include the four 1/16-NPT’s you’ll need for the 5mm plugs. Instead, they’ll sell you eight 1/8-NPT’s and shug; that’s what they use in dune buggies.
And finally, once all the machining is done and the case is cleaned and sealed up, if it’s a magnesium case you paint it. Because if you don’t, it’s going to corrode. Use regular flat-black Rustoleum. If you can’t get flat-black use gloss-black cut with a little naphtha. (If it’s an aluminum case you apply Tech-Line Coatings ‘TLTD’ thermal dispersant then bake the thing in an oven not used for food preparation.)
So what’s all that going to cost you? Dollar wise, it depends on where you’re located and how much of the work you can do yourself but at a guess, expect to pay between $150 and $400 over and above the cost of the crankcase. Here in southern California there are several good shops that do nothing but high-performance VW engines. In other parts of the country I know of guys who have paid twice as much and gotten less for their money. If you’re tooled-up to do the drilling & tapping you can cut the cost by as much as $200.
And that’s just for the crankcase. The cylinder heads, crankshaft, camshaft, pistons & cylinders, push-rods and push-rod tubes also require a significant amount of preparation before they’re ready to be used.
So hold your horses. You are not building a dune-buggy engine. You’re building an aircraft powerplant meant to deliver at least twenty years of reliable service. In future posts I’ll show you how I do it - - and why. It’s up to you as the Mechanic in Charge of your engine to decide if you want to follow suit.
Saturday, May 26, 2007
When fabricating a carb-heat box one of the trickier bits is the axle on which the flap or air-vane pivots. Most start with a round rod then file or mill it flat on one side to accept the flap. The axle now assumes a D-shape and the flap is attached to the flat part of the D with rivets or screws. A few use steel and weld flap to the shaft, trading the heavier weight for the greater ease of fabrication.
To actuate carb-heat we must cause the air-vane to move. An interesting point, often overlooked by the novice, is that the flap or air-vane need not swing through a 90 degree arc to let hot air into the box. If the hot air source is fed into the side of the box perpendicular to the normal air flow, any angle greater than 45 degrees is sufficient for the carb to draw-in 100% heated air.
The usual method of causing the air-vane to move is to attach a lever to one end of the shaft to which the air-vane is attached. The tricky bit is to do so in a manner secure enough to risk your life upon. The most popular method appears to be to continue the D-shape to the end of the rod and to make a matching D-shaped hole in your lever. At his point the novice is liable to give up because producing an accurate D-shaped hole with hand-tools can be rather difficult. Some convert the D-shape of the axle to a square shape, since it’s fairly easy to square a round hole. Of course, if they started with quarter-inch rod they end up with a rather small square - - just .177" on a side; about 11/64". With a trunion that small, if the rod is aluminum or brass you can’t put much stress on the lever, which is why a lot of guys go to a steel rod - - and weld the flapper to it. Or end up using a larger, heavier rod.
I’ve a hunch you could fill a fair sized book with the variations on this theme. And I know a few guys who have done exactly that, bouncing back and forth between the need for an air-vane that pivots smoothly, a lever that provides fail-safe actuation and the need for the thing to be fabricated with basic hand tools. The results are often blindingly complex, with stepped shafts, ball-bearings, adjustable linkages with ball-end fittings and... Most homebuilders fail to notice that in creating such marvels they’ve given up flying for designing.
The method I use is so crude it makes real engineers cringe, but it’s reliable, inexpensive, low in weight and easy to fabricate. While the 'experts' are peering down several yards of nose, talking about 'trophy points' and 'state of the art,' you're actually flying.
I make the axle out of square rod rather than round. The basic idea here is that a square rod will rotate smoothly in a round hole having a diameter equal to the diagonal dimension of the rod. This method works because the rod only has to rotate about a quarter-turn to do its job; it doesn’t have to rotate fast nor often. For a rod that is one-quarter inch on a side, a size ‘T’ drill makes a hole that is a pretty good fit. To make the square hole in the lever you simply begin with a hole 1/4" in diameter and use a rat-tailed file to give it four corners. Since you have the square rod as a gauge it’s pretty hard to go wrong. A washer and cotter key then serves to retain the lever.
There's a bit more to the actual fabrication of course, details specific to the carburetor, method of mounting the engine to a particular airframe and so forth. The purpose of this post is to make you aware of the principle - - that a square shaft can rotate smoothly in a round hole.
Need I mention that this method works equally well for cabin heat or controlling the flow of air through an oil cooler? Probably not :-)
PS -- This method isn't new. I first saw it in the 1940's, on a rather ugly flintlock rifle manufactured in the back-woods of Pennsylvania in the 18th century. The fact the rifle was still shooting 190 years later served as a nice lesson in the practical value of low-tech simplicity.
Sunday, May 20, 2007
It’s difficult to get excited about push-rod tubes, the eight little pipes that angle down from the heads to the crankcase. Nestled inside the lower plenum chambers, the push-rod tubes are out of sight and usually out of mind; properly installed, they’ll last the life of the engine. It came as quite a shock to learn they’re a part of the engine’s cooling system and a fairly sophisticated part too, for the role they play in regulating the engine’s oil temperature.
The first clue in understanding the Secret Life of Push-rod Tubes was learning that chromed push-rod tubes were part of the ‘high-latitude package,’ a kit of parts that re-configured the VW engine for operation in Arctic conditions. That’s alien stuff for most American VW mechanics, not because some parts of the States don’t have cold weather but because Volkswagen of America, the sales agent for VW products, didn’t bother to import such kits. Indeed, while Volkswagen produced more one-hundred thirty different types of vehicles, from track-laying snow-cats to farm tractors, VOA imported only half a dozen or so.
The cold weather kit consisted of chromed valve covers, push-rod tubes and sump-plate, three shutters that fit over the air-inlet on the blower housing, a different thermostat, split main bearings and an insulated blanket that wrapped around the engine’s sump. The purpose of the kit was to keep the heat in. The chromed parts were in direct contact with the engines oil and the shiny chrome surface reduced the thermal transparency of those parts by 90%, using the same principle found in shiny tea pots, chrome-plated percolators and the insides of a thermos-flask.
This puzzles a lot of folks who never took physics so let me explain that a shiny surface reflects heat. When you put chromed or polished valve covers on your engine, the heat in the oil sees that shiny surface from the inside and is bounced back into the engine. If you’ve got one of those infra-red thermometers you can prove this for yourself by simply putting a chromed valve cover on one bank and a stock valve cover on the other. Take it for a short run to warm things up then use the IR thermometer to measure the temperature of the valve covers. The shiny one will be as much as thirty degrees cooler than the black one.
Which is bad. Unless you live in Yellowknife :-) Because that’s thirty degrees of heat that is not being radiated out of the engine.
If you live in southern California and know your onions when it comes to VW engines, you’ll be running stock valve covers painted flat-black, which radiates even more heat than the stock gloss-black covers. Seeing shiny valve covers on a VW engine tells you all you need to know about the technological expertise of it’s owner :-)
As for how the push-rod tubes aid in the regulation of the engine’s temperature you need another shot of physics – the part where it tells you heat always runs ‘down-hill’ according to temperature. That means if you put something cool next to something hot, the cool thing will get hotter while the hot thing will get cooler. In designing the Volkswagen engine they took advantage of that fact by enclosing the push-rod tubes in the lower plenum chambers where they are bathed in air coming off the cylinders and heads. When the engine is first started the tubes are colder than the air coming off the engine, so the oil inside the tubes warms up. Once the engine reaches its normal operating temperature, which you want to happen as soon as possible, the oil inside the tubes will be hotter than the air coming off the engine so the oil will tend to cool down .
Like all car engines, the VW was designed for variable output, unlike an airplane engine that is designed for virtually a single level of output. Like most car engines the VW spends 98% of its service life producing an output approximately equal to a quarter of its maximum peak power. The only time it produces more is when accelerating or climbing hills and of the two, acceleration accounts for the majority. Not flat-out acceleration, as in seeing how fast you can go from a standing stop to your top speed, but relatively small accelerations, such as working your way through the gears, changing lanes and so forth. Those small bursts of power produce small bursts of excess waste-heat, most of which goes into the oil, which serves as a kind of temporary heat-sink.
Liquid-cooled engines are capable of dealing with significant excesses in the waste-heat department but air-cooled engines are not. They are designed for a relatively narrow operating window, which makes their use as a vehicle powerplant something of a challenge. Volkswagen addressed the problem by fitting an oil pump having more than eight times the output required for lubrication alone. The excess pumping capacity, in conjunction with an efficient heat exchanger and belt-driven blower allows the lubricant to serve as a coolant, which works quite well up to the engines maximum sustainable output of approximately 40bhp under Standard Day conditions. This system also works fairly well for occasional large increases in power of short duration, such as merging with traffic.
The push-rod tubes come into play for all of those small variations in output, such as working your way through the gears or when changing lanes.
The stock VW push-rod tube is the best option, being lighter and less expensive than the alternatives. Off-roaders or anyone who spends any time in the desert always carry a couple of spring-loaded push-rod tubes in case a stone get past the skid-plate but for flying Volkswagens the stock tubes do perfectly well. Assuming they are properly installed and that the engine is not allowed to over-heat.
Proper installation begins with a thin coat of flat-black paint to protect the bare steel and prevent it from rusting, since even a modest layer of rust serves as a remarkably good insulator, as any weldor can tell you. Flat-black paint because it has the best heat-transfer characteristic of any color. And a thin coat because all paints serve as insulators to some degree. But on the whole, a painted push-rod tube is a thousand times better than a rusty one.
Another aspect of proper installation has to do with the length of the push-rod tube, which is 7-9/16" (191mm to 192mm) for a stock engine. If your heads have been fly-cut of if your crankcase has been decked, you must subtract the depth of the cut from the 7.5625". In the same vein, anything that moves the heads farther from the center-line, such as barrel shims or head gaskets, must be added to the length.
Adjusting the length is done by simply compressing the push-rod tube to make it shorter and extending it to make it longer. Unfortunately, if you do either of these things incorrectly you can harm the engine. To shorten a tube you drill a 7/8" hole in two pieces of wood at least 3/4" thick. 27/32" would be a better fit but most of you probably don’t have a set of wood-bits in 1/32" increments. (For the rest of the world use 22mm). Fit the drilled blocks over each end of the tube and compress them until the distance between the blocks is reduced to that required. You may rig a simple depth gauge if you wish. (A piece of welding rod works fine.)
The idea here is to compress the accordion-pleated ‘bellows’ portion of the push-rod without damaging the portion of the tube that projects beyond the bellows. The hazard here is two-fold: The projection portion of the tube is part of the sealing surface. Any bend or wrinkle usually results in a drip. Secondly, a bend in this area is liable to contact the push-rod itself.
But the most commonly required adjustment to a push-rod tube is to make it longer. Most of the manuals tell you to simply bend the bellows-section back & forth whilst pulling on the tube. This usually results in one accordion-fold being over-extended... which then gets squeezed back together when the heads are installed and quite often leads to the formation of a crack where the welded seam dips down into the fold of the bellows.
Why? Partly because of being over-extended but mostly because of what the bellows is designed to do, which is to maintain an oil-tight seal as the engine heats & cools. It is the thermal stress of those heat cycles that provides the energy for the over-extended portion to crack.
A better way to do it is to use a tubing cutter (!!) Set the blade so that it touches only the sides of the fold, press the push-rod tube into the cutter and simply swing it around the tube. It will force that fold of the bellows apart by some small amount. Repeat as needed to achieve the required length, dividing your work between both ends of the tube.
This method works fine for one engine but it’s rather slow. If you’re building more than one engine at a time you’ll probably make yourself a tool similar to the one shown in the photos, which is nothing more than an old hacksaw blade or piece of steel strapping, stoned to a smoothly rounded edge and epoxied into a hardwood block. With the push-rod on a firm surface, press the tool against a fold and use the motion of the tool to roll the tube across the surface. The result is straighter tubes and more uniform spacing than is possible with the tubing cutter.
A key factor here is that the push-rod tube must never be longer than necessary. Some mechanics think they can prevent leaky push-rod tubes by starting out the things as much as an inch longer than required, then using the heads as a vise to achieve a crush-fit. The push-rod tubes are supposed to be straight. Crushing them into place almost always results in kinky looking tubes due to asymmetrical compression at one of the bellows sections.
Some big-bore strokers are as much as an inch wider than stock, meaning each push-rod must be extended by half an inch. The tool shown produces a consistent sixty-thou extension per pass. An extension of half an inch would take nine passes per tube, four on one end, five on the other. Since the tool goes into the valleys between the folds, and since there are twelve valleys on each bellows section, you would distribute the widening uniformly so as to equalize the stress.
The proper preparation and installation of your push-rod tubes is another example of the ‘unimportant details’ that spell the difference between a properly assembled engine and the other kind.
Mr. Hoover, Thanks for the excellent info. Should there be an
interference fit of the tube to the head/case?
No. It is meant to be a compliant fit in order to
accommodate movement between heads and the crankcase.
What type of sealant
does VW recommend if any?
None. Indeed, the use of any form of sealant on
the molded push-rod tube gaskets tends to defeat
their function, resulting in leaks.
Have you found any sealants better than what
See the preceding comment.
Have you ever experimented with increasing
the surface area of the tubes by
welding/brazing on some fins?
Could rifle blueing
be used to advantage to corrosion proof and
increase the cooling rather
Possibly. But I think a more pragmatic approach
would be to apply Tech-Line Coatings 'TLTD' thermal
dispersant according to the manufacturer's
instructions (ie, media blasting, application then baking).
Thanks for considering these questions.Bob
Wednesday, May 16, 2007
Chugger is an exercise in low-cost construction, a sturdy little bird built mostly from Box Store lumber, powered by a converted Volkswagen engine spinning a home-made propeller. The landing gear uses go-cart wheels with tires from a baggage cart.
Chugger would normally be called a fabric-covered wooden airplane but the truth is, wooden airplanes use steel fittings at every stress concentration point, such as where the landing gear attaches to the fuselage or the wings to the pylon over the pilot's head. A major cost-savings was to be the use of inexpensive mild steel rather than chrome-moly for fittings, as Erik Clutton has done with his F.R.E.D. This was the standard method in the early days of aviation, prior to the advent of malleable high-strength steels. The only down-side is that when you make a fitting from mild steel it needs to be as much as three times thicker than when made of alloy steel. And that means three times heavier as well, since all steel weighs about the same.
Still, it was obviously do-able and I forged ahead, working out the weight of every single piece of wood and every fitting too, for airplanes are the ultimate Balancing Act. Spread-sheets were a big help here because like most balancing acts you must concern yourself not only with the mass but how far it's located from the point of balance. Mass times distance becomes moment and a relatively small change can have a significant impact when the part is located far from the center of balance.
The hand-maiden of low-cost was ease of fabrication; a flying machine literally anyone could duplicate. But Chugger is also meant to be realistic and today most American's are fat so it had to be big enough to carry a fat person, which caused the wings to grow a bit. Still, it had to be small enough to build in the typical garage. I wanted the landing gear to also serve as a durable trailer, allowing the airplane to be towed on its own wheels. Accordingly, the wings had to fold. Unfortunately, I've since learned that most airport managers will not allow a towable aircraft to be erected and flown from their field unless it first undergoes an inspection by a 'certified' mechanic.
Drag increases exponentially with velocity but I've flown some real tumbleweeds that were not only safe but whose controls had a delightfully positive feel. Since tumbleweeds are never very fast I was more than happy with an expectation of 75 to 80mph for cruise. Accordingly, I paid scant attention to streamlining. But here in the western United States, for an airplane to be useful it must have a range of at least 300 miles, and more would be better. I included tankage for twenty gallons of fuel -- about one hundred and twenty pounds of the stuff. Add that to a 250 pound pilot and I found the wings were acting like Pinochio's nose, which forced me to redesign the fittings.
A ten gallon tank fit nicely in each wing but made it impossible to fold the wings without first draining the fuel. Defueling an aircraft is another thing airport managers get a bit huffy about.
Somewhere along in there it became pretty obvious that in order to fly really well my heavy, draggy, inexpensive, road-towable flying machine was going to need more power than I could get out of an inexpensive engine.
One reason I'd avoided aerodynamic drag was because I'm not very well educated and some of the equations cause my eyes to bleed. Fortunately I found some design texts from the 1930's that addressed low-speed drag in terms I could understand. With the help of a spread sheet I quickly learned that my light-weight, easily built -- but slab-sided -- empennage was generating about six times as much drag as a thicker but more streamlined design. The streamlined design was more difficult to build but the reduction in drag combined with a strict diet put the Chugger back into the realm of Volkswagen-powered aviation.
Like all diets, this one proved expensive, for it meant re-designing all of those lovely, low-cost mild steel fittings for SAE 4130 chrome-moly steel. While this appears to violate a basic tenet of the design, the cost of the 4130 is several thousand dollars less than the cost of replacing the converted Volkswagen with a small Continental. In keeping with the goal of easy fabrication I'm redesigning the fittings to allow them to be made from 4130 steel strips of standard dimensions. Which I should finish any year now :-)
In the meantime the engine is coming along, as is the 68x38 propeller. More detailed drawings and the occasional photo will be posted to the Chuggers Group on Yahoo.
Sunday, May 13, 2007
Do-It-Yourself Valve Spring Tester
Robert S. Hoover © 2003
A Volkswagen cylinder head contains seventy-seven individual components, the majority of which are capable of rendering the engine inoperable should it fail. Some of the components, such as the studs and the head casting itself are static and not subject to friction but due to the large number of dynamic components and generally poor valve train lubrication, they make up a significant portion of the engine's pumping losses. Since the pumping losses represent the engine's 'overhead,' any reduction in the pumping losses appears as an increase in the engine's output, usually for no increase in fuel consumption. By focusing on the details of those pumping losses, experience has shown that it is possible to achieve a significant increase the output of the engine.
Complex by modern-day standards, where an increasing number of engines are OHC, despite its high parts-count the VW valve train is reasonably robust thanks to seventy years of use during which the most failure-prone components have been identified and re-designed to improve their durability. That is, durability in vehicular terms. When compared to features found in aircraft engines, Volkswagen heads are something of a joke. When the displacement of the '1600' (actual displacement is 1584cc) is increased, as is commonly done when converting the engine for use in aircraft, durability takes a further hit. Fortunately, it takes only a modest amount of effort to improve its durability by an order of magnitude.
Most of the valve-train durability enhancements are covered in the so-called HVX modifications, previously posted and discussed. Although rarely seen on engines built for the Kiddie Trade and not found on any of the commercially available VW's converted for flight, the HVX mods have proven their worth through forty years of use in professionally built, high-output engines. Most recently, the use of thick-film lubricants have enhanced durability even further. (Specific How-To information for applying thick-film lubricants to valve train components will be found in the chapter on Coatings.)
Valve Train (Springs)
Poppet valves are a one-way sort of creature The cam pushes them open but they are closed by the action of the valve spring. The spring needs to be strong enough to close the valve tightly enough to make a leak-free seal but the valve spring merely initiates the sealing process. The real sealing is accomplished by the tapered sealing surface of the valve being wedged into the cone of the valve seat by the enormous pressure of combustion.
Modern-day valve springs are coiled compression springs installed around the stem of the valve and connected to it by a retainer that is free to rotate. The retainer is secured to the stem of the valve by a pair of keepers in the form of a cylindrical wedge which mates with grooves machined into the stem of the valve.
The strength of the stock VW valve spring is determined by measuring the amount of force needed to compress the spring to a height of 31.0mm (~1.220"). A number of factors can effect the strength of a coil spring and like all other VW specs, the tolerance is quite large, ranging from 117 to 135 pounds.
The valve's spring must be compressed when the valve is opened. The energy needed to compress the spring is part of the Otto Cycle's 'pumping losses' and anything that helps reduce those losses will improve the engine's efficiency. For a low rpm engine the lower valve spring value is more than enough to ensure proper operation and since the lower value reduces the pumping losses, it also serves to improve performance. Further enhancement occurs when the strength of all eight springs is equal or as nearly so as possible. For those reasons, a standard practice in any properly built engine is to use a set of springs that have been closely matched.
Matching a set of valve springs to within a pound or so can be quite difficult if you're drawing upon used parts. Not only are there different varieties of VW valve spring, each time a VW engine is stopped at least two valve springs will be compressed. In a vehicle that is driven daily this is seldom a problem but in an airplane engine that may sit for weeks between flights, the compressed spring is liable to weaken. When doing a valve job on a VW engine modified for flight, it's a good idea to re-test the valve springs.
Ideally, a new engine or a rebuilt head should include a set of new valve springs but with the number of registered air cooled Volkswagens in steady decline, it has become increasingly difficult to locate quality parts. It isn't uncommon to find after-market VW valve springs which are not square, in that the ends of the spring are not perpendicular to their axis. Such springs do not provide a symmetrical force when compressed and should not be used, an item mentioned in the factory service manual. You will also find new springs wound of lighter gauge wire than stock springs and which fail to provide the required strength when compressed. Springs longer than stock are also fairly common, often needing excessive pressure to be compressed to the specified height. Such junk is often advertised as 'racing' equipment, clearly meant for mechanically naive youngsters.
Volkswagen valve springs are progressively-wound, with the coils being closer together at the bottom than the top. Some after-market springs are not progressively-wound. (It pays to inspect all after-market VW parts before you buy.)
Twenty years ago I would never put used valve springs in an engine. Nowadays, used stock springs are often better than new, after-market stuff. If a used spring isn't rusty and shows no signs of fretting or jamming, I'll go ahead and test them.
New or used, it is extremely risky to use any valve spring without testing..
Valve spring testers are commonly available but even the least expensive model is several hundred dollars if purchased new. Fortunately, a common bathroom scale may be used to make your own spring tester. Unfortunately, inexpensive bathroom scales are not very accurate. Accuracy - - at least enough for the task at hand -- is assured by calibrating the scale with a mass of known weight, such as your own body, immediately prior to use. That of course assumes you know your own weight to within a pound. Balance beam type scales tend to be more accurate than low-cost spring-type scales. To calibrate the valve-spring's scale simply weigh yourself on a balance-beam scale then adjust the bathroom scale to read the same amount.
If you do not have access to a balance-beam type scale you'll have to create a test-mass of known weight. Having a specific gravity of 1.00, water is the handiest mass but you'd need at least fifteen gallons to verify the accuracy of your scale and the container would introduce some amount of error.
Lead is a very handy mass, having a specific gravity 11.34 times that of water and if you have a graduated beaker (which is easy enough to make) it's fairly simple to determine the volume of a given lump of lead. Unfortunately, pure lead is rather rare stuff and since other metals often make up as much as half the mass of wheel weights and other common lead alloys, it is impossible to calculate the weight of such alloys based volume alone.
If you have an accurate scale, such a laboratory type, you can of course weigh a sample of melted wheel weights, plumber's solder or other lead alloy, determine it's specific gravity and apply that to the mass as a whole.
When you are forced to create your own calibration mass without access to a precision scale you'll probably find plain old fashioned mild steel to be the best choice. This is because the amount of carbon and trace elements is typically less than 1%, allowing you to use a specific gravity of 7.93 or about 495 pounds per cubic foot ( about 4.4833 ounces per cubic inch ).
Since mild steel comes in standard sizes, even when purchased as scrap you can determine it's weight with good accuracy by simply measuring the piece, calculating its volume and applying the figures above. Then too, many scrap yards now use electronic scales accurate to a fraction of a pound, allowing you to simply buy a test-mass of the appropriate weight. Of course, being able to calculate the weight is a handy means of keeping them honest. (Hint: Weigh yourself on the junkyard's scales. Everyone does :-)
- - - - - - - - - - - - - - -
NOTE At one time it was common for EAA chapters to maintain a tool crib and test-mass for use by its members. The test-mass was usually pigs of lead- alloy cast in convenient sizes from five to twenty-five pounds, clearly stamped with their weight after being accurately weighed. The fact EAA headquarters no longer puts any emphasis on such basic needs is good evidence of their growing disinterest in supporting grass-roots aviation.
- - - - - - - - - - - - - - -
(I use a mill-end of 6" steel bar as my test mass. It is about 16" long and weighs 128 lb, 4-3/4 oz).
Volkswagen's valve-spring specification calls for a compression of 117 to 135 pounds at a height of 31mm. I made a gauge of this dimension that allows me to set the height of a bolt screwed into a pallet which sits atop a bathroom scale. The scale sits on a wooden base to which a fulcrum has been attached. The spring being tested is sipped over the bolt and a lever is used to compress it. When the lever touches the bolt I know the spring has been compressed to a height of 1.220" (ie, 31mm). And I know precisely when that happens because I've rigged the lever to turn on an LED when it touches the bolt. The LED is taped to the dial of the bathroom scale; all I have to do is keep my eye on the dial. When the light comes on I read the dial and jot down the weight on a stick-up. To eliminate human error each spring is tested at least three times. Any obvious flyers are thrown out and the testing is repeated until I have a cluster of similar values.
I try to do forty or fifty valve springs at a time. The first step is to clean them and inspect each spring visually for scratches or pitting anything that might serve as a stress-riser. They are then gauged for total length, then for squareness, both tests done on a surface plate allowing me to do a handful of springs at a time. Alas, when dealing with new, after-market springs those two tests may reduce the batch by half.
Any springs that pass the initial tests are then tested for compression height. They are then sorted according to their stick-ums and made up into matching sets, coated with preservative and put aside until needed. It isn't the Bureau of Standards but it's better than guess-work, which is what you have if you don't test your springs.
In making up a set of springs for a low rpm engine I want the lowest strength and the narrowest range. Of the two, I think matching the range is the most important factor. If I can't make up a set within a pound or two of a given strength, I'll generally keep looking.
- - - - - - - - - -
NOTE: Many fail to appreciate the importance of 'balance' in an engine. The reason professionals put so much emphasis on balancing is because the engine must use power to overcome any imbalance before any usable power can appear at the crankshaft. That means any imbalance is effectively multiplied by two. Using springs of equal strength is part of the balancing process.
- - - - - - - - - -
If you are building just one engine you should try to find someone who has a valve-spring tester. Baring that, you should cobble up your own using a bathroom scale.
So what happens if you simply buy a new set of springs and throw them in? Hopefully, not a lot. There is a chance the set may contain a spring having a radically different value but with a tolerance of 18 pounds, the odds are the engine is going to run. Sorta :-)
I should also mention that I don't know of a single non-professional engine- builder who tests their valve springs. This is another of those details they deem 'unimportant.' And when addressed in isolation, perhaps it is. But a professional engine builder addresses all of those 'unimportant' details, picking up a little torque here, better fuel consumption there, optimizing each unimportant detail for better efficiency, more power, cooler running and slower wear. No single one of those unimportant details results in a dramatic change. But add them all together and it isn't uncommon for a professionally built engine to produce up to 25% more power than a poorly built engine of exactly the same displacement. And to last twice as long as well.
Friday, May 11, 2007
Guy comes by the shop hoping to bum me outta some ammo for his machine gun and finds me stitching a spar together with a gnu-matic brad-driver & foamy glue. He's okay watching me smear on the glue. And he's still happy when I put the plywood into place, aligned by a couple of 3/4" #20 nails previously driven, now with their heads snipped off. He even helps me clamp it to the bench. But when I pick up the pneumatic brad-driver and start to stitch he begins to frown, as if he's never it done before.
"I've never seen it done like that," he sez.
"Saves a buncha time," I say as I shoot 5/8" #18 brads about every inch and a half. It's the spar for a horizontal stabilizer, sorta-copied from Pete Bower's 'Fly Baby.' It's about six feet long and
three inches deep in the middle, tapering to an inch and a half at each end. The spar caps are 3/8" square Western Hemlock, ripped out of a 2x4. The plywood shear-web is a piece of doorskin; 1/8" Luan. It's the third one I've built, the first two having been destroyed in
various tests. Not counting the glue and brads, each cost about a dollar. Plus a few hundred hours of design time.
"Is that an Approved Method?" he asks.
The brads are tacking the assemblage to the work bench, which is protected by a layer of waxed paper. (Live & lurn :-) I'll leave it to cure for a couple of days then pry it off, at which time it will
look like hell warmed over. But it cleans up nice with a disk sander. Then comes some filler blocks to be fitted and interior varnish before I can apply the closing face, which is left a bit over-size and malleted down onto the exposed tips of the brads.
"Beats the hell outta me," I tell him. "But it seems to work pretty good."
"Jesus!" he shouts, jumping back and making the sign of the cross. "You... you're EXPERIMENTING!"
That's when I turned slowly toward him and smiled, showing him my pointy teeth and red LED's for eyes. He gave a tiny shriek and ran off, shouting: "Pope Paul! Pope Paul!"
Thursday, May 10, 2007
Stock VW engine, you’re looking at more than fifty years of continuous development and production.
Not a lot of secrets in a stocker. Which wasn’t always the case.
When it was introduced the Volkswagen engine violated many Conventional Wisdoms associated with automobile engine design. For example, each lobe of the cam actuates two cam followers. Conventional Wisdom insisted the four lobes on a VW wiggle stick would wear twice as fast as the eight lobes on all other four-banger cams... unless you came up with some way to precisely control the hardness of your cams and lifters, which no one had back then. Oh, there was that new gaseous nitriding process but that only worked with steel. VW cams and lifters were cast iron and no one had come up with an accurate hardening process that was economical enough to be used for mass produced cast iron parts. Except an outfit called Krupp.
Ditto for that magnesium alloy crankcase. Never work, not for mass production. Way too expensive. Unless you can come up with a better method of extracting magnesium from sea water. Like that Dowmettal company. Same story for those crazy molded rubber parts in the torsion-bar suspension system. Just won’t work, unless you can come up with a synthetic rubber that’s actually better than the real thing. Maybe some of that Buna stuff would work... (nowadays we call it Neoprene).
Professor Porsche and his gang of engineers didn’t see such things as limitations, they saw them as challenges and came up with an engine that remains in production today. (You can get new replacement engines from the VW plant in Puebla, Mexico. Pretty good little engines.) And if you liked the torsion bar suspension system on the original People’s Car you’ll find it still going strong under our main battle tank.
- - - - - - - - - - - - - - - - - - - - -
One of the tricky bits on a VW cam is getting the hardness just right. Not a big problem nowadays, thanks to Krupp and Adolph Fry. Today you simply look it up on a chart and set the dials to produce whatever hardness and depth is required, duplicating a process that has been in common use now for more than sixty years. It’s no more difficult than, say, programming your VCR. (Yeah, I know... But there it is :-)
For those of you not familiar with surface hardening, take a look at CAMHARD01. Nowadays there are lots of ways to harden the surface of iron, steel or cast iron but one of the handiest heats the metal in an atmosphere of nitrogen gas. The depth of the hardened surface can be controlled by the temperature to which the metal is raised, how long it stays in the oven, the concentration of nitrogen inside the oven, how the part is cooled and so forth.
Hardening a cam is a bit tricker than most other hardening chores because cams have lots of corners. When hardening a part the corners are exposed to the heat & gas on two sides and tend to become harder than other areas of the part. To give you some idea what I’m talking about go see CAMHARD02.
A brittle corner on a cam can be fatal to an engine. The corners approach the hardness of a diamond (seriously! Nitriding can produce a Mohs hardness of better than 9. [Diamond is a 10]).
Talk about the perfect abrasive! Microscopic fragments of diamond-hard material being chipped off and distributed around the inside of an engine... You can bet your bean-bag it caused the VW engineers more than few headaches before they figured it out.
Corners on a cam? (Someone said.) Where the hell are there corners on a cam!
See Figure 3. That’s a picture of a typical after-market cam, fresh back from nitriding. See all those nice sharp edges? That is where the metal turns a corner. Those edges are so brittle that casual handling can cause them to chip like glass. (Look closely. See that tiny notch near the nose?) Even, worse, see the mold-lines on the cam? (Remember, cams are just high-density cast iron.) The blank comes out of the mold with a chilled hardness that is nearly as good as nitriding (although not nearly so deep). Nitride a chilled-cast surface, you end up with hardness well past 9 on the Mohs Scale and a virtual 100% guarantee of chipping those edges unless you do something about it.
Figure 4 shows the pointy end of the lobe on an after-market cam. It also shows all those un-dressed edges. This is normal for after-market parts. It is up to the person assembling the engine to determine which edges need to be chamfered, by how much and the method most suitable for doing so. For comparison, Figure 5 is a stock VW cam. (Notice that the end of the stock cam is not as sharp.) Note that all of the edges on the stock cam are nicely chamfered.
Take another look at CAMHARD02, the drawing showing how the hardness penetrates the metal. The tip of the lobs concentrates the heat during the hardening process in much the same fashion as does a corner. In fact, the tip of the cam’s lobe has to be harder than its slopes or heel if you want the thing to wear at a slow rate. And if the tip is harder than the heel, you can bet your bippie that the edges of the lobe’s nose are even harder still.
You simply can’t allow fragments from those edges to get inside your engine. Even with a full-flow oil filter such debris still gets one shot at your oil pump. And with stuff approaching the hardness of a diamond, one shot is all is takes.
So we don’t let that happen. And neither did Volkswagen. But I don’t have to tell you that because you can see it for yourself. See those nicely chamfered edges in Figure 5? That’s a stock Volkswagen cam. Look at Figure 6; there it is again. That’s a used VW cam, something I pulled out from under the bench. But you can clearly see the chamfering and, if you look real close, the VW logo cast into the metal.
See the lower lobe in Figure 6? You can see that the chamfer is a bit smaller than on the heel of the upper lobe. The chamfer doesn’t have to be very big if all you want to do is get rid of the chunkies. In fact, a chamfer of only thirty-thou or so is enough to make the edges of an after-market cam safe for society. Sure enough, there’s a picture of a lightly chamfered after-market cam lurking in Figure 7. (Wider would be better. I just whizzed this one up for the photo-op :-) If you’ve never clearanced a cam nor chamfered one, one slip of the grinder can screw the pooch in a major way, as in trashing the cam. I suggest you cover the lobes and journals with masking tape before doing any grinding.
Normally, you chamfer an after-market cam when you grind the notches that allow it to work with a stroker crank. That is, you do all your grinding - and clean-up - at one time, usually in a ‘dirty’ area of your shop. (Engines are always assembled in a clean area. It’s not an operating theater but the assembly area should be cleaner than the average kitchen.) No stroker? Then you can chamfer it any time you wish. (It’s called dressing the edges and is a standard pre-assembly procedure with any after-market part.) Just be sure to clean that sucker to within an inch of its life after doing any grinding on the thing. The idea here is to keep abrasive debris out of your engine. Grind on the cam (or anything else) then use the part without a perfect clean-up simply doesn’t make sense.
- - - - - - - - - - - - - - - - - - - -
Cams are made from cast iron because it is easy to grind to the required curves. After the lobes are ground, the surface of a Volkswagen cam is hardened to a precise degree. The result is a cam with lobes just hard enough so that the rate of wear for one cam-lobe is compatible with that of the distributed wear across the face of two cam-followers (i.e., the lifters rotate to distribute the wear). This results in uniform rate of wear allowing reliable long-term performance.
The process of surface hardening concentrates the harness along edges and thinner sections. By the time you have achieved the desired hardness in the middle of the piece any sharp edges will have been hardened to the point of brittleness.
Cast iron has a granular structure; harden it to the point of brittleness, it will chip like a piece of glass. But only if you let it. Standard automotive engineering practice is to ensure such debris is never allowed inside an engine.
When hardened debris passes through the oil pump, it will create a scratch or score. Once the metal has been scored, it will not heal. The more times such debris is allowed to pass through the pump, the more wear that will accumulate.
When building an engine, any edge capable of spawning debris is chamfered, rounded, stoned or even polished, as the case may be. ANY EDGE. Throughout the engine. The need for such attention to detail is understood by every competent mechanic. The proof of that need and the practices required is clearly evident by simply examining a professionally built engine.
- - - - - - - - - - - - - - - - - - - - - - - - -
There are no secrets in a VW engine. Or so I thought :-)
Are you using an after-market cam? Did you clean it up and chamfer the edges? Gap your rings? Stone the edges? Balance everything? That’s your job, you know; attending to all those ‘unimportant’ details the phony experts brush aside. Because when you build an engine, you’re the Mechanic in Charge.