Tuesday, July 31, 2007


(NOTE: This was written in 2003. Use it only for background information. In later posts I'll show the methods & procedures I'm presently using, tell you why some things worked better than others.)

Head Work

Die Grinder is big. Two-handed sort of tool. Home machinists usually have one from Sears, Sioux or Black & Decker. The barrel of the high rpm motor is of a standard diameter so you can use a die grinder holder on your lathe; Po' Boy Tool Post Grinder.

Real die grinders are used by Tool & Die markers to literally sculpt steel. Big die grinders, usually pneumatically powered, sometimes so large they're suspended on a counter-poise. Use one of these, solid carbide burr, you gotta dress for the occasion.

Building a big-bore VW for use in an airplane, there isn't a lot of head work; nothing like what you put into a racing engine. At propeller speeds the flow-rate of even the largest big-bore stroker is small in comparison to something designed to turn seven grand at cruise and peak-out around nine. Still, there is some work to do. Opening up the chambers to accept larger jugs leaves a wide ledge at each end of the valve recess. A flow bench will show that the heads breathe better if the ledge is set back so as to unshroud the valve. Bill Fisher covered this in his 1970-era `How to Hotrod Volkswagen Engines,' which remains in print and is still valid for such things as head work. When you get a copy of Bill's book be sure to study the flow-rate charts. Then sit down and calculate the flow rate for your engine, assuming 100% volumetric efficiency at your designed cruising rpm.

Be prepared to be underwhelmed :-)

Now go back and look at the charts. Notice that your rpm indicates stock single-port heads will do pretty well without any unshrouding or smoothing. Up to you; you're the Mechanic-in-Charge.

I always clean up the heads. Force of habit as much as better performance. The big advantage to this type of work is that the improvement ends up being built right into the engine. Like bigger displacement, better breathing isn't something you have to add on or periodically replace.

Besides unshrouding the valves there's a few sharp edges in the chambers that need to be smoothed. Ever seen air-flow through polarized filters? Comparing the air negotiating a sharp corner to one that has been properly radiused is a real eye-opener when you can see the improvement in the flow. Here again, let Bill's book be your guide. Lotsa good pictures.

Most of the head-work requiring a die grinder is simple smoothing. The head is a casting; the ports have rough surfaces, reflecting the surface of the cores, a lot rougher than the fins and other surfaces which reflect the permanent metal molds used to cast VW heads. (I'll mention those other surfaces in a minute.)

We usta think we got more flow if the passages had a mirror finish. Turns out, according to Pratt-Whitney and NASA, there's no improvement after the surface texture hits about #600. (I didn't believe them, of course. But the flow bench did :-) Why this is true has to do with the fact that fuel/air mixture is not a perfect gas. Flow bench runs straight air unless you dope it with a suspended colloid such as smoke.

Point here is that all you need to do to see a good increase in your in-flow (and thus in your VE) is to get the ports dead smooth. Don't worry about a polished finish.

The way to do that is to start with a flapper wheel or sanding drum in your die grinder and knock down all the casting imperfections. Your hand is your best gauge here. As-cast, the ports feel like rough concrete. Your job is to make them feel like smoothly sanded wood.

Once you've gotten rid of all the peaks and ground out any inclusions and smoothed the trench, you simply shift to a finer abrasive and remove the marks of your first effort. Then do it again. And again.

By the time you've gone down about three graduations in your abrasives (which are also covered in Bill's book, I think... ) the surfaces will be uniformly smooth and have an even, frosted appearance that offers just the slightest hint of tooth to the touch. Go ever everything about three times at that level then shift to your finishing grade, whatever it happens to be. But be warned: Each time you graduate to a smaller size it will take about twice as long to remove the marks from the previous grit. If you've got a die grinder and a box of Cratex hobbs expect to spend about four hours per head.

I didn't have all that stuff when I was a kid. I did my first heads using a quarter-inch drill motor. Worked okay but I think I spent about thirty hours doing a pair of heads; no big deal when you're a kid, right? :-) Nowadays, if I didn't have a die grinder I'd probably shoot myself rather than stand at the bench for thirty hours.

Which brings us to the point of all this.

Some time ago a fellow wrote to ask if he could do a set of heads using a Dremel tool. I told him I didn't know. Recently, two other fellows asked the same question and I felt that justified looking into it.

The answer is a qualified `yes.' It takes quite a bit of time but I found I could unshroud the valves and clean up the sharp edges under the exhaust valve using a hobby-type tool, an inexpensive thing I picked up at Harbor Freight. And you can smooth at least part of the ports. But the tool was too fat to get right into the ports, and it didn't come with hobbs that were long enough to reach. So yes, you can do most of the job, and you should see some improvement in your flow, but not as much as when they were properly smoothed for their full length.

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

As for the fins, I'm sure I've covered this before but I want to keep all of this together, so here it comes again.

You have to remove all of the flash between the fins. I use a pneumatic riffler for this but a sabre saw or even a hand saw will work for most of the fins. Up around the exhaust valves you must ensure several very critical passageways are not blocked. I've posted a drawing of the heads and the passageways are clearly shown but you can see them for yourself the first time you examine a head. Keep in mind that the exhaust stack and the exhaust valve guide are principle sources of heat in this area. The head is designed to have the air flow down thru the head; this is reflected by the drafting ratio in the mold. Air expands as it absorbs heat so the exhaust-side of any cooling air channel tends to have greater volume than the intake-side. Don't upset that ratio or you'll see a pressure drop, meaning the air is not picking up as much heat as before. The pressure differential in all cases should be between six and nine inches of water and this is something you should focus your attention on during your test flights. An airspeed indicator rigged with the pressure port to the upper plenum and the static port to the exhaust area, usually the space forward of the firewall, should show a pressure differential of about 90 miles per hour from inlet to outlet. Or you can rig a barometer or even an altimeter to indicate the pressure differential.

Cooling air pressure differential is not something you want to leave to chance. The Volkswagen engine was designed to use a blower having an output proprotional to the speed of the engine. To properly cool the engine using ram-air you must pay the keenest attention to a host of 'unimportant' details. Not only must the upper plenum provide sufficient pressure, your lower shrouding should provide sufficient containment to force maximum rate of flow through the hottest parts of the head. You can instrument these areas with inexpensive thermistors wired to a single gauge and read via a rotary switch or whatever; something sturdy but temporary.

Adjustments to the system take the form of changing the inlet & exhaust area. I consider the late John Thorpe to be the best authority I've read on cooling horizontally opposed aircraft engines. He wrote a series of articles for `Sport Aviation' or its precursor. See if you can track down his articles. If you can't, perhaps someone can paraphrase them or extract just the equations and post them to the archive.

Once you've cleaned up your fins, seal up the chambers, ports and valve gallery then blast the shit out of the fins with coarse abrasive media at low pressure. What you want to achieve is a rough surface. In fact, blasting the cast fins with coarse media will result in a significant increase in the surface area of the fins.

But don't blast anywhere that will eventually be inside the engine. Abrasive media has a habit of embedding itself in non-ferrous metals, coming loose as the metal goes through heat cycles. Bottom line is that if you don't want abrasive in your bearings, don't allow it to get on the engine to begin with. (Real shops use non-abrasive media for cleaning heads and the like. Walnut shells [which is what I use] or plastic beads. Frangible media such as glass beads comes under the same ban as abrasives.)

(So why is it okay to use abrasive rules for porting & polishing but using abrasive media is evil? Mostly because the ports and chambers are pretty small compared to the valve gallery, but more so because of the nature of the media. Blasted media tends to embed itself whereas the media on a sanding wheel does not. [A 30x glass and a good light will let you answer this question for yourself.])

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

Finally, one of the most difficult lessons to learn with Thermal Barrier Coatings is that the surface must have a roughness equal to #80. The best way I found to achieve this on aluminum is with plain old fashioned silica sand. And yes, it breaks the rules big-time.

(NOTE: Starting in 2005 I began using aluminum oxide abrasive rather than sand. It is more dangerous to the engine but less so to the person doing the blasting.)

To coat the tops of the pistons I was able to mask them off pretty well. After blasting I ran them through the ultrasonic cleaner then a zero-tolerance degreaser and finally into the spray booth for spraying with the TBC. They get cured in a 350 degree (F) oven and are allowed to cool in the oven for 24 hours or until hell freezes over, which is how long it seems when you're dancing around waiting to see if you've just fucked up $200 worth of pistons. Which I did, more than once, except they were gimmes; only pistons from out of the junk box. And you can remove a bad coating by blasting... but don't expect it to come out evenly. Blasting off the bad coating then taking a clean-up cut of about .0015 works. Indeed, you end up with a mirror bright beauty... which you must then carry over to the blasting cabinet and hose to a dull, frosted surface. Life is strange in the engine room :-)

The beautifully smoothed heads got the same treatment. I use solid copper head gaskets on heads that have been bored for larger jugs and you don't want the TBC to be UNDER the gasket, which means masking it off. If you can. I tried several methods. I wasn't entirely satisfied with any of them. Masking tape didn't work; after being stripped away (TBC dried but not baked) there was enough residue from the adhesive to contaminate the TBC. I ended up making aluminum rings about five thou wider than the copper gaskets and swaging them into place as a mask. The second time I did it I remembered to provide some means of pry them out of the chambers without scratching the sprayed TBC :-)

The piston tops, combustion chambers and exhaust stacks received the basic Thermal Barrier Coating. Because of its hyper-eutectic nature, baking at 350F cascades a melting process that results in ceramic- metallic alloy bonded to the aluminum substrate. Because of its ceramic nature, the surface pretty much ignores heat.

How does it work on a full-size engine? I don't know. But soon will.

The reason I mentioned it here is because of the violation of the `no abrasives' rule. Silica is definitely an abrasive. But if there was any abrasive residues left on the surface, they are now encapsulated in a cer-met alloy that you literally can not chip with a ball peen hammer. (I've spent three years convincing myself this stuff is worth the effort. I really wanted that shit to fail... save me all the trouble early on. I still don't know. But I'm starting to lean toward `Hopeful' on the self-delusion meter :-)

What's all this for? The main goal is to extend the life of the valves through better management of their heat load. I might see some improvement in power output because of a slightly higher BMEP. Or I might not. The folks who make the coatings don't have a lot of data on air-cooled engines and liquid cooled's don't even come close to the problems we have.