At the flowbench

The chambers, seats and finally the ports are finished.  What follows are some photos and video that may be interesting to some of you.

Here is a head set up to measure airflow through the intake port.

 For my purposes there is no need to have the entire intake tract set up with the head, but we do need to make sure that the air enters the intake port in a smooth stream that is repeatable from head to head.  Some guys will use a clay bellmouth on the intake stub but I doubt that I could sculpt a perfectly matching bellmouth on two heads so I simply machined one from nylon.  It snaps right on and I put a dab of clay on either side to make sure that it stays put.

Here are some photos of airflow visualization using smoke.  The way the bellmouth turns the airflow can be plainly seen.

I have been experimenting with wet flow visualization, in the following videos you can see that at valve lifts of .100", .200" and .300" the new combustion chamber contours act as very effective half-diffusers and keep the airflow attached to the head around the entire periphery of the valve head while also enhancing intake flow swirl.  As we know, a properly designed diffuser converts the flow's velocity back into pressure with minimal flow loss.  The flow in the original configuration was mostly straight across the top side of the valve in two very strong streams that tended to travel straight across the chamber with no swirl.  I am much more concerned with flow velocity at the lifts between closed and full-open than I am about flow numbers at full-lift.  Think about this, with a cam that has a symmetrical profile, the valve will see every stage of lift twice (once on opening and once on closing) while it sees full-lift at one point and then only for an instant.  This is why it is not well to get hung up on obtaining maximum flow numbers at full-lift.  Remember, the flowbench is an abstraction, as no piston-engine sees prolonged steady-state airflow at full-lift.  The flowbench is best employed as a comparative tool.   The fluid that is used for visualization is much heavier than gasoline and so tends puddle a bit but the airflow visualization is accurate and there is no explosion hazard (life is stressful enough).

There is one other item to take care of before reassembly and dyno-tuning, that is the crankcase breather system.  The engine's crankcase volume is fairly small and that, combined with the fact that it fluctuates by an amount equal to the engine's displacement twice every revolution, tends to push a considerable amount of oil out of the crankcase breather.  While it is common knowledge that subjecting the crankcase to a vacuum is beneficial, the way BMW chose causes more grief than it's worth.  Venting the crankcase to the airbox and using intake vacuum to evacuate the crankcase results in the engine ingesting a not insignificant amount of oil.  As we all know, ANY oil ingestion is highly detrimental as it dilutes the incoming fuel-air mixture, helps set the stage for detonation and is the primary cause of combustion chamber coking that airheads are famous for.  Pull the intake elbows from any large-bore airhead and you will find this.

That is engine oil dripping from the crankcase breather pipe (which itself is quite an obstruction to airflow).  I am not an advocate of removing the airbox (since the intake length is beneficial to this engine), but this crankcase breather system (along with the pulse-air exhaust injection system) is going to go.  I will use a catch-tank plumbed in series with an electric vacuum pump.  This should keep a negative pressure(vacuum) in the crankcase regardless of the piston position, give the oil control rings an easier life and maybe free up some power, but most of all, NO MORE OIL IN THE COMBUSTION CHAMBERS.

Volume check

The last part of the combustion chamber modification is to measure and equalize the volumes.  The target volume was 80cc which will yield a compression ratio of 10.0:1.  The right cylinder head measured at 79.9cc while the left head came in at 78cc and so needed some adjustment to be equal to the right head.

That's it for the combustion chamber modifications.  Next will be the port work, which I will detail here.  Here are the finished chambers.

KTM 950 Starter Clutch (sprag) Failure

 When all you get from the starter is a whizzing sound, this is why.  All of the metal that used to be part of the gear is now circulating throughout the engine in sizes ranging from dust to "Oh No".
 If your 950 starter starts acting up, don't put off getting it repaired or it WILL get much more expensive.

New guides and someplace to sit

Installing the valve guides:
Here is the shop-made driver that drives on the top shoulder of the guide, it is piloted and counterbored so as to miss the sharp-edged, tapered portion at the top of the guide.

The guides were reamed and honed to proper size and then the valve seat counterbores were machined.

The seats were machined from C63000 bronze.  This material is a superb heat conductor, has a similar expansion rate to the aluminum heads and is hard enough for unleaded use.  They are machined to a .007" interference fit in the heads and radiused (.062") so as not to broach any material when installed.

The heads are heated to 400 degrees F and the seats are frozen before being driven into place.
Here is a photo of the shop-made driver, note the o-ring to hold the seat in place on the driver.

Once the seats are installed they are contoured to match the combustion chamber shape.

Then the actual seating surfaces are cut (as well as a 60 degree back-cut and 30 degree top-cut) and zero seat runout is verified using a shop-made runout gauge.

Here are a couple photos of the runout gauge and a pilot.

Next step is to match the combustion chambers for volume (cc'ing) and then many fun-filled hours at the flowbench working on the intake and exhaust ports.

Hemi no more

Both chambers are roughly mirror images of each other.  Once I've machined new seats they will be installed along with new valve guides.  The chambers will be exactly matched for contour and volume after the valves are installed.  The intake valves will remain at their stock size of 42mm but the exhaust valves will go from the stock size of 40mm down to 38mm (more on this later).

Airhead Update

The valve seats are removed and the combustion chambers are welded.

The deck surface is trued and the squish area contour are done after fixturing the head to the lathe faceplate.

Here you can see that the deck is now below the actual combustion chamber surface.  When installed with a stock head gasket the squish area will be equal to the piston deck height, which is 1 milimeter below the cylinder deck surface.

Here is a diagram describing the standard versus the new configuration.

For reference, here is a photo of the original, recessed chamber.

T100R glamour shots

A few people have asked to see the engine in the bike so here are some photos.  The instruments are out being restored and upon their return I will also be installing the English-market low handlebars.

Airhead alterations

The usual route to "performance" with BMW's airhead engines is the typical American hot-rodder's approach, big valves, big ports, big compression, big cam, dual plugs, etc.  Bigger valves can flow more air, but, combined with bigger ports this increase in flow generally comes at the cost of velocity.
  Intake charge velocity is a prime contributor to combustion chamber turbulence and given the volume of an airhead's featureless, hemi combustion chamber, anything that can help to stir things up should be preserved or enhanced.  While those big ports and valves might flow some impressive numbers on the flowbench (which itself is an abstraction as no engine experiences steady-state flow), those big numbers generally will only provide an improvement at higher RPM, if at all.
  There are those that assume that since the larger airhead engines are oversquare, they must be designed to operate at high RPM.  Even if this were true, who would want a street-ridden engine that only makes power while you're wringing its neck?

 Let's briefly take a look at the BMW opposed twin and why it is not suited to high RPM operation.
- Heavy valvetrain?  Check.
- Foot-long pushrods? Check.  Ever heard of Leonhard Euler?
- Crank supported only at the ends?  Check.  Yes, it will flex at high RPM.
- Poor crankcase rigidity?  Check.  The crankcases will flex at high RPM. Even though the engine has a 180-degree crank and the cylinders are opposed at 180-degrees, the engine is effectively a 360-degree twin.  Both pistons reach TDC at the same time and they reach BDC at the same time, just like a British twin.  And just like a British twin, the crankcase volume changes by an amount equal to the engine's displacement every time the crank reaches BDC.  Udo Gietl discovered how much the cases will flex while building the Superbike championship engines to survive sustained high-RPM operation.
- Large, slow-burning hemi chambers?  Check.

The last item on the list is what I intend to address.  The accepted orthodoxy is to simply add a second spark plug opposite the original, back the timing off about 2-4 degrees and call it done.
This is a half-measure at best.  I think a better way would be to redesign the combustion chamber to gain turbulence which will accomplish a faster, more complete burn of the intake charge than can be had by simply lighting the original, lazy mixture at both ends.  Unfortunately there has been a lack of cylinder head development for these engines, but there have been many advances (both factory and aftermarket) in cylinder heads for the airhead's closest cousin, the American V-twin.  Yes, I said it.  Be intellectually honest and realize that, except for the cylinder angle, the engines are very similar.
  The knowledgeable V-twin tuners long ago abandoned the dual-plug, giant valves, giant ports, giant cams dogma.
Look at the chambers of any performance V-twin head, you'll look for a long time to find an open, hemispherical chamber.

To that end, here is the modified chamber that I propose.  While it may appear simple, I assure you that many hours went into coming up with this form.  The heads will be welded and reshaped to duplicate this pattern.

The ultimate goal is an engine with improved performance across its entire RPM range that will run cooler, with much less ignition advance and be ping-free on pump gas.
To those who would say "Yeah, V-twins make a ton of torque but I want horsepower" I say, ponder this:  Horsepower does not exist and can not be directly measured, it is the sum of an equation.  That equation is TORQUE x RPM, divided by a constant, = horsepower.  Torque is an actual force that exists and can be measured, horsepower is a concept that is expressed as the sum of an equation.
There are two ways to make that sum larger, increase RPM (ill-suited to an airhead) or increase torque output (the goal).

Airhead top end rebuild

It's an '88 R100RS.  Problems are soggy performance and exhaust clearances tightening up very quickly.

Here is what the exhaust valves look like at 117 thousand miles.

I will be posting the entire rebuild along with some performance rework of the heads.

T100R finished

It is ready to go back into the frame and after break-in will be returned to the owner. 
I want to take this opportunity to thank Mitch Klempf and Chris Stubbs at Klempf's British Parts in Minnesota.
These guys unquestionably are the benchmark of customer service and knowledge when it comes to parts for vintage British motorcycles, they really know their business and are an absolute pleasure to deal with.
They can be reached at:  http://www.klempfsbritishparts.com/

Here are a couple photos.

And here are a few photos of what I found when tearing this engine down in preparation for the rebuild.

Incorrect clutch basket thrust washer

Worn valve guides

T100C pistons
The bigger Daytona valves made their own clearance.

The seller claimed that this engine had never been apart, if I had a dollar for every time I heard that one....