My experiences building a 900 race motor - Kevin Bracken


ducatimeccanica.com - home



Engine cases, Crankshaft, Oil System

Plain bearing conversion with Carillo rods

Pistons

Valve clearances

Cylinders

Compression ratios

Heads

Flow bench numbers

Belt/Bevel conversion


Introduction
Like most of you in the US (and elsewhere) who are somewhere near my age (B.1948), I read all about Cook Neilsen and Phil Schilling’s escapades with Overdog, aka Old Blue. I got hooked on Ducatis and in ’75 bought a leftover new ’74 750 GT. In ’76 I moved to California to pursue both a better job and more time to ride and got to watch Cook race at Riverside, Sears Point and Laguna Seca, along with Kenny, Eddie and Freddie. Amazing stuff that made me want to race desperately. But then again I had wanted to do that since I was 17. Hailwood was God to me. I was nearing 30 at the time, but making a good living as a chemical engineer in the pharmaceutical industry at Hyland Labs near to Glendale. Because of the GT and a burned exhaust valve, I met Jim Woods, who was just getting his shop going in Glendale. He convinced me that he needed to fix the problem and a friendship began. After awhile, I joined the group of hangers-on at the shop and we started riding the Angeles Crest Highway twice a week. My poor GT had trouble keeping up with the 750 sport of Jim, and 850 Lemans’ of the other guys and in ’78 I bought a 900 SD Darmah. Suddenly I found myself leading the pack, by a lot. Jim decided he needed to build up the shop’s reputation as a performance oriented place and felt that going into the local club-racing scene would be a good way to get started. He approached me about being the rider as long as I would help put the bike together and maintain it. He would provide the shop, parts and his mechanical know-how, which was considerably more than mine. Our first racer actually was two Darmahs that were crashed insurance write-offs that Jim acquired. We put the best of the two together and I began learning the club circuits of Willow Springs, Ontario (now gone), Riverside (also, sadly gone), and Sears Point in the AFM and ARRA. I was very green, but old enough not to be too stupid. I was a slow learner, but the lessons usually stuck.

It quickly became apparent racing against the Japanese bikes that yes, the Ducati handled better, but was really down on power. So you had to ride really hard in the corners to make up the lost ground. Over the next few years until I stopped racing in August of ’84, the search was for power and reliability. I had reasonable success for an amateur racer, winning a few class championships and placing well in national BOTT races, as long as the bike held together and I didn’t do anything too crazy. I never considered myself one of the really fast guys, just a fairly fast club racer. Getting passed at Warp 9 by Freddie on the banking at Daytona in Superbike practice is really humbling.

Now I should caution everyone that what I have learned about building a Ducati racebike and engine modifications is A way of doing things. There are probably many other ways. I think that I have learned to be somewhat open-minded about this. Jim Woods started out the racing project with a lot of good ideas. I went along with them until Jim introduced me to C.R. Axtell. He and Mike Libby, his partner, repped Dell’ortos at the time and did some headwork for Jim, like making valve guides and putting in new seats. I had read about Axtell. He was always referred to as "the legendary" in the cycle magazines. I quickly found out why. During the rest of this I will refer to Ax frequently for very good reasons. Apart from his role in providing some of the HP that has been so important to many a race winner he is just one very smart and engaging guy. I have been fortunate enough to be around good and even great scientific people for most of my career. He would put most of them to shame. He is now semi-retired and we talk occasionally on the phone, usually about jazz. He designs camshafts now and has done work for some of the big car racing teams and builders.


The Engine Cases, Crankshaft and Oil System
I have grouped these together because they sort of work hand-in-hand. Most of you already know from reading about Neilsen’s bike and probably some of your own experiences that the Achilles heel of the built-up bevel drive engine is the roller crankshaft. It will fail about every 4-6 hours in racing conditions. (My 750 GT with its nicely underpowered engine had the same crank for 110,000 miles, I even wore out the angular contact main bearings at 94,000-the phenolic cages disintegrated!). All of my racing cranks were put together by Jeff Bratton, who may be the best roller crank guy around. They were all balanced by Hank-the-Crank in North Hollywood. A 53-55% balance factor was used. My racing cranks would fail just like everyone else’s. I tried different kinds of oil, settling on Redline 20w-50, which played havoc with the clutch because it was so slippery. But that didn’t really help the crankshaft. Jeff Bratton would always call me upon receipt of the latest trashed crank and new rod set and say, "Kevin, these bearings on the old rod set are purple. This thing isn’t getting any oil." For reasons that are not clear to me (stupidity?) I didn’t go to Axtell with the problem until much later. In the meantime I tried first a simple roller crank conversion that I obtained from Albert Bajohr in Germany, who I met at Daytona in ’83. That stopped the crank failures, but later I went on to a complete changeover to Carillo rods and my own pin. More on that later.

Now the basic engine design is good and by all viewpoints the crankshaft shouldn’t fail at the rate that it does. The sump carries over 4-L of oil! How come the crank is running dry? It is really very simple. Axtell looked at the internal case design and said, "Our Italian cousins are so smart in other areas, how come they made the drainback holes so small? The oil pump pickup is sure to starve. This oil has got to have a way to get back down away from all of these rotating parts. In addition to pump starvation that is lost HP." Look at it yourself. There are just a few small slots separating the sump from the crank and transmission cavities. Even hot oil just can’t flow down quick enough. The fix? Just drill a bunch of 3/8 in. holes in the pan below the transmission. The crank area has a weir (lip) in the cases that is designed to strip the oil from the crank counterweights and direct it back to this area. The pan just needs more holes to drain back to the pump pickup (See pictue below).


A simple way to improve drainback to the sump by enlarging existing holes and adding a few.


Now I went a little farther on my racing cases and widened the existing slots and oil pickup shrouding with a Dumore porting tool and put a screen over the drilled holes (See pictures below). This is common practice in wet-sump racing car engines. Roller cranks need oil volume. This fix is essential. Plain bearing cranks need oil pressure. The fix is not quite as essential if you convert to plain bearings as I did, but can’t hurt.


Right hand case and an attempt to improve oil pickup that was completely unnecessary.
It uses a hot rod Chevy style lowered oil pickup epoxied into the case.


another view of the right side case drainback screen and holes.


How crazy can you get? Screen is supposed to damp out oil impact and minimize splashback to the upper case.
Commonly used in sprintcar engines. Green stuff is 1838 3M epoxy... amazing stuff.

What else can you do to increase oilflow?

* Replace the unsealed bearing at the end of the crank with a sealed bearing. That will direct all flow coming to that cavity in the clutchside cover to the crank. This is by the way essential for a plain bearing crank to keep the oil pressure up to the 70-90 psi you need at full-on rpm. A sealed bearing has slightly higher friction loss than unsealed, but the trade is well worth it and the HP loss unmeasurable.

* Restrict the flow to the heads. Install smaller orifices in place of the dowels between the cylinders and heads that carry oil to the heads. Diameter (ID) we used for the roller crank was: 0.112 in.

* Block off the ball bearing relief valve behind the pump cavity.

* Some earlier cranks have a sleeve inserted in the end of the crank to restrict flow. Even the factory suggests removing that for racing.

* Last, and certainly not least in the way of work, is to redirect the head drains away from the rotating crankshaft. The stock head drains go through the cylinders and fall right on the crankshaft. All that oil hitting the rotating crankshaft just slows it down, somewhat like a waterwheel brake. That oil in-turn gets taken up by the transmission and this all both slows oil drainback and robs HP. Draining the rear head externally involves hacking up the fins in the cylinder so you can fit some Earl’s supply type fittings into the head drains and run braided SS lines down directly to the sump (see below).


oil line routing from rear head

* The front head is easier. You just weld a fitting on the exhaust valve cover, then run a line back to the case below the level of the crankpan (See picture below). In the overall scheme of things the oil just gets back down faster. Meanwhile you get the extra benefit of an engine that spins up faster because it is not fighting all that drainback oil. My bike in the photos shows how this is done. Others have done this too. The engine is definitely more responsive.


external drainline using stainless steel braided line. Also good view of rear header

The cases don’t take kindly to the pounding of increased HP. They will eventually crack around the main bearing webs. This is fairly well known. Inspect periodically. I fortunately have access to an excellent welder. He lives here in San Diego, and his name is Rob North. BSA/Triumph triple fanatics will probably recognize the name. I don’t honestly know if welding serious stress fractures of this type will work. Rob has welded my present slightly porous cases in a number of minor crack areas (See pictures below). They look considerably better there than the original castings! The base metal is not real good for welding. It is full of voids. You really need to have someone who knows what he is doing with aluminum welding.


welded up boss between fronthead cam drive tower casting and front cylinder casting. This is standard factory location, but this set of cases was porous, so Rob North had to fix them by filling with weld.


welding and remachineing between rear cam drive casting and rear cylinder casting. This was porous also and Rob North also welded up.

Crankshaft plain bearing conversion with Carillo rods
In my frustration with crank failures I finally decided to go with a serious plain bearing conversion. Fortunately, Ax had some leftover Carillo rods from a cancelled flattrack project for Yamaha involving the Virago engine, of all things. I am sure that this conversion can be done a number of ways, with a number of different rod sets and pin diameters. After all, plain bearing engines are the rule these days. There may be a few people out there that even do this conversion as a kit. But at the time this wasn’t common for Ducatis, and I had to figure it out as I went along. Basically there are a few things to consider.
  1. Pin:
  2. I designed a pin that mimicked the Yamaha Virago engine. Diameter is 1.9675-1.9679 in. Essentially the pin holes are horizontal (180 degrees) when the crank is out of the engine on the workbench and sitting on the counterweights. The pin feed hole is 0.200 in and the four holes to discharge to the rods are each 0.175 in. Now there has been a lot of research done by people on the best place to put the oil supply holes to counter the pressure of the piston on the power stroke and the hole diameter. I figure these people know a lot more than me. But I did reduce the hole diameters slightly (0.025) from Yamaha’s. I used 6150 steel, heat treated to Rc of 58-60 for the pin material based on a recommendation from Jorgensen Steel’s technical guru and my 75 year old German machinist friend Otto, a master craftsman, talked about in hush tones among LA machinists. Otto made the pin according to my drawing and San-Val, probably no longer in business, did the pin grinding after heat treating. There are certainly other steels that can be used. A sketch of the pin is available from me. (In Jan. 2002 the pin fractured cleanly in half. It was heat treated too hard, or more accurately, through-hardened when it should have been case-hardened. I also had not made sure that radiuses were ground into the step change diameters when it was originally made. Sharp corners create stress risers that ultimately lead to failure. I have made a new pin of 4340 case-hardened to Rc 53).
  3. Bearings:
  4. Since the rods and pin were set up for the Yamaha engine, I had a ready supply of bearings in five thicknesses. Basically any hot-rod automotive book tells you to set up the bearing clearance at about 0.001 per inch of pin diameter. This is a general rule of thumb.
  5. Rods:
  6. Carillo, one of the most highly respected names in performance rods. I used a shorter rod than the stock one. 5.550 in. vs. stock 5.709 (145mm). This gives a rod/stroke ratio of 1.89 and should give slightly better acceleration of the piston to achieve better cylinder filling compared to the 1.95 of the stock engine. Ax also had a longer rod, but suggested we go with the shorter one. The primary reason is to build toward the Ducati’s strength. That is torque down low. A longer rod will allow the engine to wind up higher, but reduce the torque at lower rpm. Since the Ducati is not a high revving engine by design, Ax suggested building for a broad power band. Now if you look at some of the books on car engines. Bill Jenkins, "The Chevrolet Racing Engine" and Waddell Wilson’s "Racing Engine Preparation, you will find a similar line of reasoning. I have found both of these books to be invaluable in learning about engines.

    This change also requires machining the base of the cylinder to compensate for the shorter rod. Since the difference in rod length was 0.159 in., then that should have been the amount that needed to come off the stock cylinders. Mine were a little shorter for reasons long ago forgotten. But I set things up for about a 0.065 in. deck height by machining off 0.149 in. See the Piston section for more on deck height.

    Normally big end side clearance is set up at 0.014-0.024 in car engines. Motorcycles appear to be able to tolerate a little less. The S-2 engine which I later copied in oil orifice sizes and big end clearances has a factory spec of 0.004-0.012 in. I eventually settled for 0.008 in. For comparison a 750 Paso is 0.0059 to 0.0138 in.

    One other aspect about using the Carillo rods is the need to machine the counterweights to clear the rod main beam. The Carillos are a lot wider than the stock rods (See pictures below). This requires serious amounts of balancing material (aka Mallory metal) to counteract that removal of metal. Also the cases need to be relieved at the bottom where the weir is located to allow for clearance of the rod bolts.

    Lastly, the rod assembly instructions from Carillo are absolutely first rate.


    assembled plain bearing crankshaft with Carillo rods installed in engine. View from front cylinder, showing external oil feedline fitting at left (in blue)


    plain bearing crankshaft with Carillo rods. Note that inner surface of flywheels are machined to clear rods.

  7. Oil Pressure:
  8. The main reason for tightening up the big end side clearance was to get the oil pressure up to the Carillo and Axtell recommended figure of 70-90 psi hot at maximum rpm. This should be about 10 psi/1000 rpm. So at idle, I read 10 psi. At cold startup the engine hits 90 psi and then settles down. The Ducati’s oil pump is designed for a roller bearing application so it is on the ragged edge for supplying the needed pressure. At first I blocked off the stock oil filter, but more recently because of concerns about getting oil to the bevel gears I put a small orifice (~0.040 in) in the base of the filter holder tube. I drilled into the boss cast into the cases in between the front cylinder and bevel tower drive (see below )


    fitting installed into external oil feed hole


    Clearer view of extra tubes welded to lower headstock and main frame backbone to control headshudder under heavy braking.


    oil line routing from rear head

    ran a braided SS line up to an Earl’s external oil filter holder that I mounted under the gas tank, down to an Earl’s oil cooler mounted to the front intake valve cover like a Monster (see below)


    mounting of oil cooler. Don’t recommend this method. It leads to cracks in the cooler. I have since switched to the Earl’s brackets that are available with their coolers.

    and then up to an external manifold with four outlets (see below)


    crowded picture showing manifold for external oil feed distribution to both heads, crankshaft, oil pressure gauge. Can also see oil cooler mounting, extra welded in frame supports to minimize head chatter during braking (Steve Wynne suggestion). External Earl’s mounting block with oil filter.


    complete view of Left hand side oil line routing and Vee-Two dry clutch.

    Two are –3 SS braided lines restricted with #65 pilot jets drilled out to 0.041 in. to feed the heads, one –4 braided line goes to an oil pressure gage mounted on my instrument panel and the last one, a braided line, goes down to the crankshaft via the clutch side housing. All this external plumbing requires machining the head cam end caps for –3 fittings and machining and welding the clutch cover for the oil feed line (see pix below).


    one way to weld a fitting into the clutch side case cover to install an external oil feed to the crank.


    the interior view

    You must also not forget to block off the internal case feed lines to the heads and crank. All of these dimensions and feed line changes were arrived at through an arduous process. I will not describe the details. I did use an external adjustable pressure relief valve at first to ensure that I did not overpressurize the system, but once I had everything working properly I removed it.

  9. Oil pump:
  10. I use the stock pump with the usual round screen strainer in the cases. The pump is speeded up with Albert Bajohr’s gears. Stock gearing is 22T on the crank and 51T on the pump, while Bajohr’s gears are 47T and 26 T, a ratio change from 1:2.318 to 1:1.808. I read recently in one of the Bevelhead posts that a Mille pump will fit the engine and has better capacity. That would be an easier fix. I will have to look into that myself.

I have disassembled the crank for inspection on numerous occasions after dyno runs and after a few thousand street miles. Everything appears to be working quite well.

Pistons
This is the key power element in the Bevel drive engine. In December 1980 CR Axtell let us bring our near-stock Ducati in for a dyno run. My dyno run sheets showed that with 40’s, Conti’s, stock valve sizes, porting and valve job by Jim Woods and standard Nippon-Denso ignition the bike made 59.6 HP. The dyno table is shown below. The first HP column is with stock cams. Later in January of ’81 we ran the bike again, but this time with the "Imola" cams. Note that the stock cams are actually better for street riding. They are geared to the engine’s strength—torque down low. The Imola cams give less power down low and only make a really big difference at and above 8500 rpm, dangerous territory for the crankshaft. I remember quite painfully ($$$) the result of racing with consistent shifting above 8500 rpm. Usually in the next race, the crankshaft would fail, taking the pistons with it.

Near-stock 900SS engine, CR Axtell’s dyno

Stock cam

Imola cam

Rpm

HP

HP

4500

47.0

42.3

6000

57.3

56.0

7000

59.6

58.0

8000

55.4

54.6

8500

45.0

54.2

9000

Not run

46.6

Note: all values are corrected to standard temperature and pressure at the time of the test

After seeing the unimpressive results on the dyno, Ax took pity on me. He suggested that some good high compression pistons would do wonders. He was right. I was to find out that this was typical advice from Axtell…being right, that is. Having had many a piston made in the US, Ax was not particularly happy about their longevity and quality control. He felt that the aluminum alloys used by the American piston manufacturers were not high enough in silicon and thus too soft, leading to dimension changes under the stress of racing. Ax suggested Europe or Japan, with Cosworth as a possibility. Luckily, Cosworth had a local site in Torrance in the early 80’s. I approached them and to my amazement they were willing to make me a few (3) prototype pistons. I drew something up and Ax corrected my misconceptions. I sent the sketch to Cosworth. It was an 87 mm diameter piston that increased the displacement to 883 cc. Ax was pleased. After some nervous moments on my part wondering if they would ever get here, Jim Woods installed the pistons (see below) and we ran on the dyno again.


original Cosworth pistons Axtell helped me design.


another view Cosworth pistons

Basically, just installing those pistons yielded 80 HP. Here is the chart again with the previous numbers side by side for comparison. The Cosworth pistons were tested with the Imola cams. I think I jumped all around the shop. When I got to Willow Springs Raceway, that change was worth 4 seconds a lap.

Stock cam

Imola cam

Cosworth pistons

Rpm

HP

HP

HP

4500

47.0

42.3

Not run

6000

57.3

56.0

68

7000

59.6

58.0

76.6

8000

55.4

54.6

80.2

8500

45.0

54.2

78.7

9000

Not run

46.6

72.5


Deck height: This and the clearance of the valves to the piston cutouts should ALWAYS be checked on first assembly of new pistons or cams. That includes any other change that might affect the piston-to-head clearance or piston-to-valve clearance on overlap. This is most easily done using strips of soft solder taped in place in the squish band areas and valve pockets. This allows something solid, but easily deformable that you can check with a micrometer and is not prone to the distortion you can get with clay. The closest I ever ran was .040 in. (1 mm). This is the usual stock setup, but if you are spinning your engine above 8000 routinely, you may get some piston to head contact from crank flex. There also may be a contribution from expansion from thermal growth of the aluminum, but that should be minor because of the steel studs and connecting rod. It doesn’t take much contact and your crank and pistons are trashed. I usually played it safe with an extra 0.010 (0.25mm) or more clearance. This can be adjusted by machining the squish band area of the piston or using extra or less base gaskets. Bill Jenkins book, "The Chevrolet Racing Engine" covers this. He indicates that deck can be increased to as much as 0.070 in. without a measurable power loss. I have run as much as 0.065 in. and didn’t see much difference, but these are much smaller engines and may be more sensitive. Waddell Wilson’s book, "Racing Engine Preparation" sets 0.040 in. as a hard limit. For what it’s worth, my best dyno runs seemed to coincide with tighter deck heights than 0.065 in., say 0.040 to 0.050 in. or so. But we were usually changing other things at the same time, so I am not sure that is the sole reason. Axtell’s personal opinion about this is to run as close to 0.040 in. as possible, even tighter is better. He used to run 0.037 in. with Triumphs. A lot depends on the flex of the crank. With the stepped pin of the later cranks, and certainly the plain bearing conversion I have done, that is eliminated to a large extent.

During the course of the year as more was understood about the piston and combustion chamber interaction, the piston dome shape was altered slightly to control detonation. The change was to radius the dome on the spark plug side. We were getting detonation because the sparkplug was too shrouded. This was seen because of the condition of the plugs and the fact that the base gaskets would blow out (ooze, really) of the cylinders. I remember almost getting a heart attack when Ax told me to file the dome and I carefully took about 1/64 of an inch off. He said "no", he would show me. He then took a huge aluminum file (large space between the teeth) and quickly filed away about 1/8-1/4 of an inch from the spark plug side of the dome. My Cosworths!! But it worked. The detonation was gone forever.

Later on I had Cosworth draw up an 88 mm piston. I still have the drawing for those pistons, but their price was too high for me to afford making up more than a few. I wanted to sell the pistons as a way of supporting my racing. After a while we tried JE, who in Ax’s opinion was the better of the American manufacturers. However, we introduced something that he had tried in the past with some success on his own projects. In the 50’s the aerospace industry was having a problem with post-forging distortion of aluminum parts. Some really bright people came up with a method of stabilizing aluminum parts such that the distortion would be minimized. Going into this in any detail gets into a technical area of structural mechanics and stress relieving that I don’t begin to understand. However, I did read the original published papers in a trip to Caltech’s library. (Being a chemical engineer does help somewhat). Basically the method involves using liquid nitrogen in combination with heat cycling to stress relieve the parts from forging and heat treatment stresses by a simple immersion process. There were a number of places that could be done in LA. It apparently works pretty well. We used it on the JE pistons prior to machining and that appeared to allow them to hold their shape considerably better than we expected. We checked skirt dimensions just about every second or third race. Too much clearance would lead to all kinds of problems with blow-by past the rings and poor heat transfer to the walls. Although the JE’s did not maintain their shape as well as the Cosworth’s, they were acceptable. By the way, liquid nitrogen stabilization has become the rage in the rifle industry to control barrel distortion.

A lot has happened in piston manufacture in the intervening 20 years. I should point out that the custom-made JE pistons I have in my Ducati now after 4500 miles show not a scratch on the skirts. Except for carbon on the domes and a nice tan color underneath, they look brand new. These were not liquid nitrogen stabilized. Apparently some progress has being made. My racing pistons never looked that good.

Returning to the development story... In the search for more and more power we increased the displacement with 88 and then 89 mm pistons (905 and 926 cc). Generally, I would say that with the pressed-in liners you really should limit the piston diameter to 88 mm. At 89 mm, ring sealing seems to become a problem, primarily because the liners get too thin and distort unevenly as they heat up. I tried many kinds of liners, boring with deck plates, stabilizing the cylinders, different ring packs etc. to try to solve that problem. Nothing was really effective. So we kind of gravitated back to the 88’s.

Unfortunately, JE’s blanks at the time were lighter at 89 mm than at 88. The 88’s required some serious milling of the interior of the piston and judicious grinding around the pin bosses to remove the excess material. The 89’s were considerably less work. Stock Ducati piston weight is about 353 g. I had high compression pistons ranging from 380 to 439 g. However I always lightened the over 400 g pistons to less than 400. Usually about 390 g was the best I could do. This was with proper training from Ax on what metal could be removed and what was sacred.

HP with the standard 80 degree bevel heads never exceeded 83-85 or so on Axtells’ dyno, however with both headwork and pistons we did manage to progressively broaden the power band. Strangely, I can’t seem to find any dyno sheets with numbers above 83 HP, but I remember hitting 85 at least once. It is difficult for me to separate the effects of headwork, piston design, exhaust system and ignition timing on HP from this point on, because we usually changed a lot of things at once. However, once the high compression pistons made the big improvement in HP, any of these other changes were really minor in comparison. I’ll get into that in more detail in the Heads and other sections. In retrospect it is kind of amusing to me that we tried so hard to find more power and the engine just did not really improve a lot more. Still it was faster on the track. Every little bit helps.

A few other items need more discussion:

Valve Cutout and Valve-to-Valve Clearances
You should have 0.060 in. (1.5 mm) on intake and 0.090 in. (2.29 mm) for the exhaust on overlap. These values are primarily to aid flow during overlap, not for piston to valve clearance purposes. You should also have 0.060 in. between the valves on overlap, although with the desmo head you can go lower. Ax has used as low as 0.040 on some spring heads. I remember that Dale Newton’s factory TT2 head had a valve-to-valve clearance of 0.025 in. Much of what you can do here is limited by material available under the cutout of the piston, cam design and the need to limit sinking the valve too far in the head. If the piston does not allow the clearances necessary, then sinking the valve(s) is an alternative. However in doing so, you run the risk of cutting down on the margin available for a good transition in the valve seat, which can negatively affect flow, as well as increasing head volume which lowers compression ratio.
Cylinders
I tried all manner of cylinder liners. LA Sleeve works ok, or did at the time. If severely damaged, the factory liners can be removed from the cylinders by heating in a >450 °F (232 °C) oven. New liners can be installed the same way. I think the best way to go is the S2 (Mille) cylinders. They are all aluminum with the Nikasil style surface. No conventional honing required. With this type of cylinder now found even in dirt bikes, there are places to get them resurfaced and diamond honed if something disastrous happens. The improvement in heat transfer from using a completely aluminum cylinder with no liner should be obvious. For those who want a detailed understanding, refer to any book on heat transfer, (Kern, "Process Heat Transfer" and Carslaw and Jaeger, "Conduction of Heat in Solids"). I will be glad to discuss this with anyone who really wants to know (trust me, you don’t). The key element in running the S2 cylinders is that the piston clearance must be tightened up substantially compared to the factory or aftermarket liners. JE recommends running 0.001 to 0.0015 in. cold clearance with Nikasil style cylinders. Normally with cast iron or steel liners, the clearance is more like 0.0035 to 0.0045 in., or even more, depending on the piston manufacturer. (They always make a recommendation with their own pistons). The tighter clearance requirement is because the S2 cylinder expands at the same rate as the piston, whereas the cast iron type liners don’t. Too much clearance once the engine warms up and you get a lot of blow-by into the case from the combustion chamber and a loss of HP.

One other issue about the S2 cylinders is the needed to run cast iron rings, or more specifically, NOT run chrome rings. Again, compared to 20 years ago, the piston manufacturers have this "minor" item figured out. JE uses Hastings rings and they seem to work well.

Compression ratio
We ran a variety of compression ratios by altering dome height, but mostly the engine turned out about 12:1. That is what I am running now in my streetbike. It might surprise you that American street gas (92 octane) can run without knocking, but Axtell has done dyno tests, which certainly stress an engine far harder than street riding or even racing. Present-day pump gas can handle that level of compression with no problem, and just as importantly, no significant difference in power compared to racing gas. I am not particular about the kind of gas I put in the engine, although Axtell’s testing seemed to show that Super Shell was slightly better than the others for power. Of course, he was limited to gasoline brands available in California. Until you get much over 12:1, say to 13:1 or so, race gas (>100 octane) just isn’t needed.
Heads
Of all the things to do for the engine this is by far the most work. On the standard heads, quite honestly, the smartest thing you can do for yourself is to get the best valve job you can find. Do not waste money on even cleaning up the ports. The HP gain is just not worth the effort. Basically the port design with the 80 degree included angle in the valves is just beyond correction without major surgery. Some people have tried "D" ports with some success. Jim Woods suggested that when we just weren’t making any progress with the standard bevel drive heads and we did install some aluminum D inserts in the exhaust ports. That helped some, but we are talking about really small improvements.

On Axtell’s flow bench the ports are tested with a simple manometer measuring pressure differential in inches of water that can be converted to flow. Many flow measurement devices use this principle. The conversion to flow is rarely done at Axtell’s because that is just another way of expressing the pressure differential. So most of the flow bench numbers I have are in inches of water. I do have some conversions, which I will show later in the section. But first let’s look at the best the standard, but modified bevel heads did for us on the dyno. I’ve also included my friend Tony Guest’s bike, a 750 SS (actual 905 cc), which was setup almost identically.

My notes show the engine with the following setup.

    • 89 mm JE pistons
    • Daeco race gas (104-105 octane)
    • 150 main jets front and rear
    • 30 degree spark advance (usually required because ring sealing was not so good on the 89 mm pistons)
    • Dyna ignition driven off the rear camshaft using a modified Lucas-Rita housing
    • Deck height 0.040 in (1 mm)
    • Megaphones (probably Bassani at that time, my notes aren’t clear)
    • 33 inch long, high-pipe headers measured around the outside of the bends
    • 40 mm Dell’ortos with long Dell’orto plastic velocity stacks and huge K&N air filters (no difference on the dyno with or without, but those expensive parts are protected)
    • Imola cams
    • Heads: 44 mm intake and 36 mm exhaust valves, "D" Port exhaust

’81 Cosworth pistons

July ‘84
K Bracken
(89 mm pistons)

June ‘84
Tony Guest’s bike
(88 mm pistons)

Rpm

HP

HP

HP

6000

68

69.6

67.0

7000

76.6

78.7

81.0

8000

80.2

82.0 (82.8)

80.6

8500

78.7

78.0 (82.5)

78.4

9000

72.5

Not run

78.0

Note: numbers in parentheses are the best peak HP that I have charts on

Not much gain for 3 years of work. But on the track the bike was definitely faster. We had some runs that peaked higher and some runs that had more power at 6000 and 7000 at different times over the years. My friend, Tony Guest’s bike, had very similar numbers from an almost identical setup. His was a 750 SS bored out to an 88 mm piston. Now we only really worked on the heads over the winter, which is mercifully short (nonexistent, well, it rains sometimes) in southern California. The rest of the time I was either rebuilding a broken engine or just maintaining and racing. So don’t be too misled about the time we all put into this. But we had to face that we had hit a wall. We weren’t going to get much more power out of the engine. So we decided to take the big plunge. This was about the time that Ax would say, "You know how to drink Scotch don’t you? Well, step over into my office." Which meant we were moving away from the bench where he did all of his porting work and over to another table about 3 feet away. This was where sketches would be made, serious discussions would be had about port flow, good designs (XR750 and Cosworth were "good" examples), poor designs (Ducati was the "bad" example) and yes, Scotch would be carefully sipped. In that state the time and effort needed to make a "good" port would not seem so out of reach.

Before I forget, here are some flow bench numbers. I will show a "stock" head or the closest thing I could find in my flow bench sheets and the best racing head we had. Everything is in inches of water. Only the Racing head was converted to CFM.

Standard Ducati Bevel Drive heads: Comparative Flow bench numbers

Valve lift
(0.001 in.)

Stock Intake
40 mm

Racing Intake
44 mm

Racing Intake Flow (CFM)

"Stock" Exhaust, (early porting work)
36 mm

Racing Exhaust, "D" Port
36 mm

Racing Exhaust Flow
(CFM)

100

0.20

0.25

N/a

0.13

0.13

N/a

150

0.43

0.63

N/a

0.38

0.38

N/a

200

0.80

1.00

107

0.50

0.63

85

250

0.88

1.25

120

0.75

0.88

101

300

1.12

1.50

132

0.75

1.00

107

350

1.30

1.63

137

0.88

1.13

115

400

1.40

1.75

142

0.88

1.25

120

450

1.50

1.88

147

0.88

1.38

126

500

1.50

1.88

147

0.88

1.38

126

Jim Woods had also tried a 42 mm Intake and 38 mm Exhaust arrangement with Tony Guest’s engine. The flow bench numbers were not much different and the dyno numbers as you can see from the dyno chart above weren’t much different either. The "D" port seemed to help the Exhaust more than the Intake, so we just did that to the Exhaust. We actually bolted a contoured piece of aluminum into the port to make the "D." Now it looks like we made some decent improvement on the flow of both the intake and exhaust. But you have to realize that Ducati Bevel drive ports are REALLY bad. Peak intake flow on a good Harley XR750 flat-tracker was about 3.30!! Ax typically gets in the range of 90-96 HP out of one of those engines (2-valve, 750). We had a very long way to go.

Axtell has a simple calculation to "predict" the performance of an engine on his dyno, based on peak flow as measured on his flowbench. According to that, my engine should have made about 84 HP. Pretty close, based on dyno testing.

Converting Belt Drive Heads to Bevel Drive Heads
First I should say that this project is not for the faint of heart. It is mega hours of work and could only have been done by me with Axtell’s and his partner Mike Libby’s guidance and considerable patience. The two-valve belt drive heads with their 60 degree included valve angle are an improvement over the standard bevel drive heads, but they still need a lot of effort to make them reach their potential.

Some of you may recall Dale Newton (and Paul Ritter) from about the same time as Cook Neilsen and Phil Schilling. (I was shocked to hear of Paul’s crash. I was in such awe of him when he was racing for Dale). Dale passed away only a few years ago. I am lucky to have counted him as a friend. We became close during this part of the story. Dale could make friends easily. Although a big man, he had an engaging, easy-going manner that quickly made you know you could trust him and he would be there to help you in any way that he could. And he loved Ducatis. About this time (’84 or so) the factory was making the TT1 and TT2 bikes. As I recall, the TT1 had the 600 Pantah engine, and for the TT2, the factory bored and stroked the engine to make it into a 750 with 88 mm pistons. Dale had a TT2 and was looking for still more power. He knew that I had a very good relationship with Ax, and Dale and I had talked at the races many times.

Dale had made some really good connections with the Ducati factory and his business occasionally took him to Europe. So he, Ax and I conferred and a wish list was made. For this project the two most important things Dale was able to get were 2 sets of belt drive heads taken from the line before any porting or valve seat work had been done and a copy of the machining print for the heads. That was amazing. I never dreamed that I would have a copy of that print. It certainly made things considerably easier as the project progressed.

First, Ax had me take a used belt drive head and port it out oversized in both the intake and exhaust. It was the first time I had ever held a porting tool. That was the patience part for Ax.

After some valve seatwork and initial evaluation, Ax told me we needed to reangle the valves to get bigger valves in and to aid in the recontouring of the port floors. This reangling of the valves moves the lower edge of the valve into the port floor. This creates more material to shape for a better flow contour. The print that I had from Ducati and a machinist that Ax used on occasion allowed this to be much easier than you might otherwise expect. We reangled the intake valves 1 degree and the exhaust 2 degrees by rotating from the tip of the valve in the normal closed position. We then put in some dummy valve seats sized for 44 mm intake and 37.5 mm exhaust (essentially Harley XR750 valve sizes. After that I spent many hours contouring the ports where Ax told me and cutting the seats with hand cutters where he showed me to. Eventually we moved to clay to establish the best contour for the port. All along the way as each change was made, I would measure things at the flowbench. Once we had things looking good, where flow was a little over 3.0 on the flow bench, it was time to move to the real heads.

That was where Frank Neimeyer came in. I don’t know how Ax found this guy. He was in his early 20’s. Ax used to call him "Frank Neimeyer, World’s Greatest Welder." Frank would take the heads and almost completely fill in the ports with weld. There was a hole about the diameter of a pencil when he was done. And the heads were still flat and undistorted. For those of us who consider welding a black art, Frank was just magical. He also filled the original sparkplug hole so we could install a 12 mm plug. In retrospect we should have done a 10 mm plug. That kind has become much more common.

I became one with the porting tool. You might have expected to find some voids in the weld material that Frank had put in the ports. No. It was solid. Now I was working for real, so doing everything right really stressed me out. I was so worried about messing up Ax’s spare Dumore that I bought one just to work on the heads. Once I got things more or less opened up we put in valve guides and seats for real and moved into the completion stage. Somehow I got both Dale’s and my heads done. That is, I did all of the rough work and then Ax would finish them up. His years of experience at this are not to be taken lightly. Along the way each change was tested on the flow bench (see below).


Front head, started out life as belt drive casting. Reangled valves, welded and ported heads, 44mm In. 37.5 mm Ex Grafted on bevel drive. 12mm sparkplug, bathtub shaped chamber. Fins removed from intake to minimize heat transfer to incoming charge.


gives an idea how much we raised the port. Notice the weld just outside the valve cover area. We also installed oversize guides


complete rear head showing , 45 degree fittings for external drainback to crankcase. Chamber is same approach as front head.


the rear head port came so high in the roof that it came through even on these heads. This is a common problem in porting the stock bevel heads. I filled in with epoxy. The front head was discovered during early porting and I had it welded. The rear head was not discovered until later. I was concerned about dislodging the valve guides, so I epoxied. Works fine.


view down the port with valve open

Dale’s heads were the easy part, basically a bolt-on when finished. Mine still had the problem of grafting on the bevel drive system and the cams. With the aid of the print and comparative measurements between the heads, it looked like the drive adaption would not be too difficult. However, novice that I was, I had no idea that the cams from the belt drive heads spun opposite to the bevel drive heads. You work on one engine all the time and you start to think that is the way all engines work. Not so. I remember driving home from Ax’s one night and it hit me-- the racing cams from the factory would not work. I had to have some cams made. Fortunately Jim Duer at Megacycle was able to make desmo cams. I meticulously graphed out (1 degree increments) the factory racing cams and asked Jim if he could match them, but in reverse. He said he could. Otto, my German machinist friend, came to my rescue as usual also, by making a free set of camshaft blanks that matched the bearings and spacings in the belt drive head, but had the drive end of the bevel head. I just gave him a sketch and he made them. Two of these (1 front, 1 rear) I sent to Jim Duer and he then made the cams. I still have the drawings for the cam blanks.

The drive adaption was actually a lot easier than I thought it was going to be. Fortunately with the ’84 heads (and for some time onward) the factory still retained the same stud centers as the bevel drive engine. Basically the job was a machine-to-match. The cam cap side required no changes. The drive side required that the rear head be machined to match the head face to centerline dimension of the bevel drive head and the front head required the same. Because of the construction of the heads, the rear head required the most material to be removed, while the front head only required that the small ring protruding from the tach-drive casting be removed. I used some 3M 1838 epoxy to fill in the extra space and bolted the drives up (see below).


view into port and adaption of bevel drive to belt drive head. Plugged and welded up tach drive.


drive side view of attachment of bevel drive to head. Bolt holes are same as standard bevel.

The holes for the lower socket head cap screws in the bevel drive housings were too wide and had to be slotted inward a bit to match the belt drive heads (see below).


mounting of bevel drive to belt drive rear head. Lucas Rita ignition drive nut in place of stock cam nut. Note the slotted lower bolt holes needed to get upper housing secured to head

Also because the cams had different shaft diameters at various points than an equivalent set of bevel drive cams I had to make up inner race bearing spacers as well. That was about it. I’ve never had any problems with the drive mechanisms.

Once I fired up the engine, needless to say, I was concerned about how stable things would be in the heads. The biggest problems lay with the cams. Apparently, to make the cams Megacycle just flipped over the lobes of the racing cam profiles. This doesn’t work because it slams the valves down on the seats too hard. I graphed out the new cam profiles and found out that this was the case. My valves would go badly out of adjustment in just one dyno run or one race. Years later, when I resurrected the bike, I called Brook Henry at Vee-Two. After talking about my situation I supplied him with all of my drawings and the existing set of cams and he supplied me with a new set of cams that has held valve clearance settings remarkably well.

Now what did all of this do? First I will show the flow bench comparison with the best racing heads and the heavily reworked belt drive heads. Then I will show what HP numbers I have from the dyno at Axtell’s. Unfortunately, we never converted the belt drive pressure differential numbers into flow numbers over the whole range. I just have peak flow at 177 CFM. But you can get the picture. It is interesting to see that the heavily reworked ports of the belt drive heads don’t really show much until 0.250 in. lift. Then things really take off. Peak pressure difference is 50% better than the best bevel head on the intake and >90% better on the exhaust. (Keep in mind that the conversion from pressure differential on the flowbench is non-linear, so the actual peak flow increased only by about 20% on the intake). Keep in mind that peak flow is not the whole story. It is area under the curve, i.e. flow over the whole range of valve lift. This was considerably better than before.

Valve lift
(0.001in)

Racing Bevel Intake
44 mm

Belt Drive Intake
44mm

Racing Bevel Exhaust
"D" Port
36 mm

Belt Drive Exhaust
37.5 mm

100

0.25

0.25

0.13

.25

150

0.63

0.63

0.38

.375

200

1.00

1.08

0.63

.68

250

1.25

1.56

0.88

1.00

300

1.50

1.90

1.00

1.43

350

1.63

2.18

1.13

2.06

400

1.75

2.45

1.25

2.50

450

1.88

2.68

1.38

2.64

500

1.88

2.82

1.38

2.64

The flow numbers are not quite up to an XR750 Harley, but a lot closer than they were before, and a vast improvement over the previous best racing heads. It looked like the effort had paid off, at least on the flowbench.

In the earlier part of this section I indicated that Ax had a rough calculation to predict the ultimate HP an engine would make for a given air flow from his flow bench. Based on that, Ax predicted that the engine should make about 100 HP. However, dyno testing the engine was frustrating. The piston rings seemed to refuse to seal. This created a lot of blow-by and forced me to run the engine with about 28 degrees of advance. When I moved to the S2 cylinders it got even worse. The breather hose would spit oil all over the back wall behind the dyno. Later on, as I described in the Pistons section, I discovered that the clearance with the pistons had been wrong. For the longest time I was convinced that the oil system was just delivering too much oil to the cylinders because of the high oil supply pressure and the oil rings couldn’t handle it. I couldn’t have been more wrong. Now that I have the engine working properly I haven’t been able to get up to Axtell’s in LA to do that "one last run". But here is a tantalizing glimpse of what is there. This is a composite of a number of runs done in 1986. Engine setup was with the Carillo rods, 89 mm JE pistons, Northwest Sleeve liners, Barnett wet clutch, Megacycle cams, 20w-50 Redline oil. The clutch slipped badly at 6000 and the breather blew oil in a constant spray. After each run the valves were seriously out of adjustment. Such is life. We had at least a 10 HP increase, but we also knew that the engine was not working properly and there was still a lot more potential.

July ‘84 K Bracken
(89 mm pistons)
’86 Belt drive head engine
(89 mm pistons)

Rpm

HP

HP

6000

69.6

76.2

7000

78.7

85.2

8000

82.0 (82.8)

93.3

8500

78.0 (82.5)

91.6

We ran the bike at a Sears point National BOTT race with Rich Oliver on board. He was in third place midway through the race and the crank sheared off at the keyway on the primary gear side. After repairing and trying a few other things to contain the blow-by, we had Danny Coe race the bike at Laguna. He finished in a strong 4th place. The overflow container could barely contain the oil from the engine breather during the race. This was obviously not a good thing.

During this time I also changed jobs and started working in a start-up pharmaceutical company, divorced and remarried, and started raising a second family. All of this made free time to work on the Ducati almost non-existent, and truthfully, I was getting burned out. I worked on the bike occasionally for a few more years. My last dyno run was in December of ’89. No real changes there. I then let the bike sit until ’96 when I decided to turn it into a street bike. Now that project is "complete" (never is really), I feel confident in believing that with no blow-by and a good dry clutch I have somewhere between 93 and 100 HP.

Post script. My wife has a 2001 Honda 996 Superhawk, good for 101 HP, and 11.0 seconds and 125.2 mph in the ¼ mile according to Motorcycle online. In roll-ons from 60-120 mph., I lose 1 bike length. (She is lighter than me!). Our bikes weigh very close to the same. Mine is 418 lbs. "dry" (actual, with oil, no gas) and hers is a claimed 423. So my resurrected ex-race bike appears to be performing quite well.