Thursday, June 17, 2010

Cone Knuckle Part 2

When we left off in part one, we had looked at and discussed four of the five issues relating to running Knuckle heads of a Shovelhead Cone lower end. We talked about rocker arm ratios, lifter blocks, cams and pushrod tube angles. That leaves the toughest issue for last; oil return from the top end.

Perhaps a review of how the stock Knuckle oil return system works is in order. Starting on one end of the system, the Knuckle breather gear has a hole that is not found in any later breathers. This hole is timed to provide crankcase vacuum to a passage (also unique to Knuckle cases) which connects to the Knuckle lifter blocks. That is what the fifth hole in a Knuckle lifter block gasket is for. In this way, crankcase vacuum is applied to the pushrod tubes and the aluminum housings which give the Knucklehead its name.

Note that this passage in the Knuckle lifter block goes to the top of the lifter, with no drain hole into the gearcase such as on a Pan or Shovel. In fact, if you have replaced your Knuckle lifters with any aftermarket lifter (with the exception of the ones from S&S, which are made correctly) there is a good chance that you have a hole drilled through the lifter which will bleed off the vacuum that is needed to evacuate the heads. This hole is not supposed to be there, and could very well be the root cause of your smoking and oil consumption on your stock Knuck.

The reason that this vacuum in the pushrod tubes and Knuckle housing is critical, is that the 'spring cups" depend on that vacuum to "suck" the oil, through the 3/16" O.D. Steel lines uphill into the housings. If you have ever had a Knuckle head apart, you are no doubt familiar with the seals that are between the aluminum housing and the tin covers. These not only seal the assembly from leaking to the outside, but they also separate the two sides of the rocker arm to keep vacuum on one side only.

Compare this to the oil return on a Shovel, Pan, or Evo. All of them return the oil from the heads via a passage through the cylinder directly to the crankcase. They basically depend on gravity aided by crankcase vacuum when the pistons are on the upstroke (or crankcase vacuum aided by gravity, if you prefer).

So, what are you to do, if you have a set of Knuckle heads, and a Cone lower end you are dying to put them on? It looks as though there are a few possibilities. One which I have performed is to rework the steel lines coming out of the spring cups so that they don't run uphill as much as in the stock application. From there grafting on soft lines and routing them externally down to the crankcase roughly adjacent to the timing plug converts the drainage from the spring cups into the gravity/crankcase vacuum style like Pans and Shovels. The oil from the aluminum "knuckle" returns down the pushrod tubes.

A second method comes directly from a comment received on what I posted in part one of "Cone Knuckle." Richard writes that he has been running a Knuckle-Cone motor since 1986. He uses Shovel lifter blocks with a brake tube brazed to the intake pushrod tube and plumbs this into the fitting on the crankcase that returns oil from the primary in stock configuration. This puts timed vacuum to the heads much like the stock system and so avoids any mods to the spring cups, etc. I think this is an absolutely brilliant idea, and though I had considered the idea of using the primary return, my thought was to go directly from the spring cups to the primary return. Richard's system is much simpler and cleaner! Thank you Richard!

But, now there is one more viable option. It turns out that I have a reader who works for S&S Cycle. After reading part one, he informed me, and got permission for me to break the news, that S&S is introducing a complete Cone Knuckle engine this summer. The S&S version will be called the Kone Knuckle, and will feature appropriate passages and machining to use Knuckle lifter blocks, and hence a stock style oil return system. Plans are to also make the cases available for those of you who have a bunch of Knuck parts stashed away for just such a project. This will make the whole process relatively painless.

Keep in mind though, that excess clearance between the Knuckle rocker arms and shafts can be a source of excessive oil in the top end, enough that even an otherwise correctly working return system will not keep up with. A rebuild of those parts may still be in order.

So, just one last thought. Whoever said that Cone Knuckleheads couldn't/shouldn't be built ....well, aren't you glad that neither Richard or S&S listened to them.

Wednesday, June 9, 2010

Cone Knuckle Part 1

Knuckle heads on a cone style Shovel bottom end. Most have seen one, so there is not much debate as to whether it can be done, but there is much disagreement as to whether it should ever be done. Those who say it shouldn't be done generally fall into one of two camps. The purist, who believes that it is sacrilege to desecrate a set of Knuckle heads by running them on anything but a set of Knuckle cases, and the quasi engineers who see nothing but disaster when a part does not work exactly as the designer envisioned.

On the other side of the issue are two likewise divergent camps. On the one hand you have the less-than-a-shade-tree mechanic who will throw together nearly any combination of parts that will physically fit together and leave the "making it work" for the next owner. On the other is the wrench who does his best to weigh the issues and address them in such a way that the finished product is functional as well as unique.

Having raced a Knuckle top/ Cone bottom drag bike for a number of years, I can attest to the fact that those who say that there are too many cam/rocker arm/geometry problems for a motor like that to run well are wrong. Two AMRA National Championships would suggest that they are very wrong. In fact, performance is easily addressed by standard methods. Making the motor streetable in regards to the oiling system is a much more difficult proposition.

There are a number of things to consider before building a Cone Knuckle. In no particular order, they are:

1. Different rocker arm ratio. Knuckle rockers are 1:1. Shovel rockers are 1.43:1. That means that a cam that lifts the valve .450 in a Shovel will lift it .315 in a Knuck. (.450 divided by 1.43)

2. Knuckle lifter blocks don't bolt onto Cone cases. Some of the bolt holes are close, but no cigar, and there is no oil port in the case to match the one in the Knuckle lifter blocks.

3. A Knuckle cam will not fit in a Cone cam cover (at least without major modifications).

4. The angle of the pushrod tubes is different between a Knuckle and a Shovel or Pan making it difficult to seal the bottom of the pushrod tube to a Shovel/Pan lifter block.

5. The oil return system from the heads is completely different between a Knuckle and a Shovel, Pan, or Evo.

Looking at the rocker arm ratios first, there are a number of ways to approach this. One is to have a cam custom ground with Knuckle profile lobes for use with Shovel lifter blocks. This solves any and all valve timing and lift discrepancies, but will be expensive and time consuming.

The second method would be to select a high lift Shovelhead cam. Stock and mild performance Knuckle cams are from .350 to .370" lift. Taking into account the Shovelhead rocker ratio of 1.43:1, you can see that a Shovel cam that is advertised as .500 to .530 lift will open the valve .350 to .370 in a Knuck. If you consider the S&S 514 Shovelhead cam, its .514 lift translates to .359' lift with the Knuck's 1:1 rocker ratio. Compare its other specs with two Andrews Knuckle cams; the N is considered a stock replacement and the S a performance cam.

S&S 514 - IO 23/IC 43 (246 duration) and EO 43/EC 23 (246 duration)
Andrews N - IO 13/IC 41 (234 duration) and EO 44/EC 16 (240 duration)
Andrews S - IO 27/IC 55 (262 duration) and EO 55/EC 27 (262 duration)

As you can see, the S&S 514 falls nicely between the specs of a stock and a performance Knuck cam. Seems like a no brainer to me that this (or a similar cam grind) would be a viable choice.

Using Knuckle lifter blocks on a Cone lower end, while possible, would require quite a bit of welding and machining as well as a custom ground cam (or even more machining to fit a Knuck cam to a Cone cam cover). Not really practical unless you are a machinist by trade and just want to show off.

If you have by now settled on using Shovel lifter blocks, the next issue is the angle at which the pushrod tube meets the block. Now back when I was racing, this was a non issue, since I ran no oil to the top end, and no pushrod tubes. This lack of pushrod tubes did, however leave me with a couple memories that will give a clue as to the solution to the problem of sealing them up. On my Cone Knuckle drag bike the pushrods appeared to be perfectly in line with the lifters. I mean straight as an arrow (at least to the naked eye).This was especially noticeable in comparison to the Shovel dragbikes which were its contemporarys. The pushrods on a Shovel were at a fairly drastic angle coming off the top of the lifters. One might never notice that if they have never looked at a Shovel motor with pushrods installed, but no covers. The reason is because the Knuckle has four different lengths of rocker arm so each one aligns with its corresponding lifter. Not so on a Shovel (or Pan, or Evo, or even Twin Cam). The misalignment on these "modern" motors is compensated for by the ball and socket arrangement on each end of the pushrod.

If you look at a Shovel lifter block, you will notice that the "seating surface" for the cork seal is NOT perpendicular to the bore for the lifter. Well, one of them is probably perpendicular, but the other three are canted to line the tube up with the rocker/box. I have not had a chance to try this, but I suspect that aligning the pushrod tubes on a Cone Knuck may be as simple as remachining the seating surface in the blocks so that they are perpendicular to the lifter bore.

Finally, we come to the oil return system from the heads. Oil return from the heads is the biggest obstacle in making the Cone Knuckle streetable. Since it is fairly involved, it will have to wait for another post - coming soon in part 2.