2000 cc Knuckster in the makin'...

I've been working on this 2000cc Knuckster for quite a while now
and I thought I'd post some progression.

So, the mechanical pie throwing contest so far;





5.250" stroke
3.850" bore
8" rods
.673" Redshift cams

The pics are from a while back, the carbs are all done now and so are the push rods.
...Working on the push rod covers right now.

Anders Nygren will be tickled to see one of his biggest dreams come to fruition !
John

I applaud all innovation, but there are reasons why this sort of thing isn't done more frequently.
1. there's no fin area relative to the amount of heat generated
2. it would take quite a bit of R&D to make the 2 intake ports flow the same considering the difference in entry angle.
3. an L with a huge, plain venturi and no accelerator pump? Without changing the venturi, the entire low to mid range will live off a series of transition holes leading from the disc toward the air-horn, but they have to be really big because the disc admits a lot of air at very small angles.
4. the combination of rotating the heads for clearance (about 20°?), and the new tappet positions vs. the original rockers make for lots of geometry problems. There are both lengthwise (2 blocks vs. 4 blocks) and lateral (rocker arm overhang vs. tappet lateral distances from the case center) angle changes, especially since (unlike pan or shovel) all 4 of the overhang distance are different. I don't know what your pushrods look like, but I suspect that the usual shallow dish and large ball aren't safe for sharp angles - the ball will escape the moment the lash moves. This really needs a small ball with 50% radius captured in the socket.
5. I'm assuming that the cams were made normally except for the swapped front cylinder X-I valve position (vs. I-X)?
6. what did you do about the rockers?

[quote="panic"]I applaud all innovation, but there are reasons why this sort of thing isn't done more frequently.
1. there's no fin area relative to the amount of heat generated
2. it would take quite a bit of R&D to make the 2 intake ports flow the same considering the difference in entry angle.
3. an L with a huge, plain venturi and no accelerator pump? Without changing the venturi, the entire low to mid range will live off a series of transition holes leading from the disc toward the air-horn, but they have to be really big because the disc admits a lot of air at very small angles.
quote]

Ha ha, yeah you sure are right about that there's a "few" resons why this sort of thing isn't done more frequently.

Ok, about the intake angles; the carbs are both mounted at exactly 14°, one up, one down
and since the floats in these are mounted from side to side, and not front to back, they are not as sensitive to be mounted at an angle angle, there's still a bunch of things to take into concideration...
Here's the intakes, and I've also done new venturis for them. Eveything in them is modified one way or another.I did run them on another motor at this very angles to see if it works, and it really does. These are the very first S&S carb (before the L-Series) I got a small business called "Rigid Hips" and I do external choke housings, idle tubes and other stuff for the early S&S carbs. http://www.giftorm.se/rigidhips.htm

I applaud all innovation, but there are reasons why this sort of thing isn't done more frequently.
4. the combination of rotating the heads for clearance (about 20°?), and the new tappet positions vs. the original rockers make for lots of geometry problems. There are both lengthwise (2 blocks vs. 4 blocks) and lateral (rocker arm overhang vs. tappet lateral distances from the case center) angle changes, especially since (unlike pan or shovel) all 4 of the overhang distance are different. I don't know what your pushrods look like, but I suspect that the usual shallow dish and large ball aren't safe for sharp angles - the ball will escape the moment the lash moves. This really needs a small ball with 50% radius captured in the socket.
5. I'm assuming that the cams were made normally except for the swapped front cylinder X-I valve position (vs. I-X)?
6. what did you do about the rockers?

Yes there's a bunch of geometry problems to look into to make this work.

The reason for rotating the heads is push rod geometry only.
And I had to move and re-angle all 4 holes in the rocker boxes.



And here's what the lifters look like after a little rigid adjustements.


However, after months and months of work all the pushrod angels are now pretty much perfect, not even close to critical at any point.
In the pic you can also see the new drilled oil support for the heads, in the cam cover.

The guy who helped me out with doing the cams first took one original Knuckle cam and measure that one up, then he took the original Sportster cam set and re-welded the Reshift lobes. Does this answer your question? ...not sure about what you mean exactly?

Here's the rockers;

Now you're really confusing me.
The guy who helped me out with doing the cams first took one original Knuckle cam and measure that one up, then he took the original Sportster cam set and re-welded the Reshift lobes.
Why did he need anything from the original cam?
The rear lobes are exactly where they belong (i.e., will time on on whatever centers they're supposed to have, and the intake and exhausts are correct), and the front lobes need to be swapped and re-indexed - but nothing to do with the original cam.

Intake port angle - good, it's just some parallax.

Very nice on the head rotation!

Still confused on the rockers - do you have details of what was done?
Sorry if you already know this, but adding rollers to the valve ends changes the geometry.
How were the ends re-clocked? Assuming no other changes, all the rockers need some angle removed (flattened), the exhausts more.
New ratio is?

very ingenious and some really nice work, but this seems to me to be so far into "because I can" territory ....

Now you're really confusing me.
The guy who helped me out with doing the cams first took one original Knuckle cam and measure that one up, then he took the original Sportster cam set and re-welded the Reshift lobes.
Why did he need anything from the original cam?
The rear lobes are exactly where they belong (i.e., will time on on whatever centers they're supposed to have, and the intake and exhausts are correct), and the front lobes need to be swapped and re-indexed - but nothing to do with the original cam.

Intake port angle - good, it's just some parallax.

Very nice on the head rotation!

Still confused on the rockers - do you have details of what was done?
Sorry if you already know this, but adding rollers to the valve ends changes the geometry.
How were the ends re-clocked? Assuming no other changes, all the rockers need some angle removed (flattened), the exhausts more.
New ratio is?

Hey, about the guy who did the cams, his name is Charley Karling and he asked me for an original Knuck cam, something about the og Knucklehead's overlap period. I never gave it much thought.
Sorry, no more pics of the rockers ...Rocker ratio on these are; 1.23:1. When I measured in everything about the rockers/pushrods/valves I used a splitted (cut open) rocker box, to see everything clearly. I've done caps on the valves for the rollers. The only anoying thing is that I have to do new tin covers since the rollers are way too high.

His questions suggests that he was either looking for the LSA or the overlap, but I'm not sure this data is at all useful considering how different the engine is from stock.
The roller tip has the effect of angling the arm up (probably closer to what's needed) because the new locus is no longer the contact pad radius but the roller axle, typically at least 3/16" higher, and the cap does as well.
The original "delta" angle between the rocker levers is different between 3 of the 4 knuck arms (1 was close enough that they decided to save some money).

In this engine, and dealing with the front cylinder, there's an angle for the valve stem to the cylinder axis: 40° CW from vertical.
There's the cylinder angle: 22.5° CW from vertical.
For the intake valve, these are added to get the angle of the longer valve-side lever: 62.5° CW from vertical.
For the exhaust valve, they are subtracted: 17.5° CW from vertical.
For the rear cylinder, same numbers but reversed and CCW.

The shorter pushrod-side levers are nominally parallel to the cylinder in the 4-cam engines, or 22.5° CW from vertical, but in the knuck engine these levers pointed at their cam lobes, which were almost centered on the crankshaft (the front tappets have less angle than the rears). The angles for all 4 levers are quite different from the XL, and also different from each other (but fairly close in pairs of mirror image).

Add to this the fact that "ideal" (mid-lift) geometry requires that both levers of each rocker be at 50% motion when at right angles to their target, more delta rotation is needed for the change in ratio, and even more for the fact that the lift is greater than stock. If the stem and the pushrod were both vertical, the lever angle would only be 180° before the corrections if the ratio were 1:1 (like the original).

This can be plotted by trigonometry using known data except for the exact height of the rocker shafts, and the X and Y locations of the rocker sockets - which you have, but as you can see it's complicated.

His questions suggests that he was either looking for the LSA or the overlap, but I'm not sure this data is at all useful considering how different the engine is from stock.
The roller tip has the effect of angling the arm up (probably closer to what's needed) because the new locus is no longer the contact pad radius but the roller axle, typically at least 3/16" higher, and the cap does as well.
The original "delta" angle between the rocker levers is different between 3 of the 4 knuck arms (1 was close enough that they decided to save some money).

In this engine, and dealing with the front cylinder, there's an angle for the valve stem to the cylinder axis: 40° CW from vertical.
There's the cylinder angle: 22.5° CW from vertical.
For the intake valve, these are added to get the angle of the longer valve-side lever: 62.5° CW from vertical.
For the exhaust valve, they are subtracted: 17.5° CW from vertical.
For the rear cylinder, same numbers but reversed and CCW.

The shorter pushrod-side levers are nominally parallel to the cylinder in the 4-cam engines, or 22.5° CW from vertical, but in the knuck engine these levers pointed at their cam lobes, which were almost centered on the crankshaft (the front tappets have less angle than the rears). The angles for all 4 levers are quite different from the XL, and also different from each other (but fairly close in pairs of mirror image).

Add to this the fact that "ideal" (mid-lift) geometry requires that both levers of each rocker be at 50% motion when at right angles to their target, more delta rotation is needed for the change in ratio, and even more for the fact that the lift is greater than stock. If the stem and the pushrod were both vertical, the lever angle would only be 180° before the corrections if the ratio were 1:1 (like the original).

This can be plotted by trigonometry using known data except for the exact height of the rocker shafts, and the X and Y locations of the rocker sockets - which you have, but as you can see it's complicated.

Hey, about the rockers, in the previous pic they're still 1:1, I cut the ball end to 1.23:1 after the rollers was in place.

Here's the "Evil Mother's" jig (that I later copied) and even thought I'm only using 2 of the 3 different different Knuck rockers, they are now dialed the same.

About the angles you're describing, I started getting in to some (not all) of that stuff when I first was starting out with this project. And needless to say this is all one big compromise but as long as the motor works within the frames of what's acceptable I'm all for it...

Here's a pic of the rotated valve outtakes in the piston dome...

There are 2 categories of rocker geo in this.

#1 is the stuff you have to live with, which includes the cross-wise pushrod angle due to overhang, the fact that the tappet centers are different from the rocker centers on the same cylinder, and the lengthwise (crank rotational axis) pushrod alignment. You can make it better, but it's still there.

#2 is things that can be erased and made to function normally, specifically the geometry at the rocker itself.
To get the long (valve-side) lever in the right place, you have to seat one valve and place the roller tip on the stem end, then open the valve to mid-lift (lobe height × 1.23 rocker ratio × 50%).
Normally, the choice of what percentage of lift to reach by 90° is between 3 alternatives:
1. Miller's "mid-lift" method [lever positions at start & finish are equally above & below horizontal]- 1/2 lift (largest area)
2. high-pivot Jesel/Johnson/Titan method [lever starts pointing too close to horizontal, ends pointing too far down] - roughly 1/3 lift (faster opening, slightly less total area)
3. low-pivot reduced thrust load method [lever starts pointing too far up, ends pointing closer to horizontal] - roughly 2/3 lift (lower guide thrust at high spring load, slightly less area, and timed too late)
In H-D engines, I would not use anything but 50% because the tappet offset already makes tappet motion BTDC faster than ATDC, and any more change makes me nervous.

To do this, a line through the shaft center and the roller axle should now be at 90° to the valve stem. Your jig appears to use the contact point of the lever, which is error with a roller, and places the lever's effective end 50% of the roller diameter too high (typically 3/16").
If the angle is off, try to measure it using trig and at least get the same angle for the same valve on the other cylinder. If the angle has passed 90°, it's not too bad. Not reaching 90° must be fixed.
The correction is to move the shaft up or down, very difficult with a knuck (and moves both rocker the same amount). Since the box position is fixed in 2 places (both the shaft centers and the casting thickness), the box can be moved on the horizontal ears easily (shim up or machine down), but moving the shafts in the vertical ears requires welding, offset bushing, eccentric adjuster, etc. Much easier to add a lash cap, or change the valve length (and perhaps the spring).

To get the short (pushrod-side) lever in the right place, you have to clock the lever's socket end to align with the tappet (which is raised to 50% of lobe height), and this is where all 4 levers will differ.
Yes, the pushrod will still have a dog-leg, because the tappet itself is still at 22.5° to vertical (rather than clocked to aim at the rocker, as the knuck was designed).

PLEASE: do not take this as instruction for any other engine, even another knuck. This one has special features, and a pan or shovel is even more different.

That's great tips, I'm on it... We'll see what happens. I'll keep you posted.