Another Knucklehead Build

Home Model Engine Machinist Forum

Help Support Home Model Engine Machinist Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I received an email asking for more detail about how the throttle spring was installed. I've included a photo showing the top half of the throttle just before it was (re)assembled to the carb body. The moly grease makes it hard to see, but the spring is wrapped around a stub shaft temporarily inserted through the top half of the throttle. A straight pin is holding one end of the spring under tension. During assembly, the top half of the throttle is carefully centered over the bottom half before being pushed down over it. As this is being done, the spring slides off the stub and onto the throttle shaft, and the stub is pushed out through the top. A slight chamfer on the top of the throttle shaft improves the chances of this happening on the first try. With the two halves together, the pin is pulled, and the spring is released onto a stop in the bottom half. After tightening a grub screw in the top half, the butterfly keeps everything together. It's really not as hard as it sounds.

There's still a handful of small parts needed to finish up the carburetor, and for the most part they make up its various adjustments. I thought it best to have the air cleaner in hand before going any further with them since it could affect their accessibility, and some extensions might be required.

My experience with air cleaners on model engines has been that they seem to create more problems than they solve. My engines aren't run for very long and certainly not under 'severe service' conditions. High revving RC aero engines seem to get along fine without them. The foam elements that I tried years ago on my Howell V-4 changed the carb's air flow characteristics enough to affect its tuning. This would become a real annoyance if the air cleaner had to be removed to access the carburetor's adjustments. In any event, the air cleaner on this engine won't have a filter per se - just a cosmetic polished aluminum cover. Eliminating the filter element and its backplate will also allow the cover to be moved inboard.

Harley's stock air cleaner for the full-size Knucklehead was a simple circular dome. I like its clean looks and was looking forward to turning a scaled-down version. My CAD model, though, showed its diameter would have to be a little too small to avoid blocking access to the main jet on the bottom of the bowl. To get around this, I stretched its ends into a more interesting oblong shape. Actually, similar air cleaners were available later on Harley Panheads.

The cover was machined from a block of aluminum in two different setups, and its exterior was polished. The photos show some of the intermediate steps. A single cover's machining time was close to five hours due to my unreasonable obsession with its interior. The interior was shelled out to create a .050" thick wall as well as a pair of integral standoffs for bolting the cover to the face of the carburetor. With the air cleaner in place on the engine, though, access to the carburetor's adjustments don't seem to have gotten any worse. - Terry



496.JPG
497.JPG
498.JPG
499.JPG
500.JPG
501.JPG
502.JPG
503.JPG
504.JPG
505.JPG
 
The photos below show what happened when I was having a couple of those days where I should have stayed out of the shop. The machining of the air cleaner cover in the previous post didn't go as smoothly as it may have sounded.

When the exterior machining of the first workpiece (the one on the left) was completed, I noticed a minor gouge on its top surface caused by a quirk in my CAM software. I've learned to check for this particular problem before cutting metal because it shows up in the displayed tool paths, and I can usually move a few parameters around to eliminate it. It occurs so infrequently, though, that I sometimes I forget to check.

I decided to try to save the workpiece by modifying an unimportant part of the design and re-milling the flat portion of the front surface another .005" deeper to remove the gouge. Since I didn't notice the gouge until after I had removed the workpiece from the vise, I had to re-reference it. This shouldn't have been a problem since I knew the coordinates of its two already drilled holes. Using a spindle microscope, I centered the spindle over one of them, but forgot the minus sign on the hole's x-coordinate when I initialized the control program with the new location.

When I started the machine, it promptly cut a much bigger gouge in the part as the cutter moved across the top of the workpiece to where it thought the new operation was to begin. If I had been thinking clearly, I could still have recovered the part by re-machining a whole new top, and this would have saved the workpiece and the machining time already invested in its periphery. The workpiece had plenty of extra stock on its bottom that would have allowed this, but instead I tossed it into the scrap pile. At the time I didn't know that the scrap pile was the safest place for it.

On the right of the photo are the remains of the second workpiece. Its new exterior g-code completed as expected. However, for one of the interior operations, I told the software that the diameter of the finishing cutter was .124" instead of its actual .249". (For some reason, my brain still stumbles over the differences between those two particular numbers.) Anyway, I started the machine and left to do some yard work. When I returned, I found a big pile of chips and a tiny carcass laying in the chip pan as well as a pair of .012" deep gouges in the faces of my vise jaws. Fortunately, the jaws are removable, and I spent the next hour or so regrinding them so I could begin machining a third workpiece.

The third time was literally a charm, and the workpiece made it through all five ridiculous hours of machining. I bolted the finished cover to a piece of wood that I had band-sawed for a close fit around its periphery so I'd have something to hold onto while polishing it. Unfortunately, the only long-enough 2-56 screws that I had on hand were SHCS's. After prepping the surface with 1000 grit paper, I took the part over to my buffer and loaded up the sisal wheel with some red rouge for final polishing. I was nearly finished when the wheel probably caught the head of one of the screws and ripped the part from my hands. The cover was flung around the inside of the wheel's circular dust collector before being launched over my right shoulder.

I was frozen for a few seconds but then started laughing hysterically. I went looking for the part and found it in a waste basket half full of discarded paper towels. I couldn't believe my luck, and my brain started questioning whether the previous few minutes had actually occurred. Although the piece of wood was pretty badly scarred, there was only a single barely perceptible mark on the part itself. Although in a different position, which will now become the bottom of the cover, the mark is strangely similar to the gouge that I started with. - Terry

506.JPG
507.JPG
508.JPG
 
Very entertaining post, but could not resist laughing... at your expenses.
Do not feel bad it took me 4 trial to make the Hoghlet cam.
1) Perfect job, just mirror image
2) I was going so fast I kept going and wiped out the lobe, a nearly cylindrical cam is pretty useless.
3) The center lobe is supposed to be twice the width but I goofed up and made it the same width as the others.
4) You know the universal saying: 4 times is a charm.
 
I have to laugh. Because so may times at shows when I say I've made something using CNC I get attitude (usually silence) that implies it was cheating and that made it easy. In fact CNC is not just slapping a piece of metal in the vise and push "go". I usually spend much more time drawing the part, generating and checking Gcode than it would have taken to make the part manually. In fact to take advantage of the CNC process, holding / clamping the piece can be a lot more difficult than a manual operation where you might be able to stop and move clamps around. My point is there is a lot of thinking that goes on when doing CNC and things can go bad really fast. I've made the same mistakes as you point out multiple times. Just a quick setup to "fix" a problem thinking you have it set so the original code will take care of it. NOT. And yes I've also scrapped a few parts after they are perfectly machined getting them through the finishing touches. LOL
Nice work.
 
I have to laugh. Because so may times at shows when I say I've made something using CNC I get attitude (usually silence) that implies it was cheating and that made it easy. In fact CNC is not just slapping a piece of metal in the vise and push "go". I usually spend much more time drawing the part, generating and checking Gcode than it would have taken to make the part manually. In fact to take advantage of the CNC process, holding / clamping the piece can be a lot more difficult than a manual operation where you might be able to stop and move clamps around. My point is there is a lot of thinking that goes on when doing CNC and things can go bad really fast. I've made the same mistakes as you point out multiple times. Just a quick setup to "fix" a problem thinking you have it set so the original code will take care of it. NOT. And yes I've also scrapped a few parts after they are perfectly machined getting them through the finishing touches. LOL
Nice work.
Those who say it's easier on cnc, just dont know what they are talking about.
 
The carburetor's Venturi is a separately machined aluminum piece that is Loctite'd inside the carb body. There isn't much clearance between it and the fully open butterfly and so, to be safe, it was positioned using a depth stop turned from Delrin. I removed the Alodine from inside the carb body with a medium Scotchbrite pad because I was uncertain about its compatibility with the Loctite. After curing, the hole for the main jet was drilled through the bowl top and into the bottom of the Venturi.

While the carb body was still set up on the mill, the hole for the idle jet pickup tube was drilled and a short length of .040" i.d. stainless tubing pressed in place. This completed the carb body's machining.

The components of the choke assembly were machined next. Their machining and assembly were much simpler than those for the throttle, and I was able to scrounge up a suitable detent spring without having to wind one. Two shallow spherical holes were drilled in the underside of the arm using a ball end mill in order to provide positive stops at full-on and full-off choke. The choke will come in handy during cold starts since, with an air cleaner covering the carb's intake, finger priming won't be possible. If needed, the spring-loaded ball should be able to hold the arm in any intermediate position even under engine vibration.

The main jet machining provided an opportunity to use a new (to me) technique for lathe turning long skinny parts. It involves taking axial cuts that don't produce first order radial forces that create tool and part deflections and their resulting chatter and taper issues. I learned about this technique in a video that Youtube 'recommended' for me a few weeks ago:



I reshaped a worn-out CCMT carbide insert in an SCLCL toolholder that I axially aligned to the headstock in order to approximate the cutting action produced by the special tool used in the video.

For the main jet's first machining step, I needed to reduce a half inch diameter 303 stainless rod to 1/8" over a length of 0.8 inches. Using this tool I accomplished it in only two passes - something that's nearly impossible on my lightweight lathe with conventional turning. Chatter was easily squelched by increasing the axial feed rate. The surface finish was very acceptable, and the taper was only a tenth or so. I've included a photo of this tool making the first pass on the main jet's workpiece. The camera shot required a flash that created some misleading shadows, but the insert's cutting edge is truly perpendicular to the spindle axis.

During earlier testing, I found it best to grind away all traces of the insert's nose and leave only a sharp corner in order to guarantee there'd be no radial cutting for near zero taper. Back clearance was also required to prevent the insert from rubbing against the workpiece behind the cut. Since I typically can't take deep radial cuts on either of my lathes, I have lots of worn-out inserts with virgin side edges that can be used in this operation. This operation would seem to be an efficient way to machine valves. A similar video in the same series applies it to turning extremely thin wall tubes.

The remainder of the main jet machining was routine except for a dozen .032" holes that were drilled radially around its 1/8" barrel. Even though I used a sensitive drill feed, it was difficult to find a sweet spot in the feed rate to get around stainless' tendency to work harden. I destroyed a couple (HSS and carbide) drills before I was done. Stainless, even 303, was a poor choice for this particular part - 12L14 would have been much easier to work with. Those innocent looking holes seemed a lot bigger on my computer screen.

I could probably have axially machined the jet needle assemblies from a piece of round stock, but for the delicate tapered portions I wanted the toughness of a high carbon sewing needle. My wife's hobby is quilting/sewing, and with the time I spend with her in notions stores I'm always on the lookout for things I can use in my own hobby. I chucked a number 18 darning needle in the lathe and re-tapered its end using a green silicon carbide dressing stick. A .046" hole for the needle was drilled in the end of a 4-40 SHCS that I used for the threaded body. In truth, there was a bit more to it. I agonized over the taper angle and its relation to the pickup holes in the jet's body before finally deciding I really didn't know what I wanted.

The idle jet needle is much shorter and was machined similarly, but it used a .020" diameter straight pin in the end of a 3-48 SHCS. The tiny hole in the end of that screw wasn't a lot of fun either, but by this time I knew to not use stainless. Although most Loctite retainers could probably have been used to hold the needles in the screws, blue thread-locker was the only one for which I could find specific recommendations for use in liquid gasoline. - Terry



509.JPG
510.JPG
511.JPG
512.JPG
513.JPG
514.jpg
515.JPG
516.JPG
517.JPG
518.JPG
 
Last edited:
...technique for lathe turning long skinny parts. It involves taking axial cuts that don't produce first order radial forces that create tool and part deflections and their resulting chatter and taper issues.

I saw the same video & tucked away the technique. One thing he didn't touch on - is there some rule of thumb relationship between the minimum base stock diameter & the finished small diameter? In your case you used 0.5" stock. For example do you think 0.375" stock achieve the same results on that small diameter & 0.8" stick-out combination?
 
Petertha,
I don't have enough experience with it yet to say for sure, but I wouldn't see why not. As long as the forces are all axial, I wouln't see why the starting diameter would matter. - Terry
 
The forces aren't all axial. This method simply takes advantage of the stiffness of large diameter stock. Virtually all the cutting force is taken up by the large stock, instead of the thin part. Attempting to make the same cut with the stock 12" from the collet would have the expected problems. Any taper in the small diameter part is still due to stock deflection from cutting forces. You'd have even better results using smaller nose radius, smaller feed, and a lot of top rake on the cutting edge.
 
Before setting the carburetor aside, I needed to know that it was basically functional. For a sanity check, I filled the bowl with isopropyl alcohol and blew 15 psi air through the carb's body. While exercising the throttle and needle adjustments, I looked for wet spray patterns on a paper towel at the carb's output. The low-flowing part-throttle patterns were difficult to see, but with a finger across the carburetor's output I was able to discern the difference between wet and dry air. I mainly wanted to make sure there were no chip blockages, that the needles had some semblance of control, and that each could be completely shut off. I wasn't expecting these tests to tell me much about the sensitivities of the adjustments - that will have to wait until the engine is up and running.

For the most part, the adjustments behaved as expected except for that of the main jet whose needle was shutting off the fuel much too early. This would have been a problem because of the limited space below the bowl for its travel. Since the taper I used was probably a bit too lazy, I chucked the needle assembly in the lathe and re-worked it with a dressing stick.

The last step was to set the voltage for the fuel pump. After some experiments, I added a half-inch long .035" diameter restrictor in the line to the carburetor. This limited the pump's more than adequate capacity inside the recirculating fuel loop. The restrictor allowed me to regulate a constant fuel level in the bowl over a rather broad range of pump voltages. The final voltage was selected to keep the bowl 80% full while unattached to the engine.

I got tired of looking at the piece of blue tape I'd been using to hold the far ends of the exhaust pipes together, and so I finally machined a proper collar for them. I wasn't able to come up with consistent measurements for the design of its i.d., and so I made an array of four parts with dimensions that I incremented by .002". After machining them in cookie cutter fashion, I trial fitted each one and found that the two middle parts best fit the pipes. I needed only one, but finish-machined both. Because of their nearly oval cross section, a clamp milled from a piece of wood was used to safely hold them in a vise for their backside machining. When completed, a setscrew was added to their rears that will slightly spread the pipes and hold them tightly in the collar under engine vibration.

I'm finally out of excuses for not returning to the point where I should have started work on the dreaded pushrod cover assemblies several months ago. There isn't much else left to do but to now go back and face them. - Terry


519.JPG
520.JPG
521.JPG
522.JPG
523.JPG
524.JPG
525.JPG
526.JPG
527.JPG
528.JPG
 
Beautiful work as always Terry ! I really look forward to your updates.
And while I know you have been putting off the pushrod covers I am certain that you will execute them with the same finesse and grace as all of the other parts in all of your builds.
They don't stand a chance against your shop "skilz" :)

And thanks again for all of your detailed posts, they are a joy to read !

Scott
 
Hi Terry

Awesome as always !!
Don't forget about Brownells, Gunsmiths make a lot of springs. I have several of their assortments. Here is one of 100 12" pcs of assorted sizes from .02" to .062" it is music wire and not SS but they work great.

https://www.brownells.com/gunsmith-...-kits/no-150-small-spring-wire-prod26217.aspx

And thanks again for such detailed posts, they really are a joy to read.

Scott

Another possibility for music wire is to contact a local piano tuner. They usually have coils of various gauges that they use to replace broken strings, and I've found them very willing to sell me short lengths at a reasonable price.

Carl
 
Hi Terry:
In post 208 partway down you have what appears to be a continuous square profile rubber seal (I believe it's the fuel bowl cap). Maybe I'm just looking at the recess for same but what did / will you use for that seal?
And if it's just going to be a standard round O-ring, how do you keep it in place on a rectangular land.
Thanks
 
Dave,
I cut that gasket from teflon sheet on my Tormach using a drag knife. It's sandwiched between the top and bottom of the bowl. There's a recess milled in the bowl top that it sits in. The recess is about half the gasket's thickness deep. - Terry
 
Dave,
I cut that gasket from teflon sheet on my Tormach using a drag knife. It's sandwiched between the top and bottom of the bowl. There's a recess milled in the bowl top that it sits in. The recess is about half the gasket's thickness deep. - Terry


Nice job. Going to have to get one of those.
Thanks
 

Latest posts

Back
Top