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.
Construction of the fuel tank assembly continued with the machining of the spacer plate. It was milled from a piece of 3/32" hard aluminum sheet that was temporarily glued down to a piece of sacrificial MDF. The top surface of this plate will close up the bottom of the pump enclosure, and its bottom face will o-ring seal the top of the fuel tank. A fabricated brass elbow, bolted to the top of the plate, connects the pickup tube to the fuel pump through a short length of Tygon tubing. An o-ring surrounding the threaded filler hole on the bottom of the pump enclosure seals the spacer plate's through-hole.

The elbow's tiny gasket was cut from .020" teflon sheet using a 60 degree drag knife on my Tormach. Although I'd normally drill the clearance holes for the 1-72 mounting screws in a separate operation, I was pleasantly surprised to find the drag knife was able to cut them out perfectly. This gasket not only maintains the integrity of the seal between the tank and the pump enclosure, but it also eliminates a wet aluminum/brass interface that might otherwise have become a site for galvanic corrosion.

A pair of spring-loaded contacts mounted between the halves of the assembly will carry current to the pump motor from a connector mounted on the bottom rear of the tank. I used commercially available contact pins from Mouser Electronics (and others):

https://www.mouser.com/ProductDetai...30-002101?qs=sGAEpiMZZMvQ0fyxs12AwCm1WaLxvbzX

These spring-loaded contacts were designed for use on fixtures used to test circuit boards. This particular 3 amp contact contains a 2x2 array of gold-plated long-travel pins. After cutting it in half, the solder side of the second pair was used for the stationary half of the mating pair. Wires were soldered to each connector before epoxying them into the mating flanges. The contact through-hole in the spacer plate was used to ensure the contacts ended up properly centered over each other.

These tiny contacts are fragile though and vulnerable to damage if the tank halves are allowed to laterally slide over one another during assembly/disassembly. To prevent this, a pair of dowel pins were added to the pump enclosure's flange. A threaded o-ring'd filler cap machined from 12L14 and painted with Gun-Kote completed the fuel tank.

After final assembly, the tank was bolted down to the engine stand, filled with fuel, and temporarily connected to an elevated dummy carb bowl for testing. After initially powering up the motor to determine the polarity needed to spin the pump in the proper direction, a cable was made to connect it to the control panel. Once completed, the recirculating fuel loop was tested for several minutes and appeared to work as expected. The motor's voltage is adjustable from the control panel and will be fine-tuned after the carburetor is completed. That carburetor seems to be my last excuse for not returning to the pushrod covers. - Terry


465.JPG
456.JPG
457.JPG
458.jpg
459.JPG
460.JPG
461.JPG
462.JPG
463.JPG
464.JPG
 
Gasoline begins boiling at about 100F, and so carburetor temperature is a concern in any air-cooled engine. The Knucklehead's carb is particularly vulnerable because it sits on the end of a long intake tube that will essentially be spit-roasted in the V between the cylinders. In order to reduce the heat conducted to the carburetor, the intake was fabricated from stainless steel which has very poor thermal conductivity (roughy 5% of that of copper). For good measure, an insulating spacer will also be added between the two.

The carburetor provided in the drawing package is fairly ambitious and has everything one would expect to see on a full-size design including a float/bowl assembly, high/low speed needles, a butterfly throttle, and a choke. Except for the throttle which is controlled by an imposing cable assembly, access to its adjustments seem pretty limited. Since a model engine typically spends most of its active time idling between occasional throttle blips, I made modifications to the design to improve access to the low speed adjustments without giving up high speed control nor altering the essence of the original carburetor.

Since it was designed to accommodate a float, the original fuel bowl is rather voluminous for the engine's scale, and its depth limits access to the high speed needle located on its bottom. Its diameter also shifts the air cleaner further away from the engine than I'd like. Since my fuel pump eliminates the need for a float, I reduced the volume of the bowl, made it square instead of round, and chamfered its front end for better access to the high speed needle.

I left out the throttle control cable which seemed to overwhelm the presentation of the engine. My throttle control will be a simple spring-loaded arm located on the butterfly shaft behind the air cleaner and finger accessible from the top of the engine. Throttle return will be accomplished with a hairpin spring located inside a rotating cup that also raises the arm above the valve box oil return lines. I moved the adjustable idle stop into the cup assembly since the original version pointed into the rear head. I also added a fixed high speed throttle stop to prevent over-travel. The idle mixture assembly was left essentially unchanged since its needle appears be accessible through the front pushrod assemblies. The high speed needle assembly at the bottom of the bowl was left unchanged for the time being, but its length will likely be shortened later.

The choke control arm was designed with two detent positions: full ON and full OFF. A spring-loaded ball will lock the choke in either position but will also supply the force needed to hold the choke in an intermediate setting. All controls will be accessible behind the air cleaner which itself will be just a polished cover with no rear plate or filter element.

The original carburetor was designed to be a multi-part soldered assembly. Because of engine vibration, I was concerned about the 6 oz lump of brass that SolidWorks predicted would be hanging off the unsupported end of the 5" long intake tube. After a lot of back and forth, I decided to change the material from brass to aluminum for a 3X weight saving and to assemble the parts using JB Weld instead of solder.

Even with the material change I decided to maintain the same close-fitting joints I had planned for the soldered assembly. However, I couldn't find any reliable data on JB Weld's minimum usable glue line thickness. I ran some experiments and determined that a .002" glue line was practical and would provide more than adequate strength between two bead-blasted aluminum surfaces. The 1/16" test plates in my peel tests bent before the epoxy failed. For good measure, though, I'll also add pins to the joints where practical. - Terry

466.JPG
467.JPG
468.JPG
469.JPG
470.JPG
471.JPG
472.JPG
473.JPG
474.JPG
475.JPG
 
I've been looking for an excuse to build up a complex bonded aluminum assembly for some time. I'd have preferred using white metal for a number of built-up parts in the past, but I always settled on using brass so their components could be soft-soldered. My concern about the weight of the Knucklehead's carburetor, whether valid or not, gave me an excuse to try my hand at building up a four-piece assembly. For adhesive, JB Weld met all my requirements with its high strength and service temperature and especially its compatibility with gasoline.

My mini project began with the machining of the carb's main body. After turning a starting work-piece that resembled a thread spool, it was moved to the mill so the front and rear mounting flanges could be machined. Except for some matched-drilled holes that will be added later, a locating hole for the high-speed jet/pick-up tube and a flat mounting surface for the top of the bowl finished up the first piece.

Although the bowl top is square, it has a number of circular features that made it more sensible to start with a piece of round stock and perform the bulk of its machining on the lathe. The remaining features of this second part including its square perimeter and a recess for the bowl gasket were completed on the mill.

Before bonding the two parts together, arrays of .010" deep grooves were milled into their mating surfaces in order to hold additional adhesive. These grooves were designed to wind up on top of one another and create a series of .020" thick glue lines after the parts were assembled. A .022" alignment hole was drilled through both parts so a temporary seamstress' pin in conjunction with the locating hole for the main jet held the two parts in alignment while being epoxied. The tiny alignment holes were drilled through the center of what will eventually become the bore for the throttle shaft. The parts were then bead-blasted and cleaned in warm soapy water before a thin layer of JB Weld was applied to each mating surface. Because the shop temperature has been in the low sixties lately, the epoxy was warmed with a heat gun before being mixed in order to reduce its viscosity.

The bottom half of the throttle assembly, a circular boss, was turned to its finished diameter, and a .022" locating hole was drilled through its center. It's bottom face was then contoured for a +.002" fit to the carb body. The face of this contour contained an array of .001" high scallops intentionally left behind during machining to provide some extra byte for the adhesive. The corresponding area on the carb body was also scratched up with a Dremel engraver. Following the surface prep described earlier, the bottom half of the throttle boss was JB Welded to the top of the carb body. After an overnight cure, its internal features were machined in place on the carb body including the recess for the hairpin return spring, the idle stop boss, and the bore for the butterfly shaft.

The last part in the bonded assembly was the boss for the idle mixture screw. This tiny and complex part has two mounting surfaces but with little surface area. Each was machine-contoured to produce a +.002" glue line. In order to augment the strength of the epoxy bond, its pickup tube will later be Loctite'd and extend through the top of the bowl and partially into the boss.

The final assembly was once more bead-blasted and thoroughly cleaned before receiving a two minute dip in Alodine:

https://www.amazon.com/dp/B0049CDP5W/?tag=skimlinks_replacement-20

Alodine is a chrome-based surface passivation treatment designed for use on aluminum alloys. It's commonly used on military hardware because of the salt spray resistance it provides. Old-school hot-rodders familiar with Holley carburetors will recognize the golden tan color that it leaves behind.

So, what did I learn from this little experiment? During the setup to machine the internals of the bonded throttle boss, I discovered it was offset from the center of the carb body by .009". I wasn't able to determine exactly what went wrong, but I suspect the error was related more to my alignment scheme rather than the use of epoxy. I decided to keep the throttle shaft centered in the boss during its machining and to later compensate for the error by offsetting the butterfly on its shaft. If I were doing it over, I'd complete the machining of the boss before bonding it to the body, but more importantly I'd use a full diameter temporary throttle shaft to hold the parts in alignment. After thinking about it, I realized my alignment scheme using the tiny needle was pretty dumb and that I had been too focused on a larger pin becoming inadvertently epoxied in place.

I learned that it was nearly impossible to maintain a consistent .002" glue line between two complex contoured parts regardless of their machining. Inherently, this would not be a problem with a soldered assembly. The viscosity of even warm epoxy is just too high, and .005" - .007" is a more reasonable expectation between a pair of non-planar parts. Although I didn't have to deal with joint strength, it's typically not a major spec in assemblies that are going to be soft-soldered. - Terry

476.JPG
477.JPG
478.JPG
479.JPG
480.JPG
481.JPG
482.JPG
483.JPG
484.JPG
485.JPG
 
Last edited:
Fascinating build with exceptional sharing of the build . Thanks ! I also have never heard of Alodined.
For my 1.414 cents worth, My friend glues up his radiators from heater cores and makes custom top/bottom tanks.
He uses BONDERITE C-IC 33 AERO (FORMERLY ALUMIPREP 33) for final clean up before the epoxy.
I have seen it in action and it truly makes a perfectly clean bonding surface on aluminum.
Thanks again!!
 
Very nice looking parts Terry. Its interesting that most epoxy specs talk about gap filling capabilities but rarely mention much about minimum gap distances. I have heard about dry joint or over clamping which is maybe another way of saying the film is very thin. But nothing very specific or quantified in terms of structural weakness. In hindsight if you knew you were going to be dealing with more precise components with gap distance within the what typical retaining compounds specify as allowable, do you think that would be a better option over epoxy? Or are there other attributes to consider?
 
Petertha,
If I understand your question, you're asking what would one do if some important dimension depended upon the previse thickness of the glue line. One option would be to build excess stock into the part and then finally machine it after bonding. I did that, in fact, with the lower half of the throttle boss. Its final height above the carb body was machined after the epoxy cured. An effective .002" glue line in a soldered assembly is inherently easy, but precise positioning of the parts are still an issue - probably more so than when epoxying them in place since you can continually measure and tweak their positioning before the adhesive sets up. - Terry
 
Last edited:
Sorry, clumsy wording on my part. If I had a stack-up of parts to make a similar assembly like your carb, where the components fit decently well; flatness, minimal gap between them, no requirement for adhesive fillet forming etc. Do you think Loctite retainer type adhesive would still be a viable way to join components, or does epoxy still offer other advantages in your opinion?

For example I made a simple holding fixture turned in the lathe & attached an aluminum part with retaining compound using light tail stock clamping pressure until cured. The light cut machining on the part went fine, nothing came flying off. But what surprising was removing the part without damaging it was more effort than I expected even with moderate torch heat. So I tucked that away in my brain for similar built up assemblies. I've just never used retainer outside of typical bearings & threads.
 
If I understand the JBW technique, the grooves are filled flush to the surface with no overflow onto the adjoining surface. Correct?
 
Petertha,
I've tried using Loctite (bearing retainer) in that way in the past but never had any luck. I always suspected it was probably because Loctite has good sheer strength but poor peel strength. I'm not aware of any low viscosity metal-to-metal adhesives that I'd trust, but really my only other experiences are with super glues. - Terry
 
Kvom,
I smeared a thin coat over the entire mating surfaces, applied pressure to the assembled pair, and cleaned up the squeeze-out with Q-tips. -Terry
 
I hadn't heard of "alodined" before. Learn something new every time I come to this thread.
Being a retired airline employee Alodined is a acid coating used on aluminum before painting it is a treatment for corrosion prevention
 
Being a retired airline employee Alodined is a acid coating used on aluminum before painting it is a treatment for corrosion prevention

You were an airline employee I worked for airline cargo systems company for 43 years before retiring.
I have been on the sidelines and must say you do some beautiful work and designing stuff out is really nice!!
 
I am having a problem deciphering the push rod housing from the print, a assembly drawing in detail I can't find. I hate to start changing from the drawing.
 
I am having a problem deciphering the push rod housing from the print, a assembly drawing in detail I can't find. I hate to start changing from the drawing.
My problem as well. Its design is important in order to not introduce serious oil leaks, and I don't see how the current design does that. I've been putting off working on it hoping my subconscious would come up with a solution. So far, it hasn't, and after the carburetor I won't have any excuse but to return to it. - Terry
 
Work continued on the carburetor with the machining of the bowl. Because of its thin walls, a boss was designed into the bowl's interior to support the inlet and outlet hose barbs. The barbs were turned from 303 stainless, threaded, and Loctite'd in place with permanent thread locker. They were installed after finishing the bowl's exterior machining but before starting its interior. The interior machining was designed to uncover the ends of the barbs and smoothly blend them into the walls of the bowl. The finished bowl was bead-blasted and Alodine'd similarly to the body that was finished earlier. Both were then lightly polished with a white Scotchbrite pad to distress and lighten their final colors for a better match to the cadmium dichromate plating on an old school carb.

My original plan was to seal both the top of the bowl and its bottom main jet exit with gaskets. I wasn't totally happy about both gaskets having to seal under the same clamping force, and so I changed the bottom seal to an o-ring. This o-ring will be compressed by a hard stop on the main jet just before the upper gasket begin to compress. The bowl gasket was cut from .015" teflon sheet using a drag knife on my Tormach. A large-headed screw was used to temporarily assemble the bowl to the carb top for leak-testing.

The components of the throttle assembly were machined next. Delrin, with its excellent gasoline resistance, was used for the butterfly disk to avoid metal-to-metal contact inside the carb body. The shaft was machined from polished drill rod. The tiny features inside the throttle-arm required a good bit of time with a 1/16" end mill, but an overly ambitious feed rate broke my only 1/16" ball cutter, and the exterior had to be finished with a file and abrasive paper.

Some experimenting with different wire sizes was required to get just the right hairpin spring for the throttle return. Standard spring wire suppliers tend to sell wire only in large expensive rolls, and so I searched local retail stores looking for some consumer products I could re-purpose. In addition to guitar strings from a local music store, I found several diameters of stainless steel spring wire being sold for use as fishing leader in a sporting goods outlet. The final spring was formed with 3-1/2 turns of .016" spring wire wound around a .111" drill bit. After cold forming the spring and several spares, they were annealed for a couple hours at 400F. A bit of grease on the inside of the cup assembly will help avoid chewing up its soft surfaces.

Because of hands and eyes that don't work so well anymore, I expected some difficulty in assembling the throttle with its internal spring. But, I wasn't prepared for the three full days it took me to come up with a spring and an assembly procedure that I could perform. In the process, several of my extra springs were flung somewhere inside the shop and never seen again. In the end, I drilled a .022" hole through the top half of the assembly for a straight pin to temporarily hold the pre-tensioned spring in place during assembly. To be honest, I didn't get that hole in the right place the first (nor the second) time I drilled it.

In the end, by offsetting the butterfly on its shaft in the opposite direction, I was able to compensate for the earlier issue that had caused the throttle shaft to wind up offset in the carb body. The size of the error was actually .0045" and not the .009" that I previously misstated. I eventually selected .007" for the clearance gap around it. The final throttle action is smooth, and the spring reliably returns the throttle to the idle stop screw with a good feel. And, I can even disassemble and reassemble it. - Terry

486.JPG
487.JPG
488.JPG
489.JPG
490.JPG
491.JPG
492.JPG
493.JPG
494.JPG
495.JPG
 
Terry, if you need spring wire I have several lifetime supplies of the following sizes.

SS 0.016 0.035
Music Wire 0.011 0.013 0.014 0.018 0.020 0.022 0.024 0.046
 

Latest posts

Back
Top