Another Knucklehead Build

[kvom]

It's not as fancy as a drag knife, but I've success in cutting teflon gaskets with the sharp point of a chamfer mill.

 

[mayhugh1]

The pushrod covers are prominent parts on the engine with important requirements in addition to looking good. They must be oil tight but allow access to the lash adjusters they will also enclose. While the engine is running they should be positively secured at both ends yet allow their pushrods to be inserted and removed without major engine disassembly.

After completing the pushrod guide blocks several months ago, my plan was to continue on with the covers, and I spent hours studying the three-piece design provided in the drawings. Although the covers looked good in the downloaded renderings, I couldn't see any provision for sealing or securing them to the rocker boxes. I was also confused by some dimensional inconsistencies that I thought might be subtle tricks that I didn't fully appreciate.

Somehow I'd managed to successfully complete the nightmarish compound angle machining needed to align the cover tubes in the bores of the crankcase, guide blocks, and rocker boxes. The thoughts of re-machining any of those parts to accommodate a redesign of the covers was a real stomach churner. In order to continue progress I went on with the rest of the engine with the intention of returning to the covers later with a fresh perspective.

'Later' arrived, and for that perspective I modeled the three tubes making up a complete cover so I could play with the dimensions and virtually assemble all their possibilities. The first CAD drawing shows the models of the three tubes as well as their cross sections. It didn't take long to realize that I'd been overthinking the drawings. The dimensional inconsistencies were simply errors, and the design expected the upper tubes to be either glued into the rocker boxes or inserted without any special means of retaining them.

If they were permanently bonded, problems would arise when trying to install or remove the pushrods. If they were temporarily bonded with a sealant, the bond would need to prevent the upper tubes from working their way out of the rocker boxes under the engine's vibration. The second CAD drawing shows cross-sectional views of the covers in two configurations: normal running and lash adjustment. With no adhesive, the sliding fit at each end of the upper tube is all that's holding it in place. In addition, the three tubes need to be capable of being assembled or broken down with a pushrod inside them so the whole assembly can be maneuvered free of the guide block and rocker box. For me, verifying this would require some actual parts.

My experience with sealants in model engines hasn't been positive, especially when used on parts that may have to come apart again. Automotive sealants don't seem to scale very well. They're thick and messy and provide debris for clogging up tiny passages. They interfere with precision fits, and they're especially difficult to use between cylindrical parts. My goal was to avoid using them by making modifications to (only) the original cover design.

The first modification was to add a rocker box seal to the upper tube. An o-ring was my preferred solution but would have required re-machining the rocker box. Instead, I added a wrap of compliant material around the top end of the upper tube for a positive seal inside the rocker box bore. The third CAD drawing shows the machining required on the end of the upper tube. A short piece of Kynar tubing is heat shrunk into the groove that's machined around the tube's outer diameter. The design is such that after the Kynar is installed, its o.d. tapers slightly positive going from top to bottom and provides a snug fit inside the rocker box. As a bonus, its compliance increases the assembly's tolerance to small alignment errors between the bores in the guide block and rocker box. Kynar is a flexible but tough high temperature material with excellent resistance to a broad range of chemicals including gasoline and oil:

https://buyheatshrink.com/heatshrinktubing/high-temperature/kynar-heat-shrink-tubing

I tested and eventually used this material several years ago to solve a different kind of issue related to the assembly of my 18 cylinder radial (post #184):

https://www.homemodelenginemachinist.com/threads/another-radial-this-time-18-cylinders.21601/page-10

I used some of the leftover material to verify the design of the rocker box seal. After partially machining a couple upper tube ends and installing the Kynar, they were gently inserted into a container of oil. Neither showed any sign of leaking even after two days.

The second modification involved altering the threaded portions of the upper and lower tubes in order to create a threaded engagement during normal engine running. This engagement provides a controlled length adjustment of the assembly as well as the axial force needed to keep the upper tube snugly inside the rocker box and the bottom tube inside the guide block. The last two CAD drawings show the modified tubes assembled in their two configurations as well as side-by-side comparisons of the modified and unmodified parts.

In order to validate the design, a trial set of parts for a single cover was machined and test-fitted in all four locations on the engine. I also went through the exercise of installing and removing a dummy pushrod with the rocker boxes and guide blocks in place. The remaining photos show the actual first article parts. The upper and lower tubes were machined from aluminum, and the center cover tube was machined from 303 stainless. Now just three more sets of parts will finally wrap up my pushrod cover saga. - Terry

 

[Scott_M]

A very elegant solution Terry, it would seem that the fresh perspective was worth the wait ! Beautiful

Question , is there provision to keep them "adjusted" to that length ? Is running vibration a concern, that they may screw together ( get shorter ) ?

Scott

 

[mayhugh1]

Scott,
I hadn't thought about them coming unscrewed. There's a good bit of friction between the upper and lower threaded tubes after they're tightened. Not so much on the cover tube though. - Terry

 

[Scott_M]

I am sure that in the amount of time it will run as a model, it won't be an issue. I was just curious how they dealt with it originally, or didn't .

They all leaked

Scott

 

[mayhugh1]

Scott,
The full-size engines used a spring to force the upper and lower tubes apart and cork gaskets at either end for seals. Due to the design of the model engine's lower tube, only an upper seal should be needed. Gravity should take care of keeping the oil from collecting in the guide block and leaking out. - Terry

 

[ddmckee54]

Gravity will take care of the droplets and bigger stuff. I assume the crankcase is vented so that any oily smutz floating around in the case will not be pressurized and seeking a place to escape? Pistons banging up and down make a dandy air pump.

Don

 

[mayhugh1]

Gravity will take care of the droplets and bigger stuff. I assume the crankcase is vented so that any oily smutz floating around in the case will not be pressurized and seeking a place to escape? Pistons banging up and down make a dandy air pump.

Don

Good point ...

 

[Lloyd-ss]

A beautiful and sophisticated project on so many different levels. I am thoroughly impressed.
I do have a question that I hope you will answer. Back in post #2 you are using the C544 for one-piece guides and seats, and you mention that you have been using the C544 for quite some time. So my question is about the longevity of the C544 as a valve seat. I am assuming that it either holds up very well or that you can re-lap it on an as-needed maintenance basis. Is that correct? Do you think the C544 would hold up in a diesel model as a valve seat, or should I use something else. I am planning a diesel build and am trying to read about proper material selection for parts that have special requirements.
Thank you very much for presenting such a wealth of information.
Lloyd

 

[mayhugh1]

Lloyd,
I've used phosphor bronze for the seats and/or valve cages on all the engines I've built except for one, and that was because I had mismarked some regular bearing bronze as being phosphor bronze. I've not seen problems with either, but then my engines don't see more than three or four hours of running time. I personally prefer phosphor bronze for seats because it has an elevated temperature rating where as plain bearing bronze does not. The yield strengths of both are similar although bearing bronze is a bit more machinable. I once considered using aluminum bronze since it seems to be popular with others, but I found it hard to turn and difficult to get a nice surface finish on the seats. I once spoke withe Dwight Giles, a builder with much more experience than I have, about seat materials and he told me he used to use aluminum bronze but found it was causing too much wear on the stainless material he was using for his valves and so he was looking something else. The truth is, if you're using 7075 for the heads, integral seats are probably more robust than either bronze. A great seat surface finish would also be easy to avhieve. I've not yet tried it myself because by the time I get to the seats I have a lot invested in the heads, and it's much easier to scrap a bad seat than a bad head. - Terry

 

[Lloyd-ss]

Terry,
I really appreciate the personal response, and do not want to hijack your thread, so I will only ask a tiny bit more, and then start a build thread so as not to impose on others. For the project I am starting (52cc diesel, horizontal, crosshead) I was planning to use 12L14 for the head (with water passages), 316 stainless for the valves, normalized 4130 DOM tube for the cylinder (with water jacket), purchased Tanaka 1.5 mm thick rings, and possibly Nitronic 60 stainless for the piston (very good galling resistance and I have some of it on hand). Your endorsement of C544 for the valve guides checks off another box in the material list. I admit that I'd rather work with 7075 for the head, but it is a simple design so the 12L14 will still be easy to work. I am guessing that you will say that with a 12L14 head, either the C544 or the 12L14 head material will be suitable. Again, thank you so much for your response. The freedom with which you and all the forum members share their knowledge is much appreciated.
Lloyd

 

[mayhugh1]

Despite a bathroom remodeling project that's been sapping a lot of my energy, I managed to machine the rest of the parts needed to complete the remaining three pushrod covers. The photos show the finished parts and the mandrels used to turn them. The one-off expanding mandrel was machined with care and indexed to a collet since it was used to turn some of the features whose concentricities are critical to the fits of the final assembles. The much less important threaded mandrel is just a headless 5/16"-24 bolt used to hold the tubes in the lathe during polishing. A second bolt was used to confirm the internal thread count in each bottom tube. With my modifications, 13 to 14 full threads are required. Care is needed while threading the bottom tubes to avoid breaking out through their o.d.'s. Part of my initial confusion with the original drawings was that it looked to me like breakout would occur if this part were threaded according to its drawing.

The slippery polished surfaces combine with the very limited access while assembling them in place to make it difficult to install the covers in the two intake positions. I was eventually able to partially install all four assemblies so I could check for carburetor interference issues, but some kind of helper tool such as a tiny strap wrench is going to be needed for final assembly.

Excluding the pushrods, there are several more machined parts needed to make up the roller lifter/lash adjuster assemblies. The lifter machining looks like it will be particularly tricky not only because of the tiny parts involved but because the shafts for the 5/32" diameter ball bearings used as rollers will each have to be uniquely canted to compensate for the pushrod angles.

Don's comment about crankcase ventilation was timely, and although I'd previously provided for pressure equalization between the crankcase and the gear case, I realized I'd forgotten all about the road draft vent I had planned for the gear case. Instead of routing more tubing over its already busy exterior, though, I decided to incorporate a vent into the dipstick which sits relatively high on the engine. I modified its previously machined handle with an array of holes that I radially drilled into an internal plenum that I also bored into its center. A new stick, more complicated than the first, was then machined for it. This one has an integral tubular screen at its upper end that will hopefully strip some of the oil from the effluent as it's pumped through a labyrinth on its way out of the engine. Since it was essentially free, I also drilled out the centers of the five 10-32 button head cap screws that join the gear case to the crankcase in order to provide some additional paths for pressure equalization between them. - Terry

 

[mayhugh1]

This post comes a few days after the one-year anniversary of the start of this build. I had hoped it would be completed by now so I could start work on the Offenhauser, but this engine has been more challenging than I had originally anticipated. At the rate I work, I probably have another six weeks in front of me; but with all the home repair/remodeling work I'm currently involved in during the day, it might take even longer. Fortunately, I enjoy the journeys much more than the destinations that these builds tend to provide.

Before starting work on the lifters, it would be wise to have some semblance of a camshaft in my hands so I can check the operations of their rollers. Although the lifters in the model are similar to those in the full-size engine, their orientations are not. The lifters in the full-size engine are perpendicular to the axis of the camshaft, and the compound angled pushrods are accommodated using hemispherical cups on the tops of the lifters and spherical ends on the pushrods.

The model engine's lifters, on the other hand, are collinear with the pushrods. This departure from the full-size engine greatly complicated the machining of the guide blocks and now requires canting the rollers on the ends of the lifters so they can ride flat on the cam lobes. Even with perfectly machined angles, the worst-case rollers will scrub across 10% of the lobe's width during normal operation. If I were starting this project over, I'd look seriously into making the changes needed to reorient the lifters in the guide blocks and cam box as they were in the full-size engine.

Although I've purchased the tiny ball bearings specified in the drawings for the rollers, I'm having second thoughts about using them. As cam rollers these bearings will be worked hard and, if they don't ride flat on the cam lobes, the twisting forces created on their inner shafts could cause them to catastrophically fail and scatter tiny steel balls throughout the cam box and oil pump.

The necessary roller angle depends upon the particular lifter's location and according to the lifter drawing will be between two and four degrees. Since the Knucklehead uses a single cam to time the valve events in both cylinders, its .080" wide lobes shown on its drawing are only .050" apart leaving little room for the lifter bodies. In order to improve these clearances, the locations of the lobes for the front and rear intake pushrods were reversed in the drawing for the model engine's camshaft. This change assigned adjacent lifters to pushrods on opposite banks of the engine. Even so, the lifters don't have much material available to support the rollers' .040" diameter shafts. Canting the rollers in the ends of the lifters will require removing precious material from their ends around the rollers and will weaken them even further.

Since I plan to eventually machine the camshaft as a 4-axis operation on my Tormach, I have some options when it comes to the contours of the cam lobe faces. Instead of conventional flat faces for flat rollers, I'm considering machining a radius'd groove into the face of each lobe that will be matched to a contour on its solid machined roller. This will eliminate the need to cant the rollers in the lifters and will distribute their loads over larger contact patches than I would have available with imperfectly machined shaft angles.

So far, no obvious problems with the scheme have jumped off the pages of my sketches but, being unconventional, I want to make a few concept parts so I can watch it in action before actually committing the camshaft to it. To this end, I machined some camshaft blanks that, although initially have only circular lobes, can be re-machined as needed. Their diameters are large enough to allow me to experiment with the groove idea but later turn them into final camshafts - either grooved or flat. If the contoured roller idea doesn't work out, I'll likely replace the ballbearings with hardened solid-turned equivalents and give the shaft angle machining my best effort. Another option that I plan to consider is to contour the rollers but use them on a flat lobed camshaft.

Before machining the cam blanks, I needed to know the exact locations of the lobe centers. To determine these, I turned a simple test shaft with a diameter equal to the expected average diameter of the final cam lobes and temporarily installed it in the camshaft's location. Using closely fitted rods with conically turned tips installed in all four pushrod positions, the centers of the lifter bores were transferred to the inked surface of the test shaft. These locations were then measured under a spindle microscope using my mill's DRO.

Centering the lobes under the centers of the lifters will leave only .035" gaps between .080" wide lobes instead of the .050" shown on the camshaft drawing. I assume, but can't be sure, that the distances on the camshaft drawing were intended to reflect the lifter centerlines. Looking for an error, I re-measured my lifter angles in the guide blocks but they matched the roller angles called out in the drawings for the lifters to within tenths of degrees.

Two complete blanks were machined from O-2 drill rod so they can be hardened after completion. In addition, I turned a third blank minus its cam gear mounting flange so I would have better visibility of the lifter during my initial experiments.

A blank for the brass cam gear was machined several months ago while I was making the other spur gears for the engine. It required only a final operation to attach it to the cam blank. The holes for its mounting screws were slotted so the final cam can be accurately indexed to the crankshaft. Without these slots, the 24 tooth cam gear would only provide 30 degrees of indexing resolution. - Terry

 

[petertha]

hoped it would be completed by now so I could start work on the Offenhauser
The Ron.C 270 design or something of your own creation?

 

[mayhugh1]

It will be Ron's...

 

[doc1955]

Very nice work every thing looks so nice!!!

 

[RonC9876]

It will be Ron's...

A vastly improved version no doubt!

 

[rodue]

If time was money us model builders would be rich. I had a good friend when ask, how long did it take to complete that model he's answer was a, " cold winter and a late spring"

 

[mayhugh1]

I couldn't put my finger on it, but something continued to bother me about the contoured cam in my previous post, and so I spent a couple days trying to create a working assembly in SolidWorks that would allow me to virtually test it. However, SolidWorks (or at least my version of it) can't mate cams and followers that have contoured surfaces, and every workaround that I came up with required changing the model in such a way to limit its usefulness. Along the way, though, I discovered a problem that I've shown in the drawings in the photos.

A lifter that's angled along the rotational axis of a camshaft will require its roller to move sideways across the lobe as the lifter moves up and down in its normal cam-following action. The problem is that this side-to-side motion won't be tolerated by the contoured cam/roller pair. The contoured cam and roller that I was considering could be designed to fit perfectly together at, say, the flank of the cam. But, as the lifter rides up on the nose of the cam, its roller will try to move across the lobe but will run into resistance created by the now misaligned contours between them. In the best case, the roller might make it up onto the lobe's contour but with a single point contact between them while loaded by the valve train. In the worst-case, parts will end up bent, broken or prematurely worn. To get around this, the end of lifter would have to be opened up around the roller to allow the roller to slide sideways on its shaft so it can remain in its original position. Removing material in this area is what I was trying to avoid in the first place with the contours.

In any event, I've decided to abandon roller lifters completely. The requirement to cant the rollers in the space available introduces too much complexity and will likely produce an unreliable result. Not only does each lifter require a uniquely angled roller, but the drawings also show that the lifters must be uniquely keyed to their positions with pins and slots in order to maintain the necessary orientations of the rollers on the cam. Contouring the cam and rollers is interesting but is really just moving the complexity around for little net gain. I have no doubt that after a lot of painstaking effort, the final result in either case would be no better than a simple contoured rod end and much less reliable.

During my breaks from SolidWorks, I put some throw-away work into designing and machining the now defunct rollers I had planned to use. For closure, I've included a couple photos of them as well. - Terry

 

[tornitore45]

Yap is all a matter of geometry. The lifter does not "lift". It is constrained to move on its own axis and therefore as it rises it moves laterally, off the cam centerline.
Cams with roller lifter have flat flanks, cams with sliding lifters have arched flanks, are you redesigning the cam? One possibility is to keep the roller straight (perpendicular to the cam axis) and have a hemispherical joint between lifter and push rod to accommodate the angle.