Ohrndorf 5 Cylinder Radial

[petertha]

Onto the rocker arms. When I machined a prototype from the plans a long time ago now, I figured maybe I messed something up. The contact pad looked just a tiny bit off relative to the valve stem top in closed position and therefore through the opening action. It chocked it up to machinist error(s). But later on, this seemed to be corroborated in the CAD assembly. I’m still not sure, it may be the way it was dimensioned or. I was modifying other shape attributes so just tweaked the contact pad at same time. Rocker arms can have a lot of fiddly radii & transitions, so seems to be a tradeoff keeping the profiles functional & aesthetically pleasing, but also considering the machining steps to yield this result. They always look chunky compared to commercial engines but I suspect some of them are forged. Anyway, the profile allowed me to design machining fixtures.

Some screen grabs of prior notes.

 

[petertha]

The rocker arms are made from 0.25" x 0.5" O1 tool steel. I purchased mine from KBC, happened to be Starrett brand. They are ground pretty close to those dimensions which I neglected to mention, I would design around the 0.25” thickness vs trimming to what was quite a close metric dimension. The thought was to harden the valve contact pad area only, presuming I could control the heat & quench in that localized area. Turns out that was wishful thinking with my simple torch method, so the whole arm would be hardened.

The various fixtures are made from aluminum. The O1 stock is both aligned & retained with very close (6-32) clearance holes. Pins might be a better choice alignment wise, but over this length it was sufficiently accurate & simpler with all the repetitive mounting & swapping. Hopefully the machining steps are reasonably clear

Drilling the blanks with requisite holes. I did some design standardization where possible using same drill size and also the holes become the fillet radii between adjoining surfaces. The axle hole is reamed.

 

[petertha]

The valve contact pad radius was profile milled by offsetting the fixture in a rotary table. Just have to be mindful of ending the endmill arc at the right position vs. over-milling into what would be an arm segment. Generally, what I found with O1 on these fiddly parts is saw off as much material so you can do the milling operation in one go at full depth - plunge & profile. If you have to make successive depth passes it increases the odds to mess up & usually the finish suffers. With 0.25” thickness and a good EM it wasn’t a problem. Showing one of the many inexpensive imported rubber abrasives I use for finishing in a Dremel type tool.

 

[petertha]

Similar milling process for the hub to the transition points

 

[petertha]

Showing the side thinning milling. Required both flat mill & ball end mill to get the profile. I leave the mounting ears on as long as possible.

 

[petertha]

Drill & tap M3 holes for pushrod adjusters.

 

[petertha]

Trim to overall length in the mill vise.

 

[petertha]

A different fixture to do the round over milling for filleted corner.

 

[petertha]

The rocker arms were hand finished with wet/dry paper, rubber abrasives in motor tool & 3M scuff pads. I wish I had the right heat-treating equipment, but I don’t. I used a regular propane torch with the parts positioned into kind of a semi enclosed corner using those beige insulating type fire bricks used in ceramics. Those really help to contain & reflect heat, not to mention make the shop safer. I basically heat to bright orange by eye & then immediately quench in warmed hydraulic oil. I found it basically impossible to magnet test, the low mass means they start cooling quite quickly, so it’s more about having the oil in very close proximity.

They cleaned quite easily with 3M abrasive pad & then went into an ex-household mini toaster oven for 20 minutes or so on some tin foil. I think its maximum temperature is ~475F on broil, so whatever hardness number equates to this shade of tan brown is what I am limited to. There were no scale issues, in fact this is how they were ultimately installed into the engine. Overall, I’m pleased with how they came out.

 

[anday1]

Peter
those pics are gold to me thank you for posting. If you have a chance could you have a look at the boxer drgs and give some advice on the conrods i am on my 10th attempt to make one (grrr) and i don't really understand the big end dimensions of the 11mm O are the split bushes just 2.5mm each side?

Kind Regards
Andrew

 

[Richard Carlstedt]

Peter, Wonderful build and congrats on your progress !

Quick question as I may encounter the same thing--Metric to Imperial conversion
When you redrew the engine to start, (1) did you draw it in metric and then let the CAD program do the conversion, or (2) did you convert the dimensions and then draw to your calculations ?

i am inclined to do (1) , but you are a man of experience on a massive project.
Thanks
Rich

 

[petertha]

When you redrew the engine to start, (1) did you draw it in metric and then let the CAD program do the conversion, or (2) did you convert the dimensions and then draw to your calculations ?

I initially modelled the parts entirely in metric to correlate to the O5/O9 drawing set. That way I could also print the drawings in metric to manually check dimensions against the plans or any other dimension of interest. Then for my shop drawings, I just revert to an imperial drawing template & it handles the conversions & sig-figs, so no difficulty in that respect. I have the ability to switch units on my DRO, but my lathe/mill dials are imperial, 95% of my measuring instruments are imperial, so I just stick with imperial for machining.

Having said that, there are still many metric features that were carried through such as fasteners 9threads), bearings, circlip grooves, radii, diameters or slots related to metric cutting tooling. So in those cases I display dual dimensions just remind me. Over time I have built up my inventory of metric tooling. And truth be told, I can generally access a wider selection of metric vs imperial typically at lower cost just as function of where I live & where I usually order from. My next engines will likely be metric too so I'm not too fussed.

Somewhat related, on the other forum someone asked about the rocker arms. They didn't quite look like some of the cad pics. I'll just paste the explanation & attach sample shop drawing since we are discussing this. My drawings probably violate most standards but OTOH I'm the only shop hand that that sees them.

I think my CAD images might be a bit confusing. They are basically snapshots of 'ideas & thoughts along the way' some of which were ultimately discarded, but I probably didn't make that very clear. I started out with the original design with a few minor tweaks. Then an idea to make the side profiling on the lathe vs the mill (abandoned). For a while the contact pad was full width (why? abandoned). A few more iterations incorporating the 0.25" stock thickness & more standardized fillet radii. Here is the the drawing of the final part. Hopefully it resembles the machined ones pictured haha.

 

[petertha]

If you have a chance could you have a look at the boxer drgs

I'll have a look & PM you. I know I didn't get very far drawing them up before I decided I better stay focused to the project at hand.

Also, I think I will expedite my post & pictures of my version of the tappet bushing. I abandoned the O5/O9 conical profile which intends to kind of 3D align to the pushrod tubing. Mine is cylindrical & uses O-rings to both center position the tube, seal it from oil bypass and act as kind of anti-vibration wear snub. I'm not sure if the boxer pushrod angles will play nice in that respect but anyways will give you something to consider if you haven't glued them in already.

 

[petertha]

I considered making the rocker arm pushrod adjusters from M3 cap screws but unfortunately the dimensions of the socket head didn’t cooperate. So O1 was selected. Basic turning of 4mm stock to 3mm for the threaded portion. I used my offshore die head for threading, put the lathe on low speed, some cutting oil & just fed the die head by hand. Once it gets close to the shoulder just let go & the head spins free on the tailstock shaft. Some finishing with the 600# wet/dry paper bonded flat stick & a rotary tool polishing spider while spinning which does a great job of cleaning the threads.

 

[petertha]

Thread the adjuster into a holding fixture, face the pushrod end to length & 3mm ball EM the socket profile.

 

[petertha]

Into an aluminum fixture to make the 0.025" wide x 0.050" deep screwdriver adjuster slit.

 

[petertha]

I ended up hardening these, the pushrod is stock annealed O1.

 

[petertha]

Back to threading die heads, this kind of prompted me to make some shop-built units which I now prefer. I think my offshore unit is intended to be used with fixed dies as opposed to split adjustable dies, is mostly what I have. My dies invariably fit the import tool very tight which may just be a rectifiable QC issue. I may be getting OCD here, but seems to me that adjusted dies must become slightly non-circular & then they might sit off cutting axis just depending on how they land in the cup. Also, the retention set screw locations never seem to match the die. I noticed on my import that once the screws were landed, the plane of the die itself could get canted off square slightly. Anyway, my new die holder prototype has a looser fit but has set screws on 4 quadrants kind of like a 4-jaw chuck principle, plus 2 more screws at 45-deg to engage the split or fit my other regular dies (pictured) dies which have a different pattern again. These seem to do a better job on the finer threads. The next best thing would be some kind of floating mechanism like those reamer holders.

 

[anday1]

I'll have a look & PM you. I know I didn't get very far drawing them up before I decided I better stay focused to the project at hand.

Also, I think I will expedite my post & pictures of my version of the tappet bushing. I abandoned the O5/O9 conical profile which intends to kind of 3D align to the pushrod tubing. Mine is cylindrical & uses O-rings to both center position the tube, seal it from oil bypass and act as kind of anti-vibration wear snub. I'm not sure if the boxer pushrod angles will play nice in that respect but anyways will give you something to consider if you haven't glued them in already.

thanks peter
i have glued all but one in but a little heat should get them out

andrew

 

[petertha]

The rocker covers are made from 6061 aluminum bar stock. They match the footprint profile of the rocker box base. They are secured with two M2.5 screws which come down on either side of the rocker arm. The internal face of the cover laterally fits snugly to the vertical segments of rocker cage which support the rocker shaft. I made a few minor mods just for ease of machining. The O5 plans show what amounts to a constant thickness shell section, which means internal ball end milling to offset the corner fillet exterior. I cheated a bit by doing a stepped profile in the ceiling area, providing topside clearance but avoiding breakthrough to the reduced width of the external fillet. They still look a bit chunky & mechanical & to my eye. Real ones I’ve seen are often more streamlined but I suspect cast or formed from sheet steel. Methods that might not scale well although I have some ideas on that front for the next engine.

Bring the blocks to dimension using face mill. Drill the hold down screw holes & partial counterbore recess. I tried to save time on certain operations by milling them stacked together, but one has to be careful here (he says from experience). What I mean is if there is even a thou difference in thicknesses; the vise will only tighten on the fattest one & others may come loose which is a bad thing. I’ve have seen setups where soft aluminum wire or some such similar squish-able material is used on one end to take up any slight difference. It all depends on how much material is being removed vs vise contact area. Ultimately, I just took my time & did them individually using vise stops for consistency.