Mercedes or BMW Inline-6 Info

And a small air pocket isn't going to be that big of a deal.  The metal will conduct heat really well.  My 8 hp hit and miss would run at 300 rpm 8 hours at a time with no water in the hopper and still be able to put your hand on  the head.  It was just running ..not doing any work.
Dave

Quote from: steamerDo you think you will use this in a model airplane?  Or just demonstration runs?.  I suspect a 4 stroke will tend to run much cooler than a nitro 2 stroke at 40000 rpm.  Just thinking outload

It would be running in beauty queen test stand demo mode, not powering an RC model. Strangely, I don't have the same enthusiasm building vintage scale RC airframes as I do their engines. My model engineering interests don't really intersect my RC interests in that regard. My (methanol glow) 5-cyl radial has typical if not more liberal cooling fins & a pretty big fan blade up front that moves a decent amount of cooling air. Even so, you would certainly toast your hand on most parts of the head. From experience, I've had to run Teflon gasket washers on the exhaust flange. Not hard running, mostly idle to 3-4K range. It just has short little exhaust elbows right now, no collector ring. But every once in a while you can see a little flame lick in there. Like the real ones!

Interesting project! Could not find an running example on youtube.

What scale are you thinking?

What about 39x46,8 mm to have equal bore/stroke ratio. Could use GY6 or Honda SS piston rings. My engine has almost same bore stroke (39x47) so you are able to use M10 sparkplugs.

I wonder how to deal with different cylinders and so on total tolerance to fit rocker arm assembly.

If you have more depth specifications of geometry of engine I could run a simulation in ENGmod

Hi Alex. I am just getting around to pondering scale. According to the Wiki link for BMW IIIa even 20% on a cylinder basis gives 30x36mm bore/stroke = 25cc /cylinder = 153cc total engine displacement. Hopefully did my math correctly? Not sure yet what the referenced bounding box? dimensions correlate to exactly but using the same factor yields >300mm in length. That is a big fellah by my standards so off the top thinking something less. The cylinders is one thing but the crankcase is a mighty slab at least on my machines. I have no casting knowledge or intentions so that factors heavily.

https://en.wikipedia.org/wiki/BMW_IIIa

Regarding piston rings, whatever build I undertake next, I am committed to try Trimble's method to make my own rings. Even if I have to outsource the heat setting & providing the spacing fixture. On my 5-cylinder Ohrndorf radial I used commercial rings from an OS-56 (4-stroke methanol glow) engine. They were of the same 24mm bore & I thought at the time this was a smart side-step. But it turns out that hitting the correct bore diameter within a few tenths AND simultaneously the required surface finish for all 5 liners is actually more work than lapping all cylinders equally to 'whatever' on the same lapping sequence & then sizing the rings to suite. Whether the bores all come in at 23.8 or all come in at 24.2mm doesn't materially affect anything from a running standpoint (key word ALL). But it makes the overall procedure of fitted rings / cylinders easier. At least that is my current opinion. I might have a different view once I get to ring making LOL

Regarding rocker assembly, yes a fussy bit for sure. Its taken me a few pics to recognize what's  going on & those were more courtesy of Mercedes pics which, fingers crossed, is similar to BMW. The camshaft resides in a longitudinal tube. The rocker boxes are spaced along the heads. It looks like they bolt onto each head via 2 underlying studs per box. I'm not clear if tube + box housings are a single integrated assembly or tube is detachable because the tube must be openly exposed every cylinder within the box to allow access to the cam shaft lobes. But appears closed from the outside between the boxes. The boxes contain axles for the rocker arms & the left/right arms are different length configuration as the contact INT/EXH cam lobes in different positions. Anyway lots of busy work in there. The Schlillings schematic suggests a simplification where the boxes are separate & the cam shaft runs open in bushings (no tube). At least that is how I interpret the sketch. The FS engine has a lever on pilot side that somehow overrides valve setting during start? have not figured that out yet. I will not go down that path for sure.

The spark plugs enter the head area in a horizontal plane from one side, but at different clock angles viewed from top. Two per cylinder both on left side if faithful to scale. The good part of this orientation is that plugs & ignition cables do not conflict with the busy rocker area real estate upstairs. But as always it seems, when you scale FS plugs to model plugs, they would be quite tiny in real life & pointing into combustion chamber from side is different from top or angled from top. I am making some headway on shop made plugs heavily leveraging off of Steve Huck design. Thus far confined to 1/4-32 thread to test against standard RCEXL plugs on an ignition test strip. They do spark just fine but have not seen the harsh reality of engine heat & pressure yet. But reducing them to them to something less, say 4mm thread will be challenging. Steve & George Britnell & maybe some others have accomplished this but the bar standard is high.

I'm just at the 'is it feasible' tire kicking stage if it wasn't painfully obvious haha. At some point I will start sketching things & see what individual 3DP parts look like.

Lastly, what is EngMOD?

Neels has created an 1D simulation software, he started with two stroke simulation but there is also an four stroke variant. Maybe for model engines to much. But he does do demo software licence to for one month. But for many years I followed for couple of years on different projects while using his software. I think N/A 4t is quit straight forward. But maybe give you an idea of set up correct inlet and outlet according to valve timing for downscaling model.

https://vannik.co.za/EngMod4TDetails.htm

Youtube link seems not working. Look at EngMod2T - Simulating Your First Model

https://www.youtube.com/watch?v=1VE4KoKDP-s

I've been running EngMod2T for years to help design 2 stroke race engines.  It comes with forever updates.  We also have a dyno to check the results.  If you know what to input, especially temperatures, the results compare well.  Without some real world testing, it's hard to know the temperatures to use.  Our 26 cc engines are small compared to motorcycle and kart engines and the temperatures are lower.  This matters mostly for tuned pipes, not as much of an issue for model four strokes.

Lohring Miller

Quote from: lohringI've been running EngMod2T for years to help design 2 stroke race engines.

Lohring, coincidentally I have been watching YouTube videos of race boats. Its not popular in my area for some reason, despite suitable ponds or water bodies. At least for 6 months & then we skate on them LOL. I've been around RC for a long time & pylon racing for a stint. I feel they are brothers from another mother LOL.

Regarding some of the prior discussion on head cooling for this Mercedes/BMW engine, do you have a head temperature guestimate a typical boat race engine is? I'm probably dating myself but the nitro boats I saw around here had a simple pickup tube in the water that force fed the head nipple & then out. I suspect the water flows in some kind of annular ring in the head, particularly 2S which is not encumbered by valves. Maybe water is pumped via internal pump these days for more cooling control? I got thinking more about the simplified head to head cooling idea via a straight hole in head. The FS engine has a larger, more tortuous surface area in the head, a result of the annular shell remaining around valves etc. because it is cast or welded together. Thoughtfully painted green in this museum picture for visualization. This would be painfully difficult to replicate on a model, a non-starter. Hence the much less efficient 'straight hole' idea between the angled valves as discussed. Not nearly as efficient as the FS engine, OTOH, with a mechanical water pump, the fluid rate could potentially be adjusted. I am just spit balling here, no practical knowledge.

My, how they have evolved. Even a turbine. But I digress...
https://www.youtube.com/watch?v=JL0V42ROFBc

Regarding bore and overal sizes if you go much below 30mm bore you then have to start looking at some very thin cylinder and jacket thicknesses which may not be practical.

If we said 30mm bore which is 1 /4.66 scale

Liner really wants to be 1.5mm wall thickness if there is to be any hope of threading it.

Water space say 1mm

Jacket say 0.5mm as any thinner starts to get damaged easily

That would give a total OD of 36mm so allowing some space between the cylinders you would be looking at say 225-230 or 9" across the six cylinders so overall engine would be 300mm plus.

If you want to start reducing the O/A length then you either get into very thin cylinder and jacket walls or have to reduce bore and keep the walls the same which does not reduce the O/A length by much, maybe only 30mm total lengthreduction if you went from 30mm bore down to 25. You can see this on the Scillings where he has 24mm bore but 35mm cylinder OD

maybe best to look at some of the smaller valve train parts and size of fixings to decide what is the smallest you want to work to and then arrive at a bore and O/A cylinder size based on that. those M5 screws don't look as easy to make when you start to get down to 1/5th scale and they become M1

I agree with you, Jason. Lots to consider. Looking at the Schillings cross section again in this regard, doesn't look too bad to my eye. Obviously compromises & tradeoffs have to be made. The uniform looking OD from cylinder jacket up through head looks about right. I'm not quite sure how to interpret the M28x1 thread. Is that the liner thread that runs up & attaches the head or also/somehow related to the jacket? The side view shows 285mm length, not including prop. Almost a foot long, that would be a nice size. I suspect he compromised on the bore in order to get this end result. The rockers are entering clockmaking domain. And I have never seen an actual build of this although very similar principles & dimensions used on his other engines which seem to run well.

Looking at a clearer image in the book I would say that the bottom edge of the jacket is threaded M28x1 and the lower part of the 27mm cylinder (24 plus 2 x 1.5) is bumped out to 28mm to take the same thread.

Top of the cylinder is probably threaded M27x1 for the head to screw on.

I'm not sure if he just uses a sealant on the threads or they are used to locate the sub assemblies and then soldered into place, either way once you start threading circular assemblies you need to pat attension to orientating the parts correctly so pipe and spark plugs face the right direction when the jacket is screwed on, one reason I showed it unthreded in my sketch.

Also reducing the bore in relation to the OD of the jackets will start to affect the valve positions and Vee distance which will have to be smaller and narrower as you need to get the valves into a smaller cylinder. Though with less volumeto flow valves can be made a bit smaller

Stumbled on this water cooled Cirrus. No other details other than what appear to be a head pipe & jacket pipe.

https://www.flickr.com/photos/15794235@N06/albums/72157636740324745/with/9616165053

Just a quick update. I suppose I am making net progress on the BMW-iii inline-6 CAD model. Many, many puzzles yet to be solved. I went through a few significant revisions before realizing that if it had any hope of physically running (the objective) then some compromises had to be made. And I'm Ok with that. Its basically a stand (way) off scale rendition. I found a slightly reduced piston size solved many real world issues. The liner was a more reasonable thickness. Bolt on heads could be accommodated.

A thorny problem was the spark plug coming in from the side vs through the top of head pointing into combustion chamber. I made an M5 thread spark plug which was as small as I really wanted to go. The idea was it would thread into an extended lip of the liner. Unfortunately 5mm is still 'big' because it limits piston TDC position, which reduces chamber volume, which  means low CR. I was coming in low 6:1 which from what I've read is marginal for gasoline. The threaded spark plug hole is already as close to the top of liner lip as I dare. The rod length 7 stroke is already quite long. I domed the piston some more to reduce volume, reduced the crown diameter a bit to allow higher piston height (and also a good thing for getting rings on & off) but it was not moving the needle much. On this engine layout, 1mm of head squish equals one CR 'point' so its quite sensitive. I was chatting with Charlie on the forum who built the 1/4-scale Cirrus, also a side plug with an interesting plug CR workaround achieving CR of +9:1. That conversation triggered a solution I really should have known myself - make a boss lip on the underside of head, dropping it lower into the liner. My Ohrndorf radial has this exact feature <head smack>. Anyway, I'm calculating 8.9:1 if I did the math right but this is always optimistic in real life.

You can see the upper & lower crankcase elements are kind of chunky Lego-like for now. I'm not getting fussed by too much detail until I work out other more pressing issues. The thought was I would mill these from solid using manual (non-CNC) machines. Current thinking is a built-up crankshaft. I have a Schillings-like axial pinned version with bearings. And another version made up throw elements spaced & cross pinned on a center shaft, similar to how Michael.K did his Bugatti & Holt. This could run on plain bushings within intermediary webs of the upper crankcase housing. Of course in the back of my mind it would be so tempting to farm out the crankcase to CNC or maybe even 3DP to capture all the scale features & other necessities still on my to-do list. But I have lots more design work to do. The lower crankcase element is a rather rectangular bathtub, but I've kept my options open to angle the side facets. I'm not sure an oil pump would reside in that sump area I've hogged out or be located externally. Many balls in the air!

The gear train cam drive system is proving to be challenging. The FS crank drives a bevel gear up to a 90-deg bevel gear on the cam shaft. But along its way it splits to drive the magnetos (distributor in my case). I initially thought the CS was 1:1, the cam gear was 1:2 & the magneto had some other internal gearing which I could not miniaturize anyway. But when I enlarge the sketch & measure, I'm not sure that is the case. there may be other tooth/ratios in there. The cam gear looks bigger but not 2x CS gear. TBC. The downward pointing shaft drives the water pump at end of shaft but has yet another takeoff point driving oil pump. So lots of gears! You can see remnants of a sketch where I'm trying different bevel gear combinations (module/sizes) to accomplish this. I am hoping to avoid making bevel gears & this is where things get interesting - matching combinations of catalog sizes to what must ultimately work. Many M0.5 bevel gears are available in a range of sizes but predominantly brass & just don't think that tooth size is strong enough. M1.0 gears are stronger, many available in steel, but the minimal tooth counts are more restrictive. M0.8 might be just right but seem less common. One option I also saw on a Schillings engine is a 1:2 spur gear combination right at the end of CS. Not at all like the FS but visually compact & solves some issues. Now everything on the drive line is 1:2, the camshaft, the distributor (but now also the water & oil pump). And now all bevel gears are 1:1 sets so in a stronger gear, say M1.0 x 15T (~15mm pd) would not look out of place compared to a 30T. So this is about where I'm at in this department. I want to order some gears corresponding to least evil designs & mock them up on some 3DP dummy assemblies.

I have some thoughts on the water jackets & related plumbing but I'll end it here for now. If you have any comments, always welcome for consideration. Particularly - has anyone ran a spark/gasoline model engine where CR was known to be too low? Like 7:1 did not cut it, once raised to 8:1 or 9:1 problems went away? What sort of problems does low CR exhibit? Like hard to start, hard to sustain running...? etc.

Just some side conversation images regarding side spark plug orientation engines, overcoming resultant 'low' CR. This is how the 1/4-scale DH Cirrus MK-1 (Merritt Zimmerman designed) engine addressed it. The plug hole is located laterally above the main combustion chamber but connected to the head volume through a smaller vertical 'spark port hole' (not sure what to call it). So the piston can run high. From the outside the plug looks eyeball correctly positioned. Apparently this configuration CR of 9.5:1 according to the SIC build article.

Schillings proposed a notch in the piston. This is glow plug/methanol which typically requires even higher CR. I'm not sure this was ever built. From what I've read, the flame front should have a rather clear path over the the piston whereas this looks to go bang more inside the piston cavity? Maybe things just work on models that are different in FS engines. I have seen some other 'tucked away' glow plug arrangements that are known to run, but usually the ignition flows down/across the piston.

Quote from: peterthaJust some side conversation images regarding side spark plug orientation engines, overcoming resultant 'low' CR. This is how the 1/4-scale DH Cirrus MK-1 (Merritt Zimmerman designed) engine addressed it. The plug hole is located laterally above the main combustion chamber but connected to the head volume through a smaller vertical 'spark port hole' (not sure what to call it). So the piston can run high. From the outside the plug looks eyeball correctly positioned. Apparently this configuration CR of 9.5:1 according to the SIC build article.

Schillings proposed a notch in the piston. This is glow plug/methanol which typically requires even higher CR. I'm not sure this was ever built. From what I've read, the flame front should have a rather clear path over the the piston whereas this looks to go bang more inside the piston cavity? Maybe things just work on models that are different in FS engines. I have seen some other 'tucked away' glow plug arrangements that are known to run, but usually the ignition flows down/across the piston.

I don't have a firm answer to the spart plug question other than there are a lot of running Mk1's out there.  In my opinion, if the pre combustion chamber around the plug is of similar proportions, I would suspect it will function and run just fine on the side of the cylinder as opposed to the head.  That said, I think most would be very hard pressed indeed, on a finished model, to recognize the difference in position of the plug if it was put in the same location as say the Mk1 versus the exact position of the BMW, and it would minimize the risk.   

Thanks Dave. If I did my math right, I'm sitting at CR = 8.9:1 with the 5mm plug. So if I saw some examples of 8:1 (or the consequences were acceptable) its smooth sailing. I could push the head down a bit more but I'm getting into valve proximity to piston. A top spark plug is just not a good fit to this engine layout. Initially I was reserving space for longitudinal head to head cooling between the valve cages like FS but I dismissed that as fussy & doubtful effectiveness. And it would intersect 2 head bolts as-is which is kind of a problem LOL. The bigger real estate conflict with top plugs & wiring is the cam shaft resides in a longitudinal tube along the head tops & then rocker boxes situated on top of tube. The longitudinal coolant pipes will probably be  something like this, flooding from lower one & cascading out upper. Not scale, but one of many 'not scale' things.