Tuesday, March 3, 2020

Finalizing Engine Geometry

When starting with a clean sheet engine design it can be easy to get lost in the permutations of the various main parameters.  Since I am not going fully clean sheet and instead using a complete cylinder head from an existing engine, some of the important parameters can be reused from the donor engine's values:  in this case a bore of 116mm and a stroke of 60.8mm.  Just a bit oversquare.  This leaves a couple of other major parameters open for interpretation: conrod length and, less common, cylinder offset.

How connecting rod length affects engine dynamics is pretty straightforward:  the longer it is, the less secondary vibration you have at the expense of a taller overall engine. With my linkage front suspension, the chassis area near where a traditional headstock would be is pretty open, so a taller engine is not a packaging problem.


Using a non-zero cylinder offset, a Désaxé engine, has a more complex affect on an engine's dynamics as introduces a significant asymmetry into the piston motion, which can have far-reaching consequences.


This image shows how cylinder offset can affect piston lateral forces which generate significant internal friction.  My design choices of a long rod and cylinder offset should provide a low friction, smooth running engine.


Setting up a dynamic simulation, validating it, and sifting through the results would be a very time consuming proposition, so I did the next best thing, turned to a subject mater expert.  Actually, since I got a good answer with minimal effort, this was the best choice!

The subject matter expert in this case is Tony Foale, most likely one of the top 10 experts on motorcycle dynamics worldwide.  Unlike the other 9 experts, Tony is not sequestered in someone's factory, he is open with his knowledge and willing to answer any questions.  His website, https://motochassis.com/, is a great source of information on characterizing motorcycle dynamic behavior.  His book, Motorcycle Handling and Chassis Design, (check out page 9-23) is required reading for anyone wanting to understand or improve motorcycle operation and performance.  His thorough analysis of the motorcycle turn-in process is second to none.  He also has several software packages available for overall chassis and suspension setup that are in use by race teams worldwide.

A quick email to Tony explained my project, and as always he was willing to help.  He 'just happened to have' some software already written to do force and vibration analysis of singles.  Well, some people crochet for a hobby.

A few simulations and back-and-forths and the results were 'an offset of close to 10 mm will reduce piston frictional losses by a useful amount over the planned rev range.'  The following graphs were provided to support this conclusion.  The absolute values of the graphs are less important than the difference between the zero offset and offset cases.

Graph showing axial piston pressure load.  As expected, not much difference between the two cases.
Graph showing axial inertia loading for the two cases.  Note how the curves are not symmetrical around TDC and BDC.  This is purely a geometric effect of the offset cylinder centerline.  It also has the benefit of increasing the intake/power cycles to about 182 degrees of crank rotation and the exhaust/compression cycles to 178 degrees.  This introduces a slight amount of Atkinson cycle efficiencies without dilution of the intake charge, which is great.
 

The combination of the two top graphs which shows a small reduction in the maximum axial piston loading, which is nice.


Graph showing the piston lateral loading, the main reason for this analysis.  In this case the max loading is not as critical (as long as it is not excessive) but the 'area under the curve', which indicates the total frictional loss, is.  In this case the area is noticeably reduced from the non-offset to the offset, which was the goal of the analysis.  So, as previously said 'an offset of close to 10 mm will reduce piston frictional losses by a useful amount over the planned rev range.'  Excellent!

To get more detailed information would have required an order of magnitude more effort in simulating the piston motion and secondary effects and still be subject to experimental verification.  As it is, this information was extremely helpful and allows me to move on to detail design of the engine support components. Thanks, Tony!

6 comments:

  1. Very interesting Chris. Hmmm..... I wonder if I should be considering the same for my next build.... I've already got the long rods planned, for pushrod v twins with limited valve rates and head breathing, maybe some offset might be useful with my planned 94mm x 92mm twin, there is perhaps room to do it, hmmm....? (haha, it never ends does it!)
    Yes, that Mr Foale is a very knowledgeable fellow indeed, and a very nice bloke too.

    ReplyDelete
  2. I have not found an end yet! These principles are useful for any engine design but become more noticeable at higher rpms.

    Tony's a blast. And still racing too!

    ReplyDelete
  3. Tony's books, articles and software have helped/inspired many people.
    He's the reason I finally pulled the trigger and downdraughted a pair of Ducati 1100 heads. A legend, in my book.

    Have you two talked about your exhaust design?
    Any ideas on the method of balancing the engine ?

    ReplyDelete
    Replies
    1. Yes, his work can be very inspirational. Or perspirational!

      The exhaust design is pretty straightforward: short primaries, big megaphone with reverse cone and large silencer. There is lots of info out there to assist with exhaust design.

      He did give some help with balancing. I am using a counterbalance shaft in a unique position and had some discussions with him about the pros and cons and how it interacts wit the offset cylinder.

      Delete
  4. This guy, Mats Clementz (has a YouTube channel and is getting some serious HP from his single racer), had a catastrophic failure when he used some sort of plastic for the driven c/b shaft gear (w. steel pinion). Angular momentum/energy variations of the c/b on every cycle overwhelmed the plastic gear in less than a min.

    ReplyDelete
  5. Wow, that's a lot of work on OEM cases! I can't imagine a plastic gear surviving for the c'bal. That is a decently loaded gear pair.

    ReplyDelete