Oil pressure (and therefore oil pump parasitic drag) building up?

je suis charlie

Why would you remove the spark plugs for such a test? You're introducing a pumping loss that wouldn't be present during normal operating conditions.

There is a small oil pump that distributes oil to the top of the engine, however all lubrication is splash. At 1500 rpm the distribution pressure is 20kPa(3 psi). Something I failed to mention is that what I am observing occurs only after the first time it is spun after the plugs are removed. Spinning the engine subsequent times, the friction remains at the greater of the two measurements. Regarding the second response, testing with the plugs has as well been performed.

Now it's time to turn on (or install) a normal pressurized oil distribution system so you can measure how well the splash system isn't working.

Mike Halloran
Corinth, NY, USA

Granted a normal pressurized oil distribution system is almost a necessity for crank journal bearings on a typical engine but the engine I'm talking about doesn't employ journal bearings on the crankshaft mains or big end. Any other thoughts?

Okay, there's a roller bearing crank and maybe a roller big end, but upstream of the conrod are a crapload of oscillating bearings that are not likely to develop hydrodynamic pressures even with pressure feed. Who checked for that?

Mike Halloran
Corinth, NY, USA

My little Honda single-cylinder motorcycle engine has gone 60,000 km (of mostly wide-open throttle!) thus far with only a circulating oil pump. Rolling-element bottom end including the con-rod big end (it has a pressed-together crankshaft with the con-rod and its needle roller bearing built into the assembly). Wrist pin and piston rings are splash lubricated just like on any other engine. There's an oil squirter nozzle in the end of the hollow camshaft and a jet on each cam lobe that squirts oil against the rocker-arm sliding face by centrifugal action from the spinning camshaft. Camshaft has ball bearings. Timing chain is splash lubricated because oil return from the cylinder head to the bottom end is via the timing chain tunnel.

As for why you have a few seconds of lower friction, the only thing I can think of is that the oil is somehow adding drag when it gets into the works, or maybe if you were quick with removing the spark plugs, the oil temperature may have gone up locally during a heat-soak period resulting in a short period of the oil having higher temperature (lower absolute viscosity) as it first starts pumping.

I'm thinking a lot of oil channels drained back to the sump, leaving air pockets in the channels. These purge fairly quickly when oil begins pumping again, but offer little fluid resistance as they do so.

Brian, I too was thinking along the the lines of lower absolute viscosity at the rings from the recent combustion temperatures, but you got me thinking about something else. Getting back to my original question, I had been wondering if what I was observing was present with other engines. Something I had not realized prior to this exchange is that maybe because my total FMEP is so much lower than a typical engine, the blip I am observing could be showing up as significant because it is a higher percentage of the total FMEP.

I have nothing to add in regards your measurement changing (oil and temperature seem likely), but this statement caught my eye:

Quote (boonebucker)

because my total FMEP is so much lower than a typical engine

If I understand your approach correctly, I believe your prototype uses the top end (pistons, rings, cylinders, valves, etc.) from an existing engine. What differences in FMEP are you seeing between your Watts Link engine and the unmodified crank driven engine?

Comparing FMEP numbers is tricky business; what is included or not included and under what conditions. The tendency to compare my best with someone else's worst. Right? Let's just say that my original predictions of being able to cut friction in half with my engine design were too conservative.

Quote (boonebucker)

Let's just say that my original predictions of being able to cut friction in half with my engine design were too conservative.

Looking at the core mechanism, I don't see how you get such a large FMEP reduction, but I'm probably missing something. When I get a chance, I'll read your old blog posts and "ask me anything" session on Reddit. Do you plan to publish your prototype test results and, if so, when and where?



P.S. I agree 100% with your Jan 17 2010 post about the need to continue refining the internal combustion engine in spite of the trend towards battery electric automobiles. Peter Senecal (Convergent Science... CFD/Multiphysics simulation software) gave a TED talk on the subject and maintains a dialog on LinkedIn.

Wow. Interesting. I count 25 spots where something pivots, not counting the crank main and big-end rod bearings, several of which may share a common axis, 20 spots if you count the spots where three things pivot on a common axis as one each. And presumably it will share the general design of a valvetrain mechanism with a normal engine, except if you use a normal cylinder head arrangement the camshaft has to spin at an axis perpendicular to the crankshaft - can't just use a normal timing chain or timing belt, it needs to turn the rotation 90 degrees - and two such mechanisms since it is an upturned flat-four thus needing two cylinder heads. Or maybe there's something unique planned for valve actuation as well, which isn't shown there.

My little rolling-element-bearing Honda single spins over with pressure from a pinkie finger until you put the piston and cylinder on (piston ring friction), and even then, it doesn't require appreciable force to spin over until you install the cylinder head and timing chain for the camshaft and valve mechanism (spark plug removed), at which point you have to overcome the valve spring pressures against the cam lobes.

When Japanese motorcycle manufacturers moved from mostly single-cylinder and two-stroke engines (all of which used, and still use, rolling-element bottom ends) into making multi-cylinder 4-stroke motorcycle engines in the 1970s, they continued using rolling-element bottom ends and built-up crankshafts for some time, because that's what they knew how to build. They all switched to hydrodynamic bearings and pressure lubrication decades ago, though.

I split the case halves on the Web for a look at the real thing. Maybe you haven't seen it.

I'm surprised pistons (not rings) are such a big contributor in this chart.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376: Eng-Tips.com Forum Policies http://eng-tips.com/market.cfm?

Quote (boonebucker)

I split the case halves on the Web for a look at the real thing. Maybe you haven't seen it.

No, I haven't. Link?

I'm assuming posting a link to my website is fine with forum posting procedures? I'll be glad to remove it if my assumption is incorrect.
Link

No worries, linking to relevant stuff is fine.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376: Eng-Tips.com Forum Policies http://eng-tips.com/market.cfm?

So the design seeks to reduce friction by replacing rotating bearing surfaces with oscillating ones. Also (I assume) by eliminating side loads and rocking of pistons.

Looking at Greg's chart, eliminating the pistons and crankshaft completely still wouldn't achieve the 50% friction reduction claimed for the Brickley Engine.

je suis charlie

Greg. The chart you posted shows ring friction in the absence of gas loads - so significantly reduced. Gas load effect is added in a curve higher up.

je suis charlie

Greg, the chart shown is(b)for diesel. Is(a)available for SI? Oil viscosities perhaps?

Different source but this is SI.

je suis charlie

Here is another link that might be helpful. Closer to my engine in displacement, but 2 cylinders Link

Attached is a brief from Ricardo titled "Calculation of Friction in High Performance Engines" presented at their 2010 European User Conference. It provides a lot of information regarding friction in the various components and subsystems of a high speed racing engine under both motored and loaded conditions. The below table showing friction of key components at 9,000 RPM was taken from slides 41, 43, and 45. Column M is FMEP (bar) when motored, L is FMEP (bar) motored, M% is percent of whole engine fiction when motored, and L% is percent of whole engine friction when loaded.

                     M      L      M%     L%
Whole Engine      : 2.10 | 3.80   100% | 100%
Top Piston Rings  : 0.04 | 0.04   1.9% | 1.9%
Oil Control Rings : 0.12 | 0.07   5.7% | 1.8%
Piston Skirts     : 0.50 | 2.25  23.8% |59.2%
Small End Bearings: 0.01 | 0.01   0.4% | 0.3%
Big End Bearings  : 0.14 | 0.15   6.7% | 3.9%
Main Bearings     : 0.20 | 0.23   9.5% | 6.1%
 

From the above, piston skirts are the largest single contributor to engine friction at 9,000 RPM. The plot of skirt friction on slide 41 shows a marked increase of motored skirt friction as RPM increases (+100%) and a marked decrease (-33%) of loaded skirt friction. I assume the decrease of skirt friction with engine speed relates to transition from mixed lubrication to hydrodynamic lubrication on the Stribeck curve of slide 6. I'm guessing the increase of motored skirt friction with RPM has to do with piston kinetics.

Per the chart below from an MIT class on Engine Friction and Lubrication, mechanical friction accounts for 3% (full load) to 9% (part load) of engine losses. By my calculations, a 50% reduction in skirt and bearing losses would result in efficiency gains of 1% (full load) to 3% (part load). That would be a pretty significant accomplishment in my opinion. Whether or not that gain can be accomplished and what increase in cost and complexity manufacturers are willing to accept to attain it is beyond me.

• https://files.engineering.com/getfile.aspx?folder=117f0a8e-6ba3-4d79-aef9-9

Organizing the crank actuation mechanism to (almost) eliminate side loading on the pistons could be expected to reduce, but not completely eliminate, the piston skirt friction, and it's a pretty big factor - the above chart is good. Other methods of addressing this in engines of conventional layout include using long rods, using pistons with minimised and optimised skirt-contact area, and so-called "desaxe" layout in which the centerline of the crankshaft is offset from the centerline of the cylinders by a few millimeters away from the power-stroke thrust side. So, there is something worth pursuing. At what cost ... is a darn good question.

Actuating the pistons through a linkage of some sort isn't a completely new idea. The Nissan variable-compression engine, in production today, does that. They did it for a different reason, but one of the side effects of their linkage design is that it reduces the angularity of the connecting rod through the power stroke (which is when it matters most).

https://www.youtube.com/watch?v=j0An3RbXcPg

I can't tell you how many times I have re-calibrated and improved my test equipment because I have not believed the results I am getting. I must say at this point that the gains are far more than 1 % at full throttle and 3% at part load.

Lack of anything else to do prompted a wee bit of digging, and it seems that you've taken this concept MUCH further than most inventors do!

https://www.youtube.com/watch?v=ihd4JcjF0Rs

I'd love to see how you did the cylinder heads and valve mechanism.

The Harbor Freight engines that I cannibalized have pushrods that run along the side of the engine. I kept all eight pushrods on the same side of the engine by turning two of the cylinders upside down. All four pushrods that meet from the two cylinders opposite each other are actuated by one cam lobe via roller followers. One lobe, two cams placed one above the other, chain driven from the crank. Two lobes drive all eight pushrods.

Quote (boonebucker)

I must say at this point that the gains are far more than 1 % at full throttle and 3% at part load.

Congratulations! That's very impressive! There's should definately be a market for those kinds of gains.

Quote (BrianPetersen)

you've taken this concept MUCH further than most inventors do

Indeed. His first patent application was filed 12/16/05, and he's stuck with it all the way to prototype. He's an inspiration for everyone who has an idea and pursues it with perserverence. That's the spirit of the individual inventor.

In my hand, on a 'good' day, I have three or four pieces to some 10,000 piece jigsaw puzzle that in all honesty I have yet to identify. Practically speaking, on a moment by moment basis, the impact of those few pieces, which I can't help but mistake for much more than three or four pieces, have the possibility of crushing what life can become. So it is my privilege and responsibility in being, having come upon this, to choose a life leaning toward a spirit of discovery.

I apologize for going abstract and philosophical in my last comment. My thanks to each of you for your kind input!

We don't tolerate philosophy here!

je suis charlie