Valvetrain: Where is the friction?

[L_RAO]

I am considering sending some components for DLC coating.

However, considering the astronomical price of DLC coatings, I need to figure out where the best "value" or "returns" on $ spent is.

So, any thoughts on where the reduction of friction is most beneficial would be appreciated.

I am considering:

valve stems
rocker arm shaft
camshaft-cam cap interface (cam caps/journals)
camshaft-rocker arm interface (rocker arm flat tappet surface)

Thanks.

 

[USAFguy22]

What about WPC treatment, it didnt seem too steep from what I researched before. When I have my transmission rebuilt I'm going that route withe the gears/syncros/journals.

 

[greenberet]

I’d guess the highest parasitic losses come where there is a reciprocating motion or oil is “scraped” off parts like between the piston and cylinder wall, the cam lobe and the follower, and windage in the crankcase. Between rotating parts that don’t scrape the oil off (like between the crankshaft and its bearings), I’d think the parts float on a layer of oil relatively well and therefore cause little friction. Maybe DLC coating the cam lobes and WPC treating the followers so that they hold the oil well would be a good idea. However, you should talk to people who have actual experience with DLC coatings.

In case you haven’t seen it, there was in interesting post on the Race Engine Technology website about DLC coating pistons here.

 

[docjohn]

would a coating of any sort be helpful in an engine with such tight tolerances as ours...I could see there being less of an issue with american large displacement motors with more microns between parts.

 

[greenberet]

DLC coatings can be really thin – about 2 microns – and are widely used in Formula 1 engines. Looking in the Service Manual, the NSX's camshaft runout design spec is 15 microns and the intake valve stem-to-guide design clearance is 25-55 microns. I’d think that a 2 micron coating shouldn’t cause problems and almost certainly not on the cam lobes.

Your wallet may be a different matter. I just got a quote to DLC coat four camshafts. Assuming no polishing or other surface treatments are required before coating: EUR 760 (USD 950). I wonder how many horsepower the reduced friction would really free up.

 

[L_RAO]

The biggest issue with DLC coatings seems to be surface prep/adhesion, as there are long-term reports of the coatings coming off.

There is a DLC provider that I spoke with that has found "their" own solution - they prep the parts with WPC treatment (no kidding).

 

[greenberet]

I found some interesting papers and presentations regarding friction in internal combustion engines. From the sources listed below:

Camshaft bearing losses were shown to make a very small contribution to total losses even at high engine speed
[Valvetrain] friction losses are dominated by losses at the cam/tappet contact​

Therefore, to answer the OP’s first question, most of the valvetrain friction seems to come from the cam/tappet contact surfaces (or in NSX engines, from the cam/rocker arm contact surfaces).

Valvetrain group contributing
~35% of total friction @ 1000 rpm
Reduces with engine speed
~10% of total friction @ >6000 rpm

Use of DLC gave a strong reduction in valvetrain friction
24% at 4000 rpm camshaft speed​

Therefore DLC coating the cam/tappet interface reduced TOTAL engine friction by about 2.4% at high rpm. When an engine is spinning at 4000 rpm and is at full load, about 13% of the energy in the fuel is swallowed by friction losses and about 25% makes it to the crankshaft (the rest of the energy is lost as heat or is used to pump air around in the crankcase). At 7500 rpm and full load, the friction losses per rpm are about 43% higher than at 4000 rpm and full load. Therefore, at 7500 rpm about 18.6% of the energy in the fuel should be lost to friction and only about 19.4% should make it to the flywheel. Decreasing the friction losses by 2.4% at that engine speed should allow the power making it to the crankshaft to increase by 2.3%. Increasing the crank horsepower of an NA1 5-speed by 2.3% would yield about 6 hp.

The DLC-coated tappets showed similar friction reduction to that of the polished tappets, suggesting that improved friction is more due to improved surface finish and not the coating itself.​

If that’s the case, WPC treating the cam lobes and rocker arms might get you the same reduction in friction (if not in wear) that a DLC coating would give, at a lower cost.

- As components wear during run-in period these losses [due to surface roughness] reduce
- This makes true prediction difficult​

So how rough are the contact surfaces of the camshafts and rocker arms when the engine is well run in anyhow? After 100,000 miles have they already polished themselves as well as is possible? Would a 6hp gain through DLC coating or WPC treating the cams and rocker arms be for a new NA1 engine? How much would it bring in a well run-in NA1 engine? Less than 6 hp?


Source 1: Valvetrain Friction - Modeling, Analysis and Measurement of a High Performance Engine Valvetrain System
Source 2: Theoretical Analysis for Friction Losses Minimization in Piston Rings
Source 3: Calculation of Friction in High Performance Engines
Source 4: Valvetrain Friction Reduction through Thin Film Coatings and Polishing

 

[L_RAO]

Another idea would be to convert the C-series engine, which is basically a larger B-series, into a roller rocker, like the K-series. Not sure how much $$$ it would involve, but I would imagine it would take some precise measurements, drawing them up on CAD, and machining them out of the proper material.

Here are some pics I took of the B-series vs. the K-series rocker arms.

 

[greenberet]

Rollers on the cam followers are great to reduce friction but they may require new camshafts if you want to keep the OEM valve lift profile.

The sliding surface of the OEM cam follower is not a flat plane but is curved. If you decrease the radius of that curvature (for example with roller followers), you’ll decrease how much lift is transferred to the valve at various camshaft rotation angles. Not at maximum and minimum lift, but at the points in between. To maintain the OEM valve lift profile, you would need to have new cams ground with “meatier” flanks. It’s not too difficult to calculate how the shape of the cams would need to change given an increased curvature of the follower’s contact surface. Also, if the roller followers are heavier than the sliding followers, you may need stiffer valve springs to prevent valve float at high engine speeds. All things considered, it would be a lot cheaper to just DLC coat the cams.

I spoke with another company that DLC coats engine parts. They said that the parts that slide against each other must not have any ridges. Ridges perpendicular to the direction of sliding can cause the coating to flake off and ridges parallel to the direction of sliding cause local pressure points that can also damage the coating. So if the engine has been used, the cams should be smoothed to as-new condition before being DLC coated and the followers should be smoothed as well before rubbing against the freshly DLC coated cams.

If you read that presentation “Calculation of Friction in High Performance Engines” I linked to above (Source 3), the top three sources of friction in a high performance gasoline engine seem to be:

1. Piston skirt
2. Cam/tappet contacts
3. Crankshaft windage

Currently, it seems to me that it would be good to:

• DLC coat the piston skirts
• DLC coat the cams (to prevent wear, if nothing else)
• DLC coat the intake valves (so that carbon deposits don’t stick to the back of the valve’s head and so that the stem has as little friction as possible. I’m not sure about coating the exhaust valves because their stems have to transfer heat to the cylinder head. Are DLC coatings good at transferring heat?)
• Lap, polish, and WPC treat the cam followers
• Bore, hone, and WPC treat the cylinders
• Knife-edge the crankshaft counterweights to reduce windage and install tungsten inserts, if necessary, to get the weight back up to where it should be
• Modify the crankcase for less crankcase pumping losses (if feasible)
• Deburr and WPC treat the gears in the transmission and final drive

Getting the camshafts smoothed, lapped, and then DLC or WPC treated would probably require several weeks of downtime while the car is in a shop somewhere. It would probably make sense to get most of the things listed above carried out at the same time, during a major engine overhaul. Since my NSX’s engine seems to be in good health, that’s not going to be for a while.

 

[chussey]

Therefore, at 7500 rpm about 18.6% of the energy in the fuel should be lost to friction and only about 19.4% should make it to the flywheel.

19.4% of fuel makes it to the flywheel - 270 flywheel hp

18.6% of fuel lost to friction - 259 lost flywheel hp

Do you really think that at 7500rpm 259hp is lost due to friction?

 

[greenberet]

I think it passes a “smell test”. On page 2 of Source 2 above they state that friction in an internal combustion engine consumes about 13% of the energy in the fuel at 4000 rpm and full throttle. Another recent study (let’s call it Source 5: Global Energy Consumption Due to Friction in Passenger Cars – see page 10) came to the conclusion that in passenger cars, 11.5% of the energy in the fuel is lost to friction in the engine – not at 4000 rpm as in Source 2, just overall.

If the assumed average engine speed of passenger cars in Source 5 was about 2800 rpm, then the amount of energy lost to friction was very similar in both studies. Based on those studies, yes, an NSX’s engine probably loses something like 259 hp to friction at 7500 rpm and full throttle. And that’s why reducing friction may be a fertile ground to increase the hp that make it to the wheels.

 

[A.S. Motorsport]

Another idea would be to convert the C-series engine, which is basically a larger B-series, into a roller rocker, like the K-series. Not sure how much $$$ it would involve, but I would imagine it would take some precise measurements, drawing them up on CAD, and machining them out of the proper material.

Here are some pics I took of the B-series vs. the K-series rocker arms.

Really like that Idea,
If you where to do a I-Vtec conversion and roller rockers could be done, coupled with RDX injection it should offer a very good modernization for the engine.

Although very expensive.

 

[greenberet]

For what it’s worth, I found this regarding the engine in the new Ferrari F12:

"The camshafts are super-finished using a lapping process that reduces surface roughness to under 0.05 Ra, thus minimising the coefficient of friction between the cam lobes and the tappets. The tappets themselves have been given a DLC (Diamond Like Carbon) coating that reduces their coefficient of friction, increasing performance and reducing fuel consumption.

The cylinder head features different conduit designs, new intake manifolds and new plenums fitted with resonators. The resonators create overpressure inside the intake tract the moment the intake valve opens and immediately before it closes, guaranteeing improved cylinder filling and generating a powerful supercharging effect which increases engine performance."​

 

[gsrboy]

Another idea would be to convert the C-series engine, which is basically a larger B-series, into a roller rocker, like the K-series. Not sure how much $$$ it would involve, but I would imagine it would take some precise measurements, drawing them up on CAD, and machining them out of the proper material.

Here are some pics I took of the B-series vs. the K-series rocker arms.

http://www.horsepowerfreaks.com/partdetails/Ferrea/Engine/Rocker_Arms/Aluminum-Magnesium/3854

 

[L_RAO]

http://www.horsepowerfreaks.com/partdetails/Ferrea/Engine/Rocker_Arms/Aluminum-Magnesium/3854

Those are for VTEC killer applications only.

In addition, they just released a "Version 2", as the 1st version wore unevenly. IMO that doesn't speak much of their R&D or QC.