17 replies to this topic

[Paul Vanderheijden]

I have run across, what is for me at least, an interesting developmental problem.

In the process of developing higher and higher levels of horsepower output, and the attendant use of higher lift camshafts, it would appear that the boundary layer lubrication interface between the rocker arm and rocker shaft, has been compromised.

Over time we have moved from cams with 9.5mm lift at the valve, to where we now regularly use cams with more than 12mm. The interface of the rocker and rocker shaft is the standard Fiat arrangement of a cast steel rocker and a ground steel rocker shaft. The shaft is fed via an oil feed through the head and pressure/flow is distributed to each of the 8 rockers through the center of the rocker shaft. Standard specifications for the rocker-to-shaft clearance is 0.002-0.003 inch. The oil feed hole for each rocker is located 90 degrees from the bottom of the shaft (where the pressure point would be when the rocker is in use). With the increased ramp speed and higher lift of these camshafts the valve spring pressures have increased 35%.

When first presented with this problem I looked at the metallurgy and found discovered that both rocker and shaft were steel with a Rockwell hardness of Rc24. I have seen several cases where rockers have friction welded themselves to the shaft. While for road cars where stresses may be much less this may be acceptable practice, generally it would be preferable to run dissimilar metals, of dissimilar hardness ratings, for a competition application. As such I have proposed to bush the standard rocker arms with bronze bushings and to make new rocker shafts from 4140 steel, have them ground to 0.005 undersize, then hard-chromed 0.010, and finally finish centerless ground to +/- 0.0002 of factory standard dimensions.

Secondly, in looking at the oil supply/feed characteristics of the rocker shaft I found that the pressure feed orifice to the shaft was roughly 1/6th of the "aggregate" total of the output orifices. This would result in a significant pressure drop even if the running clearances of all the rockers was to specification. As many rockers may have clearances that are larger, perhaps some even past the wear limit, it became apparent that there might be a pressure/flow problem as far a lubricating the rocker/shaft interface. If this interface clearance was beyond wear limits, then when the rocker was under load (all the clearance at the top of the shaft) then all of the oil would escape from this increased clearance area and NO PRESSURE would be developed to assist in oil being delivered to the actual pressure point on the bottom of the shaft. After all, oil will take the course of least resistance. Therefore I deduced that, in addition to running the proper clearance on the rocker/shaft interface, having a slightly larger oil delivery orifice to the rocker shaft would ensure that adequate flow (and hence also pressure) would be delivered to each rocker.

All of these "improvements" should make for a better rocker/shaft interface, however the problem of increased pressure on the interface, due to the use of more aggressive camshaft designs, remains the same. Here are some of the ideas that we have already considered.

Aluminium rocker - Less weight. I have such an item as an option where allowed under the rules
Roller tips - less friction - See previous item
Titanium valve spring retainers - Already in practice
Titanium Valve - possible but perhaps cost prohibitive for most amateur racers
Titanium Spring - Same argument as for Titanium Valve
Lighter push rods - Already using hollow 0.062 wall 4130 tubular push rods. Stiffer than stock ones

One idea that may have some merit is to realign/redistribute some of the spring pressure. I had envisioned using something similar to a small block "rev kit". This would put a spring platform, for a small diameter spring on to the top of each lifter, with the push rod going through the middle. Opposite each of the lifter spring platforms would be a small aluminium plate that would fit in the lifter galley and tension the springs on each lifter. This plate would be held captive by cylinder head, once the head was installed. Currently the spring pressure "over the nose" of the cam lobe is around 235 lbs. By relocating the spring pressure, required for having the lifter track the cam lobe, to the push rod side of the rocker arm fulcrum point, the rocker/rocker arm interface would no longer have to deal with this pressure. I estimate that the amount of pressure that could be removed would be in the neighbor hood of 75 lbs., reducing the valve spring pressure across the nose of the cam to 160 lbs. Further calculations indicate that this would still be sufficient to control the current "lighter than stock" components on the valve spring side of the rocker arm fulcrum point.

The total amount of parasitic loss would likely change very little, however the working of the rocker/rocker arm lubrication interface would be much enhanced.

There you have the problem and some of the possible solutions that I would propose. I welcome comments on where the process have not make sense, or other possibilities for improvement. I recognize that this is not F1 technology, but I am sure that there are inquisitive minds out there that could undoubtedly assist.

Thanks in advance

[McGuire]

Instead of bronze inserts in your rocker arm fulcrums, I would install needle roller bearing assemblies. That will absolutely eliminate the seizing problem forever, reduce friction enormously and make the entire valvetrain more stable.

The bronze bushings will probably fix it, but needle bearings will absolutely fix it and will work better too. I'm confident you can find a bearing assembly for your application straight out of the bearing catalog. It's not incredibly sensitive. It is better to reduce the rocker shaft diameter rather than bore out the rocker arm.

Also, you want to look over things and make sure you have arranged your design so it is very difficult for any needle rollers to escape in the event of a failure, where they could bounce around inside the engine causing further damage.

It would also be nice if you could rollerize the rocker arm valve tip as well. It's a significant benefit to valvetrain geometry and component life. If your rules allow it I would really look into it. As for rev kits, I have never seen them accomplish much.

I can't help wondering if any of the aftermarket needle-fulcrum, roller-tip rocker arms manufactured for a zillion other engines over the years could be adapted to your application. That might also provide a selection of rocker arm ratios.

[Paul Vanderheijden]

Macquire,

Thanks for the insight.

I have already developed an aluminium bodied roller rocker, made from 2024 and rehardened to T6. This incorporates a centerless ground steel trunnion and a steel roller. The aluminium body of the rocker is hard anodized and the both the fulcrum bore and the trunnion bore are then honed or reamed to size.

Unfortunately thing are VERY small in this motor and and there was simply not enough room to incorporate a fulcrum needle bearing assembly.

In the late 60s Abarth did utilize a needle bearing assembly, apparently when this same problem occurred. This was done by undercutting the shaft 1.95mm in the area where the rocker would run, and then boring the rocker .55mm. This allowed a series of 1.2 mm needle rollers to be used. Assembly was a nightmare, as at each rocker position the needle bearings had to loaded manually. As the needles sat in a "groove", the needles could not escape, unless they broke down.

This was precisely one of the problems encountered. As the fulcrum pressure was borne entirely by about 3-4 needles (the needles did not rotate around the shaft), their life was finite and short.

Given the advantage of modern metallurgy and surface treatments, I was hoping to find an equally elegant solution. In reading some of the literature available today from various rocker manufacturers it would appear that the bronze bushing against a hardened, chromed and ground shaft may be a more stable platform than needle bearings.

[Canuck]

I like to argue with McGuire about politics and religion but it's not often I'll counter his mechanical suggestions. I offer this only as my experience from my perspective. Needle bearing rocker arms with roller rocker tips went big in the Harley world for a brief period after Crane introduced them. At the time the only offerings were from Crane and an outfit by the name of Jim's USA. It didn't take long for rumours of needle bearing failure to spread, and it was shortly there-after that Jim's began offering a bushing retro-kit for the Crane product. Today there are a multitude of roller-tip rockers available for Harley engines and all of them run a bushing arrangment.

Harley rockers are oiled slightly different as well. Oil is fed into and through the lifter body, through the pushrod and into the rocker arm where it travels to the shaft, into the void in the area between the two bushings and the shaft, and out through a small drilled passage in the body (as opposed to the "arm") to spray the valve tip area.

I've tried using Ampco (C63000 aluminium bronze) and ToughMet (Brush Wellman material) for bushing material, both with good results but I can't say they're any better or worse than bronze at this point.

[Ollies930]

Not sure how the oil passages are laid out in your engine, but early Suzuki racing 4-valve (rocker arm) heads used external oil lines to feed oil to both ends of the heads. This allowed higher pressure and more flow to reach the rocker arms(pads in their case). More pressure was available because they tied into the main oil feed immediately after the oil pump. Hope this helps.

[NTSOS]

These look very promising for your application:

GGB Plain Bearings

John

[McGuire]

Originally posted by Paul Vanderheijden

In the late 60s Abarth did utilize a needle bearing assembly, apparently when this same problem occurred. This was done by undercutting the shaft 1.95mm in the area where the rocker would run, and then boring the rocker .55mm. This allowed a series of 1.2 mm needle rollers to be used. Assembly was a nightmare, as at each rocker position the needle bearings had to loaded manually. As the needles sat in a "groove", the needles could not escape, unless they broke down.

This was precisely one of the problems encountered. As the fulcrum pressure was borne entirely by about 3-4 needles (the needles did not rotate around the shaft), their life was finite and short.

Oops, no precession. This is why you select a catalog bearing instead of simply turning down the shaft and sticking some needles in the space.

In your application a roller or plain bearing will work. So will just hardening the shaft, to be honest about it. However, the optimum solution is a needle roller bearing: minimum friction, maximum valvetrain stability. And a needle bearing will fit anywhere a bushing will fit by using a smaller O.D. rocker axle.

Needle bearings should never be a "nightmare" to install. Petroleum jelly.

[McGuire]

Originally posted by Canuck
I like to argue with McGuire about politics and religion but it's not often I'll counter his mechanical suggestions. I offer this only as my experience from my perspective. Needle bearing rocker arms with roller rocker tips went big in the Harley world for a brief period after Crane introduced them. At the time the only offerings were from Crane and an outfit by the name of Jim's USA. It didn't take long for rumours of needle bearing failure to spread, and it was shortly there-after that Jim's began offering a bushing retro-kit for the Crane product. Today there are a multitude of roller-tip rockers available for Harley engines and all of them run a bushing arrangment.

Er, what was the actual mode of failure, or doesn't anyone in that world go to the trouble?

I don't know anything about Harleys or its shop gossip grapevine, but I know this: at the top of the pushrod/rocker arm engine world (NASCAR, NHRA Pro Stock, Corvette LeMans) they use needle roller bearings exclusively. Maybe the chopper shop people know something they don't, but I doubt it.

[Canuck]

Well it wouldn't exactly be news to suggest air cooled v-twins were not subject to the same level of development as Nascar, Pro Stock or LM Vettes. Perhaps it has to do with the horrifying pushrod angles or maybe Crane utilized a bearing that wasn't up to snuff in that installation. I was to young and dumb at the time to wonder about the cause of failure (as opposed to now when I'm middle aged and dumb).

I stand corrected too - Crane still sells a needle bearing rocker, and a bushed one. I'm certainly not trying to knock needle bearings (or you) - I'm a fan of Sonnax and have had nothing but success using their products. The needle bearings in the Crane rockers (albeit 10 years ago or more) - not so much.

[Paul Vanderheijden]

NTSOS and Macquire

I cannot comment on the "mode of failure" of the Harley Davidson example, but in the case of the Abarth implementation in the late 60s from the literature available on the subject, it would have appeared that there was a direct failure of the needle bearing. The rocker shafts were hardened to Rc58. The rocker arm itself has a hardened bush installed.

Yes, I suppose it would be possible to go to a 12-13mm polished, Rc58 hardened shaft (and either bush the existing rocker stands or make new ones with the proper shaft ID) and then incorporate a 15mm OD needle bearing insert in the existing rocker arm. This would depend entirely on whether a needle bearing with precession could be sourced. I will look into it.

I do find that there is something enticing about the simplicity of running a bronze bush ( or perhaps as NTSOS suggested a GGB plain bearing, against a shaft that has been hardened (heat treat, melaniting or hard chrome finished). This, combined with enhanced lubricant flow for both lubrication and cooling, may provide the necessary performance margin to safely run more aggressive camshaft profiles.

The improved oiling that Olies930 suggests is exactly what I had in mind in opening up the input orifice to the rocker arm shaft, so that the input volume matches that of the combined output volume of the eight orifices feeding the rockers.

Thanks all for your constructive comments.

[SCO]

If you decide to go for a bush, I would consider this material if you can get hold of it; nibron special

[McGuire]

Interesting. Everyone has a pet bushing material they would like to try. Just as a point of reference, the standard material for rocker arm bushings since time began is CDA C93200 aka SAE 660 bronze.

[scooperman]

Minis and Spridgets have similar era pushrod engines, and like your Fiat they needed improved rockers for racing. Today roller rockers are common in racing A-series engines, but one early development was to replace the stock rocker bush with an offset-drilled bush, with bespoke rocker pillars this gave an improved rocker ratio. An offset in the rocker bore might help your situation by letting you use a less agressive cam.

[SCO]

Originally posted by McGuire
Interesting. Everyone has a pet bushing material they would like to try. Just as a point of reference, the standard material for rocker arm bushings since time began is CDA C93200 aka SAE 660 bronze.

SAE 660 is apparently suitable for 'medium speed and medium load applications' strength wise it's got nothing on Nibron - tensile strength of 240MPa vs. 917MPa.

With your higher spring pressures reducing the axle diameter to accommodate the rollers may not be a good idea either...

[McGuire]

Originally posted by SCO


SAE 660 is apparently suitable for 'medium speed and medium load applications' strength wise it's got nothing on Nibron - tensile strength of 240MPa vs. 917MPa.

With your higher spring pressures reducing the axle diameter to accommodate the rollers may not be a good idea either...

Opinions vary.

[cheapracer]

I'm with Mac on this one, hard to beat a good thin walled bronze bush. Finding someone to hone it to size and squarely properly is the key although not as critical as a conrod small end, find the best rep crank grinder around and still check it yourself after. Follow British Bike guides for clearance per " recomendations or find someone who races old Nortons, Trumpys or BSA's for advice.

[Paul Vanderheijden]

Well I have taken the plunge. Ordered thin walled bushes in 660 bronze. Will bore out several sets of rockers on my Bridgeport Mill, press in bushings and then hone to size. Luckily I have a Sunnen hone in the shop, so no need to send it out.

Found.625 OD 4130, 1/4 inch wall tubing. Will have it centerless ground 0.005 undersize with revised oil hole locations and oil grooving on the underside of the shaft. Shaft will then be hard chromed 0.010 and then centerless ground back to factory size specs.

McGuire, Cheapracer ACO and others, thanks for the input. I will post here when I have some results from the dynoroom.

[McGuire]

Sounds like a solid plan, good luck with it.