Valve clearance

[yoshimitsuspeed]

Most any engine builder assumes that the cold valve clearance is what operating valve clearance will be.
I believe that under more extreme situations, possibly max load or maybe even max closed loop load that it's quite possible for valve clearances to be very close to zero.
My theory is that under operation the average temp over the length of the valve could easily be a couple hundred Celsius above the average temp of the head. Calculating this out on a given head this alone eats up a good portion of valve clearance in CTE.
In some searching I wasn't able to find anything that specifically discussed this. I did however find this study that I believe supports my claim.

http://www.ww.journalamme.org/papers_vol23_2/1122.pdf

Unfortunately it is done on diesel engines. I know a lot less about Diesel but it's my understanding that generally combustion temps and EGTs are cooler correct?
In searching older threads here I saw people suggest that the head of the exhaust valve on a performance gas engine could get close to 800C which doesn't seem too surprising to me since it's not abnormal to see 900C EGTs many inches away from the exhaust valve. In this case even if the average temp was much lower it wouldn't be hard for thermal expansion to take up all of that valve clearance.
It's my theory that when calculating for valve travel, required valve spring travel or anything like this that you should assume that in the right circumstances the valve clearance could be very close to zero. Does that sound right?

 

[dgallup]

I'm pretty sure most knowledgeable engine builders know that valve clearances change between hot and cold. The magnitude of the change however is dependent on the valve train layout and all the materials used in the head, block and valve train. The solution will be completely different for OHC, DOHC, OHV or sidevalve engines not to mention piston ports, sleeve valve, rotary and who knows what other layouts people come up with.

Diesel exhaust temps at high boost and load can be quite high. I'm going back over 30 years but I recall limiting pre-turbine exhaust temperatures to 2000F on diesel engines. The valves run cooler than that due to conduction to the seat and guide. I don't remember ever measuring valve temperature directly.

----------------------------------------

The Help for this program was created in Windows Help format, which depends on a feature that isn't included in this version of Windows.

 

[ivymike]

my experience on piston-valve clearance is primarily with diesels. EGTs around 1300F would be a reasonable figure for a higher-performance boat engine (not a trawler), upstream of the turbocharger. Diesels will often have iron heads, which would tend to expand less than Al ones & therefore offset less of the valve expansion. Piston-valve clearance typically gets smaller on these engines when the engine is hot, although it's not impossible to imagine a case where that would be the opposite given a particular set of materials and layout. Yes, they can get very close.

Regarding your specific question when calculating for valve travel, required valve spring travel or anything like this that you should assume that in the right circumstances the valve clearance could be very close to zero

I gather that you're talking about designing a new cam or piston, or machining the top off the block? If you're not changing lift/timing/piston height/deck height then you probably aren't having an effect. If you want to baseline the hot piston-valve clearance in a particular engine, you can measure it directly using roll pins fitted in the tops of your pistons where they'll hit the valves ... press-fit the pins in extra-deep holes in the pistons, run the engine up to temp and then measure what's left sticking out of the pistons afterward. You'll need to do all cylinders, as the clearance may vary due to production tolerances, temperature variations, & lash adjustments. You may want to do the test more than once to compare results at different parts of the torque curve (less clearance @ peak torque?). You'll of course want new pistons & valves afterward.

 

[yoshimitsuspeed]

As an example of what I'm trying to get to, in a recent thread someone posted that x spring could take 8.1mm of lift before binding so therefore could take an 8.3mm lift cam because the actual valve lift and compression of the spring would only be around 8.1mm due to valve lash/clearance.
It is my belief that if a spring was rated to 8.1mm that you would only want to run max of an 8.1mm cam because there is the potential that under the right circumstances that 8.1mm cam could actually get pretty dang close or possibly even hit 8.1mm of lift.
In the same light if you were designing a piston for the smallest possible valve reliefs I would calculate for cam lift and not valve lift because I believe it could be possible for the valve lift to actually hit cam lift or at least close enough to use that number over any other.

If I were the engineer designing the motor and determining initial valve lash I would expect that I would first determine what the greatest temp differential between the head and the valves would be, determine how much the valve grew in relation to the head and then add a small margin of error to that. Does that sound about right?
So then if I wanted to design a part around that valve travel and to leave myself a small margin for error it seems like you should make all calculations based on cam lift and not valve lift right?

Ivymike
Are you taking about boring small holes in the valve relief in which you would fit the roll pins?
I definitely agree that with any test like this you would want to test over as wide a range as possible.
I know on my 4age turbo the highest EGTs that I see are when I hit full boost still cruising in mid RPM closed loop and stoichimetric AFRs. It gets almost as hot at full throttle and redline but still noticeably cooler.

 

[BrianPetersen]

It would never be prudent to select a valve spring so that it will ever reach coil-bind under the worst foreseeable case, nor would it ever be prudent to select valve-to-piston so tight as to not allow for ALL of the stack-up from not only manufacturing tolerances but also wind-up of the crankshaft, wind-up of the camshaft, "stretch" / "give" of the timing chain, flex of rocker arms (if present), plus all of the thermal effects.

If you are splitting hairs, you are doing it wrong (cutting it too close for comfort)!

On an engine with fixed valvetrain geometry (no hydraulic lash adjusters) you do not ever want the valve clearance to reach zero with the engine running. If it does, the valve will leak gases during combustion and fail to transfer heat to the valve seat, which makes it run even hotter which makes the stem expand even more, and if the engine is running under load, this situation can very quickly lead to failure.

Cam profiles designed for fixed valvetrain geometry must have gentle lead-in and lead-out ramps to take up the valve clearance and start the valve moving as gently as possible, and lower it back to the seat as gently as possible. This has to allow for reasonable variation in actual valve clearance. If the actual valve clearance goes looser than that lead-in and lead-out ramp, you will hear it. (Been there, done that.) If the actual valve clearance goes to zero, you will not hear it but bad things will happen to the engine.

I have an engine with an aftermarket camshaft that isn't what it should be in terms of lead-in and lead-out ramps (not much choice for this engine outside of stock), and it's a smidge clattery when cold, OK once warmed up. I think you can see what that means. (This is an overhead-cam engine with rocker arms)

 

[yoshimitsuspeed]

BrianPeterson
This is exactly what I am trying to get at. I wouldn't expect all valve lash to be taken up but the main purpose of that valve lash is to accommodate for that possible variance right? Would it not be possible for most of that valve lash to be taken up in certain situations where the valve was significantly hotter than the head? Would it not be smart to base your calculations off the cam lift instead of the valve lift to give yourself that extra headroom?
Naturally you aren't going to design something with zero tolerance but let's say a valve spring manufacturer said the max safe lift on their spring was 10mm and let's say you have a valve lash range of .1-.2mm. Do you run a 10.1mm lift cam because that extra .1mm should be taken up by the lash? Or do you decide that cam can't be run on that spring? Sure it will probably be fine either way because the valve spring manufacturer shouldn't leave their margins that tight either but I say you consider that too close because it should be calculated at the 10mm of cam lift?

This is getting more muddled down in the details though.
I am also interested in what people think the valve expansion may actually be, how much of that lash could be taken up in real world situations and how much is there for margin for error, stackup and longevity?
If you were building a race motor might you run less lash because you don't need the margin for error or longevity? Or would you use as much or more lash because the valve might see more heat?

Just throwing some rough numbers around using my favorite 4AGE for example hot reference on the exhaust side is .26-.36mm. At this point the head should be very close to operating temp so I wouldn't expect it to expand much from here under operation. It will a little but overall shouldn't very much.
On the other hand by the time you got valve covers off and started inspecting the motor I would expect the valves to have cooled down to roughly the same temp as the head.
Now if you start it and put the motor under heavy load and the valve becomes let's say 300C average temp hotter than the head the 100mm valve stem would now be 100.36mm.
The head would likely have expanded some and it may be a long shot to think the average temp of the valve stem could be 300C hotter than the average temp of the head but that's part of what I am looking for is peoples thoughts on what could practically be taken up and what margin might be left over. The numbers above don't seem too outlandish too me but it has already taken up the high side of the lash.
To get below that .26mm low spec there would have to be less than 200C average temp difference between the head and the valve.
It seems to me that under the right operating parameters at least a good portion of that lash could be taken up. I also have to assume that's a big part of why it's there.
Therefore it seems to me that in any calculation relating to specing another part off of the valve train you would be wise to use cam lift and not calculated valve lift.
Tying back into what you were saying Bryan this could be the difference between giving a valve relief what you consider to be .5mm safe margin but the depth of the actual pocket would change by at least .26mm if you used valve lift vs cam lift.

 

[Tmoose]

"If I were the engineer designing the motor and determining initial valve lash ..... ".

"Solid lifter" ( non hydraulic lifter /tappet) cam profiles have "ramps" at the beginning or opening/closing to gently take up "lash" or existing valve clearance adjustment before the serious valve opening and closing commences. The length of the ramp should accommodate hot and cold conditions, plus some margin for wear and measuring variations.

http://speedtalk.com/forum/viewtopic.php?f=1&t=42700

Erring on the side of too tight lash can affect valve sealing AND cooling, and overheat and burn valves pretty quick.

==========================

The installed height of the valve spring and the spring design would be made to achieve particular spring pressure when the valve is closed, at maximum lift, with margin against going into coil bind, and to not overstress the spring at max lift. Arbitrarily decreasing the installed height to get close to coil bind without a genuine need for the increased spring pressure would be wasting HP and run the risk of overstressing the spring.

Some engine builders would stay at least 0.05" away from "coil bind." As measured using full spring length.

At least one cam supplier seems to recommend 0.06" minimum between each coil at full lift, but this is likely a misprint.

http://www.compcams.com/Pages/415/truth-about-valve-springs.aspx

One aftermarket US head manufacturer reports that setting the springs close to coil bind results in longer lasting springs, probably as a result of quelling spring surged and resulting fatigue. The springs would have to be chosen for a close match for a particular camshaft to make this valid. Just cranking down the installed height without a need for the extra pressure would have negative results mentioned above.

==================================

Here is one OEM (Volvo) who suggests slightly tighter valve adjustments on a cold 240 direct valve actuation OHC (no rocker arms) engine.

Valve clearance, cold engine , inlet and exhaust

Checking: Adjusting:
0.012-0.016" 0.014-0.016"
(0.30-0.40 mm) (0.35-0.40 mm)


Valve clearance, warm engine , inlet and exhaust

0.014-0.018" 0.016-0.018"
(0.35-0.45 mm) (0.40-0.45 mm)

============

Regarding minimum valve to piston clearance, for decades pushrod V8s for Hot rod use would have been built with a room temperature fully assembled minimum of .08" or a bit more. The explanation was to compensate for rod stretch, piston rock, mis-shifts and valve float. timing chain stretch etc and etc.
This means of course the engine would be highly "interference" in the event of timing chain or sprocket failure.
How many folks complain about engines with belt driven cams being "interference?"

 

[Deividas]

Valve clearance, cold engine , inlet and exhaust

Checking: Adjusting:
0.012-0.016" 0.014-0.016"
(0.30-0.40 mm) (0.35-0.40 mm)


Valve clearance, warm engine , inlet and exhaust

0.014-0.018" 0.016-0.018"
(0.35-0.45 mm) (0.40-0.45 mm)

why there is more valve clearance in warm engine than in cold?

 

[ivymike]

sorry, for some reason I thought you were talking about valve-to-piston clearance, rather than valve lash.

Are you taking about boring small holes in the valve relief in which you would fit the roll pins?
yes, if your pistons have valve relief. On many diesels they do not.

nor would it ever be prudent to select valve-to-piston so tight as to not allow for ALL of the stack-up from not only manufacturing tolerances
correct, but better to use statistical tolerance methods than simple addition of print tolerances, otherwise you end up with far more clearance than you need. you probably knew that (maybe others didnt).

why there is more valve clearance in warm engine than in cold?
regarding the question of why valve lash (clearance) may grow on a hot engine, it comes down to the geometry, materials and temperatures. If you have a well-cooled steel valve in a hot aluminum head, with an iron block and along pushrod, the overall growth of the head+block+rocker support may be greater than that of the valve+pushrod, giving increased lash. There are lots of combos of materials/layouts/etc out there with growing lash and shrinking lash both being common enough that you wouldn't want to make a guess.

 

[yoshimitsuspeed]

I haven't compared steel block engines but the 4AGE does have more lash hot over cold. Being an aluminum head this wouldn't surprise me.
Like I said above by the time you got the valve covers off and got to inspecting things I would expect the head to be very close to the same temp as the valve. With an aluminum head that means it will have expanded more than the valve from when it was cold. Once the motor starts running the valve will get hotter than the head making it longer despite having a lower CTE.
I am sure there are other small factors that can change that as ivymike suggested.

ivymike I see what you were saying. That would be an interesting test.

 

[yoshimitsuspeed]

Tmoose that is why I changed from using the 8.1mm spring in one of my earlier posts to talking about a hypothetical spring with a stated max lift rating.
There has been a good bit of controversy over the "8.1mm" spring. Some claim that is the max lift recommended by the manufacturer. Others claim that is the point of bind. I contacted HKS (the manufacturer) and they couldn't be bothered to get me actual specs so I am working with another company to make a spring that will hit the same price point with some specs that I can trust.

 

[Tmoose]

"Some claim that (8.1 mm) is the max lift recommended by the manufacturer. Others claim that is the point of bind. "

Does this suggest 8.1 mm lift?

http://www.hks-power.co.jp/en/product_search/product/218

Looks like HKS only offers an intake cam, and that is soon to be discontinued.

http://www.hks-power.co.jp/en/product_search/product/605

===========================================

It would take less than 3 minutes of measuring an actual spring to determine that.
An Engine builder of merit would not guess about such things. Some would not be satisfied until they had checked to see if all the springs are "the same."

In addition to whether the spring will coil bind with a particular theoretical camshaft (with attendant disaster), there is still the question whether the open and closed spring pressures are sufficient and suitable for the cam being used. The required spring forces are dictated by stuff other than catalog camshaft lift and duration. The cam grinder is the prime source for the spring requirements. Sometimes the info is in the form " use our spring part number XXX." But the spring specifications should be readily available.

http://www.jegs.com/i/Crane-Cams/271/10308-1/10002/-1

https://www.tegiwaimports.com/skunk2-pro-series-xp-valve-springs-honda-b-series-vtec.html

 

[yoshimitsuspeed]

Yep good find, that does look like the right one. I still have not seen any documentation that states the seat pressure or spring rate. That is the first I have seen of an officially documented lift by HKS.
HKS used to offer a number of 4A cams but they have definitely been phasing out their support for it.

I thought about measuring the springs and I asked if they would consider sending me a set to test. I didn't really want to buy a set for testing because without knowing where their limit was I was afraid of the possibility of over compressing them and changing the characteristic of the spring so I had no intention of selling them after this. They weren't interested in doing anything to work with me on that.
Just as important to me as product quality is customer service and technical support. As soon as they gave me the cold shoulder I took the idea to companies who I have a good working relationship with and found one who was interested in making me a spring so I didn't put much more thought into the HKS springs or spending my own time and money to test their product. These springs are meant for mild builds. If I were or if I had a customer building a motor with 300+ cams and 11k+ redline I would definitely be very serious about taking things into account like spring rate consistency. These are meant to be an affordable spring that can be slapped in a mildly modded motor with some 260-270 moderate lift cams and stockish redline. One would hope that a company like HKS would be able to maintain quality and consistency for a slightly above stock build but the more I learn about aftermarket product QC and consistency the more I realize I have to do myself so you may be right about at least checking some to learn consistency. I definitely will be on the batches of the new design that will be coming in. I will also feel much safer in doing so knowing the rated specs of the spring.

 

[gruntguru]

My experience with valve clearances in modified engines is that most builders and tuners tend to set them too tight. IMO the optimum clearance is one where the follower contacts only the very end of the opening and closing ramps at maximum engine output. This is not easy to determine. Probably the best method is trial and error on the dyno. Prepare to be surprised.

je suis charlie

 

[Lou Scannon]

Sometimes the info is in the form " use our spring part number XXX." But the spring specifications should be readily available.

Quite right. Comp Cams has served themselves well by publishing detailed spring specs. That enabled me to find a set of Chevy smallblock targeted springs that happened to be ideally adapted to a stock dimensioned Ford 2.3L 4 cyl. valvetrain (Comp 990 springs, for those interested). Since the target market was the ubiquitious Chevy smallblock, the price was right, even when buying 16 springs for an engine needing only 8!

"Schiefgehen will, was schiefgehen kann" - das Murphygesetz

 

[yoshimitsuspeed]

gruntguru
Can you go into more detail?
I would expect that you would want to get as much lift and duration as you safely could out of a cam. What about having less opening and closing ramp would improve performance?

Also where you suggest it barely contact at maximum output seems like it could contradict my thoughts that there would tend to be less valve lash around maximum output. If this is true and the valves are longer due to being hotter at this point then under light load and lower RPM situations the follower may not be touching the opening or closing ramp at all.
But then this would take us back to the question of just how much that dimension might change under varying load, AFR, timing and RPM.

 

[gruntguru]

If you want more lift and duration - get a bigger cam. You can't use the open/close ramps to do either effectively. All that happens is you kill the bottom end and the fuel economy.

In my experience the optimum valve clearances end up wider than normal but never excessively noisy when cold.

je suis charlie

 

[yoshimitsuspeed]

If the valve gets hotter in relation to the head under operation then how would it be possible for valve clearance to increase in those situations?
If you want to just catch the tail end of the O/C ramp then wouldn't a better cam design only include that section of the O/C ramp and maintain a tighter valve clearance?

 

[gruntguru]

I guess different expansion rates eg aluminium head and cam pillars plus cool inlet valve stem. Must admit I haven't heard of this before.

The ramp is needed to accommodate clearance change with temperature. Also to provide a tolerance for normal workshop valve setting practices (as opposed to blue-print accuracy) and for wear.

je suis charlie

 

[ivymike]

If the valve gets hotter in relation to the head under operation then how would it be possible for valve clearance to increase in those situations?

The valve is not uniform in temperature - a good bit of it would be the same temp as the surrounding material - and the Al head has about 1.5x the thermal expansion coefficient. Probably more common on a pushrod engine, where the valve and head are only a small part of the stackup.