The units are right are for the velocity curve. in/degree. Zero would be where the line goes flat. I cropped the graph small, the valve lift curve units would be on the left.

Velocity curve.. Acceleration is another curve not included in this graph.

Yes, increasing lash reduces the time the valve is open. This ramp is way too long almost 30 degrees on the closing side. A GT2 profile would have a 10 to 12 degree long ramp in comparison.
The 906 is an extreme example of a very safe profile. it takes 30 degrees to move the valve .008" vs. 12 degrees on the GT2. If I get a chance I will overlay the GT2 with a 906.

Paul - You've captured the crux of my argument (although give me a bit of latitude on "sudden" and "disproportionate"), Certainly Schrodinger's Cat applies.
I've not seen cold clearances set any tighter than 0.1mm, and the valve head temperature pic I posted is only indicative - I've not measured it myself - so can't speak as to its derivation, but taken from G&S Valves Ltd UK. I'm only talking about valve/stem length not diameter (and hence not valve to guide clearance)
Here's a rough model of what I'm talking about (these are NOT ACTUAL MEASUREMENTS) but simple estimates. (and I've omitted impact of the rocker/and rocker ratio).

The datum for the model is the valve seat/head interface (top edge)

Sign indicates whether the expansion of the element increases (1) or decreases (-1) clearance.

Lengths are estimated effective heat path lengths, in the line of valve action.

Note that the valve has been broken into 3 distinct parts, and the sign of each parts contribution to lash. (The valve head actually adds clearance as it is located on its conical seat, but the stem parts decrease clearance with increasing temperature) I had not previously considered the sign of the head as different to the stem parts)

http://forums.pelicanparts.com/uploa...1263161540.jpg

Of course this is a limited model (and I know it shows negative lash by 0.01mm) but you get the impact of the very very hot valve in the running engine, which loses its heat very very quicky when the engine is stopped.

Again, if it were as simple as "lash expands with increasing temperature", then try the acid test (set "near to zero" cold valve lash - the follower will still hit the opening ramp at the design velocity - and run the engine and you'll burn all your ex valves.)

As Walt picked up - if it's valve cooling you're after, widen the seat and reduce your duration.

John D/camgrinder - interesting REAL data on the 906 profile....which I'm still digesting !!!

John

Hello John,

Personally, I think your estimates do not reflect what actually happens in a running engine. My guess is the actual environment is much more complex than a few numbers on a spreadsheet. The heat and the mass of the parts varys widely. As I said before, my actual measurements and the published valve expansion figures would need to be incorrect by a factor of 10-20 times for the running clearance to be smaller.

I have done some more research and all of my sources agree that the actual clearances in a running aluminum block and head engine are greater, by a significant margin, than the cold stationary gap. I believe this explains the unusually tight exhaust valve spec by Porsche. They knew that the expansion of the aluminum head and other parts would result in a much wider gap under actual running conditions. The figures associated with iron engines have much smaller variances. I trust the judgement of the people who design valvetrain to know from experience which is correct. When you make some actual measurements, I would be interested in the results.

That's OK Paul - I too trust the judgement of those who design and or have bulk experience. And of course it's much more complex than that - (there's been no discussion on elasticity or dynamics to this point, and I'm certainly not qualified to launch that one...) I'm just trying to increae my knowledge/experience as I go.

So the question remains....if the running clearance just increases with increasing temperature, why not run with zero cold clearance for a nice quiet engine with extended duration as the bonus???

John

Again, if it were as simple as "lash expands with increasing temperature", then try the acid test (set "near to zero" cold valve lash - the follower will still hit the opening ramp at the design velocity - and run the engine and you'll burn all your ex valves.)

John,

I believe that this test will result in valve failure. But it does not in any way prove that the running clearance is smaller than the cold clearance. Proving the cause of failure takes much more effort. One possibility would be that the actual running clearance opens up, but the amount of seat time is insufficient to get rid of the heat before the surface deteriorates and loses it ability to shed heat. Some things are not so simple.

After resolving (in my mind at least) why the 911 engine has such a tight cold exhaust valve lash spec, my next question would be why does the aluminum head ALFA twin cam have such a wide (.020") spec at the opposite end of the spectrum ? I know from experience that the similar Jaguar twin cam also has wider clearances when hot, perhaps the sodium filled exhaust valve has an effect, but I'm afraid this may be too OT for this list.

re ALFA twin cam - the cam is located above the valve & bucket follower (longer distance between valve seat and cam centerline = greater total expansion) The cam is nested lower in the 911 design by virtue of its rockers which decreases the effective length of valve seat to cam centerline. (at least in my mind)

On that note, I'b be very curious to know the valve lash spec for a 917 or 908 engine which uses the 906 cam profile directly on bucket followers as a comparison point between the 911 (rocker based) arguments above (given the otherwise "similar" head design)

Anyone ???

John

re ALFA twin cam - the cam is located above the valve & bucket follower (longer distance between valve seat and cam centerline = greater total expansion) The cam is nested lower in the 911 design by virtue of its rockers which decreases the effective length of valve seat to cam centerline.


That would not explain it. The Jaguar twin cam used 0.004" and 0.006" for twenty years and it is a very similar design (Morin bucket pat. 1916) to the ALFA. The Lotus twin cam used similar 0.006-0.008" clearances. The Jaguar clearances were increased to 0.012" for a complex parabolic cam design with very different ramps to produce the exact lift and duration but much quieter operation. The ALFA has a similar construction, a wider VIA, does not have parabolic cams, but uses a 0.020" lash spec. Now that I think about it, the sodum cooled exhaust valve cannot be the reason, because the intake spec is just as wide.

I think in the case of Jaguar, competing with the hydraulic lifter cam noise level in a luxury car had more to do with the tight clearances than any technical reasons . But it is a useful example because the MFG was able to reduce noise at much higher clearances by redesigning the cam profile. The ALFA specs are closer to a racing twin cam hemi, 0.016-0.020", and it may just be that the market for sports cars with tuned exhaust notes can tolerate more tappet noise.

In Johns spreadsheet above, the clearance for a warm not running engine has decreased from .1mm to .09mm. Since the actual measured value increases instead of decreasing we can infer that there is something wrong with his calculations.

-Andy

Andy,

Have you found the source for "The valve lash spec is actually .1mm plus or minus .05mm (.002-.006 inches)." I have search all my (+20) Porsche manuals with no luck. I am certain that the factory never made this recommendation in their publications, and would not advise anyone to go below the factory spec, regardless of the climate. I'm sorry, "It's weather".

Paul

Sorry Paul, I haven't looked for it. The person that posted it on here in another thread was my friend Tyson Schmidt. Maybe he will see this and have a source. I agree with your comment about not going below factory spec. The question is what is that spec and at what tempuratue. The guidelines I gave I believe to be within the factory spec. Surely you agree that there must be an acceptable variation on the .1mm spec. The .1mm is not listed as a minimum but a target.

-Andy

Found the thread. I don't know how to link to it but the title is: Valve clearance smaller for more HP?

-Andy

Surely you agree that there must be an acceptable variation on the .1mm spec. The .1mm is not listed as a minimum but a target.


Andy,

No, I would disagree, the mfg valve clearance spec is a mininum, unless a range is specified. Target is not a term used for valve clearances. My advice is to try to get them all even and at or above the minimum. On the 911 this is quite easy, on a shim and bucket system it takes many hours and several attempts to get them even. It is much preferable to err on the wide side on a mechanical valve system. In my opinion, advising someone is use less than the mfg spec under any conditions is just bad advice.

I don't see anything in that thread that cites a reference. There are no auto engines I know of that have exhaust valve clearances below 0.004", can you find one ?

The only thing that has any relevance is the running of the engine and that the valves are quiet. The .1 mm is a spec and the old German mechanic always says spec is spec, anything else isnt spec.
Bruce

Bruce,

Look at it this way, Say you had an anal, know it all, tech working for you. One day he says, gee, its cold today and I'll ignore the factory spec and crank down the valves to two thousandths. The customer comes back with two burnt valves. Who's fault is it ? Gunter and Heinz ?

Here is a comparison of the same intake lobe with a hot lash of .008" and .012".
.008" would be the hot lash if you ran .001" or tighter on the lash cold. The higher end of the lift curve is slightly changed, but the time off the seat at lower lifts is increased dramatically.
I have had some customers tighten up the lash to .002" to get more top end. I wonder how hot they are getting the cylinder heads and how wide the lash is really getting.

http://drcamshafts.com/pp/valve.htm

Thanks for the input, guys...

I just set it at .004.

Heat expansion is likely linear. The difference in temp at 40 degrees vs 70 is probably zilch. I was under the impression that the valves/adjusters ran at 200-250 degrees. But a great mechanic pointed out that the valves likely run at 500-600+ degrees, as do the valve stems, etc. The difference in expansion from 40 degrees to 600+, probably isn't much more/less than from 70 to 600+ degrees, assuming its linear...

the time off the seat at lower lifts is increased dramatically.
I have had some customers tighten up the lash to .002" to get more top end. I wonder how hot they are getting the cylinder heads and how wide the lash is really getting.

Hello John,

Anyone that does this has already made the decision that the 15,000 service interval is irrelevent and they are willing to accept the damage. They probably also have narrower seats that stock and will be on to their next engine before long. There is an recent, interesting SAE paper on the sequence of events leading to exhaust valve burning. A lot has to do with the surface finish on the valve and the seat. The valve is eroding the seat surface in a welding and tearing pattern, the surface finish breaks down, the heat transfer is impaired, then the seal breaks down and the hot gas burns a trench. You can probably estimate the additional heat by estimating the percentage reduction in seat time.

Paul,
These would be racers, not street guys. If you tighten the lash on a set of GT2 cams, you don't pick up as much timing like you do with the 906 profile.

Another thing to ponder. the 993 camshafts converted to solid rockers With a cold lash of .004" you will be off the clearance ramp when the engine is hot. Doubling the valve seating velocity.

John D: put another way, with the 993 you maybe ought not to convert to solid rockers without changing out a cam designed for hydraulic rockers?

Walt

Alternatively for the 993 conversion, the "Cat's whiskers" (hot hot hot valve) takes up the clearance when the engine's actually running, and contact actually remains on the 993's clearance ramp.....just a thought.

Valve lash & tolerance
 

I assume this is a factory label, but may be applied at port of entry ???

From an 87 Carrera - bottom LH corner of the image states both clearance and tolerance

0.004" +/- 0.002" for Intake and Exhaust

In this instance, the reference to engine temp of at least 90°C may or may not apply to the lash setting - I'll leave that to your individual interpretation of the label (and I'll assume the Tech Spec book would carry some precedence - cold)

http://forums.pelicanparts.com/uploa...1263335342.jpg

John D - camgrinder

re 906 profile

Perhaps its range of applications from the factory is a testimony to the those long clearance ramps and the range of impact to duration via simple lash adjustment.
906
911R
908/917 (I assume it underwent some geometric rearrangement for bucket followers)
and others no doubt

Any thoughts or further info on this profile - maybe worthy of a thread in it's own right.

John

Wow - factory thought you could get by with 0.002 lash? I'm not going to run out and readjust, though. Should provide some comfort to those who fret that they aren't getting super accurate lash settings, though.

Walt

Walt,

That looks like an emissions compliance sticker. Kind of like the sticker on a bottle of beer that says "Don't get pregnant while drinking and driving or operating heavy machinery". Everything on it is usually there for a reason, to meet the EPA emissions test for new cars. For example, the timing spec 3 BDTC +/- 3 means that the car does not meet the test (probably HC) past 6 BTDC. Unfortunately, since many (most) state tests are different than the EPA's new car test, you fail if you are not within the tolerances, even if you pass the state's tailpipe test. A common failure is idle speed. It must be that below 0.002" the overlap spoils one of the emission criteria, but I have never heard of a smog nazi opening up the engine to check the valve clearances in order to meet the failure quota set by politicians. It states that if anything is out of spec, refer to the workshop manual. That does not have a valve clearance range and states that if the adjustment tool (0.004") does not slide in, the clearance is too small. It would be a mistake to interpret the emission sticker as the valve adjustment procedure.

The most amusing story about emission stickers involves Cadillac. The emission sticker on one of their powerful V-8 cars also said "A/C Off". The smog nazi's followed the instructions and everything tested OK. However, when you switched the automatic climate control to On, the ECU used a different fuel and spark table that fattened up the mixture and improved the idle and drivability. 99.9% of Cadillac owner bought the car with the automatic climate control On and never shut if Off. The EPA was tipped off by an ex employee, sued Cadillac, and I think they settled for a big fine without admitting it was done "on purpose".

So the engine tune-up specs on the emission sticker published by Porsche aren't really tune-up specs? Sometimes it's ok to be wrong.

-Andy

Yes, setting the valves to 0.002" is wrong. The purpose of the valve adjustment procedure is to allow for the wear between services, and 0.002" is too tight.

you could go tighter than .004" cold, but not looser. If you have more than .012" hot on most profiles, the valve seating velocity gets too high and you can bounce the valve off the seat. The exception is that 315-300 Walt has in his engine. Those clearance ramps are higher, and need more lash. .006"-.008" cold.

I think its more to do with available valve springs at the time they designed the 906 profile. They wanted reliability at high rpm vs. HP from the looks of the profile. Very conservative profile.

Thanks John

So is there a consensus?
Is lash tighter or looser in a hot engine?

I believe everyone that has actually measured clearances on a hot 911 engine agree they are wider. Aluminum block and head, significantly larger (0.008-0.012'), iron block, aluminum head, larger (0.006-0.008"), iron block, iron head, closer to zero change (+/-0.002"). The major expansion is in your head, I am still looking for the cat.

I copied this off the web...
Not sure if its true, whats the opinion?

************************************************** **************
The Problem with Using a Feeler Gage
In reality, the rocker smashes against the valve stem like a hammer on an anvil. It does so 25 times per second at 3,000 rpm. Although both are hardened steel, eventually the elephant foot and the valve stem wear, necessitating adjustment by means of the adjusting screw. Unfortunately, the rocker arm, which to account for tolerance build up is larger than the valve stem, tends to suffer concave wear to the center its active surface.
To adjust a valve clearance, a feeler Gage is usually placed between the valve stem and the rocker.
When the rocker tappet has worn, the measured gap, which is the thickness of the feeler Gage, is less than the actual gap, the feeler Gage failing to measure the concave wear point. This results in the valve opening less than it should, (adversely affecting performance), slightly off valve timing and above all, very noisy tappets.

************************************************** **************
Could adjustments made by using thread pitch and degree work better on worn (seated-in) parts?

I did a valve adjustment this summer, I did the intake first (cold) ran the engine up to temp on long spirited drive came home dumped the oil to get ready to do the exhaust next morning. I set one valve hot to see if temperature made a difference in gap, the next morning the valve I set hot had no rocker movement at all from the slight play (free float) to a bind. Say from .004 to .000, So setting @ .004 cold most likely goes to .008 hot.

http://forums.pelicanparts.com/uploa...1264283422.jpg
un-measured gap! My best rendition...

http://forums.pelicanparts.com/uploa...1264285148.jpg

http://forums.pelicanparts.com/uploa...1264285207.jpg

This is off a motorcycle engine, excessive ware on both valve stem tip and swivel foot. Clearly a feeler gage would not measure the valve lash well. The poster of the pics wrote that a .004 was more like .008. Adjustments that are to loose hammer valve stem ends, swivel foot and valve seat.

Bob

Nice pictures. Wish I coult get my modest camera to focus through a magnifying glass! If I could, I'd supply shots of some random valve stem tips. Not a scientific sample, just what is sitting on my desk. None look convex. From memory, none of my elephant feet have been concave, either.

Others, and some certainly have a lot more experience than I do, may have seen cupping of the elephant foot, or of the tip of the valve stem. Your motorcycle engine being one example. But I never have with my 911 motors. I've had an elephant foot ball tip break off, and had a couple bend at the waist, all due to overrevs. Otherwise, anything that will screw out OK I have not hesitated to reuse, and have not gotten myself in trouble with this cavalier attitude. The valve tips I looked at look fine, despite fairly high mileage on some. Only ones which look beat up came from engines which blew up, as evidenced by dents on the valve head faces. And those don't even look bad - nothing you couldn't true with a valve grinder.

Maybe on some engines this is an issue. Not on my 911s not reved over 8,000 rpm, anyway.

There are lots of things one can worry about in life. Porsche 911 valve lash should not be one of them. Feeler gauge works just fine. Factory says to use it, which should say something.

But for those who are inclined to fret, there is the back of the cam method. And one can just set up a dial indicator and do it that way. Both would take care of any concavity issues.

Walt Fricke