2.8 Short Stroke Compression Ratio fine-tuning

[YTNUKLR]

So, I've got a (not mine) 2.8 short-stroke motor on the table, minus pistons.

Carrera 3.0 Case
66mm CW Crank
2.2 Rods
3.2 95mm Cylinders
SC ported, twin-plugged heads w/ Ti retainers and Aasco springs
DC80 cams
Twin plug ignition
MFI
ARP Rod hardware and Head studs
110 Race gas

It's a full race engine, planned to run up to 8500 or so. Now I'm second-guessing myself on the compression. Originally, I was thinking 11.5:1, with these cams, would provide the necessary punch around 5k rpm. Now, I wonder if this is going to become an engine that needs to be torn down every 10 hours to be R&R'd. Ideally, I'd like to be able to get 25 hours out of it before it needs to come apart again. That may mean 10.5:1 or so...

Any thoughts welcome

Cheers,

[davidppp]

Hello, Scott.

The revs will kill it faster than the CR surely?

Those big pistons are heavy and wobble in the bores..

Kind regards
David

[blue72s]

Originally posted by davidppp
Those big pistons are heavy and wobble in the bores..

What's the maximum size pistons & cylinders do you recommend for high RPM applications?

TIA.

[YTNUKLR]

Never mind.....

[Henry Schmidt]

Hi Scott
10.5 to 10.8 to one would be appropriate for your application.
Engine sounds nice.
Where did you get that idea?


ARP Head studs, what were you thinking ?

This sounds like a candidate for a Supertec Gen II twin plug distributer.

[Henry Schmidt]

NEXT QUESTION : INDUCTION ?
PMO?
MFI ?

[YTNUKLR]

Haha, the idea was all mine!!

Henry, a 009 Pump needing a rebuild will be heading your way sometime in 2006...as will a 911SC distributor. I'm working on remaking (CNC) slide valves, we'll see how they work. If not, I've got some '70E MFI stacks that would be perfect bored out.

I'm sure your head studs are damn near perfect, but with ARP stud strength at a comparable level, the deciding factor simply became cost, and yours were a bit more .

Now, I have to ask, but I think I know the answer: is it worth buying MFI injectors new over having used ones cleaned? I'm thinking no, those things are soo expensive now.

I'll post more pics as I progress...

Cheers

[Grady Clay]

Scott,

While I defer to Henry’s superior knowledge, I use the highest CR physically possible given piston-to-head clearance and piston-to-valve clearance. Both of those you can start on the conservative side and at each maintenance rebuild creep them closer. There are cylinder base gaskets available in 0.002” increments, maybe finer today.

I have three 66x91 engines that peak at about 8300 and I would occasionally intentionally run to about 8700 rpm. There were some unintentional trips to 9000 with no bad consequences. These were Mahle forged RSR-like pistons which, I suspect, are lighter than what you will use. I also used carefully selected, prepared and tested lighter 2.0 rods.

I edged the CR as high as I could without the piston ever touching the head. It was above 11.7:1. It kept the piston and head clean in the squish area but never any contact. I was very careful warming the engine up (propane pre-heat oil and engine) and immediate shut-off for concern of expansion differences.

The issue here is to use race fuel octane appropriate for the CR and keep the head temperatures under control.

I disassemble these every 25 hours based on summing at-speed lap times. I never had a failure or even a close call. The parts looked like I could quadruple the time if I wanted. Each time the engines got new main bearings (except #8), rod bearings, rod nuts & bolts, rings and head gaskets. I would Magnaflux both the rods and new bolts. I constantly stirred around the rods to have a fresh and well tested set. I would touch-up the valve seating – that is where the power is.

I am generally not inclined to extend the hours on a very high rev engine. It is the RPM and not the CR that is the limiting factor. The loads go up with the cube of the angular speed. Stroke and mass of the rods & piston are very important. I want to take out perfectly good parts. The consequences of a failure are just too great. This past weekend a friend lost a 2.2 with a holed early aluminum case that was thoroughly prepped. The engine isn’t apart yet but I think he lost a non-Mahle piston first.

Where Porsche Motorsports wants your Cup engine rebuilt every 40 hours, I think that can be extended significantly by keeping the revs lower if not used for serious racing.


Looking at your list, the valve springs caught my eye. I am in favor of doing everything possible to lighten the reciprocating valve gear mass and then be able to use the lightest spring rate and seat pressure possible. This means the valves stay where you set them, even after 25 hours of 8000+. Additionally it takes less power to operate the valves and all the parts in the valve train are less stressed. What can you do? Titanium. The non-adjustable rocker arms that use lash-caps are also very significant.

My engines are currently single plug (then rules). I intend to make these 70.4x92 mm and twin plug. Modifying 2.4 (not 2.7) heads should let me run even higher CR. I’ll convert to twin plugs which will cost a little in CR but gain in combustion efficiency (mostly from better burn and less ignition advance). Using the longer 70.4 mm stroke will cause me to lower the max RPM.

Your engine has the advantage of a more rigid crankcase so the dimensions will hold better. Remember this doesn’t necessarily mean “longer”. Your 3.0 heads are great for your use with a 95 mm bore. One advantage of the 66 mm stroke with the 2.2 rods is the longer rod length particularly in relation to the stroke. This reduces the side loads on the piston and consequential “flopping about” of the piston in the cylinder.

The other side of very high CR is keeping the heads and cylinders cool. Up rated air cooling is important here. The “Rubbermaid Solution” is legal in everything I (my son) will run so that will be included. The turbo piston squirters will be also. With the water vaporization the 1.3:1 fan should be sufficient. I can increase to 1.6:1 or 1.82:1 if measurement shows it necessary. There isn’t an engine mounted oil cooler or heat exchangers so all the air goes to the cylinders and heads.

Please use my, Henry’s and other’s experience to build your engine so it will be reliable, put out significant power and “clean the clock” of your non-Porsche competitors.

Best,
Grady

[blue72s]

Originally posted by Grady Clay
I have three 66x91 engines

These were Mahle forged RSR-like pistons which, I suspect, are lighter than what you will use.

I don't think Mahle ever made 91's. Did I miss something?

[YTNUKLR]

Originally posted by Grady Clay
Scott,

...


I edged the CR as high as I could without the piston ever touching the head. It was above 11.7:1. It kept the piston and head clean in the squish area but never any contact.

The issue here is to use race fuel octane appropriate for the CR and keep the head temperatures under control.

Looking at your list, the valve springs caught my eye. I am in favor of doing everything possible to lighten the reciprocating valve gear mass and then be able to use the lightest spring rate and seat pressure possible. This means the valves stay where you set them, even after 25 hours of 8000+. Additionally it takes less power to operate the valves and all the parts in the valve train are less stressed. What can you do? Titanium. The non-adjustable rocker arms that use lash-caps are also very significant.

My engines are currently single plug (then rules). I intend to make these 70.4x92 mm and twin plug. Modifying 2.4 (not 2.7) heads should let me run even higher CR. I’ll convert to twin plugs which will cost a little in CR but gain in combustion efficiency (mostly from better burn and less ignition advance). Using the longer 70.4 mm stroke will cause me to lower the max RPM.

Your engine has the advantage of a more rigid crankcase so the dimensions will hold better. Remember this doesn’t necessarily mean “longer”. Your 3.0 heads are great for your use with a 95 mm bore. One advantage of the 66 mm stroke with the 2.2 rods is the longer rod length particularly in relation to the stroke. This reduces the side loads on the piston and consequential “flopping about” of the piston in the cylinder.

The other side of very high CR is keeping the heads and cylinders cool. Up rated air cooling is important here. The “Rubbermaid Solution” is legal in everything I (my son) will run so that will be included. The turbo piston squirters will be also. With the water vaporization the 1.3:1 fan should be sufficient. I can increase to 1.6:1 or 1.82:1 if measurement shows it necessary. There isn’t an engine mounted oil cooler or heat exchangers so all the air goes to the cylinders and heads.

Please use my, Henry’s and other’s experience to build your engine so it will be reliable, put out significant power and “clean the clock” of your non-Porsche competitors.

Best,
Grady

Grady, thank you, I appreciate and respect your advice very much. I was thinking along similar lines--boosting the C/R to mid-11:1 levels--but was simply concerned over the longevity vs. the same engine running 10.5-10.8:1. I am acutely aware that it is not the C/R that is going to tire this engine out, it is the speed at which it will be spinning.

The case is shuffle-pinned and fitted with Turbo oil squirters, a 3.2 oil pump (very low mileage, as-new). The crankshaft has been knife-edged, lightened and balanced to within 1g, and the rod journals have been ground down .001" (and re-hardened) to increase the rod bearing/journal clearance for more oil flow..I've seen a few alu case race 2.0 & 2.5 motors go kapooey spinning the rod bearings on #5 (and #2?) due to inadequate oil pressure/flow. I will of course still be using standard-size rod bearings.

The engine will be run exclusively on 104+ octane, and the MFI will be probably dialed just a tad rich to take care of any possible CHT issues. I plan on running a 1.3:1 fan with a 226mm (early 911SC) aluminum fan housing and fan. I need to read the "rubbermaid solution" thread again (just found it and saved the link). The car will have a second oil filter in place of the engine cooler, so the fan air will be directed at the cylinders/heads only. The car has a large front oil cooler, too . The fan ratio may be changed based on our test-runs.

The heads have Aasco springs, which apparently are not full-tilt race springs, ie. they are perhaps a "lighter" spring. I deferred to Ollie's in this case; they build many race engine heads with much success. I installed Titanium retainers, as well. At the moment, the rocker cap-adjustable RSR rocker arms are out of the question (budget), but I have forged 2.0L rocker arms that I'm planning on using. They are significantly lighter than the cast rocker arms (perhaps 10 grams although I'd need a more accurate scale to know precisely).

Based on this, and discussions with the new owner of the engine, it seems as though I am going to walk a middle ground, perhaps in the low-11:1 C/R range.

Thanks again
Scott

[davidppp]

I am most interested to read that Grady has tested very close head-piston clearances:

This is something I have also done with success.

I have one road engine running at 25 thou on the road..this is a 2.2L 84mm bore..

I wonder if grady could say what clearance he has found possible please?

I'd hazard the guess that the 91mm pistons rock a bit more and may need more clearance...but I also "know" that at very tight squish points the hydraulic forces from the mixture become very large and may stabilise the piston if it rocks...

Kind regards
David

[Grady Clay]

“I don't think Mahle ever made 91's.”
Yes, they were for the SCCA 90 mm + 0.040”
spec. They have valve cutouts for the standard
production 46 & 40 mm valves and are very high
compression. Starting with a 2.7 head they are
about 11.3:1 and you can eak it more starting with
2.2 or 2.4 heads. The pistons look similar to the
92 mm RSR but have smaller valve cutouts, 1 mm
less diameter and a higher crown. They come
matched to Nikasil cylinders. I got mine from Vasek
Polak when Arnold was parts manager (pre-Andial).

These pistons, the RSR 92 mm and others are made
in batches. This is why occasionally they are plentiful
and other times very scarce.


Scott, there is also great benefit from opening up the
sides of the cylinder spigot and matching that to the
main bearing webs. The pumping losses in a 911 are
significant. There is a (964?) check valve that fits the
breather and helps reduce the atmospheric pressure
in the case. I have considered using several of these
to vent the case above each cylinder. The 2-stroke
reed valves would work well also.

David, I have used slightly less than your 0.025” deck
but only on race engines that I intend to have apart
each 25 hours or so. I don’t trust the wear on the rod
bearing or wrist pin bushing and any stretch of the rod
or settling of the case & cylinder to not have the
clearance become “too close.” I think the critical time
is during the cool-off lap when the cylinder cools
substantially but the rod and piston are still hot. I
think the closest I ever ran was 0.50 mm – maybe a
little less. Bruce Anderson has an example (p. 178)
using 0.021” (0.533 mm).

These Mahle pistons are closely fitted to the Nikasil
cylinder and rocking isn’t the issue that it is with a
looser fit required by a JE for example. I have
measured the deck with solder on one side and
modeling clay on the other to make sure there wasn’t
some condition where the piston could contact the
head. Any contact under any circumstance is too much.

What pistons are you using with 0.025” deck?
What piston-to-cylinder clearance?



I had an interesting discussion with a very knowledgeable
SBC builder a few months ago. He was inspecting little
“pits” on a rod bearing at the location where small “spider
webs” of bearing overload failure were just starting to be
visible with a hand 10X microscope. He explained the pits
were from detonation of small bubbles of air-oil
mixture entrained in the oil. He theorized this occurs with
a combination of combustion load, inertial load, some
detonation load, too high oil temperature and too much
“foam” in the oil. I have seen this on Porsche bearings.

The classic failure is where the copper layer in the bearing
fractures and small pieces of the copper, with babbitt still
attached, flake off the steel backing. Just shy of the
failure, you can see the “spider webs” outline in the
babbitt above the fractures. I had always ascribed this
from overloading from combustion and inertia and
particularly from inaudible detonation in the combustion
chamber. I flet the microscopic pits were just more
underlying structure failure.

This, of course, is very different from oiling issues where
the babbitt is scraped off the bearing from contact with
the crank journal. This too can result in the copper
intermediate layer flaking off.

This got me thinking again about the GT3 oil system
configuration. There is a centrifugal air/oil separator.
Could this be another reason these engines can
successfully turn so fast with such long stroke and
big bore?

Here is a diagram of a, ‘02 GT3RS. It doesn’t show
an oil-air centrifugal separator but it is mentioned in
the text in several places.
"
"
© Dr. Ing. h.c. F. Porsche A.G.
"
"
© Dr. Ing. h.c. F. Porsche A.G.
"
"
© Dr. Ing. h.c. F. Porsche A.G.


I suspect there are two. One is clearly in the oil tank
and separates the air out to reduce foam in the tank.
Some other diagrams show this. The other separator
may be integral with the oil-to-water heat exchanger.
I think there is a small air return line to either the tank
or case.



I would like to see someone build a GT3 style oil tank
to fit an early 911 and a 914-6.

The key features would be:
1) Have a centrifugal separator in both the scavenge
circuit and the pressure circuit before the cooler. In the
pressure circuit it should be integral with the pressure
relief valve.
2) Have the full flow filter in the pressure circuit.
3) Have the hottest oil in the tank and the thermostatically
controlled oil cooler in the pressure circuit.
4) Mount the tank and everything else on the engine
above and where the engine oil cooler mounts. Only
the oil cooler remote to the engine.
5) Only the normal attachments to the engine at the
cooler ports, thermostat opening, scavenge out,
breather and oil pressure port.


Why would this be an advantage?
A) Better oil supply to the inlet of the pressure pump.
B) Less entrained air in the oil.
C) Improved filtering.
D) Better heat transfer at the oil cooler.
E) Lighter weight; less oil, more compact system.
F) Don’t have to drain the oil when removing the engine.

Something to think about.

Best,
Grady

[davidppp]

Hello, Grady.

Very very good thinking as usual.
I do not recall whether I ever posted my thoughts about case windage reduction, and means to achieve case vacuum.

But what I had sketched was a three-compartement case, with non return valves on each compartment.

The boxer pistons will expel air and oil when coming together and leave a high vacuum when apart.

Hysteresis loss will be low with adiabatic conditions at high revs.

A small set of scavenge pumps should suffice or may be redundant altogether.

This is a massive development excercise.

The idea of an oil tank on the engine at the oil cooler mount is also a great one:

I wanted to minimise the risk of starvation from cornering forces acting on the column of pil beteween the tank and engine.

And to maximise the vertical height of the tank above the pump inlet to asssist degassing.

The tank idea is much simpler, and I will do this Real Soon Now..

The engine I ran at 25 thou was just a rough rebuild, using a modified forged mahle VW piston ( 85.5mm) on a 2.2T bottom end..

I do not recall the piston/wall clearances..they would have been stock VW..

Kind regards
David

[Plavan]

Scott,
You sure the valves will not hit eachother with that DC80 cam? One motor with 2.7 heads with 993 intake valves cleared with a DC60. They said a DC80 cam would make them hit.
Sounds like a great motor.

[Grady Clay]

“I do not recall whether I ever posted my thoughts about case windage reduction, and means to achieve case vacuum.
But what I had sketched was a three-compartement case, with non return valves on each compartment.
The boxer pistons will expel air and oil when coming together and leave a high vacuum when apart.
Hysteresis loss will be low with adiabatic conditions at high revs.”

You are right on.

There are three opposing 2-cylinder engines spaced 120° apart. When the two opposing pistons are on down stroke, the pressure between rises. This is when the atmospheric air/oil should be expelled. On the up-stroke, there is actually negative pressure working against the back-side of the piston. Based on calculations this isn’t significant compared to the pressure pumping.

I think the critical issue is the low atmospheric pressure allows the entrained air to let the oil to drop out of the air-oil suspension. This means the oil is returned to the sump faster. The back-side of the pistons sees less restriction from compressing the air/oil atmosphere. Good ‘ol adiabatic compression losses are less. Less power lost and less heat generated.


”A small set of scavenge pumps should suffice or may be redundant altogether.”

Note how the GT3 does it. I have some important questions about the diagrams I posted above. More to come.


”This is a massive development exercise.”

No, this has been an issue since the late ‘60s. Just not many knew about it.


”The idea of an oil tank on the engine at the oil cooler mount is also a great one:
I wanted to minimize the risk of starvation from cornering forces acting on the column of oil between the tank and engine.
And to maximize the vertical height of the tank above the pump inlet to assist degassing.
The tank idea is much simpler, and I will do this Real Soon Now.”

I posted this some time ago.

Link to follow.


”The engine I ran at 25 thou was just a rough rebuild, using a modified forged Mahle VW piston ( 85.5mm) on a 2.2T bottom end.”

Cool. Post that info as others might want to use this.


”I do not recall the piston/wall clearances..they would have been stock VW.”
Those are important numbers if you can find the,

All this used to be super secret engine builder’s info. I can find all this on-line. We should have this for every Porsche engine builder so we can wax the non-Porsche competition. Sooner or later they will realize how far ahead we are and copy.

We should never be so conceited as to not look at all the other engine building info.

Best,
Grady

[Grady Clay]

Here is a good thread from ’04 about oil tank location.

”Oil Tank location“


Ok, on to the GT3 system.

EngineOilSyatem01c.jpg “
"
© Dr. -Ing. h.c. F. Porsche A.G.

“A” must be the centrifugal air separator in the sump
tank. It is shown in the X-ray diagram above.

“B” must be the ball check valve in the breather.

Is “C” in the intake after the restrictor/venturi?

What is “D”? It doesn’t make sense if that is simply
an outside vent. That would just pull in dirt and air.
Is this a breather connection to the two cam boxes?

“E”, “F” and “G” appear to be check valves. The
engineers sure didn’t want the oil to drain out while
setting. Is there another purpose?

“H” appears to be another draftsman’s error. I think
this is the filler neck.

“I” is the breather connection between the crankcase
and the sump tank.

Whoever drew the diagram missed some connections.
I put a green dot nearby each that I think are
connections (EDIT to add).

There doesn’t seem to be a connection to the #1 main
bearing. Is that an omission in the diagram or can it
be fed down the entire length of the crankshaft? I put
a green question mark ? and the
normal connection.

Note that there aren’t any oil return tubes from the
cams. All of that oil is scavenged by the four cam oil
pumps (#16) and returned to the sump tank via the
centrifugal separator. It would be interesting to know
the capacity of those pumps compared to the oil flow
to the cams, lifters and Vario.

If we add a centrifugal separator in the pressure circuit,
it should be after the oil filter (#4) and before the oil
cooler (#6). The pressure (regulating) relief valve (#7)
should be integral with the centrifugal separator so it
passes oil/air back to the tank (not to the pressure
pump as we have done for decades.) This could be a
third port into the centrifugal separator in the tank.

Centrifugal separators are very common in industry.
There shouldn’t be any lack of expertise applying this
to a new (old) Porsche dry sump system.


If someone uses the main oil pump from this engine,
you must consider the total size of the five scavenge
pumps.

I have considered doing away with the oil return tubes
and use scavenge pumps driven off the ends of the
cams (Turbo or Sportomatic) or an aftermarket pump
off the jackshaft. A small 4-section pump here could
duplicate the GT3 version simply using the oil return
tube ports in the cam housings.



One issue in a long turn is filling the entire cam box
with oil. Imagine running an engine on an engine dyno
with the engine tipped 45° with one cam box down.
This is equivalent to a long corner at 1G. The entire
volume around the rockers, cam, chain & sprocket would
rapidly fill with oil. All that gear operating under oil
would create huge losses.

The oil from the other side cam box would drain across
the bottom of the case and add to the filling.

A breather from the top of the cam housings to the
engine crankcase is something that has always been
lacking. Imagine the full cam housing truing to drain
through the oil return tubes with out the air returning –
glug, glug, glug….

How about filling the chain housing also under acceleration?
It would fill in a 914-6 on deceleration.

OilSysten45Degrees01a.jpg “
"
© 1965 Dr. -Ing. h.c. F. Porsche K.G.

With all that oil in the cam box, imagine how little is in
the sump tank. No wonder Porsche has added all these
scavenge pumps. Clearly the engineers feel that the
power cost to run the pumps is offset by the gains.

Best,
Grady


EDIT – I corrected the GT3 diagram (left off the green dots) and added content to the post.

[Grady Clay]

You know what:

These really cool old, obsolete, last century, expensive engines somehow have a LOT more life in them. It is up to us to make that happen. I am very impressed that you very knowledgeable Pelicans and many professionals choose to contribute.

There needs to be more.

There are hundreds of details that are “common knowledge” in certain circles. That needs to be available to everyone. It isn't skin off anyone’s nose.

There is a great deal that the Porsche Factory can contribute. Why don’t they provide engineering information for these long-out-of-date engines? Don’t they realize this is what helps sell new 997s and more?

I like “thinking outside the box” to help make these wonderful engines smoke the competition. With my lame old 914-6, I embarrassed the “Big Iron” 5-liter racers 20 years ago. I even took away their track records.

This is what should continue to happen today. Porsche Racers and Builders should have the combined knowledge of the past 40+ years. Add to this our collective inventiveness and track Porsches can be unbeatable.
If you know an experienced builder, admonish him to post his knowledge. He will benefit from the interaction and collectively Porsche will win.

Best,
Grady

[tadd]

Boy this thread has gone from A to Z and not quite back. It's awesome - stuff that 'I' would have never thought of, but makes perfect sense. After reading both threads given, I think they are on the mark, but no overt mention was made of the 'tensile strength' of a liquid (If this was covered, my apoligies). Like all liquids, you are good in compression, but in tension all you got is van derr wall (for non polar materials) interactions to hold things together. I have to politely poke at chucks analysis on this point. There may still be ~9psi of net pressure pushing in the 'correct' direction, but if the pump is at the source, then you are in compression for the 'long haul' and get 50+ psi in the 'correct direction'.

Just to clarify, that was a suggestion of a 'header' tank in place of the engine mounted oil cooler? I like the idea, but would this be realistic considering the rate of oil flow at 8k? It would seem to need to be a big tank (gallon maybe?).

The head scavening pump idea is awesome. Would it be worth the dollars to use the porsche parts or might one use the scavenge stages from type 1 VW aftermarket pumps?

Cheers,

tadd

[47silver]

in the commercial heating and air conditioning business they use centrifugal air separators to purge the air from chilled and hot water systems.
I do not know the pressures etc but a device like this in line may be a solution to air removal.

http://www.bellgossett.com/productPages/Parts-Inline-Air-Separator.asp

[Henry Schmidt]

Originally posted by Plavan
Scott,
You sure the valves will not hit eachother with that DC80 cam? One motor with 2.7 heads with 993 intake valves cleared with a DC60. They said a DC80 cam would make them hit.
Sounds like a great motor.

The magic of the short stroke 2.8 (66 mm stroke and 95 mm bore) is that it is not a 2.7 based engine. It is an engine based on the 3.0 Carrera or 3.0 Turbo engine. This means that the heads have 49 mm intakes and 41.5 mm exhaust with a shallower valve angle than the 2.7 heads.
These heads allow the use of any cam lift that is offered for 911 based engines.
The chamber in these 3.0 heads allow for greater compression with as many detonation issues. We pioneered this configuration years ago and with all the engines out there we have not encountered any design issues what so ever.

As I've stated many times before " The best engine Porsche never built".