MFI Pump - Open Heart Surgery

[911st]

Thx.

I suspect you are past the sweet spot for 1.5" SSI's. Carrera's near that HP range pick up a solid 10hp with 1 5/8's headers.

Not a cam expert but I believe unless one has very large ports, adding lift may not be the big goal. It seems to just come along with increasing duration or in search of a steeper approach angle for faster valve opening.

It seems most 911's by about 50% of valve lift may start to run up against the ports as the restriction.

[Flieger]

It is about area under the valve lift curve. If you ramp up steeply/quickly to moderate lift you do not need so much duration and overlap. This needs stiffer springs.

[911st]

I belive the term is somthing like Port Limited. That is the area under the valve seat is much greater than the cross section area of the port.

Thus, I suspect it is more like the bottom half of the valve lift curve.

I do not think we are trying to achive more lift as the end goal per say. We just have to go higher when we go for longer duration to accomidate the valve train -- valve, retainer, and rocker mass.

I guess we could employ steeper ramps to overcome the lack of lack of vavle overlap we can run and still make a smog cam. That would also bring with it more lift.

I could be wrong.

[kenikh]

Nope. Lift and duration are not related. These are the variables that define a cam:

Lift = Maximum valve height on the nose of the cam lobe
Duration = How long each valve stays open, in degrees
Overlap = the degress both valves are open at the same time
Lobe spacing = The angle in degrees between the midpoint in the cam lobes (I vs. E)
Cam lobe profile = The shape of the lobe curve on the lobe ramps
Ramp Speed = The velocity of the valve given the steepness of the cam lobe profile

Lift is an infinitesmally small point in time measurement and has no mathematical bearing on duration.
Duration is in degrees, as time open is infintely variable based on RPM.
Overlap is like duration in terms of time.

Thus, the higher you lift, the more potential air you can get in the motor at peak lift. This is where overlap and lobe profile come in, as the more degrees your intake valve stays open and the higher it can stay open over the entire duration of the lobe, the more "time" and thus volume of air you can get in the combustion chamber.

Cam lobe profile is actually a collection of infinitesmally small points on a curve, defining how much lift you have at a given degree mark.

Thus imagine two theoretical cam profiles at opposite ends of the spectrum: One a perfect square, the other a triangle whiche terminates into a point. Both can have identical lift and duration, since, the moment the valve lifts is when you calculate overlap. The square opens the valve to full lift instaneously, the pyramid only achieves full lift at the point at the top of the pyramid.

Both make identical lift, have identical valve spacing (mid points are same in degrees, identical durations (valve is lifted above the seat the same number of degrees) and identical overlap (both I/E lobes are open the same number of degrees), but the square lobe flows much more air as it opens the valve higher over the entire degree range than the pyramid.

Since since air volume is calculated as a function of mean lift in degrees under the curve, the pyramid clearly can't get as much air in the motor.

Thus the limiting factor in cam design is physical and material.

Since you can't accelerate items with mass instaneously and since items with mass have momentum and since this momentum needs to be countered to return the valve to a closed state in a controlled manner, lobe profile is THE secret sauce in cam specification. When was the last time you saw a cam grinder post their cam lobe profiles? The fine interplay of lift and duration in being able to close a valve reliably, given available springs (and valves/cam fingers in a desmodromic system) is what drives the limits of cam design.

Finding the right cam for a motor must take into account how much air you want to get into a motor at a specific RPM, since physics (and dollars) decide what is possible in terms of acquiring valve gear that can reliably close a valve without momentum running amok, leading the uncontrolled condition called valve float - valve float being the state where the momentum of the valve exceeds the ability of valve spring tension to maintain valve contact with the lobe, thus avoiding a catastrophic crash of the valve into the piston and valve seat.

This is irrespective of other conditions that must be accounted for like intake velocity, reversion and scavenging, which all of the above factors must be tailored to suit.

So, the net-net of this very long ramble is: overlap and lift, while related in the holistic view of cam design are entirely unrelated (as discrete entities) in terms of numerical specification of camshafts.

Clear as mud? Oh well, I love this stuff...

[Flieger]

So is it not the area under the curve of valve lift vs. degree?

[kenikh]

One addendum...duration is often measured at 1mm lift, not zero. Here's a nice couple of diagrams. The first shows three similar cams, with different lobe profiles. The second shows to intake lobes graphed with an exhaust lobe. You can see what I mean in terms of "area under the curve" and duration/lobe spacing/overlap:

3 lobe comparison:


Intake/exhaust overlap diagram:

[Flieger]

So I was making a true statement.

[kenikh]

Originally Posted by Flieger

So I was making a true statement.

Well, yes. This is the point I was making. Did I misread something?

I thought that the question was 'does lift have an effect on duration'...the answer is no.

[Flieger]

I do not know if that question was asked. In that context, I see that you are saying that they are not related other than by valve inertia. To get extreme lift, you need more time since the valve cannot instantly be opened or closed.

[Jeff Higgins]

There were criteria other than maximum power that entered into my choice of camshaft. One of the key features of this grind is a gentle ramp angle, allowing the use of stock springs, retainers, and rockers. Aside from budgetary concerns, the decision to stick with stock was based upon the desire to have a long-lived street motor, free of the rapid valve train wear inherent in heavy springs. The stock valve train can easily support 7,000 rpm, or a bit more, with the proper cam profile.

The two or three (depending on how you break them down) major components of a camshaft's design that affect its "personality" are lift, duration and overlap (those last two are very closely related, which is why I say "two or three" components). In the very simplest terms, lift determines how much power (in terms of torque) the motor will produce, and duration/overlap determines where. More lift = more power, more duration and overlap = power higher in the rev range, and ultimately then, more horsepower.

I didn't want to have to rev this motor to the moon to make horsepower. The idea behind this high lift / short duration grind is to provide enough lift to let the 3.0 liter breathe adequately, but to provide gobs of mid-range torque rather than screaming high rpm horsepower. I think the dyno chart demonstrates it worked.

There were, of course, other considerations when looking at the package as a whole. I had on hand a set of freshly rebuilt "S" throttle bodies and stacks I wanted to use. They are 36mm, a far cry from 3.0 RSR spec. The core motor on hand was an '83 SC, or small port 3.0 liter. The options were to open everything up to 44mm intakes - stacks, throttle bodies, ports - and to open the exhaust ports commensurately with that. Add a set of RSR sprint cams, and viola - a 350 hp, 9,000 rpm beast with about a 30-40 hour lifespan. I didn't want that. I wanted a torquey 7,000 rpm street motor. So, stacks and throttle bodies were left alone, and intake ports were opened to 38mm to match the throttles, with a 1mm "reversion ledge" around the base.

With my criteria for a hot, high compression, moderate rpm 3.0 liter street motor, I was up against a relative dearth of suitable cams. On the one hand were the standard "CIS friendly" 20/21 or 964 grinds. On the other hand were the GE/DC 60 or 80 series, but they are meant for higher reving motors and need heavier springs. Some guys like the "S" or mod "S" grinds, but they have too little lift and too much duration. There was really nothing out there.

That's where John Dougherty came in. I can't say enough about "camgrinder" as he is known in these parts. He understood my goals, and the dilemma those goals created. He came up with the perfect solution, for my motor and my objectives. To say I'm pleased would be the understatement of the year.

Yes, I think the 1 1/2" SSI's are choking it a bit. That's why I have a set of 1 3/4" European Racing headers and megaphones on loan from a friend. Maybe too big for my mid-range torque monster concept, but hey, they are on loan. Beggars (borrowers) can't be choosers... We'll see in a couple of weeks if there is anything to be had here.

[911st]

Again, I am not an expert.

I agree with all the above with one reservation.

The point I am trying to make is the concept of a "port limited" motor.

Example:

Our intake valves are 49mm.

A cam such as a mild GE 40 has lift of .470" lift.

This creates a curtain area under the valve seat at full open of up to 2.85 sq inch.

36mm intake port has a cross section of about 1.58 sq inch.

Thus, with a valve curtain area of 2.85" and the ports of 1.58" , the ports are limiting flow as they are only about 55% of the area under the valve.


Knowing this, what is the best way to get more air into a motor?


It is a different story with big 42mm ports. Then you get 2.85" valve area to about 2.6" at the ports.

Now we have a 'vave limmited motor' that would probably respond very well to increassing lift.

[kenikh]

Ah, I gotcha now. Let me try to regurgitate something Steve Weiner told me in this context. Generally, 911 motors are "over valved", as opposed to being port limited. Also, 911 motors, especially early cars, are over-ported, to boot. Basically, the valves are already so big, that they will over flow all but the most aggressive ports. 42mm ports are 3.0 RSR territory.

So yes, the valves greatly outflow the ports. But my understanding is that the valves size is there for headroom. This allows flexibility in design without having to cut and install new seats, which is much more intensive than enlarging and recontouring the ports.

To your point, I think that in a port limited motor (that can fully express valve diameter potential), lift won't necessarily help as the goal of increasing lift is to get more air into the chamber in less time. As volume flowed is aperture dependent as a function of maximum volume over time, if the motor is already port limited at full lift, to get more air in, you will need to increase the amount of time the valve is open.

This is the eqivalent of pouring out beer into a glass: if you can only pour so much liquid from the bottle neck (limited port) into a glass (valve, in this case), pouring (time valve is open) for 5 seconds versus 10 seconds, 10 seconds gets more beer in the glass.

Last note: you appear to be discounting the area discount from the valve stem and petal, as well as the resistance to flow the valve causes that an open whole would not. Not entirely apples to apples.

[911st]

That's it.

On the valve stem. Just guessing, it might be possible that the cross section there is less than the curtain area or the port area. Probably not on a stock head. However, if one just opened up the port and did not concern them selves with porting around the pocket that could become an issue.

Also, remember any increase in duration is most going to be at points where there cylinder volume is more static because of the rod angle and piston speed. Thus, things like larger ports or valves if they are the restriction are going to help.

Overlap generates its own potential by improving cylinder evacuation. It this case, more duration can increase overlap and its benifit.

[kenikh]

[911st]

[dicklague]

Originally Posted by jeffc280sl

dicklague,

Have a look at the afr chart on my 3/26 post. Prior to tuning I had the same sort of profile that you have. It would be nice if you had a map sensor input but I can see a lean condition occuring as you begin accelerating. Throttle opens wide and the mixture goes lean for a period of time. That can be tuned out so that afr is steady over the entire acceleration sessiuon.

Great! Can you be specific on how it can be tuned out?

Pump does not have warm up valve, but a screw for fixed setting.

This is what I have done so far:

1. all CMA settings done including the 114mm pump rod length.
2. Idle leaned out to a nice 13.0- 13.5.
3. Partial throttle 12 AFR
4. WOT 13-14

Still have the leaning on throttle opening.

What are the specific things I can to to cure that?

[jeffc280sl]

In very simple terms you want to adjust the white, black and rack screws so the stylus reads a slightly different place on the space cam. I know Porsche has specs for co during a road test at 7 degree throttle with a brake load to achieve 2500 rpms. Have you been tuning in an effort to meet this spec? I have yet to see a Porsche co spec for full load road test. MB has two partial load co specs and a full load co spec. I tuned to meet those specs and it virtually eliminated the very lean condition at the start of wot test.

Does anyone know of a full load co spec?

[356RS]

I've looked for a full load CO% and also can only come up with the partial load specs of 2400 rpm, with a 9 degree throttle: T = 1.5 - 2.0%, E & S = 2.0 - 2.5% Euro = 2.0 - 3.0% and the 2.7RS = 2.5 - 3.0%

[jeffc280sl]

Why not establish a spec for a full load test? Something reasonable to shoot for. MB full load on road test spec starts at 3k rpm in third gear and sets afr between 13 and 13.77 during the wot period.

[356RS]

I think we have already established that IMHO, at least for the modified street/DE engine. Seems like after reading all the data on this MFI thread Jeff started, 12.5 to 13.5 for a WOT run looks to be a very good AFR goal. Most of the engines built today are far from stock and it's become a challenge for the MFI pump tuning to get it right down a steady AFR path through the RPM range. So IMHO the goal has been set and all the highway speed hand held AFR meters and engine dynos are helping to get it as close as we can.