Re: Weslake Chevy heads

[Duke W.]

So Jerry - what are you looking for? A set of Moser heads? The Chevrolet prototype Trans-Am heads?

I still don't think Harry Westlake ever designed a set of SB heads, but if he did, I'd like to know more about them.

Since such heads may require different inlet and exhaust manifolds, how are you going to adapt a FI system.

BTW, the problem with the ports being "too big" will likely not be a critical issue with FI. Carbureted engines need smaller ports for higher flow velocities, so turbulence evaporates the fuel. Otherwise fuel distribution and mixture quality can be poor, even at high revs.

Since port FI systems have inherently good distribution and injection pressure causes good atomization, modern FI engines can utilize large head and manifold ports that would be unsuitable for a carbureted engine.

Atomization at idle and low speed with Rocherster FI is not as good because of very low injection pressure relative to a modern EFI system, but basically Rochester FI will keep making more power at higher revs as head flow is increased. The ultimate limitation is the air meter flow characteristics as it will eventually start to choke the engine at around 500 gross HP.

Even with just pocket ported 461s, a 30-30 cam retarded four degrees, headers and open exhaust, and the air meter tweaked to 750 CFM @ 1.5" Hg. a 327 will make about 460 gross at 7000, and head flow is the limiting factor so better heads will allow more power at higher revs until the air meter pressure drop chokes the flow.

Duke

Duke

[Michael H.]

Re: Weslake Chevy heads

The size of an intake port has a lot more to do with performance than just fuel atomization. It has everything to do with the inertia of the air mass staying in motion after the piston has reached BDC. That's why performance cam design has the intake valve open much longer/later than typical street/stock designs. The faster the air volume is moving in the runner, the more inertia it has and this causes the mass to continue moving after the piston has stopped. This is the exact reason why the GM "Semi Hemi" heads were not satisfactory for use on the 305" Trans-Am engine. Although the design provided plenty of air free flow, it lacked air velocity to continue the flow AFTER the piston stopped at BDC.

There's a fine line between not enough flow and too much. Somewhere between is an efficient engine if everything else is correct. Given a few more weeks, the intake port design would have been changed/reduced and the new heads most likely gave worked well but time ran out on the project.

At least one of these engines went to Yunick and his deduction was that they needed to redesign the intake AND exhaust ports. I don't think much was ever done with it after that.

[G B.]

Re: Weslake Chevy heads

This is the "photo" I was talking about. It is from an ad selling subscriptions to Hot Rod magazine that was published on page 15 of the August, 1975 issue. The subscription ad doesn't mention the image at all. For all I know this could be a cut-n-paste job that was meant to be some kind of joke.

This image has obviously been retouched quite heavily. It is proof of nothing.

[Duke W.]

Re: Weslake Chevy heads

Inertia is a function of both mass and velocity, and bigger ports mean more mass, which keeps inertia energy high, and you don't want port velocity to exceed about Mach 0.3 (about 300 ft/sec) as this is the point where compressibility effect losses begin to add up.

So the best inlet runner design will be big enough to not allow more than 300 ft/sec, but this is usually larger than is phsically possible on a racing engine. You can delay the inlet valve opening until the runner is essentially choked, but at that point the engine is out of breathe. A later closing inlet valve doesn't make more power - only a less restictive runner will, but if there is no more physical space, the design architectue is at its limit.

Look at the LS7 heads - the CNC machined ports are about 360 cc and they couldn't make them any bigger. The exhaust ports are intentionally compromised to allow more inlet port and valve area, and this is compesated by opening the exhaust valve early, but I doubt such large head ports and manifold runners would work very well with conventional carburetion as velocity is not high enough to create good fuel vaporization and distribution until very high revs.

Duke

[Michael H.]

Re: Weslake Chevy heads

I agree, inertia is a function of mass and velocity. However, if two equal masses of air are passed through two different size tubes, the air passing through the smaller tube will travel at a higher velocity. If that high velocity can be maintained through the length of the intake port and the inlet manifold port, the inertia of this mass will continue to move even after the piston has stopped at the bottom of it's travel. This velocity, as you well know, increases with engine RPM. At low engine speed, it works against this principal and much of the air in the cyl is forced back out before the inlet valve closes. However, at high engine speed, the air flow, or pressure, is actually higher than that of the pressure that will be building in the cyl because the piston is now on it's way back up. It's a balancing act, in a way. Too early closing of the valve and you miss the flow of air. Too late, and the pressure that's building in the cyl becomes greater than the pressure at the inlet valve opening. Intake air velocity is probably as important, if not more important, than port size, as long as the port is somewhere near the size it should be for the size and expected speed of the engine. As I mentioned previously, there is a "middle of the road" on port size for best results. Too small and the required 38 cubic inches of air can't possibly squeeze through the opening. Too large and you loose velocity, and a lot of the "after the fact" cylinder filling. This is the main reason why racing camshafts use the late closing inlet side. (actually there are two reasons. We want to keep the inlet valve open as long as possible during the intake stroke, right to BDC if possible, but we can't just slam it shut a few degrees past BDC)

One of the biggest problems with the old 1st design big block heads was the fact that the inlet ports were "just too big", at least for any street application and most race engines that were under 440 inches. One of the most important changes made to these heads years later was the reduction in size for the port. The results were quite impressive. Engineering stressed the fact that the intake ports on the early design heads should "NOT be enlarged, at all".

The inlet ports on the new LS7 are quite impressive and the size (360CC) is roughly 30% larger than the old 302/327 head ports. However, we must remember that the cubic inch size of the LS7 engine is also around 35% larger than a 302/327, so 35% more air volume is automatically required at any given RPM. The amount of air the engine inhales per revolution/sec is the factor used to determine port size.

Mechanical or electronic fuel injection engines don't require a venturi, as you mentioned. The venturi is only there in carburetors to raise the velocity of passing air over the booster in the carburetor. In the Kinsler and Crower/McKay FI units that we used, there was no venturi and the Kinsler unit for small block actually used huge Quadrajet carb secondary throttle plates at each cyl.

I used to have all the info on inlet port runner speed but I can't find it, yet. I'll dig for it tonight. I did find, however, some engineering info on the Semi Hemi small block heads. In the paperwork, development engineer Bill Howell states; "A small displacement, short stroke engine such as the 302 is easily over ventilated by increasing the intake ports to the point that the incoming fuel charge velocity is decreased. It does not allow the piston to pack it in during the intake stroke, and consequently has a negative effect on horsepower/torque productivity and how it is distributed in the power band".

I'm finding a lot of engineering paperwork on this and I'll send some to you as soon as my scanner cooperates. I'm sure you'll enjoy reading this.

[Duke W.]

Re: Weslake Chevy heads

You completely ignored compressibility effects of air when the Mach number exceeds about 0.3 and certainly 0.5. That's the whole basis of the argument. If the port velocity is so high that compressibility losses become signficant, then then engine won't make more power unless the ports are enlarged or made more efficient to reduce the Mach number.

That's a limitation for all engines. You achieve an inlet Mach index of about 0.55, and at that point delaying the inlet valve closing doesn't help. The engine needs bigger ports or better port flow efficiency or it won't make more power no matter what.

With street massaged 461s a 327 hits the 0.55 "wall" at about 3800 FPS mean piston speed, which is about 7000 revs, and a later closing inlet valve doesn't help. The heads are choked. With full race ported OE heads a 302 hits the wall at about 4000 FPS. which is 8000.

Duke

[Michael H.]

GM Engineers

Maybe we can convince the rest of the engineering staff that develped the heads and claimed the ports were too large that they don't know what they're talking about.

[Duke W.]

Re: GM Engineers

I don't think they applied fluid mechanics principals to flow back then. It's very complicated because with carburetors you have two phase flow with heat transfer between the liquid fuel, air, and manifold walls and vaporization of the liquid fuel. Mathematically modeling this is a real nightmare.

I know, because that's what I was doing at the U. of Wisconsin Engine Research Center in the '68 to '70 time frame. In a way I was probably way ahead of what GM was doing (and I might add that they sponsored my research assistant fellowship). My focus was emissions related, not power, but the basic problem is the same - how do you get the fuel, in a vaporized state, to mix homogeneously with the air to provide equal quality mixture to each cylinder through the full range of engine operating conditions.

My conclusion was that carburetors were dead, and port fuel injection would be required to control A/F ratio precisely enough to meet future emissions standards, but the same can be said for extracting peak power from any engine architecture.

It took GM a helluva long time to come around, but in the end they had no choice.

Duke

[Frank C.]

GM cyl heads

Duke...I'm fascinated with your engineering knowledge. I read all the posts about the cyl. heads and I'm curious to know why no one mentioned anything about exhaust scavenging. That was a big deal back in my drag racing days (60's-70's). IIRC that was the big reason for wanting the intake valve kept open as long as possible. Keep up the good work...Regards Frank!!!!

[Duke W.]

Re: GM cyl heads

You can only have "exhaust scavenging" on a racing engine with open exhaust where you can juggle overlap and header design to maximize the negative pressure that can be obtained at the exhaust port during the overlap period, which will get the inlet process going early.

For example my current vintage racing engine model is a FI 327 with pocket ported 461X heads and a 30-30 cam retarded four degrees that works very nicely with high efficiency 1.75" x 34" headers and 3" x 48" open side pipes. A "real racing cam" with more overlap like the first design 140 Trans Am racing cam makes a little more power, but really kills the torque below about 3500 and is not reasonably streetable. With mufflers and a vacuum advance this vintage racing engine configuration could actually be driven to cruise night and would behave about the same as an OE '64-'65 L-84 - kinda torque shy, but docile enough to be driven on the street.

On a street engine, any reasonable silenced exhaust system will create enough backpressure to pretty much negate wave dynamics from headers, and increasing overlap just increases the time that positive pressure at the exhaust port causes exhaust gas to migrate up the inlet port, which dilutes the fresh charge. Result - reduced torque/power across the entire operating range.

For a street engine were the objective is the broadest torque bandwidth the best solution is efficient exhaust manifolds, like the 2.5" OE type, minimum exhaust system backpressure (The OE C2 exhaust system scores well here for SBs, not so good for BBs) and a low overlap cam (about the same as the 300 HP cam) with a late closing inlet valve to enhance top end power by allowing inlet charge intetia to continue filling the cylinder well past BDC to enhance top end power. Exhaust valve timing is dictated by head flow characterisics and since pocket porting the heads increases the E/I flow ratio to about 0.80 from 0.65 as OE machined, we can open the exhaust valve later, which means the exhaust duration can be less than inlet duration, which is oppposite most OE and aftermarket cams.

In this end I've designed a new (hyraulic lifter) cam targeted specifically for pocket ported heads. It will idle low and smooth like a 300 HP engine. Low end torque will about split the difference between the L-79 and 300 HP cams, but it will make more top end power than any production SHP cam. This cam would be a good candidate for a 300 HP "sleeper" or to "detune" a L-79 for more docile characterisitics while improving top end power. The timing events don't look anything like an OE cam or typical aftermarket cam, but they look amazingly like the new LS7 cam adjusted for the difference in relative port flow values.

Coincidence? I didn't start out with the LS7 cam, but my research basically yielded a similar design, so GM Powertrain must be working from the same playbook as me or vice versa depending on your point of view.

I do all my research with the Engine Analyzer simulation program and back it up with actual test results at every opportunity.

My new cam design is essentially done and I'm looking to collaborate with someone who is willing to fund the cost of having the cam ground for their engine.

In the meantime another "system engineered by Duke" '65 L-76 is on the road and the owner reports SOTP more low end torque and more top end than the original configuration. We should have some chassis dyno pulls next month and the plan is to pull from 1500 to 7000 or 7200 to reveal the entire torque and power curves across the range.

I'll stick out may neck and predict that I expect to see peak "SAE corrected" torque and power to be somewhere around 270, with the 80 percent torque bandwidth starting at no more than 2000 revs.

For reference a good OE Flint built 300 HP engine will make in the 190s at the rear wheels and 220 to 230 Flint built SHP 327s.

Duke

[Michael H.]

Choose One

Duke,

In several earlier posts, you and I debated inlet port velocity and the benefit of a later closing inlet valve. I stated that, by design, the later closing point allowed inlet air flow inertia to continue filling the cyl after the piston reached BDC, especially at high RPM. Your response to this was;

"A later closing inlet valve doesn't make more power--only a less restrictive runner will" (Sunday 25 June 6:22 PM)

Then, early Sunday AM, in the same discussion, you stated that;

"A later closing inlet valve to enhance top end power by allowing inlet charge inertia to continue filling the cylinder well past BDC to enhance top end power" (Tuesday 27 June 1:38 AM) (sound familiar?)

Sooooo... you now agree that a later closing inlet valve DOES make more power at high RPM?

Well get back to the inlet port size/velocity thing later.

[Duke W.]

Re: Choose One

You're ignoring compressibility issues. Once the inlet Mach index reaches about 0.55 the inlet tract essentially "choked" and closing the inlet valve later does no good. Diddling with head port flow efficiencies, manifold design, and valve timing will tell you which of the three are the limiting factors. On OE SBs it is almost invariably the heads, which is why I recommend head massaging, even on 300 HP engines, but once the head restriction is eliminated, the manifold or valve timing become the limiting factors. The 300 HP engine with massaged heads wants more cam. A carbureted SHP engine wants a better manifold, and FI with a good cam is back to the heads because the manifold is so efficient.

High velocity is good, but it's limited by compressibility effects when the Mach index approaches 0.55. Up to about 300 ft/sec (mach 0.3) air behaves like an incompressible fluid. Above this, compressibility effects increase and have to be considered in analysis and evaluation of high output engine performance.

The dimensionless Mach index parameter was originally developed by Taylor, and is discussed in his IC engine textbook. It was a good predictor of inlet tract behavior for cams with modest timing events. In the seventies it was refined by a couple of other researchers and gives better correlation for radical valve timings than Taylor's definition.

One of the nice outputs of Engine Analyser is Mach index, and once it gets into the .50s range more cam has less and less benefit. The only way to improve flow is bigger or in particular more efficient ports.

OE 461 heads have inlet/exhaust flow efficients of about 44/41 percent as OE machined. Typical street head massasing improves these efficiencies to about 49/59 percent. The best racing engine nowadays are probably at least 70 percent on both sides, and LS7 is about 60 percent on both sides.

Improving flow efficiency is more effective than just making the ports bigger with the same efficiency, and this is a "velocity" issue. Most guys think "bigger is better" because they don't understand flow efficiency. That's what pocket porting does - improves flow efficiency without making port significantly larger, so velocity is not lost.

Flow efficiency of the port is the actual flow divided by the isentropic flow of an equivalent section size and length port. We can never achieve isentropic flow, but the closer we get the better the design, and this is directly reflected by flow efficiency.

My system engineering jobs have shown that SHP 327s achieve a Mach index of about 0.55 at 6500 with the LT-1 cam and massaged heads and the later closing inlet valve of the 30-30 cam does no good at the top end, but reduces low end torque. The limiting factor of this configuration is the inlet manifold. A better flowing manifold reduces Mach index and allows the engine to rev higher and make more power before the inlet Mach index gets back up to about 0.55. A single plane manifold or FI manifold allows even more top end revs and power, but hurts low end torque, which is why they work better on racing than street engines.

For example, my models of '64-'65 L-84 and L-76 with OE machined heads show that FI offers little more top end power, so the modest 10 HP greater OEM power rating of FI was no fluke. But once the heads are massaged for better flow efficiency, the SHP manifold becomes the limiting factor, and FI (with the same cam) will begin to pull away in power by a substantial margin. That's why I think FI is best for a vintage racing engine. The carburtor guys aren't allowed to run single plane manifolds by most of the vintage racing organizations. I discussed this in a detailed post a couple of years ago.

Duke

[Michael H.]

Re: Choose One

The point in the intake cycle that is most affected by the choke, or restriction, in the small block inlet port, is slightly after the point of maximun piston velocity which occurs at about 80* ATC. From that point, piston speed is being reduced, which also reduces the speed at which the air must pass the restriction. From about 110* ATC, the air speed is back to a levl that can more easily pass through the restriction area in the port. Negative pressure is maintained and continues well beyone BDC, which aids and assists the motion of air through the port, and this motion continues to fill the cyl. I agree, there is a point at which air speed becomes too high but there is also a point at which it's too low. As I previously mentioned, there is a "middle of the road" air speed for any port design and above or below that speed, efficiency is lost. Most of this also depends on engine size and expected RPM. As you know, either of these factors greatly affect the requirements of head port flow, but the ratio remains the same.

What most people don't understand about port flow is the fact that, in reality, it's not a constant velocity and you can't get any meaningful results using just volume and velocity numbers. If improving port flow was just a matter of enlarging the dim's, everyone would figure out a way to do it. It's not even close to being that easy. Air speed goes from zero to somewhere around 62' FPS and back to zero. (not the 3800 FPS previously posted) Inertia forces the column to continue it's motion into the cyl until cyl/port pressures are near equal. This occurs well after BDC, which is the reason why the valve is allowed to remain open. As the expected RPM range increases, so too should the number of degrees for valve closing ABC.

Because the Trans-Am rules required the use of mostly stock production line engine components, the RPM range of the 305 engines were somewhat restricted. As previously mentioned, production steel rocker arms were required and the valve springs were the 3927142's that produced 310 Lbs of open pressure. Valve float, and the resultant durability issues, kept most of these engines under 8000 RPM. Because of this, the new aluminum head with it's larger inlet ports was rejected. IF the engines could be operated at a higher RPM, or IF it was a larger displacement, the heads may have worked well. (by the way, those 10,000 RPM numbers on the tach telltale most likely come from a blipped throttle on downshifts, not under power on upshifts)

As you well know and have mentioned on many occasions, making the intake ports larger for any 327 street engine, is definitely a mistake. If that same engine is going to see open header high RPM usage, that's a different story but for the street, there's almost certainly going to be slight a loss of HP. Cleaning up the transition area is one thing but making a bigger port is not going to help unless everything else works in concert. (I'm sure I'll get no arguement here) The same holds true for the new inlet port design. IF the RPM range could have been higher, or IF the engine size was increased, the 076's would probably have been very efficient.

Unfortunately, sometimes designing head ports is a little like designing headers or cam. The things that make perfectly good sense on paper just don't always work in reality. Ask the guys that do this day after day.

In the end, it's all in the balance, or ratio. The heads had a lot of very positive changes but the few negatives killed them. Port size was too big, the exhaust ports twisted slightly forward and the coolant crossover actually ran BETWEEN a pair of inlet runners, adding unwanted heat. I've seen the dyno sheets, and I believe I have them here somewhere. I don't recall any torque numbers anywhere near the 500 mark though. (they would have been excellent for the 70 Trans-Am season if the engine size was increased to 350")

I'm definitely in agreement on the Rochester FI systems. I think I more than proved my theory and beliefs about these units several years ago at Road America. There are still a lot of people trying to figure out how that was possible. Not only did the car run as fast as it did with a competioion carburetor and intake manifold, it ran FASTER! I believed in that system in 1963 and I still do today.

[Duke W.]

Re: Choose One

Your discussion of cylinder pressure and port flow applies to incompressible fluids, where, by definition, the speed of sound is infinite and decreasing cylinder pressure due to piston down travel is transmitted instantly to the top of the inlet tract, so the only "lag" issue is fluid interia.

Incompressible flow theory works fine at low and medium mean piston speeds, but when you start pushing to the limit you have to take compressibility effects into account (which includes a finite sonic velocity) and this is were modern CFD analysis has really aided the design process by allowing very accurate simulation with mathematical models. Proposed designs can be modeled and analyzed many times to find an optimum design before tooling has to be designed for prototype castings.

I can't say fore sure, but I'd be willing to wager that CFD was used to develop the LS7 head. I know for sure that it's used in cooling system design.

F1 engines have very short ports so decreasing cylinder pressure is transmitted to the beginning of the inlet tract as quickly as possible. Air flow and peak mean piston speed is essentially limited by sonic velocity considerations. For a given port length, the faster the piston moves, the tougher it is for the air flow to keep up, and at some point VE drops off to the point were power falls off.

As I've stated, pocket porting heads primarily improves flow efficiency. Port volume doesn't change much so velocity remains in the same ball park, which keeps low end torque from falling significantly, while enhancing top end power.

On a race engine larger ports can be beneficial, but if they are poorly done, flow efficiency can drop more than flow area increases and the net result is poorer air flow and lower torque across the rev range. I think most car guys don't understand this point and fall back on "bigger is better" because they don't understand the science behind fluid flow.

FI continues to get knocked by vintage racers. FI alone won't make a better power curve. The entire package has to be system engineered to work synergistically, and this is not always the case.

If valve gear is limited to OE or equivalent, then that places a rev limit of about 8000 today just as it did 35 years ago, but some vintage organizations appear to allow any manner of aftermarket race valve gear so it's a whole different ball game.

If a vintage engine is limited to OE vintage valve gear and heads (even with unlimited porting), one has a 180 degree manifold with a 750 CFM carb. and the other is FI with the flow meter massaged to 750 CFM, and both are optimized for the inlet manifolding architecture, then there's no way the carb/manifold will outpower FI over the rev range governed by gear spacing.

Of course there are a lot of "ifs" up there, and, as always, it comes down to the details and good engine system engineering.

Duke

[Michael H.]

Re: Choose One

I don't agree. Air has mass/inertia just as does fluid, so, an object in motion tends to stay in motion ...etc etc. The sound waves that also control air movement in inlet ports and inlet manifold runners are an entirely different force. The pos/neg waves are what dictate the length of inlet manifold runners/injector stacks and header tube length, but that's a different animal entirely and has little to do with inlet valve closing time variations from cam to cam. Shorter injector stacke function at higher RPM due the the reduced amount of time required for the neg wave to travel to the top of the stack and return as a pos wave. More on that some other time.

Somewhere I have a lot of info on air flow inertia at various velocities and tube diameters. If I can find it, I'll send it. VERY interesting stuff.

On your 5th paragraph, I agree, somewhat. If the port is large but poorly shaped, it can flow a lot of air. However, a port that is smaller and correctly shaped can/will flow the same amount of air (or possibly more) but at a more desireable higher velocity. This was exactly my point when describing the change in inlet port design for the original and later release big block heads. Port volume was reduced substantally but dry air flow numbers increased slightly and velocity also increased. I suppose we could sum all of that up by saying, "use the smallest port that will supply a suficient amount of air to the cyl at any pre determined RPM range". I think that says it all. Any larger and velocity falls off along with efficiency. Bet those new LS7 heads wouldn't produce as much power as the old 461 heads on a 302/327.

If I remember correctly, valve float began around 8400 (with new springs) in my Camaro. That was a Trans-Am 305 using the 3927140 cam with the correct 3927142 valve springs. A typical upshift was in the range of 7600 RPM. As mentioned previously, there were times that these numbers would be far exceded on downshifts when the throttle was blipped between gears. The SCCA rules required the use of this cam and rocker arm. Roller cams/rockers were not allowed at that time (67-69) but I'm pretty sure they are today.

Bet everyone is getting tired of reading this. I enjoy it but probably very boring to most.

By the way, if you DO want to play around with injector stack height and header primary tube length, find the time, in seconds, that the intake or exhaust valve is off it's seat, at a given/prefered RPM and using a 64-65 FI cam. Use a factor of about 1670 FPS for the speed of the exiting pos/neg wave and from that, determine the prefered length for either intake or exhaust tube. Hope I remembered the numbers correctly.