Click on a thumbnail to see the full-size image
Performance Heads-Using Software And Templates
By Ken Weber
In days of old when men were bold and engines weren’t particular, we changed the cam and added carbs, and our performance was spectacular! Okay, so it’s a bad rhyme. But the point is that in the early years of hot rodding, factory iron was easy to improve. Milling the heads, adding a hot cam and dual exhaust were all simple changes that made significant performance increases.
As OEMs moved into the performance arena, their huge resources added increasing sophistication to production engines, which in turn have required that modifications for performance improvements be more sophisticated as well. The old approach of more cam and more compression still applies, but the more highly refined an engine, the greater the possibility that any modification will result in a step backward.
Nowhere, as much as in racing, where every horsepower is important, is this more true. The successful interaction between the various components of an engine package is vital to the overall performance of the engine. Today, it’s very easy to select a combination of parts that just "negate" each other. The wider the power band needs to be, the more difficult it is to hit it just right. I’m not implying that the engine won’t run, but the difference between a good and poor combination could mean 50 or 100 horsepower.
So, how does a rebuilder determine the best combination of parts for a particular engine application? Once the basic dimensions of the engine have been established, the interaction between the induction system, cam timing, and exhaust system should be investigated. Again, the more sophisticated the package, the more critical each individual item becomes.
In the case of a customer wanting a rocket for his street rod, someone want-ing more passing and/or hill climbing performance, or the boater wanting that extra 5 miles per hour, some increase in performance can be relatively easy to achieve by the old methods of a little more cam, more compression, etc. But creating a really "stout package" is much more involved.
Using computer power
Thirty years ago, about the only way to evaluate a "performance combination" was to build the engine and either flog it on the dyno, or build and ship it and hope it worked. Today, the processing power of the desktop computer, combined with an engine simulation program, has made the evaluation of different parts combinations a desk chair project.
This doesn’t mean the dyno is dead, just that it’s easier to come closer to the correct combination of parts and compromises before building the engine. The dyno becomes the next tool to prove the results. Of course, trial by fire is still the final determinant.
I use a Dynomation engine simulation program from Audie Technologies. It has inputs for air flow data from the flow bench, cam specs, using either SAE .006" opening and closing events and lifts, or the cam profile can be input directly from Cam Pro, Valve Pro and the like.
You can input intake runner length, minimum port area, port entry area and taper, and valve size, or you can accept the default air flow for the valve size. You need all the exhaust system specs – which can be calculated or guessed to start – and, of course, the basic mechanical specs of the engine plus fuel type and carburetion. From this, the program will generate power curves and pressure-angle graphs for any power band you choose.
For me, simulating engine combinations on the computer is as addictive as a video game. My son has dubbed my desk "the mad scientist’s laboratory." I get started with the intent to check out a couple of changes, and before I realize it, it’s well after midnight. It is easy to try different header combinations from 1-7/8" up to 2-1/2" tubes, and lengths from 15" to 50" for each tubing size, and then try these possibilities with different collector combinations.
You may say that "experience" can shortcut a lot of this computer simulation. However, the beauty of the computer is the ease of trying unusual combinations without spending a great deal of time or money. Once you identify a trend, it seems logical to take it as far as it goes.
In my current engine project – a big block Chevy – I first established some basic goals. This engine is for street use with an automatic transmission, is 555 cubic inches (4.560" bore and 4.250" stroke), has aluminum cylinder heads, solid roller lifter cam, tunnel ram-based EFI, and uses unleaded regular gas (10:1 compression). It must also idle well enough to not require a racing converter.
The desired horsepower is at least 500 with the widest rpm band consistent with the desired idle quality. Peak torque should occur around 3000 to 3500 rpm. This engine would be a good combination for a marine IO application because the power band is similar..
I chose an Eagle 4.250" stroke, 4130 crank and 6.800" steel rods, a GM Mark IV tall deck Bowtie block, and JE flat top pistons to build up the short block. The 6.800" rods were selected to help extend the rpm range of what I knew would have to be a short duration camshaft (for idle quality). They were also chosen to reduce side loading on the cylinder walls because I intended to bore the block to its design limit of a 4.560" bore.
I would have used 6.900" or 6.950" rods, but they were not available as off-the-shelf items. The longer rods also reduce the peak piston speed which occurs between 73° and 82° before and after top dead center. This reduces the peak air flow demand in the intake port, allowing smaller ports, and in turn, tends to widen the torque curve and maintain low speed efficiency. I considered a 4-1/2" inch stroke, but felt the desired torque, horsepower and power band could be met with the shorter stroke.
I have often thought that the port size of the oval port production heads would be a good starting point for an application such as this. But I wish they had higher intake ports and better exhaust ports. There are aftermarket oval port heads such as the Dart Merlin series and the Edelbrock aluminum versions, but the port location at the intake gasket surface is still in the stock location.
I selected a pair of Edelbrock Victor race heads, which have a .100" raised intake port at the manifold surface, a much deeper bowl under the valve, and which use a .400" longer intake valve. The port makes a high turn to the valve, resulting in a much straighter approach to the seat on the short turn than production heads.
The heads also have raised valve cover rails, and additional material on the roof of the intake port, allowing it to be raised even higher. The exhaust ports are raised .750" and have about the same curve from the seat to the exit as a banana.
The intake ports are rather large at 370 cc. They have a cross-section of 4.5 sq.in. in the bowl, taper to 3.80 sq.in. at the pushrod, and expand out to 4.38 sq.in. at the gasket. The pushrod side of the port is widened a bunch at the point where the floor starts its turn to the valve, resulting in the centerline of the port making an S-turn to the valve.
Plans for the heads
Based on the above information, my plan was to make a smaller cross-section raised port head, losing as little flow as possible in the process. I planned from the start to fill the intake port floor to compensate for raising the roof, reduce the volume, and raise the centerline of the port; so it’s no big deal to do a little more filling to generate the cross-sections needed, and straighten out the turns. I hoped that raising the port and making it straighter would help overcome some of the flow lost by making the port smaller. But how small was the question?
At this point the engine was simulated on the computer to develop an idea of how the parameters met the objectives, and to determine intake runner dimensions. For a starting point, intake length and header size and length were calculated from formulae provided in the Dynomation manual. This information indicated a 22" intake length from the valve face to the runner opening, to tune it for the 2nd order harmonic at 5000 rpm. The exhaust formula calculated a 2.20" ID header tube at 38.5" long and a 4" diameter x 20" long collector.
The program asked for minimum port cross sectional area and the area at the port entrance. I used the area of the port throat under the valve (3.10 sq.in. for a 2-1/4" valve) as the minimum port cross-sectional area, and 5 sq.in. for the entrance at the plenum.
A straight, constant area port supports the low end of the power curve and tapering the port (like a funnel) helps the upper end. The size, length and taper of the intake port, cam events (intake closing and exhaust opening in particular), cam duration and exhaust system dimensions are the compromises that can be juggled to shape the power curve. You can broaden it, at the expense of peak values, or provide higher peak values, in a narrower band, whichever suits the application.
The neat thing about using the computer is the ease with which these variables can be moved around until you find the combination that fits your application. For the initial cam selection, I used the shortest duration street roller lobe, 268° (224° @ .050") @ .595" lift, in the Competition Cams catalogue and entered 110° intake and exhaust lobe centers.
The first simulation met the power objectives right out of the box, showing 708 ft. lbs. of torque at 3200 rpm, and a peak of 502 horsepower at 4250 rpm. However, the power curve was full of dips and spikes, and power fell off rapidly above the 4250 rpm peak, down to 318 horsepower at 6500 rpm.
A study of the pressure-angle graphs indicated that the cam wasn’t working well with either the intake or exhaust systems in the upper rpm range. I decided that 700 ft. lbs. of torque was adequate for street use and probably would be difficult to hook to the pavement. So, I sacrificed some torque below 3000 rpm in exchange for better power above 4250.
After a series of simulations, it became apparent that 22" was too long for the intake for the wide power band desired. The combination evolved to two possibilities. A cam change to either a 276° .595" lobe at 114° separation, or a 280° .623" lift lobe at 116° separation angle. A 22" (from the valve face) intake tract is not easy to package under the hood, and like the exhaust system, a shorter runner worked better.
I simulated runners from 12" to 22" total length, with inlet areas from 3.5-6 sq.in. I settled on 14", with a 5 sq.in. inlet. It would be a huge amount of work to do this on the dyno or in the vehicle. The exhaust specification was also changed to a 2.120" ID tube, 26" long, with a 3.5" x 13" collector.
With these changes, the 276° cam produced 684 ft. lbs. @ 3000 rpm, 556 hp @ 5200, and 509 hp at 6500. The curve was smooth and had no dips or lumps, and the pressure graphs looked good all the way to 6500 rpm with no significant intake or exhaust reversion anywhere in the range.
A change to the 280° cam with no other changes made 682 ft. lbs. peak torque at 3500 rpm and boosted the peak horsepower to 580 at 5000 rpm. The engine still made 539 horsepower at 6500 rpm. The 280° combination drops from the 276’s 590 ft. lbs. to 574 at 2000 rpm.
Either combination should have a usable power band from the low 2000 rpm range to at least 6500 rpm. But the 280° cam raises a question about the stated idle quality goal. From experience, port injection can be made to idle somewhat smoother than a carburetor, so the possibility exists that this cam will prove satisfactory in that area as well. It doesn’t have to idle like a Rolls, just well enough to work with a stock converter.
On the exhaust side, the pressure waves from the primary tubes and the collectors have a large influence on the cylinder pressure that is resisting the piston travel (and crank rotation) as it rises in the cylinder during the exhaust stroke. I spent some time trying to reduce the pressure on the piston between 90° before, and top dead center, especially in the 73° to 82° before TDC range where piston velocity is highest. I did this in order to reduce the pumping losses.
By concentrating this effort in the 2500 to 3500 rpm range, it should pay off in increased fuel mileage during cruise conditions. The longer the primary pipes and collectors, the bigger the hump each creates in the power curve. But you have to remember that for every hump, there is an equal and opposite "un-hump" somewhere else in the power curve. You gain in one area and lose in another.
As an example, I once built a set of 180° headers for a small block ’62 Nova I had. At some areas in the power curve, the headers created so much reversion in the intake system, that if the air cleaners were left off, the carbs would spray large amounts of fuel on the underside of the hood and stink to high heaven. I was afraid the car would catch fire, so I never ran it without the air cleaners. These were, by necessity, long, small diameter primary tubes combined with long (6 to 8 foot) collectors.
Time to make chips
With the desired parameters of the engine established, the task of making the components match the computer specifications began. The cam was easy enough, a phone call with the specs gets the part by mail. The headers and exhaust system will be fabricated in the vehicle, and is a matter of cut and weld. This requires the engine in the chassis, so it comes later. That leaves the cylinder heads and intake.
With the high level of media attention to computer numeric control (CNC) ported heads, some customers seem to think that CNC porting is the magic ingredient that will lead to unbelievable horsepower. It’s hard to explain to them that a CNC mill only repeats what has previously been developed with a hand grinder.
Shops involved in building multiples of the same engine combination, often have CNC equipment that can duplicate a good working port time after time. But it can also duplicate a poor working port time after time, too. However, even if a shop has the equipment, it can’t always justify the time to program the equipment to do a single set of odd ball heads.
So, without computer-controlled equipment, how can we get consistent results from port-to-port and head-to-head? The eyeball is good at detecting variations in individual surfaces, but not so good at comparing one shape to another similar shape without superimposing them.
One technique is to make patterns or templates that match the contours of each side of a port and are referenced to the intake gasket surface and the valve guide. I make the patterns out of file folder material, rough cut with scissors, and do finish shaping with a die grinder. I make one pattern for each of the four walls of an intake port and likewise for the exhaust port.
Heads like a big block Chevy require two intake sets because the ports are not mirror imaged from the left port to the right port. On round ports or oval ports, I simply make a pattern at 12 o’clock, 3 o’clock and so on. Some time back, I noticed that Chrysler had patterns similar to this, for the 383-440 heads, available through its Direct Connection program. I also use inside calipers of various sizes with extended points and depth markings on the legs, to measure the width and height at different depths within the ports.
Another useful item is two-part urethane molding rubber. It’s expensive, but you can take a mold of the port and then cut it into sections and measure the area of each section. I look at the molding rubber as a tool for doing extensive port development where absolute maximum performance is the issue.
The patterns and inside dividers, if used carefully, can come close to the accuracy of the molds in the straight sections of the port. However, the molds are more accurate in the turn to the valve and the bowl because of the irregular shapes, especially when the valve guide is part of the picture. If you’re doing a single pair of heads for a motor home, you probably won’t want to make port molds.
In developing the cylinder heads, Dynomation can be used to calculate the dimensions of the intake port (or manifold), and bypass a lot of trig calculations. By entering a distance from the valve as the port length, and adjusting the port inlet area until the degree of taper matches the design requirement, it will give the cross-sectional area at a series of stations along the port, like ribs in an airplane wing. Once the port dimensions are established, they can be applied to the head using patterns and dividers; flow testing can verify the results. In these heads, I raised the intake port roof at the intake flange 3/8" more than the as-cast position, which is already .100" higher than a production GM square port head. This produced a .475" increase over the GM production square port location. The sides of the port through the pushrod restriction are widened and the floor is raised in order to reduce the size to correspond to the computer simulation’s 2.06° taper.
The port tapers from the smallest area under the seat of 3.1" to 3.55 sq.in. at 3" from the seat to 3.99 sq.in. at the manifold flange. The resulting port is 2.28" high and 1.75" wide at the intake manifold gasket surface. The intake manifold will taper from this out to the 5 sq.in. area at the plenum.
Flow results were encouraging, having lost only 3 cfm at .700" lift from 249 cfm to 246 cfm at 16" of water pressure. At .600" lift there was no change, and between .200" and .400" lift flow increased by 2 cfm.
The original plan for this article was to have the engine finished and dyno tested for a comparison between the simulation and the real thing. However, sonic testing three different blocks revealed a consistent pattern of core shift, making even a 4-1/2" bore finishing out with less than a .200" thick cylinder wall. Obviously, a 4.560" bore would be even worse. For peace of mind, I would prefer to stay above .200" wall thickness.
Several calls to GM Partech (GM’s under publicized performance division) revealed that there has been a recent run of blocks like this. If you are planning to fill the water jackets for a drag race application, it shouldn’t be a serious problem. But be sure to check the block before you use it. The GM CNC race-prepped Mark V block was to have been made available in January. Reports are that this block is better, and of course, more money. The new plan is to use one of these as soon as they are available.