Our era is often referred to as the Information Age, but not all of the available information is necessarily useful. I am beginning to think that flow benches should be labeled with a government warning: "Caution! Excessive reliance on flow numbers may be harmful to your engine!"

I'm kidding, of course. Used wisely, a flow bench can be a useful tool in engine development, just like a timing light or a dynamometer. Unfortunately, some racers believe that a flow bench is the ultimate answer machine. When the subject is cylinder heads, the four words I dread to hear are, "What do they flow?" Novice racers and magazine writers share a fixation about airflow. The mistaken belief that "more is better" is often the false assumption that produces an underperforming engine.

I learned this lesson myself when my partners Buddy Morrison and Lee Shepherd built our first flow bench in the mid-'70s. It was a great contraption that could just about suck the windows out of our rented shop on Arkansas Lane. While this homebuilt test bench boosted our racing program, it certainly didn't make us engine experts overnight - even though we initially thought we had found the key to the vault of knowledge.

We had been racing 287-cubic-inch small-blocks in various Modified and Comp classes before we decided to make the move to Pro Stock with a 331ci engine. (Students of Pro Stock history will recall that the '70s was the era of weight breaks for various engine and chassis combinations.) We were determined to be "scientific" in our approach, and reasoned that a 15 percent increase in engine displacement demanded a 15 percent increase in airflow. We dutifully enlarged the ports, increased the valve diameters, and hit our airflow targets. We set off to conquer the world of Pro Stock - but our pride and joy was a dog.

After struggling to even qualify in our initial outings, we pulled an old pair of Modified heads off the shelf. Lee worked on the ports for an afternoon, we bolted them on our Pro Stock short-block, and we qualified fifth at Englishtown in our next race.

If you went strictly by the flow numbers, those heads would hardly enough air to satisfy a respectable big-inch bracket racing engine - and yet they were magic on the race track. That was when I realized that cfm isn't everything. It's a lesson that I have seen repeated countless times in the last 25 years.

A flow bench measures air movement in a very rudimentary way - steady-state flow at a constant depression (vacuum). Obviously the conditions that exist inside a running engine are quite different. The flow bench can't simulate the effects of the pistons going up and down, the reversion pulses as the valves open and close, the sonic waves that resonate inside the runners, the inertia of the fuel droplets, and all of the other phenomena that influence engine performance in the real world. When you flow test a cylinder head, you are simply measuring how far you can move the liquid in a manometer.

The bigger you make a port, the more it flows. That's hardly shocking news. Bolt a sewer pipe onto a flow bench and it will generate terrific flow numbers. So should we use ports as big as sewer pipes on our race cars? The flow bench says we should - the time slip says something completely different.

If airflow were everything, we would all use the longest duration camshafts we could find - after all, more duration means more flow. In fact we know that there is a finite limit to how long the valves can be open before performance
suffers. That is because the valve events have to be in harmony with the rest of the engine.

The same principle applies to cylinder heads. Simple airflow capacity should never be the first consideration in evaluating cylinder heads. Characteristics that are far more important include air speed, port cross section, port volume and shape, and the relationship between the size of the throat and the valve seat. If these attributes are wrong, you can work forever on the flow bench and not overcome the fundamental flaws.

Here is a do-it-yourself example: Turn on a garden hose and the water will dribble out a couple of feet. Now put a nozzle on the hose and the water will spray across your backyard. The water pressure and volume haven't changed, but the velocity has increased dramatically. Now think about the air and fuel going into your engine's cylinders. Which would you prefer: slow and lazy or fast and responsive?

An engineer will tell you that an engine requires a prescribed amount of air and fuel to produce "X" horsepower. In a perfect world, that may be true - but we race with imperfect engines. The shape and cross-sectional area of
the runners are absolutely critical to performance. For example, I have two sets of Pro Stock cylinder heads that produce nearly identical flow numbers, yet one pair produces nearly 150 more horsepower at 9,200 rpm than the other. The flow bench can't tell the difference between them, but the engine certainly can.

There are software programs that claim to be able to predict an engine's performance based on airflow numbers. Unfortunately, a critical shortcoming of many of these programs is that they are based on inaccurate information or false assumptions. A computer is an excellent calculator, but it is not an experienced engine builder. The software doesn't know whether a port's short-turn radius is shaped properly, whether the flow is turbulent at critical valve lifts, or whether the flame speed is fast enough. Racers have a tendency to believe that computers are infallible, so they accept the software's solutions as gospel, when in fact they may be badly flawed.

Textbooks would lead you to believe that an exhaust to intake flow ratio of 80 percent is ideal - yet a typical Pro Stock head has exhaust ports that flow less than 60 percent of the intake runners. You can improve the exhaust flow
tremendously with about 40 minutes of work with a hand grinder - but the supposed improvements will just about kill the engine's on-track performance. I know because I've been there.

We have also learned that low-lift flow (meaning anything below .400-inch valve lift in a Pro Stock engine with a .900-inch lift camshaft) is relatively unimportant. Think about the valve events in a racing engine: From the point when the valve first moves off its seat until it reaches mid-lift, the piston is either going the wrong way (that is, it is rising in the cylinder) or it's parked near TDC. The piston doesn't begin to move away from the combustion chamber with enough velocity to lower the pressure in the cylinder until the valve is nearly halfway open. Consequently it is high-lift flow that really matters in a drag racing engine.

The shape of the combustion chamber also has a significant impact on performance. A conventional chamber with deep reliefs around the valve seats and a relatively flat valve seat angle can produce terrific flow at .200 to .300-inch valve lift. Today a state-of-the-art chamber typically has 55-degree valve seats and steep walls that guide the air/fuel mixture into the cylinder to enhance high-lift flow.
This doesn't mean that every racer needs state-of-the-art Pro Stock cylinder heads - along with the high maintenance they require. The heads have to match the application. Conventional combustion chambers and 45-degree valve seats are just fine for a dependable, low-maintenance racing engine that will run a full season between overhauls.

The classic Hemi combustion chamber is capable of producing impressive flow figures, but it's not going to make impressive power. Engine technology in all forms of motorsports is converging around smaller, high-efficiency combustion chamber designs. You can see the result in lower brake specific fuel consumption (BSFC) numbers, which indicate improved engine efficiency. Twenty years ago, a racing engine with a .48 BSFC was considered very good; today's competition engines produce BSFC numbers in the neighborhood of .35. This means that a given quantity of fuel is being atomized and burned more effectively to produce more power. A cylinder head's combustion efficiency can't be measured on a flow bench, yet it has a huge impact on performance.

I am not against flow benches; in fact, we use computerized flow benches daily at Reher-Morrison Racing Engines. What I am against is over reliance on flow numbers as the primary measurement of a cylinder head's performance. A flow bench is a valuable tool that can help a racer fine tune a combination - but it is not the ultimate authority.

I was speaking with (a very well known cylinder head porter) a few days ago about alot of these other heads on the market. He has had a few sets of (brand X heads) in his shop and has kinda seen what we have found as well: they flow well on the bench (as good as his in house head), but don't seem to make the same power on the dyno,etc. From talking with him, he was mentioning that (brand X) removes material in areas that will show a gain on the flow bench, but will not make anymore power on the dyno. He says they are big in the wrong spots and too small in others.

As an example, a set of (brand X) stage 2s 5.3l heads were installed on a car that had all the bolt ons and a TR224 cam. It made 379/370 or so before adding the heads. After the heads, the car did 413/385 after tuning. A week before, a set of (the well known porter's) stage 2 5.3l heads on a car that already had the TR224 cam as well and the IDENTICAL mods as the other (brand X) car (FLPs, LS6 intake, ASP, stock rear/4:10s,etc). Both dynoed the exact same before swapping heads on (the curves practically overlayed one another). This car pulled 443/418 with the (well known porter's heads) . On the flow bench, the (brand X) actually flowed better from .100-.550, while (well known porter's) heads flowed slightly better past this. Same thing on the exhaust side.

This just goes to show that a flow bench is an excellent tool, but not something that should be used as the determing factor in how a head will perform when put on a motor. The CC volumes were identical on the two heads, so compression difference wasn't an issue.

Subject: An idea
Date: Wed, 6 Nov 2002 07:50:40 -0600

Hi,

I am going to run this idea by you, and I guess you can forward it out if you feel it has any merit.

I'd like to do a test, but I am going to need some help. What I would like to do is have a shootout. My idea was to solicit each of the major vendors for a set of heads and their camshafts, and have a H/C shootout.

Very simply, we flow all the heads on the same bench. We have folks look at the ports and review the heads from an overall standpoint. That way there is no variation.

For the second portion the vendors provide whatever camshafts they wish to have tested. We put them on a dyno mule LS1 engine and run them on an engine dyno. This would eliminate all the other variations. I know that we have the facilities to do the work. It would be a matter of whether or not they would be willing to allow us the use of their facilities. I am willing to volunteer my time to the project at no cost, as I would just like to see this sort of test done.

I didn't want to post this publicly yet, I wanted to run it by you and a few others and get your opinion of the viability of such an idea. I know that the shops have a business to run, so I don't want to interfere with that. I figured that a lot of this could be handled off hours by some of us, as long as they will give us access to the facilities. If they want to help, great. I also thought that if the folks who did the test were not associated with a certain shop it would help reduce the possibility of people raising the issue of fairness or partiality in the test.

I think most of the gear out there is close, but it would be nice to be able to say, here is what X, Y, and Z did...

And, if this is just totally off the wall and unworkable, then let me know. But as I said, before I went very far with it, I just wanted to run it past you for your input.

You could pretty much give up on that idea. The results will never be accurate. Shops can (and will) work up a badass set of heads, which may or may not represent "off the shelf" versions available to customers. Also, this type of testing has more potential to hurt than it does to help. What does it possibly do? It might bring one shop to the cream of the crop, thus

making all the competitors suffer. I don't think many shops will gamble that, and like I said those test samples might not be accurate representations of what is available.

You are just better off going with the shop you trust, and paying attention to results on an average. You will notice some shops only have good running

"shop" cars that belong to the owners of shops, while others have extensive customer representation with good results. I know which ones I'd be more inclined to believe are "normal".

I can't speak for the shops involved would think about this. I think they feel just fine and confident with their products and reputation without needing a test like this. They are sitting pretty right now, they have nothing to gain here and only the potential to lose. Its the little dogs out there that can gain from this. Make sense?

Maybe I am the only one around here old enough to remember the canonical counter example - Ford's old 302 heads.

Ports that looked like tunnels and valves that looked like manhole covers. Flowed like a dream. And if you could twist the engine to 12000, they made monster HP. But everyone that ever made any power and kept the engine together, filled the ports in and reduced the flow.