Tech Talk #70 – Airflow Fallacies: Avoiding the Pitfalls of the Flow Bench

DavidTechArticlesBy David Reher, Reher-Morrison Racing Engines

“The pursuit of a big cfm number has ruined countless cylinder heads.”

“What’s it flow?”

Whenever a conversation about cylinder heads begins with that question, I cringe. I know where this discussion is going, and it’s not good. When a racer wants to distill the performance of a highly developed cylinder head down to a single number, I know I’m dealing with someone who is fixated on the flow bench.

I can speak from hard-earned experience, because there was a time when the flow bench was the center of my universe. When my partners Buddy Morrison and Lee Shepherd constructed our first flow bench in the ’70s, it was a revelation – or so we believed. We were addicted to airflow, and like three flow bench junkies, we convinced ourselves that big flow numbers translated to quicker elapsed times. But that was more than 30 years ago, and since then I’ve learned to avoid the pitfalls of flow bench testing.

Unfortunately many racers coming into the sport haven’t been taught the lessons that Buddy, Lee and I learned the hard way. Cylinder head manufacturers, porting shops, and engine builders constantly advertise flow numbers – and I confess that I’m sometimes guilty as well. In this environment, it’s understandable that some racers think it’s all about maximum airflow. They shop for the biggest cfm number at the lowest price, like finding a screaming bargain on a 52-inch TV at WalMart.

The strategy to win the “Biggest CFM Contest” is simple: Grind the largest port that will physically fit in the head, use the biggest valves that will fit the combustion chambers, and test it on the biggest fixture you can find. That head might win the prize for airflow, but it won’t win on the dyno or on the race track.

The factors that determine the performance of a cylinder head are complex. A head that is ported without considering air speed, the size of the engine, the rpm range, the location of the valves, and a dozen other parameters isn’t going to be the best head, regardless of its peak airflow. And yet I see racers who are seduced by big cfm numbers bolt a pair of 10,000 rpm cylinder heads on a 7,000 rpm short block and then wonder why the engine won’t run.

The most critical area in a competition cylinder head is the valve seat, and the order of importance works its way out from there. There are many questions that are much more important than airflow: How far are the valve heads off the cylinder wall? What’s the ratio of valve size to bore diameter? What’s the ratio of the airflow to the size of the valve? What’s the size of the port, what’s its taper, how high is the short-side radius? The answers to these aren’t as simple as comparing a flow number, but they are what really make a difference in an engine.

Airflow is simply one measurement among many that influence engine performance. With the availability of affordable flow benches and computer simulation programs, it’s easy to fall into the airflow trap. A builder works on a cylinder head, sees some bigger cfm numbers, and keeps working for more flow. But if he doesn’t stop and test the engine on a dyno and on the drag strip, it’s very likely he’s gone down a blind alley. What the manometer on a flow bench sees at a steady 28 inches of depression is not at all what the engine sees in the real world. The pursuit of a big cfm rating has ruined countless cylinder heads in terms of what will actually run on an engine.

I put more faith in dyno pulls and time slips than I do in flow benches. I’ll cite an example from back in the day when Buddy, Lee and I were winning Pro Stock championships. Lee came up with an idea for a tuliped exhaust valve. He filled in the back of the valve with Bondo, and tested the new design on our flow bench. It was killer. We instantly saw a tremendous improvement in airflow with a small exhaust port, a nice tight radius below the seat, and much more stable flow. So we had some titanium tulip exhaust valves made and tested them on the dyno – and the engine didn’t run well at all. We had great airflow on the bench, but the engine didn’t care.

We were working late one night, and Buddy decided to yank the heads off the block and have Lee open up the exhaust throats. Well, Lee kept grinding and Buddy kept taking the heads on and off, and eventually we picked up 30 horsepower that night. We were porting from the dyno and not from the flow bench. When Lee finally flow tested the heads the next day, they were down 30 or 40 cfm, but that’s not what that engine saw.

The final test of a cylinder head is on the track. Frank Iaconio was our chief Pro Stock rival, and he was a smart racer. Frankie used to change valves at the track — he’d make a run, come back to the pits and switch from valves with a 30-degree back angle to a 20-degree back angle. We did similar tests on the dyno, but he did it at the track. I was impressed.

I’m not dismissing flow benches. In fact, we use them daily at Reher-Morrison Racing Engines. But a flow bench is a tool, and it’s really not much different than a micrometer. A micrometer can measure the diameter of a piston, but you have to run the engine to learn the correct piston clearance. Knowing the sizes of the piston and cylinder bore doesn’t tell you if the piston is going to gall or collapse a skirt until you run it. And knowing the airflow of a cylinder head doesn’t tell you whether it will make good power on a given engine until you race it.

Experience is the most important tool in cylinder head development. A person with extensive dyno and track experience has been through it all before, and knows how to avoid the flow bench fallacies.