To some, it’s a mysterious piece of equipment that seems to hold all the answers for cylinder head-porting. If it were really that simple, everyone would do it. The flowbench is a useful tool which allows engine builders and cylinder-head specialists to measure gains or losses in airflow. Besides cylinder heads, airflow can be measured through other engine parts such as carburetors, throttle bodies, and intake manifolds. Much experimentation and testing are still required to make changes that result in better performance on the street or track.
“The flowbench isn’t some magical device,” says Bret Williamson, senior technical support engineer for Power Test, makers of the SuperFlow brand. “It doesn’t tell you, ‘Make a change here, make a change there.’ That’s something you have to learn on your own.”
“It just tells you, ‘This is what it’s flowing,’ and if that’s good, fine. If you can make it better, let’s try. We’ll grind here, grind there, make a different valve angle there, whatever. And if it improves, the flowbench will tell you the improvement. If it doesn’t improve, it’s back to the drawing board.”
How A Flowbench Works
A flowbench is a relatively simple testing device that forces or draws air using electric motors — much like a vacuum cleaner — through cylinder head ports or any engine part for testing. It measures the pressure drop above and below a calibrated orifice, which is merely a small hole in a metal plate.
Above this orifice is a test cylinder of the bore diameter found on the engine and a mounting plate, both airtight. To measure the pressure drop, early models used vertical and inclined manometers, which are devices using fluid-filled tubes that must be monitored, with the findings manually recorded.
Newer SuperFlow models replace these analog gauges with electronic pressure transducers and computers to chart the results. SuperFlow makes various models, sized for their intended jobs, from 160 cfm (useful for small engines) to 1,000 cfm. Because of the way a flowbench measures flow, it is not affected by altitude, which Williamson says is a common misconception. “It’s a ratiometric device, which means regardless of where I test, I still get the same result.”
Increasing Airflow Is Only Part Of The Story To Increasing Power
Casey Snyder manages the cylinder head department at LS/LT specialist Late Model Engines in Houston, Texas. He says it’s overly simplified — at least these days — to think that a shop merely places a cylinder head on a bench, flow-tests it, gets some CFM numbers, and then tries to improve them.
“That’s really only a small part of the story,” he says. “In reality, maybe that’s how it started, but now there are many uses for cylinder-head developers. One of those is to find turbulence and find areas of the port that are very high velocity.”
Pitot tubes can be used with additional manometers to detect the high-velocity areas. He notes that some operations a cylinder-head specialist performs can hurt airflow but increase power.
“It’s easy to take an exhaust port that flows great, and on the right engine combination, go in and open up the throat area (or what people call the venturi) and lose a bunch of airflow, but pick up a bunch of power. So, the bench is just a tool. Sometimes, you have to understand what you’re trying to get out of it. It’s another set of data to look at, but it’s not a primary design thing,” explains Snyder.
“Normally, we already have a pretty good idea of what shape and size we want our ports. We start there, and we learn that over time. It’s just R&D. We make a little change here, this works, so we kind of head in that direction, and we’ll flow-test along the way. Sometimes, larger changes, such as trying various shapes of modeling clay in a runner, may surprisingly show no change. Meanwhile, something the thickness of a layer of paint near the valve job on the short-turn side of the port is easily detectable on the bench,” he says.
Snyder gives one example of where some analysis on a flowbench can yield quicker times on the track. For an engine making peak horsepower at 8,000 rpm, a racer might look at the dyno’s power curve and plan to shift at 8,200 rpm. “We focus a lot on overrun power, which is past peak power. Let’s say it made 1,000 horsepower at 8,000 rpm, but the curve drops off really fast. We try to work on that other side of the curve.”
Doing so, allows the driver to move the shift points up by 500 rpm or more, Snyder says. “We get a much bigger chunk of the power curve that way, even though we’re spending a lot of time past peak power. That’s when you go to the flowbench and find you have cylinder heads that are going turbulent, which means you get to a certain lift or velocity and they start making noise.”
When a port goes turbulent, basically, the air doesn’t stay attached to the short-turn side of the port, and you’re not very effectively using the area you have available. “A lot of it is experience, but a lot of it is just putting the head on there and listening to it. Does it sound smooth, or is there fluidic switching? Are there red flags?” Snyder explains.
As good as late-model LS cylinder heads are from the factory, Snyder says it’s still pretty easy to improve them. “Most of those factory heads are not designed to flow to the lift that we generally get with aftermarket cams. We generate a bit more lift than those heads were ever intended to see. They generally have traits that promote mixture motion, and it actually works against us to get airflow through it.”
But Snyder explains, increased airflow does not necessarily correlate to an increase in power. He recognizes it can be confusing to the enthusiast shopping for parts to put together an engine at home. Without additional information, the higher flow numbers can seem to be the deciding factor.
“At the end of the day, we put out a product we’re proud of, and flow numbers are part of that. We want to show everybody that, yes, we can make cylinder heads flow great. But, we also want [to keep] the very reliable thick port walls and ports we know make power, that are dyno- and track-proven,” says Snyder.
“We have port sizes we like, we have the general shapes we like, and from there, we will tune them in on the bench. The valve job makes a huge difference in how cylinder heads flow. So, one good place to really dial-in your valve job is in the back cuts of the valve.” Snyder adds that valve jobs at Late Model Engines are “pretty basic” with a seat, bottom, and top cut, and under that, usually a bowl cut.
Standard Flowbench Settings Have Changed Over The Years
“Industry standards, at one time, were 25 inches of water column (a measure of vacuum) on the manometer, but once General Motors started testing at 28 inches — or about 1 psi — everyone else followed,” Snyder says. If one shop still tests at 25 inches, and another tests at 28, it’s possible to apply an equation to mathematically correct one to the other, but it may not be exact at higher velocities.
“There are people who think we should be flowing at higher pressure-differentials, maybe 50 inches of water. That would be more representative of what the engine [vacuum] is, but as it stands, 28 is the standard.” Snyder says. Flowing at a higher pressure-differential can yield advantages, such as detecting port turbulence, which wouldn’t show up at the standard testing pressure.
Using The Flowbench As A Validation Tool
At EFI University in Lake Havasu City, Arizona, instructors encourage students to learn to use the flowbench as a validation tool to compare previous test results from a known engine combination, says founder Ben Strader.
“For us, the flowbench is a metric,” Strader says. “It doesn’t necessarily represent the true airflow into an engine, because the actual pressure differentials in the port while the engine runs are quite different than what we can generate on a flowbench. However, there are some consistent correlations between the head’s performance on the bench and how they typically behave on an engine. So, we can often use the flow number from a bench to closely approximate engine performance before committing to a combination for a project.”
Strader says the flow numbers on a bench are typically used to match the cylinder head’s capacity to its application: engine size, target horsepower level, and operating RPM range.
“We consider the engine to be a balancing act of a basic study in economics. In other words, when the ‘supply’ is well matched to the ‘demand,’ we find the engine performs as expected. However, if a cylinder is much larger or much smaller than required, the balance of ‘supply and demand’ is upset, and the engine tends to disappoint. We use math to get us in the ballpark, and we use the flowbench to validate that once we know what we want.”
Snyder says Late Model Engines has “tons” of different port CAD programs cataloged for CNC-machining. Each new project, such as for its new line of LT heads in development, start as raw castings. With all external machining and port development done in-house, they will start off mainly by hand until it works well and flows well on the bench. The port will then be digitized in-house so it can be duplicated by the CNC machine.
“Just because you got one to flow great hand-porting it, it doesn’t mean it’s going to come off the CNC machine and do the same thing. Heads are just too sensitive in some areas for that to be realistic all the time,” Snyder says.
With the digitizing, two files can be overlaid and compared to see where the difference is, that caused a performance loss. “We always work off of our last best-set of heads. We take that port, and then we improve upon it until we get the next best one, and onto the next best one. But we always end up at the flowbench to confirm we’re not out in left-field from where we intended to go with our next version of that cylinder head.”
It’s just one reason flowbenches endure in the age of CNC machines. “Ten years ago, we thought that flowbenches wouldn’t sell anymore because people were able to duplicate existing heads,” Williamson says. “They could go out and get somebody else’s head, take a machine, measure it all out, and cut that head themselves to duplicate. But, we still sell flowbenches. We still sell a lot of flowbenches.”
It just proves that even in the age of supercomputers and robotic machines, technology is still dependant on human ingenuity and reasoning, and data is still king.