Testing Intake Manifolds + Heads - Discussion of Intake design

[ThomasVaught]

As many know, I work for the Blue Oval guys.

I was having a discussion with an engineer the other day about
testing of intake manifolds and cylinder heads.

I have an old copy of Ken Speery's Flow Bench book he published in 1992 where described the GM Flow Bench lab and how they did their testing.

Their bench was based off of a GM Roots 6-71 Supercharger that
could flow 400 cfm and it used a variable speed controller. It used
a Laminar Flow Element as the flow measurement device.

The book gives some good information on how they combined computer testing, Engine testing, and Flow bench testing.

They felt intake testing was a three step process:

Flow the intake by itself and read the flow restrictions in the intake.

Flow the intake with a cylinder head at max valve lift and check
for swirl and tumble. He stated that the heads would mask a poor intake runner.

Test the intake and heads on a motoring dyno with the valves moving to determine transient flow.

I would like to know if anyone ever looked at flowing all of the runners at the same time with measurement devices at each runner like Smokey suggested years ago? Larry M?

Tom V.

ps The engineer I was talking to considered a manifold having
a 10% spread in balanced flow between runners acceptable and I thought that was a poor intake. I have always tried to be with-in 2% for all of the runners.

[Tony]

The big problem with testing intake systems is that each cylinder operates in sequence, and can rob flow, or otherwise interfere with adjacent cylinders. Hence measuring steady state flow of just one induction tract, or all tracts flowing simultaneously in steady flow, is going to be far removed from what actually happens in a running engine.

At least with EFI, the nightmare of fuel distribution and wet flow becomes less of a problem, but the whole thing is still vastly complicated.

Then there is the added complexity of various resonances caused by pulsing flow. None of that would show up on a flow bench, but it could create serious dips and peaks in airflow at various rpm in individual cylinders.

I do know someone that says he hooked up a DOHC head to an electric motor to drive just the inlet cam at various rpm, and then measured the respective individual induction flows, while simultaneously evacuating all cylinders.

Even that did not produce useful results. Apparently piston motion (rod ratio), valve timing, and various thermodynamic effects, especially from exhaust reversion alter things so significantly that the cold air test was all but useless, at least that is what I was told.

I believe the first step is to first get your cylinder head working as well as possible, and then check that cylinder head plus intake runner combined does not restrict things too much at maximum valve lift.

Everything else, especially cylinder to cylinder air distribution at full power needs to be done on a fully loaded running engine at full operating temperature. I doubt if there is any real way to "cheat" because of the complex constantly changing interactions going on in a running engine.

If there was an easy way to do this, it would have become well known by now. Some of the cleverest engineers and racers on the planet have been thinking about all this for over a hundred years. It sure ain't simple.