I've been looking at head flow data for a little while.

It seems that most heads will jump considerably around the 0.400 valve lift
mark.

Is this mostly because of valve shrouding, or is there another major factor involved?

Lift__A___B___C___D___E___F__
.200 144 125 141 130 147 121
.300 208 177 201 184 197 174
.400 244 220 247 222 230 216
.500 262 254 258 242 250 235
.600 261 226 257 252 243 236

Theoretically, for head "A", if I could achieve full duration at 0.500 lift (240 degrees),
I would not need to lift the valve any higher.

Why does the flow drop off for some of these heads at 0.600" lift?

If the valve is out of the runner area, shouldn't the flow remain fairly constant?

Since the runner volume, depression value, or bench flow has not changed,
what would cause the drop in flow (for heads A, B, C, E)?

I can almost visualize how the valve may be used to create some sort of
venturi effect by keeping the opening of the runner covered.

In other words, if the valve was removed totally, what is the suggested
result?

Would the air entering the chamber become more turbulent, or laminar?

 

#1. What's the valve sizes for each head.... That's going to help explain where the ports "jump" in flow.

#2. Just because the port stalls doesn't mean it will make more AVERAGE power or peak power with a cam that stays below that. I've proven that wrong numberous times, so have many stock classes.

Bret

 

They are not actually "jumping" around that .400 lift range but rather they are actually falling off after that lift usually. 90% of the time it just means the port can't keep up with the valve job and the port is going turbulent. Basically this would be a novice or econo ported head usually. Basically something will always limit the port and it will always level off at some amount of lift or at least approach an asymptote of cfm. If the restriction of the system is not at the throat and valve job area the head will back up like that 90% of the time and it will resemble the lift curve that you posted.

 

The most common raeson would be the shortside too tall or not layed back enough and/or a valve job that focuses on flow in one area and suffers other places.

If the valve job is descent, the flow can be coming around te valve fine but as it tries to come over the shortside it is moving too fast and can't make the turn and stay stuck to all of the walls of the port. Once the air comes detached from the wall, it starts bouncing around (going turbulent) and killing flow.

Laying the shortside back a lil will help but if you do too much, you start losing flow below .400 lift. I have found that raising the roof ALOT and widening the port here will allow the air to slow downa nd make the turn and stay in contact with the short side.

Best bet is to just keep playing with the port and record the findings. You will find a sweet spot once things are shaped correctly and you can gain high lift with out losing low lift flow. Once you go too far, you kill low lift flow. Just keep going and so not be afraid on runner volume.

Lloyd Elliott
972-617-5671
Eportworks.com

 

Let me understand how a flow bench works to relate to these answers.

The manometer is set to 28" of water, based on the amount of air flowing
through the bench and the size of the port (runner cc). The manometer
value is varied by the flow of the bench controls.

The valve is open at 0.100" lift at this point.

You record the value stating cubic feet per minute @ 0.100"

Open the valve to 0.200"

Check the manometer for 28"

Adjust the flow bench controls to maintain 28" of water.

Record the value stating cubic feet per minute @ 0.200"


Repeat for the remainder of valve lift increments.

Correct me where I'm wrong, and the questions will follow

 

You are right so far if I understand you.

 

Alright, if my understanding of bench flow operation is somewhat clear then:

Each time the valve opening is incremented, it is analogus to having a larger
runner.

What I mean is, if the valve was completely removed from the beginning,
to expose the complete intake runner, the manometer would drop from 28" of
water, to much less...possibly 22" for sake of an example.

In order to restore the 28" of water, I must increase the bench flow, correct?

By doing so, I am also increasing the velocity of the air traveling through the
intake runner. At some point, the runner will be too small in volume to support
the bench flow and begin to "choke" the numbers.

This is what we're witnessing at about 0.600" lift in the first post?

I'll stop there and get your corrections.

 

hmm, sort of.


It is MUCH more likely that the port is poorly shaped and that there is boundry layer flow seperation. The flow seperation effectively changes the shape of the port, reducing the cross section compared to the actual cross section. Reread Lloyd's post and think of valve angle. the ports are nowhere near straight, and remember how much everyone raves over shallower valve angle heads.

Notice i wrote about cross section and not port volume. Port volume is nearly irrelevent for steady state flows which the flow bench is measuring. It is relevent in a running engines dynamic flow because by measuring length of cross section you can figure the corresponding parameter to the total volume. Then you can have a better insight into the runner length tuning effects (harmonic tuning, wave tuning, whatever you want to call it). If we are talking flow we are interested in the cross section area. Varying runner length of a straight section of port with constant cross section will have very little effect on aerodynamic drag; drag being relevent because it correlates to flow for a given differential pressure. In other words cross section and runner length are different parameters.

-brent

 

Well, I opened a can of worms, so let's eat them!

I found a cool little photo from Chevy High Performance
http://www.chevyhiperformance.com/techarticles/83138/

Bret,

The head flow was taken from a comparison table using 23 degree heads,
2.02 intake, 200 cc runners.

I'm am all new to the terms used in your replies, so let me clarify with pictures:

Valve angle I'm pretty certain is reference to the block, or piston.
The "shortside" I'm guessing would be the side closer to the deck, because
it seems to be the shorter length of the intake runner.
http://www.gmthunder.com/tino/angleandshortside.jpg

When Brent says, "flow seperation" and Lloyd says "sticking to the wall"
I envision eddy currents getting trapped in a pocket swirling around creating
turbulence. This would create a "flow seperation"...at least some air continues
to travel into the cylinder, while some air feels the need to spin aimlessly?
This bit of turbulence is cutting down the effective runner volume as it impedes
air flow into the cylinder.

http://www.gmthunder.com/tino/turbulence.jpg

Lastly, I think I have cross-section figured out. We're referring to the opening
dimensions of the port and not the overall volume of the port.

In other words, the runner can be 200 cc with a cross section of 1.6 inches x
2.5 inches


or

the runner volume can be 220 cc with a cross section of 1.6 inches x
2.5 inches

 

Interesting thread.

Would anyone care to make a guess,
how would the resulting product be different if:,
the designer was given the instructions...
optimize for 4 psi, not 1 psi{28 in. water}.

 

AHH, i just had a huge repy typed and i wiped it out with the back button on my mouse!!!

Anyways,

Just a couple minor thoughts to add:

I led you slightly astray in my original post, Port length for a given type of head casting is pretty much fixed. Therefore changes in overall runner length are achieved by changing the intake manifold. Therefore, port volume is more of an indication of varying cross section than of varying length.

Your dramatized numbers where geometrically sound, you just held the wrong term constant. I held it constant because i was trying to discuss different flow numbers for similarly sized ports.

Thoughts on the word turbulence:
According to the engineering definition of turbulence=/=boundary layer seperation. Nearly ALL real life flows are turbulent because air just isn't that viscous. You can have a tubulent flow with OR without boundary layer seperation; which is when bulk vortices to form.
The common or laymans definition of turbulence is when bulk vortices, eddies, possibly flow reversal appear, in other words turbulance=boundary layer seperation.

The fact is that flow through a pipe elbow (or port) gets disrupted the most in two areas. The flow backs up/slows down/ builds pressure at the beginning of the outside of the bend, and the flow lessens around the inside radius of the turn. Even thought the molecules can't brake or steer, it is surprising how nicely they 'apex' the turn similar to racecars. They crash into the outer-wall on the entry and go into a spin which slows them down; aka building stagnation pressure on the outermost edge which in turn pushes the inner molecules into the turn initially keeping them on the inside radius. Then they overshoot the inside radius on the exit(the low pressure area on the inside boundry layer seperation). The only thing they can't do which keeps them from perfectly apexing the turn is 'swing wide' due to the uniform traffic on the straightaway.

Even though the boundary layer likely seperates on both the inside and outside radius, it is the inner radius that usually causes more of the problem like LE said.....

I'll let the experts with more real hands on experience with port design comment on the higher differential pressure.

-brent

 

Thanks for the lengthy replies. I can't say I understand most of the terminology,
but what I picked through in all of the replies makes me understand the head flow data a touch more.

After reading your theories and physics behind air flow through the port, I'm
betting valve timing (specifically intake centerline position) is a very critical
adjustment.

I want the max valve lift point to occur at varying stages through out the
RPM range.

There must be an equation to calculate air velocity at each lift point?

That would help a cam grind tech figure out the best intake centerline for
a certain window of RPM (based on piston speed and simulated engine
air velocities/RPM)?

 

The port would have to be huge compared to the valve area in cross section to keep the air speed over the short turn down or it would back up bad. Also you could straighten the port out a lot but thats not always an option in real heads.