Flowing ports.....maybe we have it all wrong

[Donnie1126]

I can give you the reasons why I think we have it all wrong, but then it might seem impractical even if it is logical. By impractical I mean for the time or measures it would take to do a set of heads the way i'm thinking they should be done.Does anyone else feel that the accepted standard of porting heads and flowing them is outdated?I have quite a few ideas as to why this is so.I'm just wondering if anyone else does?I know about the wet flowing ideas and testing heads at depressions beyond 28"s.But i'm thinking along different lines that would take porting and flowing to a new level.These ideas have been with me for quite a while and for some reason I believe there are others that know what i'm getting at.

[DavidHarsay]

I'm listening....

[SStrokerAce]

Me too....

Depends on the actual pressure drop that the port sees in use, problem is it's not very easy to figure out what that pressure drop is! It can easily be MORE or LESS than 28" in running condition and if you give it MORE airflow on the bench the depression can be less dynamicly.

Speak on brother, I think you are going the right way here.

Bret

[Ape]

donnie:
I have no idea what you´re contemplating, but im very very curious, but did you know that weslake did non steady flow test allready in the 50´s with a gas volume counter.

Hhmmmm. donnie do you mean the possibility of adjusting the bench to actual pistonspeed and thus port speed??

Well lets let us know

[highVE]

I too am very curious as to what will be stated. AT first i thought he was saying we were just wasting time porting and such. however after reading on i figure this could be going places.
Ok dude, lets here it!!

Mike Theroux
www.mikesportingservice.com

[Donnie1126]

Well fellas I don't have a whole lot of time this morning.I'm already late for work.That's the problem for me.I do something else for a living and I wish I had more time to contemplate and experiment with this stuff.The reason i'm posting here is to get some opinions and feedback on alot of my thinking about this stuff.Maybe i'm wrong. If so I can hang my head in shame and go off and kick myself in the butt for leading you fellas on. Ape hit the nail on the head on some of what it's all about.Flowing ports to at least get an average number over a whole range of depresssions and not just at a single number.The thing is the bench is continually being adjusted to bring the number to some value like 28" but it never shows at what power range or rpm the engine will see that depression.And that's just not between two different heads,but also between two different ports especially a hand ported set of heads.And also between various lifts on the same port.I think that is the reason some heads can outperform others and not even show the advantage on the bench over the other head.I'll give you what I think is a very good example of what i'm talking about.Take a head and flow the port at whatever depression.Then take an intake manifold designed to match that head and flow a port at the same depression.Bolt the two together and the number will probably drop at the same depression.That's probably because the two won't coincide with the same depression at the same rpm.Now let me give you an idea.I haven't tried this yet but I do have the port molds for the head port and the manifold port that goes with it.The idea is to make a port model of the head port and intake port.To do this I have to cast a model of both and then cast around them to get an actual port that is flexible using something like silicone.I now would have two flexible ports I can connect together and put on my bench to adjust.Any reading I get I can be assured is the two working together.That's just one idea among a whole lot I hope to try.Another is to build a new flow bench that measures an average number over a whole range of test pressures.I could keep rambling on but I better get to work.I'm sure these ideas have already been contemplated and some maybe even tried.If there's anyone that knows of any let me know if they worked or not.

[Ape]

Hi donnie,
clear perplex is available for building clearescent flowboxes. Thats how i came to decide that my SBC-like (motorcycle) port needed serious downdraft, in order to raise the Coe.D. for the same valve size in comparison to just stuffing a big valve in it and destroying velocity for gaining not even equal results. But to not overstrain oneself on the task of doing 3dimensional flowbenching i would assume (manually non realtime)checking a port during the most critical events would do it. like highest piston velocity etc.
..........
But the culprit lies within the overlap since you will have different draft (horiz.+vert.)vectors and thus different port outflow vectors and also port speeds (added product of pistonvelocity and induced overlap suction) what might make the task very very difficult without real engine software(ricardo, avl, lotuseng,..) i.m.h.o . But for a beginning i would assume checking during critical flow situations would do a magnificient service. But a mean number should be rather easily achievable through interpolating back to one testpressure and looking at the then comparable CoeD. I assume but i might be wrong?

kind regards

[OldSStroker]

Well fellas I don't have a whole lot of time this morning. I'm already late for work. That's the problem for me. I do something else for a living and I wish I had more time to contemplate and experiment with this stuff.

The reason i'm posting here is to get some opinions and feedback on alot of my thinking about this stuff. Maybe i'm wrong. If so I can hang my head in shame and go off and kick myself in the butt for leading you fellas on.

Ape hit the nail on the head on some of what it's all about. Flowing ports to at least get an average number over a whole range of depresssions and not just at a single number. The thing is the bench is continually being adjusted to bring the number to some value like 28" but it never shows at what power range or rpm the engine will see that depression.

And that's just not between two different heads, but also between two different ports especially a hand ported set of heads. And also between various lifts on the same port. I think that is the reason some heads can outperform others and not even show the advantage on the bench over the other head.

I'll give you what I think is a very good example of what i'm talking about. Take a head and flow the port at whatever depression. Then take an intake manifold designed to match that head and flow a port at the same depression. Bolt the two together and the number will probably drop at the same depression. That's probably because the two won't coincide with the same depression at the same rpm.

Now let me give you an idea. I haven't tried this yet but I do have the port molds for the head port and the manifold port that goes with it. The idea is to make a port model of the head port and intake port. To do this I have to cast a model of both and then cast around them to get an actual port that is flexible using something like silicone. I now would have two flexible ports I can connect together and put on my bench to adjust. Any reading I get I can be assured is the two working together. That's just one idea among a whole lot I hope to try. Another is to build a new flow bench that measures an average number over a whole range of test pressures.

I could keep rambling on but I better get to work. I'm sure these ideas have already been contemplated and some maybe even tried. If there's anyone that knows of any let me know if they worked or not.

No offense, Donnie, but I have trouble reading one long paragraph. I like your ideas.

I suggest that CFD (Computational Fluid Dynamics) may already have been used to do much of what you suggest. The higher end CFD is probably restricted to OEM and F1 teams with huge budgets, but your method seems to be a practical approach.

If you had a known good head and manifold that had produced great torque and power numbers, that might be a good one to use as a baseline.
Keep at it!

[putztastics]

Sounds something like this;

Some thoughts about Traditional Port flow testing.

Most all benches are Good ONLY for testing valve seats and valve grinds up to .15" lift but open port flow testing (high lifts) should be done differently.

The real problem is port flow Separation (from the wall) which you will not see at mild 28" depressions. Port flow separation will cause the flow to be suddenly restricted to a smaller section, with increased velocity. This effect is not the same as the mach related choked-flow and occurs well before you get high mach velocities.

You really need to ask, how does the intake port compare to a perfect pipe. A perfect pipe will flow as the square root of the port depression.

First you need a bench that will pull at least 100" H20 at open port flow. We have seen good heads that start to flatline even at 40". Most commercial benches will only do about 40" because most use only a Single stage centrifigal sucker.

For Port testing you CLAMP the intake valve at about 80% of full lift and leave it there. Now you Step the vacumn. Start measuring flow at 20" and then increase in steps (like 10" - - -20,30,40,50,60,70,---). Does the Flow increase as the square root of the depression? (like perfect pipe flow)? If not find out where in the port the flow is not attached to the port wall.

You will find a pressure/flow level where the port flow detaches from the inside wall. The flow is restricted to a smaller section of the port. Any port mods that moves this point to a higher depression/flow is GOOD. Also any mods to reduce its limiting effect is GOOD. Improvements here should show Real gains in engine performance, since an engine is essentially an air pump.
_________________
Dave Manzolini
DEPAC Dyno Systems
201 Mill St
Rome, NY 13440
315-339-1265
FAX 339-5052
www.depac.com

That post is from page six of the Flowbench calibration thread. No comments on that idea over there though.

[panic]

Re: higher vacuum tests.
Smokey said decades ago that:
1. 28" shows the closest relationship to flow = power.
Lower vacuum tests would show only minor changes in results for major mods - you had to cut 1/8" away from a port to see an improvement.
Higher rates showed violent (non-linear) changes to very small mods - 10% increase for a surface texture change that did nothing useful for power, leading you down the wrong path.
2. he built a large, complex and expensive fixture with an huge electric motor to cycle a V8 to simulate the actual pressure and flow changes closely (except it was dry, and not at operating temp).
Result: same as 28" bench test.
His conclusion: waste of time.

New ideas are always worth examining, just don't re-invent the wheel.

[Donnie1126]

I'm back from a hard day at work. (actually only worked about 3 hours)
Well i'm not sure I really made myself real clear.The post I made this morning was in haste.The point was not to measure flow at greater or less depressions as much as it was to compare the changes(in valve lift,port changes.manifold to port differences etc.It's really hard to explain what i'm trying to say so i'll try as simply as I can.
Take the manifold to head comparison I made this morning.The manifold port is flowed at 28"s.The head port is flowed at 28"s.Let me suggest merely for the sake of simplicity in my example that both will flow exactly the same cfm.I'll give them a value of 300 cfm.So what is the difference between the two?Even though they both flow the same cfm at the same depression it is more often than not that the flow bench air valve will have to be adjusted differently between the two.This suggest one thing.That each port will only flow 300 cfm at two different engine speeds.That would mean that when one sees 28"s in a running engine,the other port is at more or less than 28"s.
This example is the whole basis behind my thinking.And it doesn't stop at just comparing manifolds to heads.The list goes on and on.
Another example:A set of heads have been meticulously hand ported to achieve exactly the same flow at 28"s for every lift of the valve.The only thing is,the port volumes vary from port to port.This would probably not allow each port to flow equally at the same speed of the engine.The same example can be found on the bench.Because each port volume is different,the air valve has to be adjusted to each port to get 28"s.
This is just a few examples but I could go on and on.

[Ape]

Well, ....donnie,

how about goin on??
Im allways very curious and openminded about new ideas.
Like your ideas but couldnt follow the manifoldrunner, head thread completely since they should be one integer union anyways? Otherwise i think this thread could get great.

kind regards

christian

[Rick360]

The only time a manifold port should be flowed is when it's bolted onto a head. Then it is the entire intake runner that has 28" depression, not just a manifold.

When I have flowed at higher depressions and the port didn't have turbulence at the higher depression, it would always flow the same at lower pressures and converted. Not a huge number of tests overall, but that's what I've seen.

I'm not sure if this is what you're getting at but Engine Pro from RSA has an output screen that shows what the flow bench test pressure should be for each lift based on theoretical cylinder pressures and flows etc. I'll attach a sample to show you. I don't have a bench that can pull these numbers YET.



Rick

[MadBill]

That is exactly the thought that came to me Rick, as I read Donnie's first post! Logically, the test depression should closely match (or at least be based on) the actual port depression in the running engine at each lift point through the valve open period at the operating RPM of interest*.
Simple modifications to a computer controlled bench (or maybe just pushing the right buttons, since I've re-invented a lot of wheels in my life) should even allow the test depression to automatically change with the valve opening in a manner that mimics the pressure profile of the relevant cycle of an operating engine, yielding data much closer to an engine's real world needs.
*(easier said than done re exhaust during the blow down period, but I have some ideas re that...)

[Joe Mendelis]

If airflow is so important, why do some of the best Cup heads make more power than heads flowing less air at all lift points?

There are SB2 heads flowing 400 cfm at .700 lift puttin the smack down on 430 cfm heads of similar cross section.