The Top Mitsubishi Galant VR-4 Resource

Join the best E39A 1991-1992 Mitsubishi Galant VR-4 community and document your GVR4 journey.

  • Software Upgraded - Reset Your Password to Login
    In order to log in after the forum software change, you need to reset your password. If you don't have access to the email address you used to register your GVR4.org account, you won't be able to reset your password. In that case, follow the instructions here to regain access to the forum.

Head flow numbers

curtis

Well-known member
Joined
May 4, 2003
Messages
11,892
Location
Clarksville TN
Not my site and not advertising for any represented in the table. Just have never came across flow numbers of that many brands all at once. Some I've heard of and some I haven't. Some of the names i know from me messing with Fords years ago. Thing is I didn't see Head Dynamics and Chris Beren's work represented. /ubbthreads/images/graemlins/dunno.gif But that shits Top Secret not for public consumption anyway, you guys can't handle those numbers anyway so see below.


Stan Weiss Mitsubishi flow data
 

curtis

Well-known member
Joined
May 4, 2003
Messages
11,892
Location
Clarksville TN
Oh remember .500 or .550 or .600 lift means nothing unless you have a cam that will open that much.

If anyone with mad excel skills wants to do it plot each one on a graph so we can figure the area under the curve. HP is made with a larger area under the curve on a line graph because the Valve is open once and closed once but goes past all the other numbers twice. Cars with heads that flow big numbers in the mid lifts always are faster than cars with heads that have lazy graphs and peak huge numbers up top.
 
Last edited:

Our APM CNC 1g head flows 304 and 284 @.500 the Evo head flows 286 and 280 @.500. The Evo heads are a lot more prone to cracking of the exhaust ports which was really surprising. The only problem with a fully ported head is you really need a big turbo and a lot of boost with BIG cams to take advantage of these things. If you make less than 600 whp it is almost a waste for a full cnc job, just have the bowl area cleaned up. We also found if you run a 1mm over intake valve and a stock or .5mm exhaust it works better than a +1mm intake and +1mm exhaust valve.
 

toybreaker

iconoclast
Joined
Apr 30, 2006
Messages
3,581
... As an old school machinist, I'm curious about how these tests are performed ....

The entry and exit from the port can be "simulated" many different ways.

The saavy guys use an intake manifold set-up, complete with throttle body, that is similiar (if not exactly the same) as what that particular head will be running, to ensure thay are not going to get a biased set-up in service.

They also use an exhaust manifold that is representative of the type and style that will be used on that particular engine.



Back in our racing days, we saw a lot of heads that had awesome laboratory flow numbers perform very poorly in the real world when intake and exhaust mani's were bolted up.

It's also helpfull to remember most of these numbers in the chart Curis posted are being presented by the shops that did the work, or a facility that *may* have a vested interest in how the product performs.

While I can't blame a shop for presenting the best possible scenerio for their products, I think it creates an unrealistic basis for comparison. Going thru a bunch of cores and cherry picking the best one is just good business, but it leads to results that will be hard to duplicate in the real world.


(Most) Magazines also flat suck at presenting unbiased information (ever notice how people that advertise heavily in a certain rag always do really well, while competitors that don't advertise have a poor performance index?... /ubbthreads/images/graemlins/idea.gif )

Bias, baby, bias... /ubbthreads/images/graemlins/banghead.gif



In addition, there's a whole slew of other factors that will seriously affect the numbers.

Core shift is a bitch.

Mitsubishi's 4g63 castings are better than some, but there's some pretty sizable variations in port centers from head to head and mani to mani. (we're not talking about a thou or two, we're talking sizable variations up to 60 thou or more)

Something as simple just swapping intake manifolds ("stock for stock") can make a port's numbers go from dog to super hero status, if the core shift of the mani compliments the core shift of the head.

If the as tested combo is port matched (just common sense!), then the variations really show when compared to a combo that has not been matched.

So, always take flow numbers with a grain of salt!!!




Most shops production pieces won't get anywhere near the "test" numbers...

I played with this stuff when I was a kid, and I learned a whole sh*t ton about reality... (and my limitations ! )

Many heads show a bias, and/or favor certain cylinders based on the basic engine design/manifold layout.

It's just simple physics.

Mitsubishi's heads and mani's are no exception.

Entry and exit areas of the intake/exhaust manifolds can be just as important as the work done to the ports in the head.

Minimising the bias (unequal flow to each individual cylinder) is mission critical to putting a package together that will live a long life and be tunable.

> remember, your wideband is reading the average of all four cylinders <

(and your egt is probably only reading one cylinder...)

A good "average" tune will lead to melt down/crushed ring lands on the weak cylinder when you turn things up. /ubbthreads/images/graemlins/wink.gif

The closer you get to the edge, the more this matters. /ubbthreads/images/graemlins/wink.gif





You can only tune to the weakest cylinder



All that said, I've seen a buncha "worked" 4g63 heads as a machinist...

...and I haven't been all that impressed with many of them...

...The old adage of "hog that baby out"... "lets get things flowing" is simply retarded when it comes to ports the size of the 90-94 ports...

It's my opinion that 90% of the benefits of a professional port job can be done at home, using standard tooling.

Just focus on the seat/bowl area.

Smooth the seat/bowl transition, and you're most of the way there.

Clean up the match between the mani's and the ports, and for all intents and purposes (for a street car), you're golden.

Sure, more *could* be gained, but there's a lot of other areas that could benefit from your attention before trying to squeeze that last lil bit out of the head.

If you are building an all out set-up, then it's best to let a professional handle the head work.

 
Last edited:

Core shift is a big problem with the mitsu heads. We took a sample from 5 different heads to establish our cnc program centerlines . The nice thing about CNC is there is nearly zero difference from port to port. One port is designed and then is just replicated for the other 3. With the mitsu head you also cannot hog it out. On our cnc program there are some parts of the wall that will not clean up. Almost all the power that can be unlocked from these heads is in the "bowl " area especially the 1G with its poor entry angle. I prefer on a street/strip car people run the 2g head with its better design and smaller runners. The 1g outflows the 2g and evo head because the ports are larger but this also why the 1g heads make an awesome methanol head.
 

toybreaker

iconoclast
Joined
Apr 30, 2006
Messages
3,581
A much more consistent product can be achieved with cnc work, and it's really cool to see the new guys applying these tools to the job!

It also leaves me without much to say, as my age is showing... /ubbthreads/images/graemlins/blush.gif

Back in our day, cnc machines were reserved for aerospace apps, (and they were run on punch tapes.)

Now, harbour freight sells them... /ubbthreads/images/graemlins/rofl.gif

We did it all by hand, and verified our results with a crude, home built flow bench. (not quite steam powered, but archaic, none the less /ubbthreads/images/graemlins/blush.gif )

However, the basic principles still applly, and as an old school guy, I still think mixture distribution is logrythmically more important than raw numbers.

Quote:
I prefer on a street/strip car people run the 2g head with its better design and smaller runners. The 1g outflows the 2g and evo head because the ports are larger but this also why the 1g heads make an awesome methanol head.



This is good advice!

Proven time and time again with real world results.
 

curtis

Well-known member
Joined
May 4, 2003
Messages
11,892
Location
Clarksville TN
I knew this would wake toybreaker up. /ubbthreads/images/graemlins/grin.gif


Cheating on a flowbench is the easiest thing to do. I've seen people flow the stock head or a head thats already been ported then they'll take putty and build up a ve stack on the inlet of the runner and form out a radius they like. The numbers can change just by the radius of the curve on the putty and never touch the port. Running the numbers up can also be skewed but how many inches of water the bench is pulling.

Like was said the only fair way to test is have all the head have the same valves and all the seats cut with the exact same valve job, the angles on the seat and the width of the angles can change the flow dramatically. Also every head should be tested with the same intake that is perfectly centered. Those kinds of test would cost thousands and the logistics of making it happen is mind boggling. I have a friend here in town that worked out the port for the Neal Brodix BB chevy head, he worked on one set of ports for months changing angles, margin widths, seat installed heights etc. When done had one of the fastest flowing heads in the industry up over 600 on the intake If I remember correctly.

Main reason I posted this is for a rough idea what these heads can flow from the intake manifold thread I posted Buscher was talking about the ports flowing 1200 and 1300. As you can see no head can flow this kind of number and the thing is not sure what he flow the intakes at as well. Most intakes / carbs are flowed at around 20 inches of water where heads are at 28.
 
Support Vendors who Support the GVR-4 Community
Boosted Fabrication ECM Tuning ExtremePSI Fuel Injector Clinic Jacks Transmissions JNZ Tuning Kiggly Racing Morrison Fabrications RixRacing RockAuto RTM Racing STM Tuned

Recent Forum Posts

Top