Here are some links to Airflow testing I did years ago, if anyone is interested.



The velocity graphs are calculated from the airflow passing through the runner area, the smaller the port, for the same airflow, will show higher on the graph



Flow of std primary ports

http://photo.livevideo.com/photo/allsize.a...AC6E&size=o



Velocity of std primary ports

http://photo.livevideo.com/photo/allsize.a...B18C&size=o



Flow of std secondary ports

http://photo.livevideo.com/photo/allsize.a...A508&size=o



Velocity of std secondary ports

http://photo.livevideo.com/photo/allsize.a...E007&size=o

Interesting. Its good to see some hard data with comparisons. Were the cosmo irons JC? If so, that would explain the low velocity. Where abouts in the port or runner was the air speed probe positioned? Was it the same position for every test?



You can really see the late port closing of the 6 ports.

You can also see why people had great success with 3B irons in the past.

The Cosmo was a sedan just before the S6, it's basically a S6 13B with very large runners, I think they are like the 20B. Mazda realised their mistake and made the ports smaller in the S6



The velocities are "calculated" from the CFM and the runner area. Pitot testing was not that common back then(10yrs ago), I've only started using pitots for velocities in the last 2yrs



Another point of interest is comparing the 3B RX3 small centre port to the very large(tall) RX5 centre. The RX5 centre is the same as the early RX7 large centre ports. I didn't have a std tall port centre plate and had to use a mild ported one. You can see from the graph it opens 10º before and closes 10º after the 3B plate .If you subtract these points and put it back to a std timing plate the flow would be the same as the 3B plate



Another point is the S2 RX7 port, the port has a very flat turn that directs the flow straight into the rotor face ,the rotor face hurts the flow, thats why the curve has a dip at the top. After this port Mazda made all their ports with large late turns to direct air 90º from the plate face so they are not shrouded by the rotor. Mazda's best flowing ports have the late large turn, S4, S6. The best of the 12A plates is the 12A Y Turbo plate, it has the best shape. It doesn't look that good on the graph because it closes at 40º if opened up to the std 50º closing point it would flow more than the S2 and not have a dip.

very cool! so based on that, the best motor would be something like a 3b centre with 6 port outers?

Not familiar with the 3B irons, what are they from other than the RX-3, or is there not that much of a difference from the S2-S3 stuff after porting? Also, would the earlier centers (S2-S3) still be superior to the later centers (S4-S5) even after both were ported? I suspect I know the answer to the second is yes, but wanted to throw it out. Finally, is the "6PI Aust" flows for both the aux and secondary ports on the 6 port irons, or just the aux ports? It looks like it pretty much has to be both, but want to be sure.



Thank you very much for posting this stuff; been wanting to see flowbench numbers, but as far as I know noone's ever posted this kind of data.

I would think 3Bs are close to S1



The 6PI is with secondary and aux ports



Here are some pictures of what ports look like inside



Primary centre plate

http://photo.livevideo.com/photo/allsize.a...C002FDD93609E7E



Secondary ports

http://photo.livevideo.com/photo/allsize.a...8B6752448AA2173



Notice how the S6 and Cosmo ports turn up in the plate to match the manifold making the turn more gradual and the actual port area is smaller than the large opening at the manifold suggests

thats really neat.

Here is a flow graph of a PP Vs BP Vs std 12A turbo



http://photo.livevideo.com/photo/allsize.a...2378&size=o



Calculated Velocity in the runner



http://photo.livevideo.com/photo/allsize.a...8604&size=o



Looking at the PP flow graph you have to account for the BP and std 12A being only one(secondary) port. When you add the primary port the flow will go up about 80%, just imagine the curves are 80% higher and the same timing(width)



With the velocity graph when you add the primary port you have to imagine the curves to reduce in height by about 80% due to the extra area



Look at when the PP opens, even though the port opens at 100ºBTDC it takes 30º before the flow has the potential to start in earnest, even though it probably doesn't

Always remember the flow is mainly dictated by the rotor demand, even though the port can flow 270cfm @ TDC (T on the graph) the chamber will be at it's smallest and there will be little flow because the rotor is not asking for any flow. There will actually be slight flow due to pulse tuning etc.



With the PP the flow seems to be a direct correlation with area, ie increase area by 10% and flow goes up 10%



The BP EYE lines are tests on the size of the eyebrow notch in the Alum housing which uncovers the bridge of the BP, ie BP=4 EYE is a 4mm chamfer in the housing, 10 EYE 10mm chamfer

Nice. You must've done a fair few hours of testing! According to my gas velocity chart the PP diam is around 44mm. Close?

so by this test a bridge port does not increase cfm over std ports?? or am I just looking at this all wrong.

Its not unreasonable to think so. Even with a good port, the air column still needs to make an almost 90 degree with little to no radius. Also, there is very little you can do to keep the air column from just skipping over the eyebrow at any appreciable volume.



But that advantage it does have is an increase in port timing. Its proven that bridgeports make more torque and hp than side ports, everything else being the same. Its like having a long duration cam with the same lift as stock on a piston engine.

you wanna look at not only the peak # but the area too. the bp only flows about 25 cfm more, but it hit peak flow earlier than the sp and hangs in later.



i would have thought the pp would be closer to the bp, but maybe thats because the bp didnt go into the water jacket

Ooh, I like the additional photos and the peripheral port flow numbers... eesh, a bit impressive those. Really need to build a flowbench of my own; there's various kind of non-conventional numbers I want to look at at some point, but god knows I've enough projects.



I think I see what you mean with the way the ports on the 3B and RX5 curve towards the end... going to have to change some details of how I grind my ports, I think. Always learning! Thanks again for all the data and pics.

The PP from memory was big , I think 52mm



The BP is a std port just with the bridge cut only, the original port is still std. The bridge is the same height as the port and 8mm wide and does not cut the orange water O-Ring. The increase is only due to the bridge. It is not the most achievable, as an estimate, bridge 25cfm+port 176cfm+primary 80% =362cfm ish





When you add 80% to the flow numbers for BP because of the primary port it catches up some toward the PP







Like was said the flow didn't increase much but the timing did. There is not much actual flow (even though the flow bench says it can) before TDC(T) and after BTC(B) except for wave tuning, so why have more timing. The most air you can get in a engine without wave tuning, inertia tuning is VE 100%, you need the tunings to get higher VE and thats where the extra port timing comes in. VE, what affects VE?, it seems velocity has an effect, you need an air speed of 700ft/sec to get to a VE of 125%, if you had a port speed of 820ft/sec you could achieve VE 135%, if you could overcome all the losses. I'm kinda trending these days that the velocity is equal if not more important than the flow.



If you port an engine to 300cfm and the rotor/piston only demands 200cfm you will have only 200cfm flowing through a port that can take 300cfm, this makes the flow too slow at that rpm, you have to make the port smaller and take some flow out to make peak power at that point. You could rev the motor harder and the rotor/piston will demand more air which will make the port velocity higher, if you can keep you motor together. Actual velocity probing will avoid this problem.

If you have a HP curve that won't quit, is still rising past the point where you want peak HP, it probably has too much flow and not enough velocity. If you could get it to peak earlier it will have a fatter power curve, more area under the curve. You won't have HP bragging rights but you will be quicker.

Hence "The builders job is to get the MOST flow out of the SMALLEST port."



If you have two ports that flow 150 cfm at a certain pressure drop but one is smaller, the smaller one will actually make better power over a longer range.