:)

Thanks for jumping in and clarifying, Steve. Steve has been the master architect of my motor and specced and sourced the head work.

GOD, I wish I could assemble this motor soon - the flow bench numbers on these ehads are just rotten good. Too many irons in the fire.

Great work on the flow bench - this thread, along with Mark Jung's MFI dyno thread and Jeff Higgins MFI pump threads are why I come to Pelican every day.

Update on Inlet Porting
 

Its been a while since i updated on my progress. I, like Kenikh above, have way too many irons in the fire :)

I have not changed much on the inlet port since the last update, only port matching to the inlet manifold. I had to open the port face to 42mm to match the PMO manifold and blended it back to the port original size in 1/2". This increased te flow to 275 CFM and the velocities as indicated in the port map below. I am satisfied with the flow and velocities and will go with what i have for now. Next step would be to test these in real life.

I found another interesting formula for calculating port velocity based on engine size, port area and RPM indicating that the current flow is sufficient. Here goes:

LPV= 0.00353*RPM*S*B2 divided by CA

LPV= limiting port velocity
S = stroke (2.771)
B = bore (3.74)
CA= min port cross section @ 40.3mm (1.97sq-in)

LPV @ 7525 RPM = 278 FPS


http://forums.pelicanparts.com/uploa...1242331139.jpg

Below is the Entire inlet flow map without air filter


http://forums.pelicanparts.com/uploa...1242331208.jpg

I am not making any more changes to the port until i have the final valve job completed. I have in the mean time decided to not use the SI stainless steel valves and will go another route with 8mm valve stems. There is still some flow to be gained especially in the SSR as the velocity in this area is pretty high. I will work on laying the SSR back some after the heads come back from the valve job.

I am copying the design to the other ports using cross sectional templates to get things as close as possible between the ports.

This is what the port looks like currently ( I'm not giving up my day job just yet :) )


http://forums.pelicanparts.com/uploa...1242332029.jpg


http://forums.pelicanparts.com/uploa...1242332077.jpg

Cheers

Gert

Update exhaust port
 

I started on the exhaust a while back and the main goal was not to make the port too big and from the above program output used the recommended Port CSA of 1.82sq-in (D of 38.6mm) was about 3mm bigger than stock. I also used the public domain flow numbers for a ported exhaust port as a bogie to chase with Max flow of 201 CFM @.5" lift @ 28" of H2O. (from Kenikh's post in the S registry)

When I started flow testing the first thing I noticed is that the ex port is extremely turbulent and the flow on the floor of the is almost non existent until almost max lift. I placed a short string in the port floor and it pointed anywhere but out of the port the most of the lift range. All the flow happened on the long radius of the port (roof) and the velocities were quite high +300 FPS range and it flowed no more than 162CFM at max lift.

I opened the port to the 38.6 mm diameter and didn’t gain anything at all. This baffled me. I changed valve head shape to a smaller angle i.e. the smaller radius between stem and valve head, lost another bunch of flow and the port still stayed very turbulent.

I did some more research and in Graham A Bell’s book on tuning 4 stroke engines he suggests the inlet port size to be the same as valve head in this case 41.5mm, this is very close to the inlet port size. This is quite a substantial amount of material to remove and was a little hesitant to do it. In Bruce Anderson’s book all the factory 3.0L racing engines had 43mm/43mm port sizes for inlet and exhaust which gave me some sense of comfort.

After opening the port to 41.5mm the flow went up to 191CFM with around 330FPS on the roof of the port the floor flowed about 150FPS. Still very turbulent. I started probing the port and discovered that the flow increases if I put an obstruction on the floor. I played around with that and got the optimum shape and the flow increased to 203CFM. This was with the stock valve shape. Changing the valve shape to anything but the stock shape also lost flow, up to 10CFM. The valve shape and seat configuration seems to be critical for good flow on the exhaust port and i will do the final shaping after the valve job is completed.

http://forums.pelicanparts.com/uploa...1242351633.jpg


http://forums.pelicanparts.com/uploa...1242351383.jpg


http://forums.pelicanparts.com/uploa...1242351472.jpg


I am considering to have the port floor welded up and shaped as the ones from Extreme Cylinder Heads which should have the added benefit of reducing reversion.

this thread has already answered so many questions. I can't wait to read more.

Can you compare the flow velocities? The initial results that show a stock head that has a flow velocity of 290 fps and a slightly ported head of 340 fps. At what location is that measured? I would like to compare it to the current map of the intake tract flow velocities.

The bump in the exhaust tract is quite interesting.

Does this include the carb venturi?

The exhaust port work is interesting, but I'd be cautious. Compared to intakes, exhausts are very different animal. Specifically, the temperatures and pressures (and thus the energy level) are much higher, I believe that the density will also be dropping quickly as the charge leaves the cylinder. I don't have a clue about the level of turbulence, but suspect that it's also pretty high coming out of the cylinder. My observation is that exhaust gasses will do just about anything to get out of the cylinder -- including burn their way through gaskets. You most likely don't have to help them a whole lot to get out the exhaust pipe. Just don't put anything in their way.

My uneducated opinion is that acoustical tuning of exhausts is most likely far more important then the port shape. The overall volume of the exhaust port (combined with the volume of the exhaust primary) will -- I believe -- have a far larger impact on the acoustical tuning of the exhaust system then the port shape will. Personally I'd be afraid that your addition to the exhaust port will end up setting up an echo towards the valve which might break-up the overall system's tuning at the valve face.

Yes the 41mm diameter is the venturi in the carb section, the drawing is a little crude but at least gives some view.

It should be comparable except for the last modification at the port entry that was a big departure of from the stock and slightly modified port entry diameters. I measured that velocity 0.5" into the port right around where the blending of the manifold now ends. Most of the mods were in the bowl area and just some casting flash cleanup on the port leading up to the bowl.

Adding the carb and the inlet manifold reduced the overall flow by about 23CFM. I was a little surprised bu this as the carb manifold combination flowed over 300CFM.

BTW -- as you're considering the flow numbers, see if you can calculate the peak flow velocity at the smallest point in the track at your engine's peak HP engine speed. Then also look at the flow velocity at the carb venturi and the valve face at your peak torque engine speed.

Obviously you don't want your velocities at the peak HP engine speed to go too high. But it's also important that your velocities at the intended peak torque engine speed do not go too low, or else you're going to end up with driveability problems and a "peaky" engine as opposed to an engine that pulls strongly across it's rev range.

You are right, there is a lot more to exhaust system than just the port and a completely different animal all together. The pressure delta, cylinder to port puts these ports on initial low lift close to sonic speeds and these ports are even more of a dark art than the intakes, even guys like Darren Morgan and gurrus from Speedtalk dont really have an idea as there is no real means of testing them.

In this case the shape is pretty strait forward with almost no change in the shape after the guide. In my case i will be using venturie merege collectors from Burns SS and from what i read in Bell's book this type of collector reduces acoustic tuning effects and relies more on momenutm to perfom scavanging.

The port that i ended up with is also in the 2sq-in range and quite a bit bigger than the CNC port maps Bullet Bob posted where his ports peak out at 1.8sq-in if i read his plots correctly.

As far as the floor addition goes, when blended correctly could be viewd as D shape on its back, which is somewhat in line with what i have seen and read for other heads. (Vizard and Bill's port shape on the Extreme Heads) this rizer in the ports are also used by some of the Harley porters that has a hemi head configuration and should also have a negative effect on reversion as this is on the slow side of teh port. If it is not cost prohibitive then i will try it, good thing about it is that you can just grind it out if it does not work. :)

I have already done this with the LPV formula a couple of posts above, it uses the smallest CSA in the port, the RPM you are choosing and the engine size. At peak power (7525 RPM ) the velocity is 278 FPS.

If i use the same formula and assume 5525 as peak torque RPM ( i am not sure what that real RPM is at this point) the velocity goes down to 205 CFM which is a little on the low side and will have to see how this affects drive-ability. It can not be worse than the stock engine as the smallest port size has not changed much only the carb venturi has. The only item that i can change in this case again to have a smaller CSA is the carb venturi to get the torque peak velocity up.

I did this (changed ventuies) in my ITB miata and it was a trade-off between good top-end power and reasonable lower torque or great torque but quick drop-off the power curve at high RPMs. It probably will come down to the same thing here, evaluating the track and tuning according to the requirements. This is a pain but doable. One thing that will be necessary will be to change my driving style and gearing to make use of the power band that will be not as wide as it used to be.

Update
 

The heads are done and here is the latest update. The head volume ended up to be 86cc with 52mm inlet and 42.5mm exhaust valves. Both inlet and exhaust flows quite well with the following numbers:

Lift Inlet CFM Exhaust CFM
0.050 29.9 13.7
0.100 70.5 48.5
0.150 103.0 81.2
0.200 138.1 104.3
0.250 178.8 126.6
0.300 212.2 151.4
0.350 241.9 173.8
0.400 265.7 191.5
0.450 276.8 202.2
0.500 288.7 209.4

Now for the final test.

http://forums.pelicanparts.com/uploa...1279481420.jpg


http://forums.pelicanparts.com/uploa...1279481471.jpg


http://forums.pelicanparts.com/uploa...1279481519.jpg

Interesting. Keep us posted!

Gertvr,

Have you used a velocity prob in that port to measure port velocity rather than calculate it? I ask because the stock port has anything but a uniform velocity profile with the stock port shape.

The insert in the exhaust port floor, is it welded in place? It looks like a separate piece.

Regards,

PFM

Check your PM
 

Bullet Bob,

Check your PM please.

PFM

The ports are welded up and then machined down see pictures of the process below. I used a pitot tube to measure the port velocity.

http://forums.pelicanparts.com/uploa...1279629123.jpg

http://forums.pelicanparts.com/uploa...1279629182.jpg

http://forums.pelicanparts.com/uploa...1279629469.jpg

Cheers Gert

Just read this thread....very educational. Thanks for sharing (teaching)........way cool stuff.


Doyle

Any news here?

Very interesting research...

The combination made good power, the car is very drive-able even with a very radical cam and pulls strongly from 3.5K through 8K. I am more than happy with the outcome.

Below is the result of one of the dyno pulls.

http://forums.pelicanparts.com/uploa...1344981672.jpg

That's really wonderful power there. Congrats. PMO 50s right?

Thanks. Yes, ended up with 45mm venturis and not 41 as tested above


Cheers

Gert

Amazing wealth of information, thanks for sharing!

Oh, and congratulations on the outcome.

Was there a reason for the A/F discrepancies between left and right banks?

TIA,
Ben.

Sensors
 

We have 2 channels on the dyno for sensors, we used only one for this and when the other is inactive or unhooked thats the value.

Mike Bruns JBRacing.com

110 HP/liter with carbs very nice. http://forums.pelicanparts.com/support/smileys/clap.gif

Question on dyno sheet: Does "JE new design forging" = their FSR or forged side relief forging?

Je
 

Cupcar, those were just before the production of the FSR began, these were the newest P forging with the 4 angled supports at the pin boss, we had done some testing for Je on the prototype FSR that were made from billet on our 3.8 and on a 3.0 race engine. They worked very well. I am going to dyno our 3.8 next week with the newest Xtreme heads.

Mike Bruns

Mike...can you please post here so I don't miss the info you hopefully are wiling to post on the 3.8L w/ FSR pistons & Extreme heads you dyno...this info will help me a lot in trying to nail down an engine build...
thx in advance...Bob

Thanks.
I am building a 3.8 with 130 mm GT3 Ti rods on GT3 crank and was thinking of using the JE FSR forging (-3 mm compression height) with a factory GT3 Goetz ring package and LN cylinders.

Any other experience or thoughts? Thanks again.

Cupcar.....

excellent question :) Looking forward to the response. I haven't ordered my p&c's yet so may just wait to see what happens here.

OP: Amazing work here, very impressed with your process and willingness to jump right in. And your results are super impressive, you should be proud..... http://forums.pelicanparts.com/support/smileys/clap.gif

Show us some pictures of good port job.

Is this good? I have also seen pictures where the rib in front of the guide was removed completly.

http://forums.pelicanparts.com/uploa...1346481457.jpg

3.8 Dyno Sheet
 

Here is the dyno sheet on our 3.8 with Carbs/ Dist. w MSD, we just replaced the modified OEM heads with a set from Xtreme, we picked up a little power and torque from the best numbers over the last 4 years. I suspect with the improved flow data we could use a bit more camshaft but it is so driveable that will wait for later

Mike Brunshttp://forums.pelicanparts.com/uploa...1346961184.jpg

3.8 rsr
 

This is what is lives in
Mike Brunshttp://forums.pelicanparts.com/uploa...1346962005.jpg

I found a little trick in the process of building engines for Formula Ford.
We put anti-reversion cones in the headers right at the head...and this "mellowed" out the peakyness of the torque curve slightly.
The cones were made as gasket plates (take the place of gaskets) between the head and the headers...with the "cone" part reducing the diameter of the header by about 10-15%.
The cones extend down the headers about an inch or so.
The cones stop or reduce the back pressure from re-entering the head and give you a slight suction effect to help expell the exhaust gasses.
Feel free to experiment.
Bob

Interesting stuff Mike.

What sort of gain to you see going from modded heads to the Extreme heads in this case?

Heads
 

We ran this engine in March at the 12 hr SVRA event and had it on the dyno then at 411 hp. That was the 3rd set of OEM based heads in 4 years, they make around 35 hours and split from port to port through the chamber. The Xtreme heads bare weigh 1/2 # more than the OEM due to more material where its needed. I am going to test with different cam and exhaust combinations soon, with the flow bench numbers being a good bit better than we had I bet it will respond to that.

Mike Bruns

Mike - these are engine dyno figures? Or chassis?

With that nice flat torque curve I can easily believe it when you say it drives like a dream.

Having heads split at 35 hours sounds discouraging, except for Extreme's business.

Bob - a header maker from Florida some years back made anti-reversion headers for 911s. I borrowed a set once, but never put anything on a dyno and didn't have any other way of comparing them with the Bursch's which took their place when he wanted them back. . A friend modified a set of Bursch headers to have these, but again I've not done back to back tests to see what the effects might be. Though I like the idea of making them as a separate piece.

Walt...we found the things to be of benefit especially on longer tracks like Mosport.
It just sort of raised the torque level from 4500 - 6500 (seat of the pants and lap times).
The car pulled better coming out of slower turns.
The cones themselves are easy to make....just imagine a gasket (thin steel plate) with the cone extending out from it.
We whipped them up in a couple of hours...and being separate...easy to install.
BTW....we NEVER told the competition about them..and they looked very carefully to see why we were faster...even the tech inspectors missed them!
Bob

Dyno
 

Walt, that is engine dyno info, It does pull like a truck, we have been running a 6 into 2 into 1 system that is stepped on the primary's as well, power is good and we don't have to use a muffler. The 964 castings are a problem, all the earlier heads are not a problem.
When the 964/993 RSR's ran Daytona if you didn't start with new heads chances are you would not make it 24 hrs. plus the test days, BTDT
Mike Bruns

Mike - this makes me think we got it as close to right as could be expected for PCA Club Racing GT classification purposes. 110 hp/liter from NA 2 valve 911 engines. Same as the factory stated for its NA race motors.

Now, if I can only tweak my motor up to that level. I think it is hard for us homebrew guys to bolt together a bunch of parts and come up with that HP. Though we keep trying.

bob - the 911 anti-reversion heaers I had used had a bell where the primary attached to the flange. That allowed the anti-reversion piece to be same diameter as the port, with the gap below it, and the space above the gap in the bell, to do (supposedly) the magic.

But I may try making a slip-in cone for grins. If I can get quick on pulling headers, I could afford an hours chassis dyno time to test one with, and then without.