Homebrew Head Porting: Attempting the dark art .....

[jluetjen]

Hi John. I agree on your interpretation of the last table. I'll reiterate that I think that this data is far more meaningful if you're looking at the peak torque RPM, or maybe across the rev range for a given CSA. As far as your cam question, keep in mind that both lobes don't share the same profile, even though they are mounted on the same shaft.

Another thought that might be interesting Gert, is to run a couple of known engine configurations through your model to validate it -- for example your stock engine configuration. You should be able to correlate the outputs from your model with your engine's factory HP and torque graphs. Chances are the numbers won't match exactly, but hopefully the peak torque and peak HP numbers should occur at roughly the same points on the rev range as your model predicts. Hopefully the rough order of magnitude numbers will be similar to the stock configuration, and trend in a similar fashion. If these don't happen, you'll know that you need to do some more work to understand the variations from the model's predictions.

[Gertvr]

Originally Posted by john_cramer

Gert, great stuff!

To add to John's questions, how can the intake and exhaust lobe rpm be different given that our engines are SOHC, i.e. both lobes are on the same camshaft. Maybe that doesn't make a difference in the calculations.
.

John,

Thanks,

I am not sure what his data means I can only infer that it relates to the RPM @ at which efficiency peak for the exhaust and inlet lobes are achieved based on the rest of the engine components like valves, L/D ratio etc. I dont hink they mean they will be spinning at different RPMs as they are, as you mentioned on the same camshaft. I will email the owner of the program (Larry Meaux) for clarification.

Gert

[jluetjen]

John; in response to your question about why the intake and exhaust valves would have different ideal engine speeds. I remembered a couple of equations that I worked out when I was doing the "How to select a 911 Camshaft" thread. Specifically there were two formulae that I derived by doing a regression analysis against a number of the 911 engine configurations.

Peak torque engine speed = -3151+(Intake Duration degrees * 32.53)

Peak HP Engine Speed = (exhaust duration degrees * 66.62) - 9083

Looking at what Gert's model produced:

------- Operating RPM Ranges of various Components -------

Camshaft Intake Lobe RPM = 7123
Exhaust Lobe RPM = 5889

The thing that is strange at first blush is that it would appear Gert's program suggests that the intake cam lobe has a bigger affect on the HP peak (which you would expect to be up in the 7000 RPM+ range) and the exhaust lobe on the peak torque RPM (usually a couple of thousand RPM less then the peak HP engine speed). My analysis suggests that the relationship is vice-versa.

Curious...

[engguy]

You can have a head that flows real good high numbers that doesn't make the hp of one with proper designed ports that flow less. Flow isn't everything.

[Gertvr]

Originally Posted by jluetjen

Hi John. I agree on your interpretation of the last table. I'll reiterate that I think that this data is far more meaningful if you're looking at the peak torque RPM, or maybe across the rev range for a given CSA. As far as your cam question, keep in mind that both lobes don't share the same profile, even though they are mounted on the same shaft.

Another thought that might be interesting Gert, is to run a couple of known engine configurations through your model to validate it -- for example your stock engine configuration. You should be able to correlate the outputs from your model with your engine's factory HP and torque graphs. Chances are the numbers won't match exactly, but hopefully the peak torque and peak HP numbers should occur at roughly the same points on the rev range as your model predicts. Hopefully the rough order of magnitude numbers will be similar to the stock configuration, and trend in a similar fashion. If these don't happen, you'll know that you need to do some more work to understand the variations from the model's predictions.

Response from the creator of the program:
What is camshaft intake and exhaust Lobe rpm? Is that the speed of the cam? And why a range?

Lobe RPM estimations are equivalent to how much
"Area Under the Curve" they represent from a
combination of Valve Lift and Duration @.050"

From these "Lobe RPMs" estimations the User can see
at what RPM Range those particular Cam Specs will give
you the greatest Power

or

Lobe RPMs are the "BallPark RPM estimation" at which each Cam Lobe
would naturally like to operate.

Lobe RPM = is a combination of Lift and Duration and Centers
...similar to Area Under the Curve, or Lobe Area

As suggested I ran the same program with the stock SC configuration and came up with this:

I used the cam timing and HP data from Bruce A’s book and tweaked the VE% until I achieved the quoted HP number. Tq also came close i.e quoted values to the program’s calculation. Peak Hp RPM was adjusted to be the same as the choke RPM of the exhaust valve andd was not the same as BA book, closer to the plot in Johns post on selecting Cams.

What is interesting and maybe the reason Porsche reduced the port sizes again:

Based on the highlighted areas below it is clear that in this configuration i.e. 39mm (1.85 sq-In) SC ports were way to big and could have been prone to reversion due to low port velocities


182.651 Cubic Inches @ 6250 RPM with 94.0 % Volumetric Efficiency PerCent

Required Intake Flow between 127.6 CFM and 133.9 CFM at 28 Inches
Required Exhaust Flow between 99.7 CFM and 108.0 CFM at 28 Inches

600 RPM/Sec Dyno Test Lowest Low Average Best
Peak HorsePower 171.7 178.8 182.3 185.9
Peak Torque Lbs-Ft 160.0 166.6 169.9 173.2

HorsePower per CID 0.940 0.979 0.998 1.018
Torque per Cubic Inch 0.876 0.912 0.930 0.948

BMEP in psi 132.1 137.5 140.2 143.0
Carb CFM at 1.5 in Hg. 310 345 363 380

Target EGT= 1486 degrees F at end of 4 second 600 RPM/Sec Dyno accel. test
Octane (R+M)/2 Method = 86.3 to 88.9 Octane required range
Air Standard Efficiency = 58.05734 % for 8.500:1 Compression Ratio

Peak HorsePower calculated from Cylinder Head Flow CFM only
600 RPM/Sec Dyno Test Lowest Average Best Potential
Head Flow Peak HP = 166.3 205.6 244.9

----- Engine Design Specifications -----
( English Units ) ( per each Valve Sq.Inch area )
Engine Size CID = 182.651 Intake Valve Net Area = 2.824
CID per Cylinder = 30.442 Intake Valve Dia. Area = 2.923
Rod/Stroke Ratio = 1.816 Intake Valve Stem Area = 0.098
Bore/Stroke Ratio = 1.350 Exhaust Valve Net Area = 1.996
Int Valve/Bore Ratio = 0.516 Exhaust Valve Dia. Area = 2.094
Exh Valve/Bore Ratio = 0.437 Exhaust Valve Stem Area = 0.098
Exh/Int Valve Ratio = 0.847 Exh/Int Valve Area Ratio = 0.717
Intake Valve L/D Ratio= .233 Exhaust Valve L/D Ratio= .242
CFM/Sq.Inch = 43.6 to 45.8 CFM/Sq.Inch =47.6 to 51.6
Curtain Area -to- Valve Area Convergence Intake Valve Lift inch= .482
Curtain Area -to- Valve Area Convergence Exhaust Valve Lift inch= .408


------- Piston Motion Data -------
Average Piston Speed (FPM)= 2886.46 in Feet Per Minute
Maximum Piston Speed (FPM)= 4703.61 occurs at 75.584 Degrees ATDC
Piston Depth at 75.584 degree ATDC= 1.2228 inches Cylinder Volume= 220.1 CC
Maximum TDC Rod Tension GForce= 1960.54 G's
Maximum BDC Rod Compression GForce= 1113.87 G's


------- Current Camshaft Specs @ .050 -------

IntOpen= 7.00 IntClose= 47.00 ExhOpen= 49.00 ExhClose= -3.00
Intake Duration @ .050 = 234.00 Exhaust Duration @ .050 = 226.00
Intake CenterLine = 110.00 Exhaust CenterLine = 116.00
Compression Duration= 133.00 Power Duration = 131.00
OverLap Duration = 4.00 Lobe Center Angle (LCA)= 113.00
Camshaft Advanced = 3.00 degrees

-Recommended Camshaft Valve Lift-
Minimum Normal Maximum
Intake = 0.395 0.426 0.468
Exhaust = 0.368 0.396 0.435
Max-effort Intake Lift = 0.491
Max-effort Exhaust Lift = 0.456
Minimum Intake Valve Lift to prevent Choke = .426 Lift @ 6250 RPM
Minimum Exhaust Valve Lift to prevent Choke = .396 Lift @ 6250 RPM


------- Operating RPM Ranges of various Components -------

Camshaft Intake Lobe RPM = 5819 Exhaust Lobe RPM = 5551
Camshaft's Intake and Exhaust Lobes operating RPM range = 3752 to 5752
Note=> Lobe RPMs are only BallPark estimations

Minimum Intake Valve Lift to prevent Choke = .426 Lift @ 6250 RPM
Minimum Exhaust Valve Lift to prevent Choke = .396 Lift @ 6250 RPM

Current (Intake Valve Curtain Area -VS- Time) Choke RPM = 6608 RPM
Current (Exhaust Valve Curtain Area -VS- Time) Choke RPM = 6236 RPM

Intake Valve Area + Curtain Area operating RPM Range = 4787 to 6787 RPM

Intake Valve Diameter RPM Range = 4608 to 6608

Intake Flow CFM @28in RPM Range = 3995 to 5995
__________________________________________________ _________________________

Best estimate RPM operating range from all Components = 4050 to 6050

Note=>The BEST Engine Combo will have all Component's RPM Ranges coinciding
__________________________________________________ _________________________


--- Cross-Sectional Areas at various Intake Port Velocities (@ 28 in.) ---
112 FPS at Intake Valve Curtain Area= 2.727 sq.in. at .450 Lift
105 FPS at Intake Valve OD Area and at Convergence Lift = .482
129 FPS 90% PerCent Rule Seat-Throat Velocity CSA= 2.367 sq.in. at 6250 RPM
350 FPS Velocity CSA= 0.874 sq.in. at 6250 RPM Port Sonic-Choke with HP Loss
330 FPS Velocity CSA= 0.928 sq.in. at 6250 RPM Port Sonic-Choke with HP Loss
311 FPS Velocity CSA= 0.984 sq.in. at 6250 RPM Smallest Useable Port CSA
300 FPS Velocity CSA= 1.020 sq.in. at 6250 RPM Recommended Smallest Port CSA
285 FPS Velocity CSA= 1.074 sq.in. at 6250 RPM Recommended Smallest Port CSA
260 FPS Velocity CSA= 1.177 sq.in. at 6250 RPM Recommended Port CSA
250 FPS Velocity CSA= 1.225 sq.in. at 6250 RPM Recommended Port CSA
240 FPS Velocity CSA= 1.276 sq.in. at 6250 RPM Largest Intake Port Entry CSA
220 FPS Velocity CSA= 1.392 sq.in. at 6250 RPM Largest Intake Port Entry CSA
210 FPS Velocity CSA= 1.458 sq.in. at 6250 RPM Torque Loss + Reversion
200 FPS Velocity CSA= 1.531 sq.in. at 6250 RPM Torque Loss + Reversion

--- Cross-Sectional Areas at various Exhaust Port Velocities (@ 28 in.) ---
123 FPS at Exhaust Valve Curtain Area= 2.026 sq.in. at .395 Lift
119 FPS at Exhaust Valve OD Area and at Convergence Lift = .408
147 FPS 90% PerCent Rule Seat-Throat Velocity CSA= 1.696 sq.in. at 6250 RPM
435 FPS Velocity CSA= 0.573 sq.in. at 6250 RPM Sonic Choke at Throat Area
350 FPS Velocity CSA= 0.711 sq.in. at 6250 RPM Port Sonic-Choke with HP Loss
330 FPS Velocity CSA= 0.755 sq.in. at 6250 RPM Port Sonic-Choke with HP Loss
311 FPS Velocity CSA= 0.801 sq.in. at 6250 RPM Smallest Useable Port CSA
300 FPS Velocity CSA= 0.831 sq.in. at 6250 RPM Recommended Smallest Port CSA
285 FPS Velocity CSA= 0.874 sq.in. at 6250 RPM Recommended Smallest Port CSA
250 FPS Velocity CSA= 0.997 sq.in. at 6250 RPM Recommended Port CSA
240 FPS Velocity CSA= 1.038 sq.in. at 6250 RPM Recommended Port CSA
225 FPS Velocity CSA= 1.108 sq.in. at 6250 RPM Largest Exhaust Port Exit CSA
210 FPS Velocity CSA= 1.187 sq.in. at 6250 RPM Largest Exhaust Port Exit CSA
190 FPS Velocity CSA= 1.312 sq.in. at 6250 RPM Torque Loss + Reversion
180 FPS Velocity CSA= 1.385 sq.in. at 6250 RPM Torque Loss + Reversion

Next post i will use to model my planned engine to create some gasflow and velocity targets.

[CBRacerX]

Subscribed!

[Gertvr]

Theoretical model to start head porting with from the PipeMax program:


CFM # is based on VE% required to fullfill cylinder demand of 222.6 CFM at peak RPM between 70 and 80 deg crank angle as per the screen shots below

Target inlet flow : 222 – 234 CFM (Carb, Manifold, head with valve at .502” lift & WOT)
Target inlet port velocity: 275 FPS down centre of port bare head
Target inlet port CSA 1.92 (sq-in) for above velocity
Target 0 turbulence at all lifts especially convergence lift for both intake and exhaust
Maximize swirl by port shape (better combustion, need to find a way to measure this)

Will start flow testing by next week.

Screen shots of program output.

[Bullet Bob]

Here are some interesting tidbits on Porsche head shape. These graphs show the 3.0 exhaust port's cross sectional area from the port (left) to the valve (right). The Y axis is the area in in^2. The first graph is of an unmodified 3.0L exhause port and the second is a port that has been smoothed by machining off the protruding valve guide etc... I do cnc porting and the coarsness of the ridges are determined by how much the tool steps over during machining. To make the ridges 1/2 as big it would double the time it takes to machine the port. These graphs are taken from the CAD files that drive the CNC machining process so they are extremely accurate.

http://www.youtube.com/watch?v=2lsS05SbYyI

[Flieger]

The x axis is arc length along the "centerline" of the intake port tract?

I assume the big dip in the stock graph is the valve guide boss intruding into the port?

[Bullet Bob]

Here is the same port that has been further refined.

[Bullet Bob]

Right, the X axis is a line down the center of the port pointing in the flow direction. The area accounts for the valve boss, guide, and valve stem and it is normal to the flow direction.

[Flieger]

Interesting how the area just behind the valve opens up quicker in graph#2 and then smooths out again in graph#3. Is this a "5-angle valve grind" job?


Thanks for posting these. I find this information very interesting and useful in visual form.

[Bullet Bob]

No, the walls have just been stretched and massaged a bit. It is very difficult to get it just right and tiny differences really show up on the graphs. The difference between smooth and not smooth is very subtle.

[jluetjen]

Just a data interpretation point, note that the graphs do not have their origin at 0,0 which tends to accentuate the difference.

For example, it looks like a stock port loses about 9% of it's area at the valve boss, so the velocity through that venturi would increase by almost 10%. I don't remember off of the top of my head how much the pressure will drop as a result of this. The question is -- is this a good thing or not? Obviously obstructions are bad things. But 9% doesn't strike me as a huge issue, especially if an engine isn't port constricted except for potentially at peak revs. A couple of hypothetical mental examples...

1) It might be advantageous to increase the charge's velocity prior to it flowing into the combustion chamber since this will increase turbulence downstream of the venturi and the valve -- specifically in the combustion chamber. Thus it aids the mixture of fuel and air in combustion chamber which can improve combustion quality and performance.

Alternatively...

2) You might want a straight shot down the port to allow the smoothest possible flow past the valve, and depend on the combustion chamber shape to provide the turbulence needed to ensure a high quality combustion event. In this case, just grinding the whole boss away is easy enough. But then the question is what does this do to the flow past the valve stem? In many cases a smooth transition may be preferable to the reduction in cross-sectional area. I suspect that Bullet Bob is actually opening the port up around the boss and leaving the boss in place, thus leaving the boss to smooth the flow around the valve guide while still keeping the same cross-sectional area.

Either approach might be successful, but they both depend on making port design do what is best to support the most efficient and powerful combustion in the chamber. So the design of the piston crown and combustion chamber in the head also come into play. So my point is that I think that we need to be careful to assume that either "More is better" or "Straight is better."

[gtc]

Am I reading these wrong, or does the stock port (top graph) have more cross sectional area behind the valve?

[BReyes]

Like graduate school work without the school, actually more exciting because it is automotive/motorsport.

Comphrehensive, and with good planning.

[jluetjen]

Originally Posted by gtc

Am I reading these wrong, or does the stock port (top graph) have more cross sectional area behind the valve?

Good point. How did that happen? Are you adding material as well as cutting it away Bullet Bob, or maybe using a different head as the raw material for your porting excercise?

[Gertvr]

My apologies for the slow updates, been busy at work with new projects.

I finally got time to calibrate the flowbench and get the first flow tests done. I used 3 sharp edge orifice plates that were sized for 100, 200 & 300 CFM @28" of H20. The flow numbers measured on the bench were 99 , 201 and 300 CFM for the respective plates and i was very happy with the numbers for a home built bench.

I fabricated the attachment in the pictures below to connect the cylinder and head to the flowbench and the mechanism to open the valve and proceeded to do my first flow tests.

I had done some light porting on one head previousy and did a comparative test to one of the stock heads. The ported head used a +1mm inlet valve.





The heads flowed a respectable amount of air/sq-in of valve area (82CFM stock and 87CFM modified) and compares favorably to other flow numbers posted on the forum. This doesnt really mean a lot as the flow alone is not the begin and end-all, but at least gives an idea. Low lift flow is still very close to stock and there was a slight lift in the flow numbers from .150" up to the end of the lift range for the modified head.

The logged velocity measurements were taken with a pitot tube 1.5" into the port in the center, I also measured the airspeed just off the SSR and measured 390FPS. From the research i have done, indications are this might be too high and it will be one of the areas to focus on in the next steps together with the low lift flow.

I have also done flow tests without a radius on the port and that yielded lower flow numbers (~-6CFM) but higher velocity numbers. I have not logged them but will in future. This might be useful in future to get the port velocity up for lower lift numbers.

I also flow tested the 50PMO carb on a manifold and then the PMO connected to the slightly ported head all at 28"

The carb and manifold flowed 305CFM and the carb and manifold connected to the head flowed 250CFM @ full lift (.502") and introduced 15CFM loss over the bare head. The loss could be due to the fact that the manifold was not mached to the head and the head port edge was in the airstream.

I will next map the port velocities and pressures at various positions in the port to get an idea of the flow patterns to plan the next modifications.

Head mounted on flowbench with digital flow readouts:

[jluetjen]

Excellent!

[Bullet Bob]

Originally Posted by gtc

Am I reading these wrong, or does the stock port (top graph) have more cross sectional area behind the valve?

I'm not really sure how this happened. I ran these tests over a year ago and it was more out of curiosity than anything. The cnc porting I do is to convert 3.0 small port heads to big port heads and T heads to S heads. These are an exact match to the factory heads to comply with the 911 spec racing class.

Since I had accurate CAD files for these heads I played around with them to see what the characteristics were and the impact of minor modifications. In modifying the heads I removed a bit of material right around the valve boss on the outside of the port and I eliminated the valve guide protrusion in the same manner as the pictures of the machined heads shown earlier in this post. If I remember correctly I think I even tried valves with different size stems.

One thing that stood out was the factory 3.0 heads had a much smoother transition than the T and S heads leading me to believe that this was the direction Porsche took as they refined their port shapes.