4-VALVE HEAD - OPTIMUM PORT VELOCITY?? 4

[v114]

I just recently found out about this forum and am very impressed so far with the topics and discussions. I decided to post a question.

Quick background is that I am designing a head for a V twin from a clean sheet of paper. It is a DOHC, 4 valve design. I have a mechanical engineering background and I have experience with flow simulation software and would like to utilize it to design the proper port sizes and shapes for this engine. The engine has 4.25 bore and 4.00 stroke. The rev range should be 7000 to 8000. I would like to see it make peak power well above 6500.
Most of what I find for literature exists for 2 valve designs.

What equations /rules apply for max port velocity at 28 inches H2O ( intake and exhaust)for a 4 valve design (2 intake, 2 exhaust?

Any other suggestions on good port design practices would be appreciated.

Thanks

 

[SMOKEY44211]

Max port velocity should not (cannot) exceed the speed of sound.------ Phil

 

[willeng]

Phil is right, the thing to consider also is the total valve area of the design & this will govern what the port area etc will be. Everybody has there own idea's on things like this, but for me it is better to design the ports with the intention of porting to the correct size & shape to get the engine characteristics required. In other words start with the smallest ports in the design stage & test from there to get what you need, allow ample port wall thickness.

Don't forget the velocity around the valve head as well, there is a calculator at

http://www.rbracing-rsr.com/machcalc.html

There is also another calculator for runner area vs torque

http://www.rbracing-rsr.com/runnertorquecalc.html

Of coarse i wouldn't take these as being 100% correct but it might help get a better understanding of whats required as starting points only!

 

[flowmaster]

I have had good results with single and twin cylinder engines by keeping the average intake port velocities around 360 to 380 fps based on port cross section .do every thing you can to keep the port straight to reduce localized velocities .

 

[EngJW]

A number of sources quote a figure of 50% or so of the speed of sound. This is based on average piston speed, the temperature of the inlet charge (which determines the speed of sound), and the flow area. The flow area might be the port area, the area under the seat, or through the seat depending on who is quoting the figures. There is at least one good SAE paper on the subject.

I think these rules of thumb are very usefull to get one started on a design. Simulations may have their place, but eventually you will have to slug it out with a good 2d or 3d cad program to figure out what will fit within the bore, how you will run the ports, how to get the spark plug in, and how to get enough cooling for the exhaust. This is what will make or break a design, and there will be lots of compromises.

 

[andyv8]

3.0L V6 4valve - max power at 6850:
Mean inlet port speed = 100m/s
Mean exhaust port speed = 120m/s
Based upon average port diameter (exc. throat area which was CNC machined)

Any higher and the powerband just went higher and torque dropped off, and vice versa.
This was also shown to be optimum based upon 1D engine simulation.

 

[v114]

Thanks for the info.

For Andyv8:

I assume that the velocities of 100 intake and 120 exhaust were measured/calulated based on flow bench work using 28 inches H2O? Is that correct?? If so, these are well in line with what I have been using so far.

One more question for all: We have built a prototype of this head design and flow tested it. The intake port is very "noisy". Piecing whistling sound. I am working with an experienced builder and he suspects turbulence. The port flows 300 CFM at .200 lift, 360 CFM at .500 lift at 28 inches and the ports are not overly small. Problem is this engine made best power at only 5200 rpm. Could the noisy port be responsible??

Any thoughts??

Thanks, Jim

 

[andyv8]

Jim,

The 100/120 port velocities were calculated from mean piston speed.

port vel = (bore^2/port_dia^2)*mean_piston_speed
where port_dia is the equivalent dia of the combined 2 inlet ports. (all in mm)

I'll try to dig out some flow bench data from the same heads.

 

[richdubbya]

My experience with four valve engines comes from experience developing maximum useable power in existing engines rather than design. I have found some basic differences between 2 valve and 4 valve designs though. One of the basic differences is the ability of the 4 valve design, because the lighter valve train weight, lower valve spring pressures, the lowered requirement for valve lift (more low/mid curtain area) and the resulting ability to open and close the valves faster, to require much less valve open duration for an equivilent rpm/cylinder size than the typical ohv 2 valve design. I have found that in long duration 2 valve applications you need a higher port velocity to get the flow moving at the low lifts when the piston is still moving slowly and you need high velocity when the valve is closing to prevent blow back. I believe that on well designed 4 valves the shorter duration and much faster opening and closing lessens the need for this high velocity. Furthermore this fast action (and the multi valves much higher low/mid lift flow) require a relatively larger volume behind the intake valves (less velocity). The same applies to the exhaust side, the blowdown occurs so much faster that larger ports (compared to a 2 valve and less velocity as derived from an engine size formula) are required. IMO velocity has a time element (duration) along with the engine size/port size normally associated with it. I have found 4 valve designs with smaller ports and more duration can develop more rpm but with less torque and do not perform as well. The larger port and shorter duration engines I have experience with are superior even at the lowest rpms. I have no knowledge on how this would affect any low emission requirements though.

 

[v114]

Thanks for the info.

Couple of more questions:

1. We are using 246 duration and .460 lift for both the intake and exhaust. Does that sound reasonable for peak power near 7000 rpm?

2. "Furthermore this fast action (and the multi valves much higher low/mid lift flow) require a relatively larger volume behind the intake valves (less velocity)."

Do this mean that the intake runer length should be longer or bigger in cross section than on a comparable 2 valve??

3. Basically, this engine has very good low end torque but will not make power past 5200 rpm. We are tying to figure out why. Both cylinders now are sharing a single 48mm carb and we suspect that we may need more. The engine is 4.25 bore and 4.00 stroke.

thanks

 

[richdubbya]

These are only my opinions, and they are based on smaller cylinders than yours and much higher rpms. I would suggest listening to everyone and drawing your own conclusions. I would think that comparing equal engine sizes and rpm ranges a four valve would require at least 20 degrees less duration, .100 less lift and a small 10 to 20% increase in port cross section compared to an optimum 2 valve. I have only worked with a throttle body or carb for each cylinder so I have no knowledge of manifolding or carb size for multiple cylinders on 4 valve engines. We successfully use 41mm throttle bodys for each 300cc cylinder with port injection. Carburetors this size will result in a loss of low end torque and are limited to 38mm. I can't give an opinion on your specific numbers not knowing how the duration was measured or valve/port sizes etc. I have seen good power peaks at 10,500 with less than 220 degrees duration @.050 lift in admittedly smaller cylinders than you are working with. My experience favors equal or slightly less duration on the exhaust side. Camshaft phasing is very critical in these engines and can alter the rpm range considerably.

 

[andyv8]

I needed a 74mm throttle body on the V6 for best power - 65mm wasn't too bad, but still reduced power by ~6bhp @ 7000rpm.

Found the flow bench info (25" water):
Hope it's of some use.
Bore = 87mm
Stroke = 82.6mm


Bare Cyl. Head with plasticine bell mouth
Cylinder 1 - Forward Flow
Lift (") Flow % Range CFM
0.050 13.10810811 4 38.8
0.100 27.90540541 4 82.6
0.150 40.97972973 4 121.3
0.200 52.63513514 4 155.8
0.250 61.68918919 4 182.6
0.300 68.31081081 4 202.2
0.350 71.99324324 4 213.1
0.400 74.69594595 4 221.1
0.450 75.70945946 4 224.1

Cylinder 1 - Reverse Flow
Lift (") Flow % Range CFM
0.050 4 41.9
0.100 4 81.4
0.150 4 120.5
0.200 4 147.7
0.250 4 166.8
0.300 4 177
0.350 4 183.5
0.400 4 188.9
0.450 4 191.9

 

[sreid]

Years ago a Cosworth BDA four cylinder 1600 cc (about 100 cu. in.)had a 50 mm carb per cylinder (two each DCOE 50).

One 48 mm carb may be starving your twin.

 

[maverickgt]

i like you're idea, one thing I have always thought about is trying to maintain a constant port cross-sectional area, the best port velocity is going to depend on your application and design, start modest and get bigger while testing power output. Use numbers as an initial test, when it comes to engines, even the best engineers int he world rely on research and development. I'm sure I didn't add much, but good luck.

 

[Hasbro]

Here is some additional input you may investigate. Although not relating to the cylinder head, the 4 cylinder oval migit engines have found for some time now, that the single most rpm extending area of attention is, the increased taper in the injector manifold to the butterfly. The port stays the same in the head, albeit the size of the butterfly has become remarkably larger in relation to what one would concider adaquate. It has nothing to do with airflow in, but affect the pressure wave in reverse. This wave is succesfully dampened enough to allow the flow to reverse forward into the port with decreased pumping loss, allowing these engines to hang on conciderably longer than without this taper. Although the power band your looking for may not apply, there is certainly a finger pointing in the manifolding direction IMO.

John Haskell
Aire Research Engr.

 

[v114]

In general, how does a 4 valve differ with respect to port velocity as compared to a 2 valve? I am looking for some specific guidelines as far as proper port velocity (feet/sec) using a pressure differential of 28" H2O. This a 7-8000 rpm potential motor.

My goal at this stage is to use a port flow simulation program (which I have access to) that I can set up to simulate flow performance on a flow bench. Input to my program are pressure differential (28" in this case) and also port velocity.

Thanks again

 

[AboveRedline]

V114,

No one is answering our question about flow bench velocity because thats not the way most of us approach the problem. The demand placed on the head comes from piston motion (and exhaust scavenging). You need to calculate the demand first, and then apply the demand across the cross-sectional area of the head, and this gives you the velocity in the head. Most people look for intake manifold mean port velocities in the 300-360 ft/sec range.
The flow bench can be misleading because when you put a bigger head on it, the pressure drop goes down, and common practice is to crank the motor up to restore the 28 in pressure drop. Does an engine do this? No. The piston motion generates the available potential energy, in other words you only get so much to work with.
Shrieking ports on the flow bench are not good. Because you have super high low lift flow, and not much gain as the flow increases (what size valves?, the short side turn may be too high, however this usually is not noisy.
360 cfm should be plenty to take a Harley to 8,000 rpm. Are the exhaust pipes proven with regards to length and diameter?

 

[v114]

AboveRedline:

"Shrieking ports on the flow bench are not good. Because you have super high low lift flow, and not much gain as the flow increases (what size valves?, the short side turn may be too high, however this usually is not noisy."

Only the intake ports seem noisy. There is a divider where the single runner divides into separate ports leading to the valves. There may be something there that is causing the problem with noise. I just don't know whether that noise has anything to do with the limited peak power rpm we see on the dyno. Can the port be limiting rpm because of turbulence??

"The demand placed on the head comes from piston motion"

What is the correct calculation for the demand from piston action?

"Are the exhaust pipes proven with regards to length and diameter?"

Nothing on this engine is proven. It is a completely new design. We are learning what it needs. On the dyno, we tried two completely different exhausts and the peak power would still not go beyond 5200 rpm.

 

[WilliamH]

v114,

RE: Carburetor

Your description of the carburetor as "48mm" is somewhat vague. The airflow limiting feature of a carburetor is usually the venturi. Assuming you have a 48mm butterfly, the venturi ID would be smaller in diameter. You may need a larger venturi ID. You probably need a much larger carburetor as well.

I used to build V-8 road race engines in the early seventies with 48mm Weber carbs. I used four 48mm Weber IDA's (butterfly diameter). The engines were about 300 cubic inches displacement. These carburetors had removable venturi's and various venturi ID's were offered. These were two throat carburetors, i.e. one throat per cylinder. We had excellent power to around 8500 RPM.

If your carburetor has a venturi that is too small, you could be going into choke flow easily. Hence, your 5200 RPM upper limit. We used to size the venturi's to match the track. Small relative diameter for tight "torque" tracks. Larger venturi's for a track like Riverside International Raceway (long track). Venturi size will move the torque curve like a dial on a radio.

Flow test your carburetor and compare the test results with what the engine needs at say 90% to 110% VE or so.

I have a Porsche 996 GT-3 Cup head (four valve) in my shop that I would flow test if I had a bunch of extra time. Whatever the Porsche factory did to this head, they did right.

Will

 

[v114]

"Your description of the carburetor as "48mm" is somewhat vague."

The carb used is an HSR48mm Mikuni flat slide. No buterfly. It measures about 48mm at the bore.

"Flow test your carburetor and compare the test results with what the engine needs at say 90% to 110% VE or so."

Mikuni states that the carb will flow 270CFM at 12".

I calculated the theoretical flow rate needed for 114 cu.in. at 8000 rpm to be 264 CFM (assuming 100% VE). Given this, what is the recommended intake/carb flow for the application? Is it safe to design to only 264 CFM? I assume there is a multiplication factor needed for a design flow rate based on theoretical flow required.

If the theoretical flow is 264 CFM
The intake flows - 360 CFM at 28"
The carb is capable of 270 CFM at 12"

Could the carb and intake port be a limiter????