2789cc valvesize - 48x38 or 50x40

[-Alex-]

2789cc valvesize - should i go with 48x38 or larger 50x40 for TP billet heads?    All valvetrain stuff will be light, JPM std dual springs, JPM ti retainers, alu pushrods, Udo's light lifters.  I intend to use stainless T4 lenght valves with 3 groove locks. I have not yet decided the cam, some of the two cams, either 280 or 282@0,050" from JPM.

I am aiming  to 250~260hp. 

I do have 48x117mm manleys allready, but Scat only sells at web 50mm 117mm valves. Are the any good, or equal as std Manley valves for VW?

 

[Torben Alstrup]

If the intake port can support it, definitely 50 x 40. AFAICT the Scat Valves are fine. I have not seen any problems with them worse or better than Manleys.

Another option is DelWest or even Ferrara, but they are most likely not exactly cheap.

T

[magic]

2789cc... with JPM's 282 camshaft and 1.4:1 + 51.5 IDA or 48 IDA's with 44mm. JPM Raptor Venturies...  
YOU SHOULD BE FULLY COVERED with 48x38 valves up to 250-260 hp. !!!  
 
With 50x40 you either block the air in front of the valve, or you make the indtake canal larger, And loses velocity.
 ie. you lose top hp, without getting torque... OR're losing torque without getting more top hp.....  

Regards, Magic  

Correction ... You will lose torque but of course not lose peak hp.! 

[magic]

ups ... different ways of looking at it ..... sorry Torben! ;-)

Hilsen Mads

[Torben Alstrup]

No problem. Its a free country.

You are sort of correct. That´s why I wrote what I did. But with almost 700 cc per plug he has the dispacement to exploit the larger intake surface.

T

[magic]

Hmm ... Torben why did't you stop me ....    Of course you're right!

Ok, what I was trying to say was .... you can easily achieve your goal with 48x38.
It is of course easier with 50x40 and you will end up with more HP.
But with the 48x38 you will be able to reach the 250-260HP. AND get a stronger bottom to mid-range (much better I'll say)

Mads ibsgaard

[-Alex-]

I am running fuel injection with 2 x drla style 48mm throttlebodies, no IDA:s 

Heads do have somewhat large intake ports,  chamber side port opening is 42mm and manifold side port opening is 46,5x36mm.

Valves would look like this (chamber is littlebit different)

[magic]

That is a (hot) dedicated street engine if I ever saw one!!...  
 So you are hunting for street-ability!??
 And then your 46.5 X36mm would be pretty close to ok with your 48mm. throttlebodies right?!

[-Alex-]

Yes, dual 48mm throttlebodies and i am getting loong velocitystacks.  I saw here one 2276cc engine making 265hp with MS250 46x38 heads   

[BeetleBug]

I saw here one 2276cc engine making 265hp with MS250 46x38 heads   

And

- A 40mm intake valve making 194hp on a 1600ccm engine
- A 40mm intake valve making 206hp on a 1776 ccm engine

What can the intake channel support? There is no need for stuffing a huge valve in there if not the rest of the channel is able to match.

-BB-

[magic]

Yes, dual 48mm throttlebodies and i am getting loong velocitystacks.  I saw here one 2276cc engine making 265hp with MS250 46x38 heads   

Cooool 

[magic]

"Looong velocity stacks" = torque...
So it IS a street engine you are building there!...       48x38!!!

[magic]

My 84x102 type 1, with JPM 230 heads, 48x38, 11,2:1, 282 * Raptor cam, 48 IDA and JPM 44mm. Raptor Venturies.....  and CRAPPY jetting (waay too rich)  
gave 247hk. and 328nm.  It could easily have made ​​over 250hp. with proper jetting!!
But my IDA: s are sold now, and I'm also switching to throttlebodies, on my hunt for more street ability....... And fun......

Good luck with your 2,8L !!!!!  

Magic

[GetBackOnTrack]

I have both 50mm and 48mm stainless intake valves "for type 4, 117 mm lenght" in stock.

Im Building a 2805cc type 4 myself, using 50/40 valves and shooting for 275-280 HP.

Just a Little comparible info 

Regards, Jakob.

[-Alex-]

Yes, it will be somewhat a street engine, pushing the germanlook 1300 to the limits on track/strip and cruising at highway and beach roads (yes we have beach roads here in Finland) 

Johannes recommended me that 282@ cam few years ago, i will be using 1.4 pauter rockers, so there is about 15mm valvelift.

Wonder what is the usual 48mm valveseat ID?  42-43mm?

I'v been looking CB performance stacks:

http://www.cbperformance.com/ProductDetails.asp?ProductCode=6379

[morkrieger]

Hi,

If you have a proper CAD model, you are not far off from running a proper CFD simulation on it to get some flow numbers and toy around with valve size and channel volume.
If your input is right, the differences with the flow bench are almost non-existent, at least on the stuff i tested so far. I used Solidworks Flow Simulation for this set against a 'analog' SF-300 Superflow.

A dynamic simulation requires programs in the likes of ANSIS FLUENT which are near impossible to understand (especially for internal combustion engines) without proper training.
Though a virtual static, constant depression test, can give you insight regarding velocities and volumes in ways you can't see on a flow bench, like slicing the head and probing velocities on several points at once.

[-Alex-]

I have done few airspeed test with 129 000 Pa pressure at intake port and 100 000 pressure at cylinder (Solidworks  FloXpress program doesnt allow suction?) .  This is the head what i have designing for Finnish cnc manufacturer:

[morkrieger]

Flow Xpress is very limited, do you have access to the full module 'Flow simulation'?

If yes, then setup an internal flow sim, and mimic a flow bench (including a test cylinder). Insert the 'lids' to seal the ends and setup the boundary conditions.
Now on the 'goals' insert a surface goal on the bottom surface (the lid) of the cylinder (like the flowbench orifice), for 'total volume', this will plot in your simulation then.

Make sure the mesh 'size' selected matches your model, too large of a cell makes the calculation unreliable.

I can help you out here, but perhaps not over this forum ? Send me a PM if needed.

[Udo]

I am running fuel injection with 2 x drla style 48mm throttlebodies, no IDA:s 

Heads do have somewhat large intake ports,  chamber side port opening is 42mm and manifold side port opening is 46,5x36mm.

Valves would look like this (chamber is littlebit different)

For that port size-do not know how they look inside - 48 are big enough

Udo

[-Alex-]

This is cad cutaway from Thorstens head, port is from intake flange 46,5x36mm  to chamber 42mm.  I dont have solidworks flow simulation.

http://img7.imageshack.us/img7/139/newchannel.jpg

[Jon]

The sharp corner in the roof of the canal just before the valve seat, where all the lines separate from the wall, is that intentional?

[-Alex-]

I am not sure what you mean?  Sharp corners at port are not intentional, some if not most 3D cad programs make like that port. I'll explain later today with pics how ports are made at 3D cad.

[Jon]

If the flow simulation is decent, the top half of the canal looks ineffective (red arrows)
The flows follows the shape much better in the bottom (blue arrow)

[ Attachment: You are not allowed to view attachments ]

[morkrieger]

Hi,

Alex, look at options-> add-ins are you sure there is no flow simulation there? (it needs to be 'enabled' seperately).

The FloXpress stuff is very basic and the flow paths (or worms-nest as i call it) doesnt show much except pretty pictures, what you need is data with set goals and a decent Mesh

Send me a PM if you want more info or ask here.

[-Alex-]

Yes, its not intended, i dont know if the intake port near at valve doesnt work at best or just the simple FloXpress doesnt give best results. You can choose tubes and balls to show flow, and with balls    it shows continous moving flow.


I found the Flow Motion, i thought first that it is extra (pricey) package, but i got it on.  It definetly looks more complex than floXpress  

[wph]

I am not sure what you mean?  Sharp corners at port are not intentional, some if not most 3D cad programs make like that port. I'll explain later today with pics how ports are made at 3D cad.

This is why many of the CNC heads are scanned from a hand ported prototype ports.
Cylinder head professionals design and modify the ports to a casting and sell their work to the CNC company.
Top end aftermarket castings flow very well out of the box (in general) if they are designed and manufactured by
technology available today (3D CAD, scanning, flow simulation and rapid prototyping)
 

[wph]

What you see between the red arrows is commonly called short side radius, SSR.
You want the fattest curve / largest radius to this point unless you're running the latest
four valve head designs. You need a flow straightening section just before the seat and
a combustion chamber wall supporting the ”flow cone” out of the seat area.

In the picture flow is separating from the port wall at a point of left red arrow creating a wake &
turbulence behind it to the most sensitive place in the intake port, 1” before and after the seat throat which
should be the smallest cross sectional area in a port at least in a 2- valve (racing) engine.

This area is so sensitive that a cut in a port 6 mm wide and 0.5mm deep can gain you 10-20 hp worth
of flow in a 4- cylinder head if the port's cross sectional area for the intended rpm & displacement is adequate.
Sometimes it is better NOT to blend the cut to the rest of the port.
Just a food for thought.....

Yes, its not intended, i dont know if the intake port near at valve doesnt work at best or just the simple FloXpress doesnt give best results. You can choose tubes and balls to show flow, and with balls    it shows continous moving flow.





I found the Flow Motion, i thought first that it is extra (pricey) package, but i got it on.  It definetly looks more complex than floXpress 

"You can choose tubes and balls to show flow, and with balls      it shows continous moving flow."

If you are preferring to adding objects in front of SSR in a simulation, that's a common flowbench practice for "seeing"
what the flow "wants" or needs 

[morkrieger]

Hi guys,

with Flow Simulation, the full module, you can now easily measure the real volume flow, or even preferred, the mass-flow through the port, at a given valve lift.
You can 'iterate' through several configurations/port designs and look for trends, classic flowbench tools like the straw, or even pitot tube are pretty much replaced by hard numbers.
I must press on the fact that the Mesh (or 3D 'air cells') needs to be sound for reliable results.

You will also be amazed how much difference a good/bad manifold makes, it's a bit more than 'just a few potential HP'.

I'm pretty confident Alex can get this going, the on-board help in SW is pretty good. Good luck  And also, listen to the pro's out there that do the real job (flow/porting), a computer is a very 'dumb' tool.


(and as added bonus, you could dive into thermal simulations as well, even get close to the amount of heat the head is able to dissipate, or improve your fan-shroud etc..etc...it's so much fun )

[-Alex-]

I know that short side radius must be big as possible, but basic modeling at SW doesnt allow it, at least not with easy work.  Thats why i am thinking to leave the chamber side port size a bit smaller on heads which will go into production (so you can hand finish after seats are installed) on Tommis Billet:

http://cal-look.no/lounge/index.php/topic,21032.0.html


Port is done with Cut Loft  with two sections (sketches) only, at the chamber and intake flange + ofcourse center line.  I'v tried to ad sections and it works - but its hard to find correct shape of section and find correct place for it - this would be also way to get venturi into port if needed?

[morkrieger]

Not wanting to sound like 'the wise guy' here but....
I strongly suggest looking into the 'lofted cut' (with the start-end constraints function) and 'surfacing' stuff on solidworks, it's daunting at first but it allows you to 'clay' the port, and also have smooth convergent/divergent geometry in the port.
CNC is the easy part, a well suited/well tested model is 90% of the work. Though leaving some room for hand porting sounds interesting as well....so many options .

attached is a quick-n-dirty example of what i mean, this is done with working the start-end constraints of a loft.