Flow rates of available stock heads??

[333pg333]

Does anyone know the cfm's of the 951, 944s2, and the 968 heads? I believe that the stock 951 is 188cfm's but I wouldn't stake my life on it. My modified 8v is 247cfm but I'm led to believe that the 16v pretty easily surpasses this. I am deliberating getting a 16v head on a spare engine but I want to see if the extra costs are worth it.
Thanks for any advice,
Patrick

 

[George Elliott]

Patrick,
I'm no expert on flowing heads, but your question reminds me of the quite detailed info on Lindsay Racing website on this subject. There must be significant additional flow with 16v. Porsche went to 16v with 3.0 capacity, they must have figured a 750cc cylinder needed it. Flow is only one advantage of 16v, Audi went to 5v heads for a while. Honda & Mercedes went to 3 valve heads for a while. With an odd number of valves, the inlet is always the greater number.

By my reckoning 3000cc = 1181cubic inches = 0.684cubic ft.
At 6000 rpm thats 0.684x6000= 4104cfm
On a 2.5 litre at 6000rpm its 0.570 x 6000=3417cfm.

If you work out the port opening, -say a 40mm valve with a port below it of say a 36mm open dia, the opening is 3.14x18x18=1017mmsq, less an 8mm dia valve stem, is 3.14x4x4=50, then the Port has an opening of 967mm sq.

If you have two slightly smaller valves as there is less space for them, then assuming a valve of 80% of the 8v equivalent, say 29mm open dia, the numbers are, (3.14x14.5x14.5)x2 valves=1320 less the two stems = 1220mm sq. Thats a 20% increase in port area over the 8v setup. The 2.5 to 3.0 litre capacity represented a 20% increase.
I do not know the actual valve sizes.
I also believe a 3.0 will work fine with a 8v head, (eg 968 Turbo) Cam profile is another variable.

The cfm figures used by the head specialists are measured at a level of pressure to simulate back pressure in the system, and enable them to get a reading. These numbers are just mathmatical facts which I consider are relevant and may be of interest to you. Hopefully they are accurate, if not some-one will surely correct me.

Personally, I prefer 8v engines, for reasons of durability & simplicity. A 968 has long fragile Valve stems not to mention the valve drive mechanism.

George

944t
964rs

 

[333pg333]

Thanks very much George for your detailed and thoughtful reply. It is by far the best one I've had on all the sites I've asked the question of. As with most decisions there are the pro's and con's to be considered. This has grown out of not wanting to go through this ridiculous waiting that I've endured with my car off the road due mainly to a conspiracy of circumstance. It's been out of action since Nov 06 and not through lack of funds or trying on my part. So I decided to get a spare block to have as backup. This then grew to a long block, that has now morphed into an engine and 'Hey Presto' out of left field we now have an S2 that we're going to convert into a track only car so this is now looking like a 3.4L 16v dry sumped beasty!! As you can imagine the costs are spiralling out of control and hence that's why I'm asking the questions re 8v v's 16v. I could just stick with 3L and use my 8v head and this would save $1000's so it's becoming a matter of indulgence as much as anything. I am most likely going Link standalone and this engine will be in the road car for a while to do some tuning so that could prove a bit of fun. There will be some custom work done on the crank, rods, pistons, block, head, cam, itb's, manifold so unless we can make the valves stand up to it I may also stick to the 8v. Much to deliberate and that's why I wanted to know the stock cfm's on the 16v as a starting point v's 247cfm's.
Thanks again
Patrick

 

[333pg333]

Woops, just one thing George. You say that 3000cc = 1181 cu "?? Maybe I'm missing something here but isn't a 5L V8 something like 327"cubes? Different rate of measurement?
EDIT: Ok i've just checked. 3000cc = 183.07 "^3 = 0.105 '^3

 

[Diver944]

Don't know if this will be any help but Peter Empson recently posted that his S2 head is being breathed on and rebuilt by SimonP at the moment. I don't know if the flow will be measured afterwards but maybe he'll post here when he gets it back (hopefully next week in time for the Silverstone evening [8D])

 

[333pg333]

Interestingly Peter responded on the 968 uk site and said that his had been ported/polished etc and returned 197cfms which I find quite strange?? That's why I was hoping someone had the stock flow rate for these heads as that sounds waay under what I'd expect. In fact I am wondering if he meant 297? Does anyone know Simon well enough to check on this? I have also received info from an American that he has an 8v that has been relatively cheaply modified using Chevrolet valves and has 280cfms which I also find very hard to believe an the inverse fashion? From what I believe there are flow benches and there are flow benches. A bit like the subjectivety of Dynos. I made mention on Rennlist that US dynos are viewed around the world as being a bit optimistic. It didn't go down so well with them. I was accused of jealousy and mechanical inadequacy so to speak. Haha what did I expect?

 

[PJS917]

Patrick,

PM me with your email address and I will mail you my figures and see if the measurment techniques are comparable.

 

[Peter Empson]

ORIGINAL: 333pg333

Interestingly Peter responded on the 968 uk site and said that his had been ported/polished etc and returned 197cfms which I find quite strange??

Hi Patrick,
I don't think that was me, unless I was really, really drunk []

My nice and shiny head is in a box awaiting to be fitted so dont know what to expect till next week. Simon didn't mention any flow figures so I'm not sure if he felt it necessary to measure it, but I know he visits here so maybe he might post if he has an opinion.
Best regards,
Peter

 

[SimonP]

Hi Patrick and Peter
I can't add anything to help I'm afraid, however, I will be doing some development work on an S2 16V head soon for another project and as a consequence will probably discover the flow figures of the head in standard spec.

 

[333pg333]

Ah different Peters. Sorry for the confusion.[&:] Anyway thanks for any input guys as it's a big dollar decision and I have a few choices. With this new addition of the S2 that is going to be the final recipient of whatever engine we come up with there are now new costs to be considered. You would think that the basic flow rates of the Porsche heads would be available somewhere?
pgarvan@bigpond.net.au for PJS917 Peter.

 

[Neil Haughey]

ORIGINAL: SimonP

Hi Patrick and Peter
I can't add anything to help I'm afraid, however, I will be doing some development work on an S2 16V head soon for another project and as a consequence will probably discover the flow figures of the head in standard spec.

I have often wondered if anyone has considered fitting bigger inlet valves to the S2/968 head?

 

[thingo]

ORIGINAL: Neil Haughey


I have often wondered if anyone has considered fitting bigger inlet valves to the S2/968 head?

My understanding is that the 968 has bigger inlet valves than the S2(and smaller exhaust- 39/33 v 37/34)

 

[Neil Haughey]

If true that makes sense because I always thought the S2 had funny valve sizes. The angle of the engine means the flow path for the inlet is like you get on motorbikes and may probably suffer from some of the same problems. This is without concidering the fact that compared to many other 16v engines the S2 valve sizes look like the inlets are much to small or the exhausts are much to big. Still can't understand then why the word on the street is that the 968 doesn't make good power on throttle bodies.

Maybe the ports have a bad shape? On motorbike engines tuners often have to add material i.e. you have to build up the bottom of the port to get better air speed and higher CFM by creating a more continuous longer curve into the combustion chamber.

 

[George Elliott]

Patrick, apologies for giving you incorrect cfm rates. Per your reply a:

2.5L flows 529cfm @6000rpm
3.0L 635cfm @6000rpm (+20%)
3.2L 678cfm @6000rpm (+28%)
3.4L 720cfm @6000rpm (+36%)

You will have difficulty getting stock rates as all the heads will flow well in excess of these numbers.
The (flow related) performance improvement comes from the number of times the flow capacity is greater than the flow required.
To get a reading, the tuners have to restrict the flow by applying a back pressure which will be applied before and after for comparison.
The whole tuning subject is so complex that even huge companies who have been making cars for decades are unable to make a decent engine. Personally I would have a lot of respect for Cosworth, Yamaha, Honda, Lotus plus the F1 specialists in England. For Turbo applications, I have to hand it to Porsche. There may well be a "Barnes Wallis" type genius who can bring all the variables together for you to make a great performance package.
Let us know how you get on.....
George

 

[Neil Haughey]

A mate of mine from years back was a CFD expert. He was into modelling flame propogation and all that stuff that the major motor manufacturers are obsessed with. Any way after about 5 mins my eyes started to glaze over. I thought modelling nuclear reactors was bad enough. Maybe the bigger you are as an industry the more complicated you can make the subject.

 

[knightma]

ORIGINAL: George Elliott

Patrick, apologies for giving you incorrect cfm rates. Per your reply a:

2.5L flows 529cfm @6000rpm
3.0L 635cfm @6000rpm (+20%)
3.2L 678cfm @6000rpm (+28%)
3.4L 720cfm @6000rpm (+36%)

It's still incorrect George, a 4 stroke engine only has an induction stroke every second revolution so the values are actually half what you quote above. Also, these flow rates are swept volumes, which for any given engine speed will be greater than the actual air flow through a n/a engine (not sure about a turbo-charged engine as that is forced induction so it will depend on the boost pressure).

Actual flow rates are measured on a flow bench. A flow bench tries to simulate full power engine conditions and measure the flow rate through the head. To do this it applies a pressure differential of usually 28 inches of water (about 0,07 atmospheres) across the head and measures the pressure differerential across a calibrated orifice plate set in series with the head. A calibration table for the orifice plate then allows a flow rate to be obtained for the measured pressure differential. This is done for each individual intake duct with the intake valve in the open position. The values from each intake duct are then added together to give a flow rate for the head.

 

[333pg333]

Thanks George and Mark but I have to say that I think those numbers must be out of whack still. Based on a properly measured flow bench my custom head (2.5 turbo) flows just under 250cfms. My understanding is that the stock one flows somewhere between 188 and 200cfms.

 

[DivineE]

^^^ That easy hey[8D]!!

But very interesting to know. How do you calculate what the cylinders actually require?? So you can work out how much of a restriction the head is to the engines potential? Also when arguing the 8v vs 16v argument I've often heard that 8v engines will always produce more power at lower revs, something to do with the pattern in which the air and fuel enters the cylinder (I think I read it myself in a Dave Knight book but it was a long time ago). Therefore on a 944 engine which I don't think will ever be a high revving engine because its a 2500-3000cc 4cylinder with a Titanic size crank and a massive stroke surely a well sorted 8v head would outperform a 16v even if the peak flow figures were not quite as good?

...Just thinking out loud by the way

Regards,

Ben

 

[DivineE]

ORIGINAL: 333pg333

Based on a properly measured flow bench my custom head (2.5 turbo) flows just under 250cfms. My understanding is that the stock one flows somewhere between 188 and 200cfms.

Surely that figure must be per cylinder? A standard toyota supra mk2 head flows 1000cfm (just a random figure I remember)

That would be 166.6cfm per cylinder which sounds about right as the bore is much smaller than a 944.

 

[Neil Haughey]

ORIGINAL: DivineE

^^^ That easy hey[8D]!!

But very interesting to know. How do you calculate what the cylinders actually require?? Regards,

Ben

That is the easy part since its just the swept volume x number of induction strokes per second per cylinder if one is considering a single cylinder. An engine that breathes exactly that much air has a volumetric efficiency of 100%. Problem is that (as I have been informed no idea if true) an F1 engine will breath more then that at its most efficient rpm's and load, perhaps as much as 120%, whereas a road car engine will be down around 80% perhaps reaching 95+% at the point where the pulse charging from the inlet manifold works best (should be tuned to be the point of max torque so 4K ish rpm's in an S2).

What this all means of course is that on a stock road car engine it is possible that a turbo/sc bolted on reading 0 psi boost could be making much more power then without the turbo/sc applied.