Guess what my wife did for a complementary T thing, Thats her putting one on a suburban in that infamous T ad. By the way thats ALL the interest we have in the company, and we sold the suburban. But it did work, and dyno testing showed an increase in power!

And there are technical, proveable reasons why a fluid spinning will be more effecient in flowing thru certain openings. It also helps mix up the gas with the air better.

As the T thing I didn't beleive it but I have personally seen the before and after results of a couple of dozen cars, on a chassis dyno. It really does add 3% to 8% more power. And I have seen it after it was removed and saw the power drop back down.

Whats even more interesting is I have seen it work on Fuel injected cars. Put in just after the flow meter, it still works, not for all but most it does do some good.

Theoretical power not achieved
 

The flow rates and even the crude guesstimates would suggest the heads are not the limiting factor on the 911 motor's output.

What is it?

I wonder if we have a major gain if we can reduce the air pumping losses in the case, or if the revs need to be much higher?

I'd be very interested in the views of the very expert bunch round here please..

david

Snowman, does the dyno program want the smallest cross-section in the port?
There are formulas to predict the maximum airflow the port will allow based on the smallest cross-section. Too small a cross-section and the engine will act as if it has a small carb. The HP peak will be earlier in the RPM range. If you build a few engines, record this data and compare the dyno sheets.
A mistake I have seen time and again is assuming a smaller port will have higher port velocity. You can map a port with the velocity probe attached to that SF-600 at the college and prove it either way.

I have a Superflow SF-600 in my garage at home if anyone needs any flowtesting done. Since 1992 I have done about 1000 tests and I still cannot guestimate airflow numbers by just looking at a port.
The clay radius is the most common method of testing the intake port. In a small percentage of cases I believe it has increased airflow. The only way to do it right is bolt up the entire intake system and then test.
Changing the backside radius of the valve can have a dramatic effect on airflow. The valve seat area "venturi" effect
is another place which can be improved and verified on a flowbench.

I see my fair share of flow sheets, dyno sheets etc. One of the best Porsche head porters is Michael Stimson. He has been porting Porsche heads for 20 years or more. He is one of the "secret" suppliers many racing engine builder have. If anyone is looking to have heads done you can reach him at 650-361-8636

Sorry for getting back to you so late.

The dyno program uses the actual MEASURED head flows. So it dosen't matter whats in there, its the actual flow, consequently very very accurate. ITs also important to include every thing in the intake system when measuring flow. On my 912 race engine I measured the flow with EVERYTHING connected. thats the head, cylinder, piston, intake manifold, carbs, air horn, headers, EVERYTHING! I was able to get almost the same head flow with or without everything connected, thats the way it should be. I also measured the entire intake volume, including air horns (closed intake valve and poured water into it , until it was at the top. That volume is 110% the displacement of that cylinder. exactly what it should be. The length also corresponded to the HP peak at 7800 rpm.

Its 9-13-04 and I can now do flow measurements again.

Re: Theoretical power not achieved
 

David I think this is an important point, one will find that despite our egos the engineers have done a good job with the design of this type of motor, and it is difficult to "significantly" improve upon it. If we were to make such headway it would take some effort, such as in one of the previous entrys Swirl was mentioned and to promote that (flow all the way around the valve at low and high lifts) one would have to offset the port from cyl. center line, which would be most difficult with a grinder.
Also the current cams which are limited by the journal dimensions,rocker arm ratios etc. and the deccel rates inherent in the Porsche design are not signifficantly different from profile to profile after you exit the street profiles.
I think the pursuit of improved air flow is always important and any increase in CFM without hogging out the ports while maintaining a resonable velocity, such as a slight increase in valve size without going overboard on your modified port dia. and dealing with improved combustion through a slight compression increase can provide a reasonable percentage of h.p. gain

Best regards

My guess as to the major limitatin in HP in a 911 Porsche enging is RPM. The max RPM is limited by the stroke and bore, journal diameters. Given these are basic design parameters, one cannot do much to change them. Unfortunately they limit a Porsche engine to about 8500 RPM or so. Given a specific displacement, the ONLY way to a lot more HP is a lot more RPM. You can't make a Porsche engine go to 12,000 Plus RPMs, therefore you can't make a whole lot more HP with one, except by going the blower/turbo route. Then you can do some incredible things with them, but not as much as a good V8 design can do.

Have anyone of you all seen the Glenn Curtis Museum in NY?

He designed, built a flat, opposed 4 cylinder.air cooled, engine way before the Germans did, its in the museum. Looks just like an early VW engine, but made for an airplane back in the 20's or so.

I have seen a 3.3L dyno sheet with peak HP at 9200 rpms. I have seen a few 2.0 litres as high as 9500!
The advantage other engines have is net valve lift.
To get near .560" net lift in the 911 using factory rockers is tough. The profile uses the entire pad of the rocker arm. The base circle has to be small enough to keep the nose of the lobe under the bearing journal diameter.
The port size in most 911 engines could be smaller without a loss in HP. HP by modifications to the ports... I have seen some high flow numbers with welded/ported 3.6 heads. Up to 20% over stock heads.
"Swirl " is usually bad for HP. Its normally done to promote better burning in the combustion chamber. A high swirl port will sometimes stall sooner than a nicely designed port without swirl.
Of course the port should flow around the entire valve, short turn side and long side etc. But most the air flow is "line of sight" and flows accross the backside of the valve into the cylinder. At high airflow speed the air will not want to turn down the short side. By shaping the floor correctly (usually by welding)and helping the air around the short turn the air flow will increase.

Snowman, Thats good about your intake system. Sounds like its working properly. Usually when adding length to an intake system the extra mass of air will slow the port down. If the port runners in your intake are shaped correctly and have a taper sometimes the airflow can even increase.

Great thread, Guys.

I still think 100 bhp/l low ( OK 110) for a 2L six full race motor at around 8k rpm.

Especially with the valve sizes we have.

My 1700 Ford Kent 4 cylinder (restricted formula) race motor with 38mm chokes in the 48 Webers and 40.2mm inlets lifting to MUCH less than 0.450 makes 180 at 7,600.. and so do lots of others... claim more!

I do take Camgrinders point about lift and geometry.. this restriction is however similar to quite a few DOHC and SOHC engines which make more specific BHP..

I suspect combustion is poor in the hemi-head.. especially single plug..and driving close behind any race 911 in an open car will convince you that plenty fuel is unburnt..

One of the thinks I'd like to know about is swirl.. and yes, this may in some cases reduce cylinder filling, but since Grady Clay has actually done the tests and failed to confirm the claimed advantage of Boat-tailing.. we are not left with much to explain the deficit.

Has anyone knowledge of swirl-inducing experiments?

Kind regards
David

In the evolution of a street engine to a racing engine we face several compromises when budgets are limited. Which is precisely where we all are. Johns staement in regards to cams and the swirl issue are I believe correct, If one were to investigate swirl for example he could not be confined to a half hearted attempt and would need to signifficantly offset the port, this could be costly and possibly if it had any benefit at all the benefit could be out of the competition zone of 4,ooo to 8,000 rpm and there goes our bragging rites to improved h.p. The thermal efficency on a good running Porsche is 28% to possibly 30% although I seriously doubt the latter. This is not good and the surface to volume ratio of the Hemi Head is the culprit a Trans Am Chevy has the same bore stroke
as a 3.8 air cooled RSR albeit with two more cyls and is in the 32% to 34% with the modified Wedge. This situation has been rectified by the current generation of Water Wagons Porsche has produced with a much reduced combustion cahmber surface ratio and a great utilization of the fuel burn. If we are to emulate any racing power plants we would probably need to find a situation simular to our own, I am not sure but I think Honda had a air cooled Hemi type chamber 15 years ago in some type of competition environment. It could just lead to tunning of lenghts
but maybe something can spark the tuners on this forum to a better mousetrap.

Best regards

David; I think that you have to be careful in believing that 911 engines have left something on the table and are "underpowered". I'm not arguing against trying though. But as I believe Helmut Bott once pointed out; "Lots of other manufacturers claim to make more HP then Porsches, but we we manage to win more races". You can draw your own conclusions.

As far as your comparisons to a Kent (known in the US as a Capri 1600 or FF engine), I'd be interested in understanding the basic specs of your motor. Bore, stroke and peak torque engine speed of your motor.

I've been collecting published data for a larger number of engines (over 200) ranging from M-B motor used in the WWI Albatros (9.4 liters, 102 HP at 1288 RPM, BMEP at peak HP of = 108 PSI) and Zeppelin (87 liters, 1200 HP at 1600 RPM, BMEP = 110 PSI) up through the Nissan Supertouring engine (4 valves BTW, 1.96 liters, 315 HP at 8250, 203 lb-ft at 7500 RPM, BMEP = 248 PSI!!!). In the global scheme of things, Porsche's air-cooled 2 valve engine doesn't do too badly against other non-inline 2 valve engines. The RSR (2.8 liters, 300 HP at 8000 RPM, 217 lb-ft at 6500, BMEP of 173) and the 917 4.5 ( 4.5 liters, 580 HP at 8500 RPM, 376 lb-ft at 6800 RPM, BMEP of 197) look pretty good, but are outpaced by some more recent engines such as a CAMS Ford V8 (4.86 liters, 580 HP at 7000 RPM, BMEP of 218).

Yes, the advent of the successful 4 valve head a la Cosworth's engine's have to some degree overshadowed the performance the air-cooled two valve. This is why the 3 liter 908 engine (3 liters, 350 HP at 8400 RPM, 235 lb-ft at 6600 RPM, BMEP of 180) never would have made a successful F1 engine as the contemporary DFV (3 liters, 430 HP at 10,000 RPM by the early 70's, BMEP = 187).

Now in all of this, the stock 2.0S made 170 HP out of 2 liters at 6700 RPM, 134 lb-ft of torque, and a peak HP BMEP of 165. Ford's Kent 1600 put out 88 HP at 5500 and 97 lb-ft at 3200 for a BMEP of 148, out performing only the afforementioned WW1 technology and the stock Triumph TR8 (!). The stock spec Cosworth BDA (with a Kent bottom end and a magical Cosworth head) was putting out 120 HP at 6500 RPM, 110 lb-ft at 4500 RPM and a BMEP of 141). An '83 Loyning FF spec of the Kent 1600 engine (114 HP at 6250, 105 lb-ft at 4750, BMEP of 148) was starting to get into 911S territory when it comes to performance.

Even a carb'd F2 spec BDA (215 HP from 1.6 liters at 9200 RPM, 110 lb-ft at 6000 RPM, BMEP = 178) is just about equaling a carb'd 906 motor (210 HP at about 8000 RPM). The earlier FVA (again with a Kent block)was more efficient, but generated less HP then BDA with 200 HP at 9000 RPM and a BMEP of 180. So the Porsche engines are no sloutches, it's just that the 4-valve heads of other race engines allow them to spin faster thus generating more HP. But in regards to getting the most power per liter per rev, 911's seem to be right up there!

Hello again, all.

While researching the web I found this site:

http://www.pumaracing.co.uk/pp03.htm.

The numbers make interesting debating points I think.

The engines which I am most familiar with, the 2 and 2.2L, have "predicted" maxima of 239 BHP for the smallest available inlet valve..the 39mm!

By the way, the Ford Kent 1700 race motors have an 83.5mm bore, but we are limited to 40.3mm valves...the formula predicts only about 160 BHP with the inline valve configuration...

Kind regards
David

David;
Yeah, I'm familiar with that site and have actually been keeping track of the "Predicted HP" versus the reported for all of those 200 plus engines. It's an OK rule of thumb, but the distribution is pretty wide and so I wouldn't use those formulae for much more then discussion or a really rough order of magnitude check.

There are a number of engines that I've seen that seem to beat the formula you referenced. For example a Vauxhall Supertouring 2.0 4 cylinder by rights should make about 266 HP, but in fact it's pretty clearly documented that they were putting out at least 285 HP.

The latest Ford Fiesta Super 1600 Rally engine should be putting out about 209 HP when in fact it's really putting out about 220 HP. (Wouldn't you like to have one of those for your race motor?!)

Earlier you had pointed out that the 911's hemi combustion chamber (and the compromises that it entails) most likely limit it's performance potential. You should talk considering the Kent's "Huron" head configuration with the combustion chamber in the piston bowl. The result is small valves and a poor heat path from the combustion chamber. cc for cc, the WORST performing Porsche engine was the 924's which had.... (wait for it!)...

Huron combustion chambers just like the Kent's.

The best numbers seen for the un-turbo'd 924 2.0 engine was 180 HP, maybe wheezing it's way to 190 HP with slide valves and MFI. Still, that's not bad considering the aforementioned formula predicted all of 168 HP based on the valve size. Getting 180 HP from a 2.0 911 motor is a piece of cake by comparison.

BTW, the numbers that you listed for your 1700 Kent are pretty good. In round figures it works out to a peak HP BMEP of 181 which puts you in such company as the following:
* Cosworth FVA
* Ferrari 312 PB flat 12 (early spec)
* Toyota 2000 GT (race spec)

and just behind...
* '59 Auston Martin DB3S
* Corvette LS1
* Ferrari 360 GT (in FIA/ACO trim)
* A late 70's Datsun 280Z C-Production spec.
* Dennis shaw's SCCA GT5 Spec Mazda GLC (180 HP from 1.3 liters!)
* SB Opel/Vauxhall 1600 (less HP then yours with 164, but at a lower 7250 engine speed)

The best performing engine in this regards that I've found to date was the Nissan Supertouring engine which was putting out 315 HP at 8250 for a BMEP of 248. But I guess this isn't surprising given the effort that went into tuning those engines for maximum HP out of 2 liters and 8500 RPM rev limit.

Thought I'd ask this mostly on topic question.

Since there seems to be a few people familiar with porting, and the 911 port characteristics...

The question is:

'82SC engine ( 34mm port) bone stock with SSI and new full engine rebuild.

I'm converting to PMO carbs. The question is should I stay with the stock port config and go with 40mm PMO or should I go and have the intake ports enlarged and run 46mm PMO?

The car is a street car that may occasionally be taken on a large high speed track. But it's by far a street car.

I want the peak HP at 6000 rpms. Let the TQ fall where it may.

Will the 34mm port do the job? Will the 34mm port cost me a lot of power and if so, what are the estimates on power differences at 6000 rpms between 34mm intake ports w/40mm PMO and 39mm intake port and 46mm PMO (remember.. stock cams.. stock pistons, stock engine)

I like the nice midrange power of the car. I want more midrange and more top end (don't we all). I don't need to spin the motor past 6500 rpms.

Recommendations? (ps.. I'm placing the order for the PMO's tomorrow).

Thanks,

TonyG

Stay with stock!!! The larger ports and intakes are ONLY for very high rpms, ie more than 8000. The 34 will do fine, and will help on the low end torque, ie mid range of 4500 rpm or so. Peak hp will still be above 6000 rpm. The stock Porsche sizes are best for ANY street use, maybe a little above normal street use.

The 46mm carbs will be awful with stock cams.

Agreed.

Christian,

What are you doing about valve head cooling with that radiused seat? That's the problem I've always had with the fancy-schancy valve angle cuts... Less contact, less cooling... or am I looking at the wrong side?

PS stay with 40mm PMO carbs.

Just as an update, (I'm not trying to pick any fights or insite any arguments), here is an updated head flow graph. This is not intended to represent a "best case" nor provide any secrets of porting. It's not even a representative sample. What it is is just some different heads picked at random which Jack (Snowman) was kind enough to flow.
As such it's informative in so much as it suggest the different charactoristics of the different head/valve sizes for the 60's and 70's 911 heads.

The standard disclaimers apply: your results may very, differences in flow benches, techniques, etc. etc. etc.

http://forums.pelicanparts.com/uploa...1102424328.jpg

Some notes and comments:
1) The numbers in the table which are offset to the right are just plug numbers that I added to smooth out the lines between Jack's datapoints.

2) It's not clear if the big drop in flow for the '66 head at low lifts is due to the way that Jack tested it (he had to use a VW cylinder since he didn't have a 911 cylinder) or because of the significantly smaller valve size of the '66 911 head.

3) A kind of "duh!" observation -- flow at maximum lift seems to be determined by port size combined with valve size while flow at lower lifts appears to be be more associated with valve size. This seems intuitive.

4) These were all used heads prior to valve jobs and such.

Enjoy!

John-
Thanks for posting the flow chart. Does this information match your predictions concerning flow and gas speed for 32mm vs. 36mm ports?
-Scott