100hp/L - How to questions?
 

I am currently doing a top-end refresh on my 3.0 thanks to broken head studs.

I was playing around with Desktop Dyno the other night trying to estimate my newfound output thanks to my "while-your-in-there" syndrome. While the numbers are relative they are a healthy bump from stock internals. Then I started thinking...

What does it take to make a NA 911 motor hit a target output of 100HP/L? I know there are many builder secrets to achieving this output goal. I would love to have details. I know the whole package must work together. I think 300NA HP out of my 3.0 would be great.

A few questions for the discussion:

How high rpm is required? Can you build the motor to limit at 7500 or do we need 8500? What is it about high rpm besides charge momentum that increases VE to the point of making 100 HP/L?

Is Intake velocity or air flow volume more important? Intake velocity helps VE due to charge density. You must be able to support the volume of air based upon displacement

Lift or duration more important on cam design? Lift only gets you to the point of matching the intake volume flow. Duration must take over after lift is limiting factor.

How much overlap is really required for savaging effects that help charge density?

Tuned intakes? We know equal length runners are important. ITB setups work better than common plenum. Steve W. has mentioned ITBs under a common plenum (3.6) intake manifold.

Tuned exhaust? (How much do headers help compared to SSIs?) What about individual tracks compared with cross-over designs. I think it is understood that a cross-over tube helps with low speed torque curves by providing balance to the exhaust pulse. What about high rpm.

I am hoping to open a discussion that we can all learn from.

The magic number is usually 8000 RPM with cams that can breathe big enough. It usually requires a lot of overlap to get that much air, but on a 3 liter, the displacement tends to overcome the loss of low end torque from big overlap. Look at 906 cams as a baseline, although there are more modern cams with more lift/less overlap that can do the same thing. I think a properly built 3 liter on possbily GE80s or GE100s built to rev to 8K would do it. I'd like one of Henry's 9 bolt 66mm cranks to build a short stroke motor on a 3 liter case. His short stroke 2.8s make quite a bit over 100HP/L on RSR sprint cams.

I can say with 100% confidence that this is not the case. I have driven a 2.8 short stroke. Smooth and powerful, all the way through the powerband.

We recently built a 3.0 Twin plug, 10.5: 1, DC80 w/104 lobe center MFI engine for a european rally car that achieved a reported 310 DIN @ 7260 rpm. This dyno number was achieved by a tuner that had no vested interest in the build. Stated another way, he had no reason to fabricate this number.
The engine produced a flat torque curve from 3800 to 6400 268ft/lb @5800 . We ran a 1 5/8 conventional header and 36 mm Venti port. That's correct a 36 mm intake port. The valves were 50mm in and 42.5 mm ex.
38 mm exhaustt.

http://forums.pelicanparts.com/uploa...1218590580.jpg
http://forums.pelicanparts.com/uploa...1218590601.jpg
http://forums.pelicanparts.com/uploa...1218590617.jpg

Nice dizzy ;-)

A question for Henry... can a 3.0L be made to perform, as you state like this example, with (what would be the set up be ?? ) the least expensive induction & ignition set up and not have to use the expensive $$ MFI set up ?? Can this be done with carbs or a modified cis or ?? what is the least expensive induction & ignition set up to allow this hp/torque development ?? What are your thoughts on the Bitz Racing ignition set up just reviewed in the recent issue of Excellence magazine ??
Thanks, Bob

Jamie, back to your questions about specific output, here is some historical data from Bruce Anderson's book about the engine types over the years. This is a good starting point for the discussion about power-to-weight-- the next step is to look at the cams, intake ports, valve sizes and other modifications necessary to achieve these numbers.

http://forums.pelicanparts.com/uploa...1218636882.jpg

Henry, Thank you for posting the specs on the engine. Again, nice work as usual. John Thank you for posting the list. Let's see what we ideas we can pull from the information presented so far. Let's start with the port size and resulting velocity.

Henry posted the specs of an engine that achieves 103 hp/l. What peaks my interest is the intake dimensions. He describes smaller SC ports (36mm) with larger valves compared to stock. My calculations come up with an intake velocity of 119.3 m/s at 7300 rpm. This is a function of displacement pumping air through the intake port. I am not sure how a "Venti port" changes things or what machine work has been done to the heads. I recall a previous description of proprietary head work to achieve this number. I respect that but would really love to know how the valve pocket changes. It makes me wonder if Henry has been reading the MototuneUSA site. :)

John Luetjen posted years ago that factory 911 engines achieve optimum output with an intake gas speed of approximately 100 m/s at the peak HP. Henry's engine exceeds this value but a good margin (20%) and yet still makes good power.

So far, the key is to have a cam that has peak power at a high rpm and yet keeps the intake velocity sufficiently high. The question is: what is the rpm limit? Kenik thinks this is 8k. Henry showed it can be done 740 rpm less.

Now if we look at the cam selection we see that the DC80 cam has a lobe center of 104. Usually I see a lobe center of 98 or 100 deg for this cam. Increasing the lobe center has the effect of reducing overlap. This results in increased low-speed torque but less high-speed scavaging. I do not think this is unusually wide but an effort to lower the expected peak HP. I think this is the case since the DC80 cam would have an expected peak HP of around 7900 rpm based upon exhaust duration. (I used John's formula for calculation). This raises the question if the peak HP is lowered due to the high intake velocity and the inability to support air flow at 7900 rpm.

I think I have port diameters for some of the listed engines. Time to look at port velocity at the rated peak HP rpm values.

Moving the lobe centers from 100 to 104 will widen the usable powerband, but lose a little peak torque. Between the two, I would expect more top end from the 104 lobe center cam vs. the 100. But a slight loss in peak torque. Also, I would imagine the power curve on Henry's engine didn't fall off like a rock after the peak.
On Henry's engine I am amazed it comes on the cam at only 3800 rpm. The heads must have very good velocity numbers.


Here is a straight forward build that makes 95? hp/litre.. at the crankshaft anyways.

http://forums.pelicanparts.com/porsche-911-technical-forum/246563-another-rsr-clone-project-dyno-day.html

A little more compression, and more camshaft and maybe some head work and He would be real close to 100/litre.

This is either Sherwood's or Thom's but I think it's originally Bruce's (actually, it's originally the Factory's :))

http://forums.pelicanparts.com/uploa...1218667469.jpg

i was actually epxecting closer to 100hp/L out of my engine... but i guess cam choice (not done by myself) really limited that...

the agruement my engine builder held was engine life... how long do our f-6's live @ 8000+rpm?
the other point was stress on the case, as my unit is holding onto 100mm (3.5L from 3.3L, 100mm bore with 74mm stroke) barrels, is there potential for case sagging even with good studs etc...

given that this unit is using 46mm webers and nice big headers is it simply a nice cam away from higher hp?

edit: using a mod-s cam at the moment, made 278 at the rear wheels. I am willing to loose some mid-range for more up top gustor, especially for a few more revs... jut got a civic Si with the k20a and reving to 8200 rpm is run in itself!

Facey,

You are Cam limited and Carb limited to get 100HP/L on that big motor.

Bob

Jamie,

I would expect to see 100 hp/liter starting around 7300-7500 rpm provided that you are able to achieve 100% volumetric efficiency at the given rpm. 110 hp/liter @ 7800-8000 rpm.

Some of the controlling factors are;

camshaft selection
port size
rod ratio

Henry's engine has a minimum port size of 36mm. The length of this area is relatively short. My guess is that there is enough velocity to help lower rpm cylinder filling yet does not impede maximum flow. the Dyno result was very surprising considering we were shooting for 285 Hp.

how do u calculate rod ratio?

rod length/stroke

130/66=1.969 good high rpm stability/velocity.
127.8/70.4=1.815 good solid street ratio. higher piston velocity sooner in the rpm range
127/74.4=1.706 bottom end of good street ratio. Even higher velocities. larger ports recommended.
127/76.4=1.662 WTF....requires big ports and valves. high side loads. Stuff you see on truck motors.

i am assuming u had to have measured the rod length before u assembled the engine...or is it evident in the stroke? as in all 70.4mm strokes have a 127.8mm rod length?

of course it could be changed, but is it typically?

The numbers I gave are what Porsche used on all of their motors up untill the GT3 cup.
So a 2.7 and a 3.0 have the same rod length.
It depends on budget and application. Engine builders may not choose to do it to keep component prices and turnaround at a minimum.
Engineers at Porsche increased the rod length for their GT3 cup engine to 130mm in an attempt to reduce sideloads of their engines at 9300 rpm.

130/76.4=1.701

Their sprint rods are even longer but I do not know that dimension.

To achieve this, Porsche had to do away with the second compression ring. They were truely at the limit of their current engine design for the stroke they were useing.

I am interested in what they have done on the future engine design.

so in my engine using 100mm Mahle P&C's w/ pauter rods it is likely that i have the 127mm rod length? given that i believe non-stock rods were an option i'd guess that is the case, given that i am using a 74.4mm stroke.... what would be your guess at the rev limit of that setup? I mean for safe street use, but it does see the odd (maybe not so odd) rev up!


this is neat to finally know why the 66mm stroke is so desireable...i've heard about the importance of rod/stroke ratio but never had it explained... next build = 66mm stroke with 102mm bore...something like 3.24L? should be able to rev way way up! (with the right cam , v. spring, retainers..)...

Thanks to work travel and other things in life that get in the way I have not had a chance to repost on the subject. I did have some time to think about what others have said.

Let's start with a collection of data. The following is compiled from a listing of cars and their outputs by John Luetjen. The list has many makes and models but I have limited it to the 911s. I have also added the information for Henry's 3.0. There is more to the data and the spreadsheet calculates many engine parameters that might be of interest but fall outside this discussion.

http://forums.pelicanparts.com/uploa...1219425749.jpg

Aaron mentioned that the specific output would be rpm dependent and as such a range of 7300- 7500 rpm would generate the majic 100 hp/l. So I calculated the specific output and graphed it vs the rpm of peak power. I fit a linear curve for reference. Don't ask about fit statistics... They are not great in such scattered data. It is interesting that the 100 hp/l output falls right at 7500 rpm. This would lead us to believe that a cam that provides peak HP at about 7500 rpm would be a good choice.


http://forums.pelicanparts.com/uploa...1219425760.jpg

As I looked through the data I found something interesting. There were a few engines that had relatively high outputs compared to the fit line. When I find the individual points I notice that the intake port sizes are slightly smaller than what the factory did. This is the 2.0 and 2.2 E Production specifications using a GE60 cam. Both of these engines produce peak HP right at 100 hp/l and 7000 rpm. Of course, they have a more favorable rod ratio but I think I might have found the ideal combination.

The smaller intake port size will increase the intake port velocity at peak power. So I scaled the rpm value by a ratio of the intake port to intake valve size and came up with a new dependent parameter.

http://forums.pelicanparts.com/uploa...1219425771.jpg

What started this discussion is my building of a new 3.0. It will have a Mod_s-108 cam, EFI, 9.5:1 JEs and 78 SC heads. Using desktop dyno I get a Peak HP value of 260 at 6900 rpm. This corresponds to a specific output of 86.6 hp/l. The strange thing is that if I use the ratio of port size to valve and then scale my engine fits right on the curve with a calculated VE of 87 resulting in 261 HP. I guess large data sets create the general formulas for baseline values.

If I was to reduce the port size I might expect more power according to my graph. For example, using a 36mm intake port would give me almost exactly 100 hp/L. If only I had the funds to build another set of heads and a second engine to find out... If only I knew the build secrets that keep me on the upper size of the curve.

YMMV.

Get yourself engine analyzer pro ($$) or dynomation ($$$) and program in a stock engine or something you know the exact figures for. I prefer Engine Analyzer Pro because you can get every last nitty gritty detail in there, which is also it's downfall as if you don't, it will skew the results (usually on the low side at least). I've designed a few engines in EAP and had the results be within +/- 1.5%. EAP is especially useful for choosing a cam or choosing a compression ratio based off of knock index & dynamic C.R.

"Here is a straight forward build that makes 95? hp/litre.. at the crankshaft anyways."

http://forums.pelicanparts.com/showthread.php?t=246563

The motor in my RSR is close to 100 hp/lt and it is pure joy to drive on both the street and on the track.

John D made me a set of DC-60 cams and the motor is very tractable from 1000-4000 RPM and just screams from 4000-7500 RPM.

With twin plug, RSR heads and 10.5 CR, it was not inexpensive to build but it is well engineered and great fun.

advance timing?
 

Tom,

I had read your enginedyno post previously. Congrats on a great build. I have a 3.2 of similiar specs. I am juststarting to drive it and am sorting it out and have not had it on a dyno yet. I have read several posts about twin plugged motors and the correct timing advance. Are you runing 25 - 26 degrees which seemns to be the common wisdom, or have you found something else that works for you. thanks tom

I should be at the track this evening but I have to pipe in. I have had five motors in my lifetime that have made 100HP/L. One built by R. Ginther-rebuilt by Andial. One built by Woods-2.9L. Three built by me-3.0L, 3.4L, and last 3.6L. Three of these were dynoed on Woods dyno to allow good comparison and all had Carbs.
All of these motors were expensive and tuned on a very conservative dyno.

These numbers I see quoted and the Cams used are very questionable. Trying to get Torqued spread over a large RPM range usually means that Peak Power is also reduced over the upper range. A peaky cam will make power over a narrow range and get boasting rights, but won't win many events. But now I see boasts of wide torque and high HP with low Compression. Yes 10.3 to 1 is low compression trying to make competitive power unless all you are doing is running against other Porsches with the same parts.

Guys are being mislead to think they can build 100HP/L with standard heads, small ports, not very aggressive cams, and inexpensive add-ons and very liberal dyno readings. The Cam makers are being too kind to blast some of you on your postings of unbelievable claims of wideband HP on small lift cams.

Bob

Hello all.

This thread is revealing quite a few bits of hard-won tuning tricks!..typical of Henry's generosity..

A litle math shows us the Factory ports are generally bigger than needed for peak power, even at 8000 rpm.

For example, the gas speed of the 2L FIA race engines is only about 390ft/sec in a 30mm port.
Plenty of these engines make well over 100bhp/L...

Put another way, gas speed at intake charge temperatures becomes a blocking factor on flow at about 690 ft/sec..at 8000 rpm a 2L motor will pull enough air to just reach this speed thro' a 27.7mm hole..

Note I am neglecting boundary layers, and obstructions in the port here..

And we know thwe Factory used much bigger ports, and this was surely the product of some dyno work..

But its pretty clear we are wise to look elsewhere than simply bigger ports to make more power...and torque spread, gas mileage, and mixture quality all will surely benefit from keeping up gas speeds.


kind regards
David

Small, high flow ports seem to be a key component in our unexpected results.

Reread my post "36mm intake port" on a 3.0. 3.0 factory RSR intake ports were 43mm. yikes!!!!!

sub'd

36mm ports are early 911S spec?

I'd venture to suggest that port size is overrated. Who cares what the aperture size is?

Every infinitely small increment behind the intake port is an aperture when you cross section a head. Performance is about port design which maximizes velocity as a function of air volume, not a static diameter reading. The below heads measure at any given diameter on the intake port between 39mm and 40mm. The air flow was modeled to my specific displacement (2.3L, 85mm x 66mm), camshaft (DC44 on 102 lobe centers) and top engine speed (8000 RPM). I provide this example as a counterpoint to the above discussion on "small" ports: it ain't the ports as much as what is behind them (between the port and the valve).

Credit goes to Herr Weiner for the work below.

http://i25.photobucket.com/albums/c7...n/IMG_0487.jpg
http://i25.photobucket.com/albums/c7...n/IMG_0493.jpg
http://i25.photobucket.com/albums/c7...n/IMG_0502.jpg
http://i25.photobucket.com/albums/c7...n/IMG_0497.jpg
http://i25.photobucket.com/albums/c7...n/IMG_0498.jpg
http://i25.photobucket.com/albums/c7...n/IMG_0499.jpg
http://i25.photobucket.com/albums/c7...n/IMG_0500.jpg

Here is another good thread on the subject.

http://forums.pelicanparts.com/911-engine-rebuilding-forum/343951-race-motors-steve-henry-lot.html?highlight=excel

Amen, to that!! :)

Its all about throat shapes and volumes between the port openings and the valve seats,.... :) :)

Beautiful head work Steve. Given these are heads destined for a track engine, would you care to comment on the +/– issues of abbreviating the length of the valve guides on a 911?

Thanks,
Sherwood

Herr Weiner, Could you elaborate on throat shape and volume? Obviously from the pictures of Kenik's heads a streamlined shape around the valve guide boss is important for flow. How does intake port volume help? I assume it is related to engine and intake pulse frequency tuning. This would be why you want to maintain a constant diameter throughout the section.

IN addition, how much does the loss in length of the valve guide affect service lifetime?

I am learning a lot.

Sherwood, those heads are for a street motor. It may never see a track...ever. The cams and heads mean it will pull harder than a stock 'S' low and as hard a 906 up high.

Thanks, it's about time someone sheds some true light on the Head/port size myth. This is why a good Race Engine isn't cheap because there usually is a good head guy waiting in front of a good Cam Grinder! Someone spent hours on those ports.

Novak, your Desktop dyno made lead you astray. I have 60hours of work with that software but I also have three engines to compare the results against real dyno data on Wood's dyno. Someone before told you that you have to do baseline work on a known engine. You are going to spend more money experimenting than paying a good Race Engine builder to build you a motor. The real Builders are not sharing their secrets just sitting back laughing at the bull of
Thermodynamics 101 and Fluid Dynamics 202. Past Racers running 3.0L engines trying to make 100HP/L were not using RSR Cams or those who were making power were not using them! If they were that good, Elgin, Woods, Schrick would not have been needed.

Bob

Sherwood,

Shortening the guides a little (as pictured) has not proven to reduce their lifespan. As long as one uses a good guide material, installs it properly, and hone it to fit, guide life is really excellent.


Jamie,

Without dilvulging specific data, I can tell you that such heads represent many thousands of hours on a flow bench and engine dyno with the objective to maintain sufficient velocity at all valve lifts and volume for improved HP and torque. One uses special pitot tubes to measure boundary layer airflows at many points in the port and around the guide boss as part of the process. You get to ruin a LOT of heads during the course of development, too. Patience is a big virtue in this endeavor,....:) :)

Those heads are equally at home in a race engine as well as a high-performance street one,...:)

Mister Barnes speak volumes of wisdom; making a real 100+HP/litre is an involved process involving custom cams, pistons, headwork/valves, induction, exhaust system configuration/design, windage and extensive oiling modifications so that the whole thing may live. Years ago, we had a 3.0 RSR motor making over 400 HP but it was a 24 hour engine and FAR FAR from stock RSR spec,... :)

hows does honda achieve 100hp/l at their price point?
just new design and stuff like vtec?

First, scale - they make a LOT of cars. Scale allows them to drop Titanium rods cheaply into an S2K, after all.

Then there is technology; it has moved on. To name a few things:

1. Water cooling; easier to make power on a cooler running motor
2. VTEC - variable valve timing gets around SO many of the power generation issues in older cars
3. Multi-valve heads - It is easier to spin high revs on a car with two lighter valves on the intake and exahust side versus the monolithic single valves on old Porsches (and the 4 valve heads are much more modern and volumetrically efficient)

Then there are things like bearing technology, modern casting and forging techniques, direct fuel injection, etc...things that cannot be economically back engineered into a 50 year old engine design.

I could go on, but you get the picture.

Are those valves 8 or 9mm?

6mm even!

Perhaps, Not 9 for sure, there is a lot of meat in those guides.

Which valves? Mine? Stock! Perhaps not for long, depending on some news. We'll see.