and I thought my degree wheel was the largest available.. Cool stuff!! Love the flow bench!!

 

Colin,
Thanks for that.I think that with the degree of compressability that we are considering here, my observation that flow equals volume divided by area is accurate enough.
I am afraid I have seen too many times, flow bench wonder heads failing to deliver on a motor. As we all know flow results do not necessarily equate to bhp results ! I view flow testing as an interesting piece of data but no more.

I am puzzled by your observation that stacks and throttle bodies give no benifit on a pretty standard motor.I have seen 911 motors on engine dynos that would disprove that theory , certainly when using correctly mapped Motronic.
As a simple example try putting a combined manifold and single throttle on an old 2.7RS and see what happens !! On second thoughts dont !!!!!!!!!!!!!!!

I am sure that the 9M heads work well and I look foward to seeing how Geoffrey gets on.

All the best

Geoff

 

I think one of of the things that is missing here is that an engine is a dynamic operating environment. What "optimizes" lower RPM will come at the expense of the upper RPM range and vice versa. So, there are a lot of trade offs that are made when designing and building an engine.

Geoff, I agree with you that I have seen heads that flow large numbers on flow benches but don't work well once installed in the engine. I think it takes years to learn how to evaluate ports and port flow numbers, and I certainly am not an expert and I wish I had more experience with it. I measure port flow as well as velocity at several points in the port. This along with visually inspecting the ports and measuring the port volume all help to evaluate what is happening within the port. This data helps a cam grinder optimize the airflow through the cylinder head and work within your operating range.

 

Geoffrey,
Couldnt agree with you more. The stuff that makes a good head is soooo complex.
I think that all us observers of this project are really keen to see the end result ! but appreciate the information as you go along.
Thanks for taking the trouble .

All the best

Geoff

 

My goal here is an engine that produces 380hp, operates in a range of 5000-7800, has a nice flat torque curve in that range, and runs on 93 octane street fuel. That means that on the dyno that I use it will need to see 323rwhp based on what I've experienced for drivetrain loss of a G50 using engine dyno and chassis dyno results on the same car. That is my goal and I expect to exceed it. The factory 993RSRs were delivered with 350-360hp depending on your source of information on 98RON fuel. The last 3.8l engine I tuned with ITBs, 13:1 compression produced 342rwhp on racing fuel. The last RSR engine with cams and ITBs and race fuel I tuned produced 325rwhp. A good general guide on compression ratio is that 1 point of compression is about 4% of the total power.

 

Geoffrey,
The factory RSR intake system was a bit restrictive ! With the restrictor sorted there was more !!!!!

Its always a fine balance to match fuel quality with dynamic compression , which itself ,is a cam duration modification of static compression .
A few factors to match to get the result you want !
Looks like you are doing all the right stuff to get there. It will be very interesting to see what you finally achieve - I guess you are quite interested yourself !!!!!!!!!!!!!!

All the best

Geoff

 

I know the sprint configured engines (30hr engines) with the 11.5:1 compression and 54mm intake valves produced 408hp on the engine dyno, so as you point out, there is a wide range of HP the engines were configured for. In some series they ran restrictors and others not. It is hard matching the engine configurations with the HP rating given. The general consensus is that 400-415 is about what an unrestricted RSR engine "can" make and consistent with what I've seen, even from higher compression engines. The most I've seen is 354rwhp from a 13:1, mild ported 993 heads, ITBs, GE100 camshaft engine which is about 416hp with the above mentioned drivetrain loss.

 

Geoffrey -
You should combine all your posts to create a "How to build a 964 Dream Race Car" site! Just when I think the car is at the uppermost level you come up with something else that blows me away!

 

Does anyone know if Robert Linton has managed to hit the 450HP he was hinting at?

 

Geoffrey,
Yes, agreed. All kinds of national restrictors used . My other thought is that there must come a point where the cost of evermore development is greater than going over to modern 4 valve heads and making big power.
I had a quick look at what would be involved in using 996TT heads . Small pain to organise head water cooling but very feasible .
I will never do it ( too busy etc ) but that could make a serious project .

Anyway good luck and please keep posting such good stuff.

Geoff

 

This is the power curve from my 3.8 engine running the very same set of heads that Geoffrey now has:



......at 12.75:1 compression, Optimax pump fuel, standard 993 ignition system, standard 993 heat exchangers and 100 cell cats. Maybe with a set of headers, no cats and CDi ignition we could make another 15-20bhp at the wheels but there is only one way to will find out for sure, hence why two M&W CDi units & coils are on order and a prototype set of headers are on the bench.

Like I said, I think that the port areas are pretty close to optimal, but I'm sure that everyone else will make their own minds up.

 

According to the engine modelling programmes that I use, our race engine is running at 126% VE which is equivalent to running the engine on 4psi of boost. I see similar VE numbers with our 355bhp package on the 993VR engine, which makes sense because it makes the same power & torque as an identical supercharged engine.

I agree regarding the flowbench results. On the flow bench our heads do not make big numbers, whereas on the engine they do.

As for stacks on a 993 engine, I only have one direct comparison which was the same totally standard engine run on Motec in two different cars, where it made roughly the same power on the stacks and NVR manifold but a greater peak and spread of torque on the factory manifold. Motec ensured that the results were fully optimised and I am happy that there was no more power to find from either test.
Change the cams for ones with greater overlap and the result would no doubt be different, but then it would no longer be a standard engine.

 

Colin, it is interesting that your dyno shows a 14.5% drivetrain loss for your car. I know that the one engine I had on an engine dyno then immediately on a chassis dyno showed a 15% drivetrain loss. Both are G50 transmissions and I like that the data correlates.

The same day I flow tested the head, I was also working with a BMW e30M3 head and it is interesting to see the differences in flow when the length of the intake trumpet is changed. I know this will directly relate to the powerband of the engine as you lengthen and shorten the stacks. This is the principle of how the varioram intake works in simplistic form.

 

ve ?
nvr ?
we not all got "Bachelor of Engineering" behind us .

 

Colin,
Took me a moment to work out what a NVR manifold was !!
Cant comment on Motec but using Motronic ,there is a significant difference between a NVR manifold and stacks.

I dont want to go round the Motronic/Motec stuff but I would believe that even your modelling software should show a larger peak bhp with stacks . Its the way motors work ! For example you may have noticed that this year F1 are not allowed variable length stacks .

I will shut up now as I am sure this must be boring everyone and I dont want to get banned !!

All the best
Geoff