3 litre 944 Turbo on track

[Hilux]

So with a turbo you can use the boost pressure to increase the CR and compensate for decreases in volumetric efficiency.

All turbos do this. Especially modern little diesel engines with tiny turbos. They positively pressurise the charge air as against utilising atmospheric pressure to overcome a partial vacuum and exhaust pulsing and rely heavily on a cool dense supply of air (hence the use of air/air intercoolers) Atomised fuel molecules stick to these air molecules much better so if you can cram efficiently spaced fuel molecules into a chamber under pressure you can get good power from efficient burning. All turbo engines will perform better on cooler damper days than dry hot ones.

I was extremely encouraged to see the results of my theory about torque etc working and also that the rear wheels can drive high torque if the acceleration is smooth enough and the extra weight loading from the torque reaction can then sustain grip "" and this has encouraged me to build a much more powerful version with a good torque spread.

I entirely agree (not that I have half your expertise Barry) and thats what I have been saying in this thread and others for a long time. A good spread of torque (flat curve) means linear progressive (useable) power. Look at V8`s for example. Electric motors and steam engines have near maximum torque immediately they move.

Power is simply a function of torque and all engine graphs generally without exception will show the meeting point at 5250rpm when torque is already tailing off so its how it gets there that counts if you want a quick car. The headline power figures for engines may be impressive but its near max revs and with less torque due to rotation but stalls quickly if challenged.

 

[barks944]

My analogy about the car was aimed at the power stroke in the engine. I mean't that in the power stroke wouldn't it be more difficult to apply a force at high RPM as the piston is moving faster down the cylinder? Essentially as the piston is moving faster in the direction that the gas molecules are impacting upon the piston to create torque, the impacts have less effect. Therefore you would have to have a higher cylinder pressure to get the same torque, maybe adding a further decrease in efficiency. Maybe this would have very little effect. Just a random thought that occured to me really. No real bearing on this convo . I think a piston moves about 25 metres per second at 7000 rpm assuming it moves 0.2m in one rev. No Idea how fast a molucule would be moving in a combustion chamber though! Probably well over 1000 metres per second!

Hilux, I was thinking more about actively controlling the boost pressure to overcome the changes in the volumetric efficiency as the RPM's change. For example as the revs rise you might need to increase boost pressure to get the same effective CR and maintain a constant torque. But as Baz says, you would probably need a big turbo to achieve this and that would in turn take longer to spool up. I guess variable vane or twin turbocharging systems would help with this issue.

 

[sawood12]

Tom, your analogy is right. The expanding gases in the combustion chamber exerts a constant force upon the top of the piston causing it to accellarate, so as it accellarates the force acting reduces. An analogy I prefer to use is riding a bicycle. As you start riding downhill and the pedal rpm increases the force you can exert on the pedals reduces as you're legs cannot work fast enough - the speed of your legs being analogous to the rate of expansion of the gasses in the combustion chamber. This is also why diesels rev lower than petrol engines as diesel burns alot slower so cannot rev as high- the vmax of the piston is lower.

On a 'typical' N/A petrol engine that showed that under WOT the piston was only accelerated over about 45% of its stroke (on a level surface), it then had a period of constant speed. This was to do with the expansion of the gases in the combustion chamber. The piston was initially accellarated but at a certain point the velocity of the piston decending was equal to the speed of the expansion of the gasses. This obviously has a negative effect on torque as when the constant velocity stage occurs the expanding gasses are no longer pushing on the piston with any significant force - the gases are still expanding and driving the piston but not accellarating the piston as the load the engine is under (due to vehicle weight, aerodynamic forces, general drag of engine and transmission) is balanced out by the power of the expansion of the gasses acting on the piston crown. This could be remedied by either increasing combustion temps (not feasibly on a car due to limitations of materials), increasing compression ratio in an N/A engine (hence modern engines are always pushing for higher compression ratio's)or turbocharging.

The fundamental problem here is that engines are of a fixed configuration. You ideally want an engine to change configuration as the revs and conditions change - and that is what is trying to be achieved with things like variable valve timing and lift, variable geometry airboxes etc, however these are minor tweaks to the configuration of the engine. The next generation of engines will have variable stroke length adjusted by an eccentric big end bearing, so they can vary compression ratio continually to suit the demands on the engine at any given time. Mercedes are working on these currently and have them working on test rigs, and reckon that they will combine the advantages of petrol and diesel engines - in fact they will work like diesels when you are pootling around - longer stroke, lower revving, no spark ignition, more efficient combustion therefore better fuel economy. And then when you want performance they will change to a spark ignition cycle. The tricky bit is managing engine management during the transition. The limitations of modern map based ECU's (which have changed little in 20 yrs, just been tweaked) just are not man enough for the task - too many parameters to control, things happen too fast with too many permiatations so is impossible to map. Therefore they are working on a completely new type of engine management system that employs artificial intelligence to manage everything during the transition from diesel cycle to spark ignition cycle. Problem is by the time they crack it and bring it to market we'll probably all be riding around in driverless pods powered by nothing but the good thoughts of do-gooding environmentalists.

 

[Neil Haughey]

Need to caveat that as my understanding is that increasing combustion temps or increasing CR (which by side effect increases combustion temps), increases the production of oxides of nitrogen. ISTR the theoretical efficiency maximum is at 15:1 but not that much better beyond 12:1. Like all these things changing one thing has side effects that require other technology. Mind you they are putting urea injection systems on diesels now to reduce the NOx output.

By AI I guess you mean they will be running some form of self adapting neural network system on the test bed which will then become the 'program' load for the ECU's. Surprised really they haven't cracked it by now, they published papers on use of AI for self learning engine management getting on 10 years ago now.

 

[333pg333]

I think we all realise that torque is what gets you out of corners and this in turn should (providing there are no traction issues) give you a higher top speed down the following straight. This is the basis of good lap times. The issue of just what sort of torque and hp curve are best suited in one's car is also highly dictated by the type of driving and track that you frequent. A good torquey road car doesn't automatically mean that's the best combination for the track. Having driven a 2.5L with stock (tired) internals and a too large turbo forced me to rev the car out to approx 7000rpm on the track. Obviously this isn't always ideal, but also on the track you are usually above 4000rpm anyway. So having the curve on Baz's existing motor would not necessarily be 100% ideal no matter how much we all respect torque.

Personally I like the idea of increasing EVERYTHING and seeing what happens. My looonnggg overdue 8v 3L build with just this setup saw 14psi at 2700rpm on the engine dyno last month. This is with a Borg Warner S362 turbo (or close to it) which for comparisons sake is closest to a GT35. Of course you could generate high load on the dyno to induce early spool. I am not sure what the load was set at, but I feel that this was very promising. My point being is that I think there is still cause to believe that you can have cake and eat it.

I will read with interest about your next motor you do with 16v Baz. As you stated, there are a lot more variables available with this setup. I think you will be amazed at the potential.

Patrick

 

[TTM]

Torque is boring without a suited level of horsepower - While I passengered in a 968 turbo S back in 2004 I noticed that 368lbs/ft at 3k rpm with a peak of 305bhp was not the last word in exhilaration, and thought that a reasonably higly sprung 2.5 was actually more fun if perhaps not as easy/fast/efficient.

Since my 3.0T has been running, and although it still runs the restrictive airbox that I feel seriously hinders spool/response (seeing 15psi at 3100rpm at the mo), I still can't quite get to grip with it as it is almost impossible to drive it on a relaxed mode anywhere else than on derestricted Autobahn, and even there I've already had a few trousers moment when flooring the throttle in fifth at 120mph as the rear end squats just like my chipped 2.5 did when it hit full boost in 4th gear.

What I'm saying here is that as soon as people get into this 3.0T thing and decide to remove the factory restrictions to reach a suited level of horsepower they may be in for a level of weaponery and obligatory attention/caution they might not have expected - believe it or not but at the moment I have more fun driving to work with my 55bhp fistful of Renault shed than starting the 944T as the thought of taking off for crushing another speed wall settles into my mind.

 

[Barry Hart]

I am glad I started posting on here as the technical level of the responses is extremely high and interesting.

Sawood is right that what we all need is an engine that works like a kids "transformer" and changes with revs - and in about 1983 - I was building and dyno testing engines with variable port areas, pressures and timing (both inlet and exhaust) but just then it was impossible without massive resources to get a control system built to handle the changes. The mechanical part was not too difficult.

Control systems are available now so the whole arena opens up again to low cost development and interest once the hardware has been built.

I was fascinated reading the discussion about fuel and piston speeds - something I know very little about and had not considered. My gut feeling is that it is not a real issue in an engine that I specifically do not want to raise the maximum revs of - instead preferring to find more torque - and I am also puzzled - that if it indeed it is a restricting factor, how a F1 engine manages on a N/A inlet at 18,000 rpm with controlled fuels?

Yes I know the bmep is relativley low and they have infinitely variable cam timing (and possibly lift?) - but even so it must be better than the porportional increase in bmep at lower revs - or there would be no point in doing it.

Regarding driving on a track being always over 4K revs - I think this is just a mind set. In any gear the torque curve at the rear wheel follows exactly the one at the engine (just higher or lower in size and proportion to the gear ratio reduction or increase) however the amount of torque that converts to linear force at the tyre - that the tyre has to grip without slipping alters as you change gear - the higher the gear the lower the rear wheel torque - so even if below 4K you may have a little less torque at the engine you may still have more at the tyre in a lower gear and this could result in the tyre spinning up.

Similarly the rotation circumference of the tyre that transmits a pulse of torque is longer in a higher gear and this influences the way that the rubber grips and spreads the load over a longer surface area. So I think it is just possible that the benefit of changing down one less gear and smoothly accelerating from a higher gear at lower revs without upsetting the ballance of the car or the dynamics of the engine gas flow or the traction rate of change at the rear tyre - may just make some corners quicker to exit from in the higher gear if the torque is exceptionally high at low revs - which few will have built an engine suitable to explore.

Despite my obvious past involvement with engines and gearboxes and my resurging interest in Porsche turbo charged engines - we are still only a small private business concentrating on doing the best we can in standard form for typical owners of standard cars. I think the fact of my background experience and our machine shop have resulted in us solving the problems of rebuilding Boxster and 996 engines at reasonable prices (and while we are doing it strengthen some weak areas) and it is this (and a desire to educate our staff and improve their experience) that has resulted in us merely dabbling in something more interesting and challengine than we worked on before.

This turbo was relatively easy to build because we had all the parts and also fitted in nicely with my theories on engine power/torque characteristics and therefore was worthwhile exploring. The interst at work now has extended to what else we can do and I think that by involving the staff in what we will be trying to do is very valuable for them and is more importnat to me that the outcome.

We cannot afford to go racing (time and money constraints and anyway just too busy looking after customer needs) but we can afford to build something unusual and use that practical experience to learn more and provoke some interesting technical discussions - just like we have on here and other forums recently - at a much higher level than most postings - which I hope some readers find interesting and informative.

However we are not trying to grab any glory, pretend to be the experts, make any special claims, set the world on fire or change the face of engine design - just merely trying to explore various theories and try and test some of them to establish what in the real World - is important and works. In doing this the thoughts and experiences of others is valuable and appreciated.

Baz

 

[edh]

One interesting point Barry - you, and others, have mentioned the difficulty in putting down the power with a 2.5 running higher boost. I think that may have a lot to do with the suspension setup.

I ran a 250 turbo, slightly modified & running around 1 bar boost to produce around 320 lb ft. I drove the car on road & track in all weathers & I only ever remember 2 occasions in 4 years where I got significant wheelspin (on tarmac at least - kerbs at Donington are a bit slippy in the wet!) - and you had to be very brutal in a low gear to unstick the rear end. On track it was very controllable coming out of corners - maybe I was just a bit too cautious & was too late in getting on the power in the wet? [] ) My car was running standard 245/45/16 rear tyres and Leda dampers. I have read other comments about much improved traction when people move to aftermarket suspension such as KW's, which supports my theory.

 

[333pg333]

"Regarding driving on a track being always over 4K revs - I think this is just a mind set. In any gear the torque curve at the rear wheel follows exactly the one at the engine (just higher or lower in size and proportion to the gear ratio reduction or increase) however the amount of torque that converts to linear force at the tyre - that the tyre has to grip without slipping alters as you change gear - the higher the gear the lower the rear wheel torque - so even if below 4K you may have a little less torque at the engine you may still have more at the tyre in a lower gear and this could result in the tyre spinning up."

Well it's a mindset in as much as we tend to drive cars on feel and sound as much as we imagine that we watch the tacho etc. So we generally use the gear that is most suitable for each particular corner. I guess if you had a massive supercharged Diesel motor which had gobs of torque the second you started her up, then you may indeed stay in lower rpms, but so far I have found as a generalisation, we tend to be at higher rpms on the track most of the time. Essentially you have the foot flat to the floor most of the time, except for braking events. Of course we balance the car on the throttle too, but this is for a smaller part of the corner. I hope to be able to drive on more of the torque in the future but even then, you can tune the same engine to change it's curve to suit your needs.
Let alone all the other components that have an effect on the traction situation.

 

[Hilux]

My car was running standard 245/45/16 rear tyres and Leda dampers. I have read other comments about much improved traction when people move to aftermarket suspension such as KW's, which supports my theory.

Better ARB`s reduce weight transference and improve grip even more.

 

[paulf968]

mind you we need a good dose of weight transference in powerful cars like this to stop wheelspin !

 

[Neil Haughey]

ORIGINAL: paulf968

mind you we need a good dose of weight transference in powerful cars like this to stop wheelspin !

Interesting thing is, as Baz has talked about the greater the acceleration the greater the loading onto the rear of the car, and hence less wheelspin. The key is definitely to make sure the acceleration builds up slowly/smoothly in just the same way as when you chuck a car at a bend you get a high sudden weight transfer that overloads the tyres with the ultimate result being slower cornering.

 

[barks944]

If you had a good electronic boost controller you could specify exactly how the boost comes in too allow for weight transfer before hitting full boost. Rather than following a natural boost rise curve or using the boost controller to boost as quickly as possible. EBC's that are setup to get boost up as quickly as possible would be making this issue worse as there is less time for the weight transfer to take effect before you hit high boost and get loads of torque. This would be a big issue when shifting at higher RPM's as the boost would probably build a lot quicker then when you are coming on boost from below the threshold. You would need a boost controller which could actually open the wastegate when it otherwise wouldn't to do this. Not sure how you would do that as you would need a pressure source available when boost wasn't or another form of wastegate.

I think the main issue when specifying a power curve for a car is gearing and what the purpose of the car is. With the right gears you can produce power anywhere in the rev range and utilise it to good effect. Gearboxes are torque converters remember. If you take the same engines producing the same power but at different RPM's you can use the gearbox to get the same torque output at the wheels at the same vehicle speed. The difference then is the rev range over which the engines produce power and whoever can stay in the lower gear for longer wins. This wouldnt be an issue with a CVT and I don't think the two cars would perform any differently. I think the main difference between the two engines would be the efficiency, the one producing power at lower RPM's should be more efficient.

Barry, regarding the issue of piston speed, I think the speed that the piston is moving is slow compared to the speed of the molecules in the gas. Probably well under 1% so it probably wouldnt make much odds even at 18k rpm.

 

[Barry Hart]

I have been given a list from rennlist of all the people (mainly in the US) who have also built similar engines - so now feel a bit embarrased that we thought there were just a few who had done it. However I think so far ours is the esiest and least expensive and the results are not that far behind the others and probably more user friendly to drive - so I hope it still had some value.

The good thing is that I will now contact them all and ask specific details of thier builds and results and use this to compile a chart to see any correlations that will short cut some of the work specifying the next engine (quite a few have used the 16 valve head and some list special cams etc so should reveal a lot).

Meanwhile we have a few more tests to do with the old turbo and will start work on the next one in a few weeks.

We have a chassis from a 1986 curved dash 944 2.5 that could be suitable. Does anyone know of any specific differences in the basic chassis that would prevent it being suitable to build as a hybrid 944 turbo/968? (apart obviously from running gear etc).

Now that we know the results and the map for our existing turbo we are thinking of going to the other extreme with the next car and build a top end fast and powerful animal and map that and then use the info retrieved to build the car we are excited about to try and achieve both the good ends of output graph each in one car. However we might be able to get enough information from the other builders to avoid that middle stage and go straight for the ultimate engine straight away - time will tell.



Baz

 

[DivineE]

Just realised I have been reading this thread with great interest but haven't posted an oppinion yet (very unlike me) []

Again a little off topic to the 3ltr build but following on from discussions about variable compression ratio. Surely as a turbo has the ability to lift C/R a much simpler solution to altering the stroke of the engine (and all the complications relating to that) would be to have a C/R somewhere between that of a regular turbo and N/A engine and use a sequential turbo setup controlled electronically to give a boost curve exactly opposite to the engines natural torque curve.

i.e. maximum boost at low revs to raise the compression, reducing as the engines natural curve rises and then increasing again as the engine struggles to get good cylinder filling at higher revs? The desired affect being that the engine would be as close to pinking as possible / safe for the maximum spread and the practice being that the tiny first turbo would boost from idle, the slight drop in pressure as it runs out of puff and the big turbo takes over would occur as you require less boost (to avoid a peak) and the big turbo would be able to maintain pressure enough to maintain the C/R right to the redline? Even if it was 20% less than perfect its got to be cheaper and easier to produce than a perfect system?

 

[DivineE]

p.s. as I said on the other forum Barry:

To my knowledge a conversion with standard parts is unique and v.interesting to compare to the other high spec options as it seperates the benefits of the extra capacity from the other alteration that were made. Not many people would have been able to do the work without being tempted to change other parts like the turbo etc when they're in there.

 

[barks944]

DivineE, I agree about using boost to keep the CR the same up to the redline. I think most EBC's would do RPM based boost control which is what you would need for this to work I think.

 

[sawood12]

ORIGINAL: DivineE

Just realised I have been reading this thread with great interest but haven't posted an opinion yet (very unlike me) []

Again a little off topic to the 3ltr build but following on from discussions about variable compression ratio. Surely as a turbo has the ability to lift C/R a much simpler solution to altering the stroke of the engine (and all the complications relating to that) would be to have a C/R somewhere between that of a regular turbo and N/A engine and use a sequential turbo setup controlled electronically to give a boost curve exactly opposite to the engines natural torque curve.

Yes, this has an effect to a point, but is only a half-way house when used on it's own. This will determine what force is produced onto the crown of the piston due to expanding gases, but how you use that force depends on the mechanics of the moving parts of the engine. For example, if you have a constant force but increase the radius arm it acts on, the torque is increased at the pivot - so same force acting over a longer distance and more torque.

ORIGINAL: Hilux

My car was running standard 245/45/16 rear tyres and Leda dampers. I have read other comments about much improved traction when people move to aftermarket suspension such as KW's, which supports my theory.

Better ARB`s reduce weight transference and improve grip even more.

That is true, to a point. It is like everything - too much of a good thing is bad. ARB's are a necessary evil and interfere with the main suspension by interfering with it's independance and fighting against the springs. Generally it is the challenge of a chassis designer to minimise the strength of the ARB's or even do away with them altogether, just as it is to produce a suspension system that is hard yet compliant - classical engineering paradoxes. An ARB that is too stiff will actually cost you grip as it will try to lift the inside wheel thereby reducing the force on that wheel and therefore the grip.

I definately noticed a significant improvement in grip when I installed the KW and body roll was significanlty improved so never felt the need to upgrade ARB's. But on semi slicks on a dry track i'm sure stiffer ARB's would have had an overall beneficial effect.

 

[barks944]

I think that the lower your geometric compression ratio the higher the effective capacity of the engine when on boost and the greater the torque/power, as you can have a higher volume of air/fuel mix inside the piston and still maintain an effective compression ratio lower enough to prevent knock. If you increase the geometric compression ratio you have to run less boost to maintain the effective CR and prevent knock. You have to balance between performance on and off boost though so you dont want it too low.

 

[Neil Haughey]

Very much so. The Miller cycle takes things the other way, it works by having a lot of retard on the inlet timing such that for a part of the compression stroke the engine pushes air back into the inlet manifold. This effectively reduces the CR but its doing it by effectively reducing the capacity. The benefit is that the power stroke is then effectively longer then the compression stoke. It would be very tempting to supercharge an S2 or 968 motor and run it as a Miller Cycle engine but I reckon the capacity is then reduced to roughly 2.5 litre or so.