300bhp, 350lbs feet torque 944 3 litre turbo

[hartech]

For continuity I wanted to add information to previous postings on this subject - but the search facility is currently down - so I cannot link them up - but I just wanted to draw the attention of those interested to the fact that we have finally finished the car and had it dyno and road tested with remarkable results (graphs of output included).

Without being able to add this to the previous postings - I don't see a lot of benefit in expanding the details too far today - except to say that they are all posted on the 968.uk.com web site (including bhp and torque graphs etc) for those who may have been following the story and are interested.

Baz

[hartech]

Thanks Diver944 that contains our expectations and the graphs uploaded confirm it.
& .

Baz

[hartech]

Sorry - I tried and it didn't work so I will try again to send the graphs.

Baz

[hartech]

It is not only quick but flexible and manageable - a very nice road car to drive.

On the 968.uk register - over the years - there have been many lengthy arguments about the merits of power and torque (and probably some on here) and - which is more important. Much of this has continued on that web site but I didn't realise that it was not easy to access that site if you have not joined.

Basically - to put it very very simply - it is the torque at the rear wheel that accelerates a car. Torque mutiplied by revs creates the bhp reading. If you have a lot of torque at low revs - from the engine - this will result in a lot of torque at low revs in any gear. If you can change the ratios and the diff ratio then you can alter the torque at the rear wheel - so a high revving engine with lower torque - when the torque is aplied through a greater speed reduction may have higher rear wheel torque - but if you cannot change those ratios - then just revving an engine higher to create more bhp may not make it faster since the gear ratios are the same and the relative torque at any revs must be higher to increase acceleration - and you usually reduce the power band as well.

The higher it revs the less time there is for the intake to receive the quantity of air that results in the compression pressures needed to create torque - so although it is easy to raise revs and bhp - this may lower torque.

With a fixed 5 speed gearbox (in the 944 turbo) with high final drive ratio - increasing the torque in mid range - in my view - should create a quicker car and as tuning a 2.5 with more boost usually results in a harsh power curve and difficulty controlling rear wheel spin - I thought that this creation might result in a nice quick car to drive that is also very manageable - and that seems to be the case.

Baz

[hartech]

The biggest problem would be the pistons because the S2 piston does not in my opinion have enough taper for the performance planned. The 968 piston is more suitable.

The 2.7 cylinder head fits on the 3 litre block and has larger inlet ports, the 2.5 head does not fit and the inlet valves are smaller but should still work well. The 2.5 head can be welded and modified to fit. The 2.7 engine has the right bores and head and the pistons may be suitable but I doubt have enough taper either - but I have not checked out the piston ovality and taper on these.

The 968 block also has spray jets to spray oil under the piston and onto the bore for cooling and lubrication (but they could be fitted to a 2.5 or 2.7 block).

Whichever pistons are used they would have to be machined to lower the compression ratio for use with a turbo (but not if a naturally aspirated 3 litre engine was planned).

I do not have any 2.7 heads available.

We are considering manufacturing replacement liners to enable any 2.5 to 3 litre block to be repaired and if so any block could be fitted with 4 liners to suit 104mm pistons for changing a 2.5 to a 2.7 (with a 2.5 crank or 3 litre with the S2 or 968 crank). You also need bigger injectors and a re-mapped ECU.

Without that you must use a 2.7, 3.0 S2 or 3.0 968 block with 2.7 or modified 2.5 head, a 3 litre crank, probably altered 968 pistons, shorter head studds and the whole 944 turbo auxiliaries.

If youa re not using a 3 litre block (but sleeving a 2.5 with liners) it may be neccesary to machine some slots in the block to keep the oil level down with the bigger crank throw.

It begins to sound quite complicated - but was actually quite simple - but then we had all the various bits in stock and we have a machine shop to modify things in.

I think the best route might be a 968 car and an old crashed 944 turbo as a donor for various parts or a 944 turbo and buy in (or modify to) a 3 litre block and 2.7 head if using a 3 litre or 2.7 block.

Baz

[hartech]

I guess Billy that it is not too bad if you already have one car or the other or can wait while you source the parts from various breakers.

The performance is superb (into 996 territory) and all I wanted to convey was that you can get that relatively easily and relativley inexpensively without a huge number of special parts being manufactured, or great technical difficulties and without the frustration often associated with the technical problems getting many other project engines running and performing acceptably.

I have grafted in Stuart Cooksons reaction to driving the car last week - posted to another web site.

"Today I needed to run some errands so a perfect excuse to use the 968 and one of those errands had me in north Manchester, so I thought I'd pop over to see Barry's 944 Turbo project. I've been to Hartech before but while I was there Barry gave me a guided tour round the new features of the premises, parts and spares all indexed and itemised in racking and the new engine building department which is being made bigger due to the amount of work they are dealing with on Boxster and 911 engines(and some of the problems are factory generated).

While we were chatting about the development of the Turbo Barry suggested we go for a drive, Barry drove the 944 out and down the straight from the industrial estate.

Barry used the accelerator to good effect between the speed bumps.
Initial response from me was 'bloody hell'(or words to that effect) this is quick, I mean very quick.

We got onto the main road, keeping in mind we were on public roads so some amount of caution was necessary but again Barry used the accelerator to good effect.

We got to a layby and Barry turned the car round to let me drive it back.
I waited for the traffic to clear on the road ahead and then gave it some beans. I would say that the car behaves quite normally below 2000rpm and then you get a steadily increasing surge up to 3000rpm and that's when you really notice a huge amount of torque and the acceleration far far exceeds that of a 968.

We were running out of clear road very quickly and weren't really able to explore the full potential of speed but the massive amount of torque through the midrange gave an idea of what the car was capable of so I really wanted to see how the car would do in higher gears but still low revs so we continued with the 'seat of the pants' road test accelerating from around 2500rpm in 3rd and then in 4th. Again the amount of torque is outstanding from 3000rpm. The problem is that in the higher gears we were doing silly speeds so curtailed the test.

I had said to Barry that I reckoned there was a huge amount of torque available over and above the 968 and that we had at least 50% of that available torque at 3000rpm so when we got back to his office we checked and sure enough the Dyno figures confirmed it. Max torque on a 968 is 225lbs ft, and on this turbo it's closer to 350max, with about 210lbs ft available at 3000rpm as opposed to around 180 for the 968.

Over the next 1000rpm up to 4000rpm is the big difference, the 968 gains another 45 up to 225 whereas the Turbo gains around 130 to get close to 350.

So the car is able to be driven around town very smoothly at pretty low revs but with lots of reserve. The car is very sedate, no that's the wrong word, it isn't sedate but it is under control and feels pretty normal to drive until you start accelerating hard.

Barry wanted an honest opinion on the car and what it feels like to drive, I think I hit the nail on the head when I said that you wouldn't know it wasn't a production car(apart from the extra torque and power) and I think Barry appreciated the comment. I said this because the engine bay looks just like any other 944 Turbo and because it drives like a normal car, no dramas, no running problems etc. After all, if you can put together something that works very well in normal driving conditions and there doesn't appear to be any problems with it's running, yet has the potential to obliterate almost anything else on the road, and do it with standard parts, that has to be an achievement. As Barry has already mentioned he is going to do something about the suspension because the car does pitch down at the rear quite a bit under heavy acceleration. He also wants to get it onto a track and explore it's performance a bit more, there really is no way you can do it on a public road without endangering your licence. It will be interesting to be in it at high revs in high gears although from the figures around 6000rpm the torque and the bhp is on a par again with a 968, but in a straight line drag race Barry's Turbo would already be way ahead. The overtaking potential in this car is phenomenal.

We discussed the timing of the completion of the project and it is apparent that had Barry completed this project years earlier he would have had a whole lot of people qeueing at the door for the conversion, as it is people inevitably move onto newer models and the big market for it may have diminished significantly.

Well done Barry and thanks for the drive.

Stuart.

Our track test day is now booked but we don't expect it all to go smoothly as the car is bog standard apart from the engine and stronger lsd - and not at all prepared for track testing.

Baz

[hartech]

Sorry no but I would not expect it to compare - his should be (and I am sure is) much more powerful.

Baz

[hartech]

I don't have a set of printed results for a chipped twin port waste gate 944 2.5 turbo but if you E-mail me yours I can graft them on to the previous graphs for comparison.

However none of the similar cars I have driven have had anything like the bottom end torque or indeed bhp and were relatively difficult to drive quickly because the turbo response had built in lag both spooling up and after backing off - which this engine does not have - being more like a naturally aspirated very big engine.

Baz

[hartech]

Hi Ben - yes everything else is standard and in my opinion it really has produced a quite remarkable car for comparatively little money and nice to drive as well (as you will already know from yours).

If you refer to my answers up the page you will see that we probably will make a liner to suit all models - as the pattern is not so expensive and it would solve a problem long term for those seeking to repair and existing engine or modify one.

I don't see why the standard 2.5 head would not work if we sleeved a 2.5 block and there are plenty of 944 turbos still around at reasonable costs.

I also think for those more adventurous - a 968 with perhaps a damaged engine - turned into the equivalent of a turbo S - would have some high value when finished - although I guess it may be better not to graft a 944 turbo set of auxiliaries on to it but to fit a larger turbo and intercooler and a more modern engine management system from the outset?

Baz

[hartech]

A lot of very bizarre comments recently – all very difficult to deal with without offending people as many of them are not only misleading but completely wrong. I am frankly not sure if a group of people are trying to catch me out and have posted a load of cr*p to test me as their comments either reveal a serious lack of understanding of torque, power and gear ratio graphs or is this just the typical reaction one might expect on here from a small minority - or is it that people really don’t understand simple basic issues – but feel they should never the less post their criticisms as if they were the experts for others to absorb? Sorry if I seem bhy - but I am rather gob smacked about some of those comments from people I imagined knew better!

For those who don’t know who to believe – it is important that someone explains the real issues properly and as no one else seems to have jumped to correct them – it falls to me to do so. So - for those not sure what to believe - I will patiently try and explain (although for people whose understanding of dynamics is so far out – I doubt it will do any good and probably just provoke more uneducated criticism).

Firstly let us please remember that we set out specifically to construct a relatively simple, relatively inexpensive conversion that we planned to provide more bottom end torque, about 300 bhp top end and reduced turbo lag and we achieved all that.

Anywhere between 3000 and 6000 rpm this engine has anything between much more power or torque than a standard 944 250bhp turbo, S2 or 968 or finally at 6000 rpm about the same.

You have to drive the car from the revs you hit when you change gear to the maximum revs before changing up again.

In this gearbox from 1st to second you drop to 3500 rpm from which you must drive it again all the way to 6000 rpm. Just look at the torque or power graph from 3500 to 6000 rpm. Where is it less than the other engines and is it not plain obvious that it is massively more throughout? It is on average about 50% more!

From 2nd to third you drop to 4100 rpm which is the point of peak torque (and torque provides acceleration through the simple formula that Torque = (Inertial, drag and rolling resistance) * acceleration or put another way acceleration = Torque divided by various forms of resistance (which are the same for all the cars I have compared) and therefore acceleration is proportional to torque.

From 3rd to 4th you drop to 4450 rpm which is almost on peak torque and 100 lbsft more than a standard turbo.

From 4th to 5th you drop to 4750 rpm which is just coming on to peak power that is almost flat to 5500 rpm by which time you are travelling at about 147 mph.
These graphs show a superb match between the ratios available and the power and torque delivered.

Throughout all these gear changes you are constantly driving with much more torque or bhp than the other engines. This means that you will be accelerating to that speed very much quicker and get there very much sooner. 6000 rpm is about 160 mph in 5th and as you drive on to that speed – you are still driving with more power and torque than a standard 944 2.5 turbo but you got there much quicker. It is only at 160mph that the power then – for the first time – is slightly less than a 968 and std turbo (but still more than an S2 until about 6300 rpm or 168mph (where an S2 would have to rev to about 7200 with the same power to match it). It may well be that on a very long straight road – several miles long - a 250bhp turbo or a 968 may eventually reach a couple of mph more in top speed – but would they ever catch you up? – I doubt it.

I don’t know how NJH reckons my curve would only give 140mph – totally beyond my understanding that one! 6000 rpm in a 944 turbo is about 160 mph.

Another reverse lack of understanding is Ajax50’s comment that a steep power curve = loads of torque – as really the opposite is true (although to be fair he did support the advantages of torque – perhaps just got a bit muddled up about the graphs). Since power is proportional to torque * revs, if the torque was say a flat line – then the power would increase in a straight line as the revs rise. Since maximum torque is always achieved before maximum bhp, and the maximum possible is in proportion to the breathing capabilities of the engine – any increase in torque before top revs or peak bhp, will make the line higher at lower revs and therefore the bhp line would not be steeper but flatter. Anything that increases top revs and bhp at high revs more (which is the common tuning result) almost always lowers the torque at lower revs and the result again is always a steeper power curve. So once again the stated criticism is actually arse about face!

Those of you that don’t seem to understand the first thing about these graphs or power and torque really should try and resist the temptation to make firm statements on such a forum because many others simply don’t know what is right and wrong and it will not only mislead them and confuse many but also is very unfair criticism when in most cases it seems to be wrong and even the opposite of what actually happens.

Fortunately many others who have contacted me directly fully appreciate the results and some even can imagine driving the car from viewing the graphs and in one case has done both and found the graphical description and the practical experience the same. For those that quite rightly immediately recognised the achievement and the benefits of such a pair of graphs – can I add that although acceleration is proportional to torque (if resistance is constant or the same between compared cars) there is a phenomenon inside an engine and its transmission that relates to the stored energy between individual power pulses and strain energy lost as a result. It means that in actual fact there is a slight benefit from slightly reduced torque if the number of revs is increased and the strain energy lost is reduced. Put into more simple terms it means that a small increase in bhp at high revs can offer a small improvement in acceleration even if the torque is slightly lower and putting it simply it means that the best torque or power curve for the fastest acceleration, is one in which the torque is dropping off while the power is still increasing between the gear change revs.

However you must qualify that last point by stating that such benefits can only be exploited if you have the freedom to alter gear and final drive ratios. Unless you can alter these (and most of us are limited to the ratios we inherited with our car) for most cases these ratios are far too high to use on a race track or a drag race. Since - as you change up in the gears - the gear ratios gradually reduce the rev drop as you change up to higher gears - it means that most drivers on a track are using the lower gear ratios that require an engine to produce a relatively wide torque or power curve and not a narrow peaky one (even if that produces slightly more bhp to boast about). This is not true of GP racing or non restricted competition – but for most of us it is a fact of life we have to understand and adapt to if we are to get the best out of a road based car or gearbox.

I fully accept that there is more potential from this engine, and if it was easy to change the final drive ratio – there would be a benefit from different internal ratios as well and a different power curve may be better (although sometimes too close a ratio does not suit a turbo due to reduced time to spin it up under full power and heat etc) – but we don’t usually have that luxury and if not and if we are stuck with the gear ratios of our car and want to use it on the track or on the road (who needs more than 160mph on the road?) then looking at the rev band we are forced to work with at the speeds we are likely to encounter in the lower gears we will be using – it all means that a torquey engine like this one may be faster accelerating compared to a peaky one with more bhp but a narrower power band – and what’s more the torquey engine will be easier to drive out of corners on the limit as it will have more power control and therefore should lap quicker. There is a problem with the exhaust capacity - mainly when the back pressure on the valve overlap fights against the incoming charge pressure - but look we just put together a handful of parts to see what would happen and it achieved our objectives perfectly.

So compared to those objectives I don’t think a lot more development work is needed Ian UK1 and you really need to look at graphs showing the engine revs between the gear change points to fully understand that there is a power band you are forced to work with and the average power or torque in that power band is what is important and you also need to work out the speeds you will be reaching at various revs to understand that most road cars are massively over-geared for track use often only using two or three gears 2nd 3rd and 4th.

I am trying to develop a system to superimpose the test results onto a graph tracing the torque and power in each gear as the car accelerates and then to compare the areas under the graphs to see if a general rule of area to acceleration can be determined (I think it can).

Unfortunately all our actual road dyno test results have been carried out in second gear (as they are on public roads and for general safety) so at the moment I only have records for acceleration in second gear to compare between cars. The road dyno takes readings every 1/24th of a second and the results show that our 3 litre turbo in second between 3500 rpm and 6000 rpm is quarter of a second quicker (or 10%) than a standard 944 turbo, 0.6 of a second quicker (or 20%) than a 968 and three quarters of a second quicker (or 25%) than an S2 (and our car is geared slightly higher anyway).

It doesn’t matter if you understand graphs or engine characteristics – this is simply a lot faster in second gear and proves all the theory and who is right or wrong beyond any doubt. At the moment I can only speculate about how much quicker that will be in the higher gears – but all records (including Porsches own printed graphs) show that an S2 takes 28 secs 0 - 120 mph, and a 944 turbo takes 23 secs for the same 0-120 mph.

If you plot a graph of all 944/968 Porsche’s (from the manufacturers figures in the various handbooks) of say 0-120 mph and the time you will find an almost straight line graph for bhp and time and a very slight curve plotting torque against time. In both cases if you plot the torque and bhp of our engine on this graph is predicts a 0-120 mph time of 12 to 13 secs. I am not claiming this but I will test it out - although it feels very similar and when I prove it is similar - I hope it will put to bed the criticism that has recently seemed to imply that somehow this engine is all wrong, is slower than an S2 or somehow is undeserving of any recognition.

To be fair there is more on the 968.uk web site on this subject.

Thanks also for those of you who recognised our achievement with encouraging comments and support.

Baz

[hartech]

Thanks for that very useful information and support John Mitchell. You are recognised as the expert for building 944 turbo derivatives and I can understand why and your experiences are invaluable. I know that such knowledge comes at a high price (25 years ago I gave up trying to beat everyone else by building the fastest motorcycle racing engines after I recognised the bad impact it was having on my private life) – thanks for sharing it.

Since then we have concentrated on good workmanship keeping engines standard (as Porsche intended) but recently ventured back into engine work to solve the many problems of the Boxster and 996/997 engines to help customers stuck with potentially a very expensive failures and this has lead to an increase in equipment and a return to my love of “engine” engineering again.

Statistics or graphs are very valuable and anyone who has used them knows that a small comparatively inaccurate sweep of any set of figures still provides a strong guide to any trend or conclusion. Anyone can create such graphs from published information and they will show that you can predict rates of acceleration, top speeds, standing quarter mile times etc from a graph plotting power or torque against times or speeds etc with remarkable repeatability. You don’t need to understand the science of engines to do this and it is still very valuable.

At a higher level - although understanding the rules for engines is relatively simple, I don’t think it is being taught by people who understand anything other than the academic side and nothing like the practicality – resulting in the many strange ideas people have – of which the notion that torque is something you calculate from using calculus and engine power and revs is typical.

Yes – for any given revs and power you can derive torque – but what it is – is lost in the academia because practically – torque is by how much a force is being applied to twist something and all the other figures are mathematical calculations derived from it. The average pressure on the piston on the power stroke (or bmep) applied on the area of the piston crown (derived from the unit capacity) twists the crankshaft through the altering radius of the con-rod and the rate of change of speed (or revs) is the result of how much of that pressure (or the resulting torque) is applied to the resistance holding the system back. If you want to argue academically then if torque really was “the rate of change of acceleration” then there would be no torque at constant speed. You could argue that the excess torque needed to maintain constant speed is proportional to the rate of change of acceleration that results - but then that simply agrees with the torque/resistance = acceleration equation again. Can you imagine your young son asking you what torque is when you are tightening a bolt and you answer “it’s the rate of change of acceleration I am applying to the spanner”?

Engines are simple and a good way to simplify what makes them tick is to forget for a minute that pistons go up and down – and imagine it is a system in which air goes in and comes out the other side hotter and that the amount we can get in and out is proportional to the resulting performance. Going into a bit more technical detail – the “time areas” of the ports, the pressure differences on inlet and exhaust sides and the areas of the pipes in and out – govern the performance and looked at like this - a lot becomes easier to understand.

Advanced mathematics do have a value. I have found that if you calculate the “time areas” of the inlet and exhaust valves (for which integration is a useful tool knowing the cam ramp figures and valve sizes etc) and divide by the capacity (to create a specific time area) you will find a remarkably accurate correlation between the resulting torque, power and performance of vastly different engines, enabling output to be predicted at the design stage and showing that it is mass air flow that governs most performance criteria not the specific designs of an individual engine.

Even more useful are the sorts of result John Mitchell provided, from actual tests because if these were plotted on various graphs they would show a similar and straightforward predictability (or correlation), based on actual results.

You could create graphs plotting output against boost pressures, turbo sizes, inlet sizes etc and if you were comparing the results against the same cars with different settings - all the results would show some simple, logical and straightforward rules to reliably predict the results before you embark on risky expense to create something new. This is the essence of science and the core reason for it and using practical results is by far the safest method. The problem is knowing what to try before investing time and money without any results and for this you need to understand the basics that actually work but are not just theory.

If any one part of an engine system limits the resulting performance then changing that one part would always have a big effect while changing something else that is working within its capacity will have a lesser affect. Working out what the limiting factors are is a good guide to where to spend your time and effort and John’s experiences are a big help in confirming some of that.

Basically the bigger the holes and the bigger the pressure differences, the more power you can create.

The problem then is not so much how to create a lot of bhp (which really is relatively easy) but realising that this in itself is not necessarily all that you need to create a car that will perform well.

If the aim is good controllable acceleration between a minimum and a maximum speed – then just seeking to maximise bhp will not necessarily provide the answer as this will inevitably be achieved quite easily with large holes (or valves, valve timing, inlet and exhaust pipes, large turbos or high boost pressures and true compression pressures) at high revs. All these will contribute towards higher bhp – but the aim must always be to first establish the max and min speeds you seek the engine to work within. These will be lower for a short circuit car or a drag racer than an indy circuit car. Then the number of speeds you can use and finally to calculate the rev band your engine has to work within.

After that you can argue about what shape power curve will be best within that restrictive rev band that you must use to the greatest effect – although everyone seems to understand that the best overall acceleration you can achieve by maximising power (or torque) between those revs.

Whether that is best achieved from a flat torque curve or a rapidly rising power graph etc – is much argued about – but never the less each individual creates a shape they seek and the ones that get it right provide the fastest cars.

As maximum torque will always come before maximum bhp - there is always a trade off between them – but the crucial issue is that practically the more you change fixed parameters to increase air flow at high revs, the more you reduce their efficient flow at lower revs – or put more simply, the more you try and increase maximum horse power the shorter the power band and the lower the torque at lower revs – you will achieve. So picking the right parameters to maximise the performance between the revs you are at when you change gear – to the maximum revs you can reach before changing gear again – is the secret and just going for the highest bhp figure you can achieve will not necessarily achieve that.

Porsche have always understood this and created the most torquey engines with relatively flat torque curves and excellent acceleration and understanding the issues have developed turbos and variable engine parameters (varioram, variocam, variable valve lift etc) to find a way to combine big holes and air flow for maximum performance at high revs with good torque at lower revs.

Usually to change a car designed for fast road use to a track car is a big adjustment and to tune an engine requires a lot of expense and patience.

We thought that a standard 944 250bhp turbo had the basic strength to tune and the potential to make faster.

The common route is to increase boost pressure but the downside is too harsh an increase in torque to control easily and a car not so pleasant to drive on public roads or tracks (for track days) unless many other track mods are combined.

With only 5 speeds and ratios designed for top speeds rarely found on tracks, the engine would need a wide spread of power to maximise the relatively large rev range it must be used in and then on top of that it should not spin up the rear wheels too uncontrollably. So an increase in capacity seemed a logical approach since it is cheaper than many other solutions, should increase mid range torque and make the engine more driveable yet if boost pressures were standard should also be relatively reliable and as max revs are not increased, should also be mechanically within limits.

Since with increased revs inevitably comes smaller “time areas” – if you were seeking more top end power - it would be necessary to compensate for this downside by increasing throttle sizes, inlet pipes, exhaust systems and turbo capacity – to increase the revs that the increased capacity worked in - whereas in contrast by aiming to increase power (or torque as they are directly linked) at lower revs - the existing valve time areas, inlet systems and exhaust pipes should be able to cope since there is more time available at lower revs to move the increased quantity of air. And since we were seeking more torque and a wide spread of power – this also fitted that requirement.

Planning to increase torque and power at lower revs and with a wider power band would be no good if we originally were trying to build something to loop a fast race circuit at high speed – but we were not - so the proposition fitted all the criteria and seemed an inexpensive solution to find a huge increase in performance.

Thanks NJH & Diver 944 for totally understanding all this and for expressing it in such a logical and accurate way.

I appreciate all the excellent work done by others like John Mitchell to push the boundaries of the engine even further - but we simply wanted to see if our predictions were right if we put together something we feel Porsche originally probably had in mind – and that others Worldwide could probably copy relatively easily without too much expense to provide something we felt would transform the driveability of the 250 bhp turbo – and so far it seems to have worked out OK.

Mind you after being forced to defend what we have done by the criticism from others – sod’s law would predict our track test will all go wrong and apply a liberal amount of egg to my face since the car has so far only been driven a small distance on public roads and is not set up at all for track driving or even developed – it has literally been bolted together and tried.

For the interest of 333pg333, the car is a standard 250 bhp turbo. The block is 968 as are the pistons (with just scooped out crowns to lower geometric c/r). The gearbox has a std LSD with thicker plates operating at higher torque. It has a 968 clutch plate. The injectors are larger and the ECU has been remapped. The waste gate spring has been replaced with a standard new one. The head is a std 2.7 head and 2.7 camshaft).

Our liners (used in other engines) are cast high strength aluminium heat treated, with Nikasil bores and our fitting system is a cross between wet and dry and our secret for now. I cannot see a problem fitting them to these engines and we are presently making a pattern in house to go ahead.

Getting as bit apprehensive now about or “test session” – no doubt you will all find out how we went on soon.

Baz

[hartech]

zcacogb, the problem is that turbos have a dual response - they spin up relative to exhaust flow and pressures. Because you have to lower the geometric compression ratio to avoid detonation (when the engine is running on turbo cylinder filling) - the performance "off turbo" at low revs is usually relatively poor. Then when the turbo eventually finds enough gas flow to start working it sets off a complimentary system whereby the increase in intake delivery is rapidly followed by an increase in exhaust flow that spins it up even harder and if the pressure in the system is set high this change is rapid. Different waste gates and maps can contribute to an improvement but the basic rules remain fixed - if the turbo is powerfull enough to deliver enough air at high revs to produce good power despite the reduced time that the inlet valves are open - then at lower revs - when there is more time - it takes more to spin it up (so is delayed) but once it spins up there will be a shock to the system (which incidentally is partially lost through rapid strain energy entrappment).

Fitting a bigger capacity engine with auxiliaries from a smaller engine will always accentuate the power band towards torque at lower revs and restrict power at higher revs.

Our interest was to see by how much it would flatten the power delivery. We thought that the engine while running almost naturally aspirated at low revs would never the less produce more power at low revs due to the increased pressure drops from the larger capacity engine and also spin up the standard turbo quicker as a result of the increased exhausr flow form the bigger capacity engine at low revs - to increase bottom end torque and reduce the sharpness of the turbo delivery.

We knew that the std turbo can deliver enough air to produce 300 to 320 bhp at higher boost pressures - so although our engine was bigger - that same volume of air - after ignition - would be pushing on an 8% bigger area of piston though a longer stroke so should still produce about 300 bhp at the top end which was ok if the lead up to those peak revs was improved.

We thought the extra capacity would improve driveability and mid range torque by about 20% and reduce the "turbo effect"- but the results - while similar in principle to what we expected - were better in magnitude.

333pg333 our liners have top support flanges to rest on machined faces (similar to those pictured on our web site for 996's www.hartech.org buyers guide etc) and ribs around the outside of the casting to improve hoop stress resistance and heat transfer through greater surface area. The aluminium used has similar strength to mild steel and better heat transfer and the Nikasil bonds better to cast aluminium while the cast surface provides effectively some oil retaining porosity and the expansion rates allow tight piston fits.

Thanks - that's right Niffty 951 - the car should be much more noticeably quicker in the higher gears - partly of course because it is in those gears for longer before changing up and therefore any benefit will apply though more time (more seconds) and be much more noticeable and partly because the resistance is increasing and the extra torque has more of a chance to do its work. Even so I still think that three quarters of a second advantage just in second gear (compared to an S2) should enable anyone to imagine how many seconds in a lap the car should gain.

I have really tried to be modest about all this throughout. I don't think we have done anything extraordinary and I am well aware that others have done more and deserve more recognition. I specifically invited Stuart and others to drive the car as soon as it was running because I didn't want "my opinion" to be downgraded by people anticipating exageration and I wanted a "second party" opinion.

I don't expect the track test to be reported as extraordinary because this was planned as a fast and pleasant road car - improving on the already superb 944 turbo - and its limitations on the track will become obvious and also the drivers are used to cars set up for track driving and may be unable to assess the engine - particularly as it doesn't feel to me all that fast.

I think this is because one thing Ajax 50 said was relevant - about the rate of change of acceleration - and how we interpret the sensation. Once moving we cannot detect constant speed (if our eyes are closed) and steady acceleration is easy to get used to whereas a rapidly changing rate of acceleration feels dramatic. Our turbo does not have that sudden kick from modest to rapid acceleration - it just steadily builds up and then feels pretty constant - which doesn't feel that quick. It is the same problem I had years ago with the 350cc Armstrong GP bike that the riders felt was slow even though they were lapping faster and achieving higher top speeds because it had an even torque spread.

I don't think my initial report that we had completed the engine was too dramatic and I simply expected some interest and responses. I didn't expect reports that it would only do 140mph, there was something wrong with having more torque and bhp up 6K revs by a mile than standard models or the various arguments about what torque is etc and in defending all that and trying to convey to those who may be interested or confused - what actually is important - I think I have inadvertently raised the profile of this whole exercise way above what I intended (or really what it deserves).

It really upsets me to read that so many people who clearly know quite a lot about engines and are interested - have misunderstood so many of the basics and I enjoy trying to explain when they have made mistakes to help them improve their understanding. I appreciate that this makes me come over as a big head and therefore a good target to be shot down - but you cannot be knowledgeable and at the same time convey to others who are arguing with you - without appearing arrogant at times - which I regret.

They say "those that can do and those that can't teach" and it is rare for those "that can" to also try and teach as well and I don't find it easy. However there are some of us that have built engines that performed very well and interestingly all those agree with me about the important issues and technical requirements. It seems only those that think they know it all, but have never put their knowledge to the test, argue so vehemently at times.

I am just really chuffed that the type of performance we anticipated was achieved and with the same ease we expected and there are only a handfull of people who were there before we started and knew what I predicted and they are soon to test out the results.

The ability to predict outcomes and then achieve them is perhaps the greatest test of an engineer and in this respect I am pleased with our creation and that it could be reproduced elsewhere quite easily. I don't feel the need to prove anything as it is a matter of record that the first time a 350cc triple I had built was run around a track (Barry Sheene at Brands Hatch) it broke the lap record, the first time a 500cc engine was raced it was the fastest through the speed trap at the TT, the first International race two of my 500's raced in they were 1st and second (NW 2000), the first time Neil Mckenzie raced my 350 (reported as uncompetitive by team riders) he won by half a lap, the first time my 500 cc square four was fired up and in its first 3 laps round Silverstone (for the British GP), it qualified ahead of the multi million pound Honda four stroke also being raced in public for the first time but after months of extensive development and in one TT sidecar race my engines won both classes.

Yes I am bragging (sorry) but only to reinforce to those interested in learning about engines and fascinated (as I am) by their development, what it is all about and who to trust when the "Internet experts" start stating their theories that are often misleading or plainly wrong. I didn't have any financial resources (other than what I could earn), very little development time, and had to make a living as well as invest time and money in ambitious projects - never having sponsorship back up, the best tyres, the best fuel or the very top riders - so getting it right at the design stage was essential and required a proper understanding of what was really going on.

I am therefore especially pleased some thirty years after these past engine exploits - because we cover these sort of issues here at work in training sessions and it helps my staff to believe in what I tell them (which is often different to what they read, were taught or believed) when things like this work out OK.

I am just a little sorry that what was always been a relatively modest project, that I was happy to slowly progress until someone else with more enthusiasm took over, has attracted so much controversial attention - but you cannot back off in this position once you have inadvertantly got there - and I suppose it makes for good internet reading?

Baz

[hartech]

Andy97 that would take some working out and for many different scenarios.

All would require a strip and rebuild - so would be over £1000 before doing anything else.

Starting with a 968 would need pistons modified, a 2.7 head and then all the 944 turbo auxiliaries.

Starting with a 2.7 would already have the head and bores but would need different pistons (possibly modified 968) and a crankshaft, rods etc and all those auxiliaries.

Starting with an S2 would require pistons and the 2.7 head + aux etc.

Starting with a 944 turbo would require cylinder liners, crankshaft, pistons but you already have all the auxiliaries.

Then it depends upon whether we are doing everything or just the engine and all the parts are just delivered for modification and reassembly.

Then again many parts required would have to be sourced second hand and there is a question of finding them, prices etc.

All in all it would be very difficult to fix a price at this stage - but I think the first stage is to manufacture some liners and id someone wants one building with just an engine initially - to take it from there.

Apart from modifying the pistons (that require a lathe) and suplying and fitting liners (if it has a 100mm bore) and finding or supplying shorter head studs in some cases, the rest is well within the capabilities of a "home mechanic"/enthusiast to do themselves.

I expect the mapping could be done reasonably now the parameters have been set.

Baz

[hartech]

Perhaps I should just add for clarity that our "liners" are more like replacement cylinders - as the whole of our casting fits in the crankcases and the top (where the coolant contacts it) contacts the outside of our liner directly and not through a part of the old casting in which we have slipped a liner. That's how we can add ribs to increase the strength and the cooling contact area and a solid tube is stiffer than two tubes inside each other. Photo of 996 "liner" on our web site explains better.


Baz

[hartech]

Some more interesting issues to tackle - but linked by torque.

Firstly – look – torque is simply a force – a twisting force – it has no connection with speed or time unless you decide to use it in a formula to create another measurement system using it. Like if you use a few constants and speed to create a reading for power or use some differences in speed to calculate acceleration.

Just because after doing that you can then reverse the formula to work backwards does not mean that when asked what torque is you could say it is something that you get when you have power or acceleration and apply mathematics or calculus to it – it is the other way around – it is the force that results in that power or that acceleration and unless you see it for what it is you will always get muddled thinking about how to use or change it for the best.

I still think the best analogy it to imagine removing the engine and fitting one of those torque spanners that bend against a scale to the prop shaft and then you sit in the engine bay and twist the torque spanner while someone with a fast shutter speed camera, photographs the torque reading on the torque spanner.

If you do not apply enough torque the car will not move – so no work is done, there is no power to calculate and no speed is achieved.

Once you apply more torque the car starts moving, work is done and at any speed you can calculate a figure to compare it to other cars (called power) to establish why some cars in this set up travel faster or further than others.

When you cannot turn the spanner with enough power to accelerate the car – the resistance to motion has equalled the torque applied (i.e. top speed).

When the torque applied is greater than the amount needed to keep the car travelling at a constant speed then the spare extra torque speeds up the rotational speed of the spanner and with it of the car – which is then accelerating (i.e. the rate of change of speed).

A more powerful person may be able to twist the spanner with more force and as a result it will accelerate quicker (as will the car) but this doesn’t mean that the twist he applies is a mathematical figure derived from the power or rate of acceleration – the twist he applies depends on the strength in his arm! Or even the tension in his muscles – just as the pressure acting on the area of the piston through the angle and radius of the con-rod does in an engine.

How much it accelerates depends upon how much extra torque is applied and as T=I*a (or acceleration = Torque/resistance), the more torque you apply the more acceleration results.

However in the above scenario – if we changed up a gear the car would accelerate more slowly because it is the rear wheel torque that accelerates it not the gearbox input torque (or engine torque). It is of course still proportional and anywhere the engine produces more torque it also produces more torque and power at the rear wheel in some proportion – but still the amount available to accelerate the car is increased or reduced by the gear ratio and final drive ratio in use.

So if an overall combined internal and final drive ratio is 2 to one say – the rear wheel torque in that gear will be twice the engine torque at any particular engine or rear wheel speed in that gear (if in this discussion we ignore internal losses as the same for each model or tuning scenario).

Fitting an S2 crown wheel and pinion shaft to a turbo gearbox (or just using the S2 gearbox) improves the rear wheel torque in any gear and at any revs by about 12.5% - so if you don’t need to travel at the original top speeds – is a good move (as the LSD can still be fitted) although it is not as strong a tooth profile and will fail sooner (which is irrelevant for racing I suppose).

However in our particular scenario – because we produce much more torque at lower revs – while a lower overall gear ratio would still improve the torque we have by 12.5% - we already have something like 75% more torque anyway – so it is not so relevant.

Indeed if you compare an S2 with our Turbo – say both changing gear from 3rd to 4th (at around 87 to 93mph) the S2 picks up at about 200 lbs ft torque and our turbo at about 340 lbs ft (engine torque) or 800 lbs ft at the wheel with the S2 or 1186 lbs ft with the turbo).

Similarly if you changed from 4th to 5th at 115mph in the S2 or 127mph in the turbo you would pick up at 200 lbs ft (engine torque 761 lbs ft rear wheel torque) in the S2 and 325lbs ft engine torque (1134 lbs ft rear wheel torque in the turbo.

This is an interesting comparison because if you went from 3rd to 4th to 5th in the S2 but missed out 4th all together in our turbo and went from 3rd straight to 5th (dropping to 3500 rpm) you would also pick up at 300 lbs ft (engine torque), the rear wheel torque would still be about the same (about 840 at the rear wheel in 5th at 3500 rpm compared to about 800 lbs ft with the S2 in 3rd) and therefore you should get similar acceleration to the S2 to start with and then still get an increasing acceleration all the way out of a long unwinding corner – to your top speed - without changing gear – such is the advantage of a torquey engine.

Someone described their 3.2 litre turbo as 4th gear feeling more like 2nd in a standard turbo – so I have looked at that too to see if the figures and calculations support that feeling.

The standard turbo in 2nd has an average rear wheel torque of about 1500 lbs ft whereas a 3.2 turbo with an average 400 lbs ft engine torque in 4th would produce 1400 lbs ft at the rear wheel (only 6.5% lower) and therefore very similar acceleration – yes it would feel almost as quick in 4th.

By comparison our engine would be about an average of about 1100 lbs feet in 4th so not as quick although it should be as quick in 3rd as a std car in second (producing an average of 1420 lbs ft in third) and we will test this out.

Now the argument about flat torque curves.

If theoretically there was a completely flat torque curve - then since power is torque * a constant * revs – the power would increase as a straight line with revs – i.e. would be double at double the revs. If you deviate from that straight torque curve to increase the torque at the top end – this will make that part of the graph steeper. And if as a consequence of trying to increase the torque at the top end you lose some of it lower down (which is always my experience) then that earlier part of the power graph will be lower and therefore also be steeper – so the change to a more top end power curve reveals a steeper power curve. In contrast if you change the engine so that it achieves more power at lower revs it can only do this if the torque is greater at those lower revs (so the point on the torque graph is higher) and if this has the inevitable consequence of reducing the power a bit at the top end then the torque at that top end of that graph will be lower. So the net result on the graph is a flatter graph.

So in general a steeper power graph will indicate a peaky engine whereas a flatter power graph will indicate a flatter torque curve.

Of course if you take an engine and somehow manage to increase the top end power without losing anything at the bottom – then in this case a steeper power curve will indicate a more powerful car and there will be more average torque – but practically – if you know how to do that – you are a very cleverer man. Similarly - if - as suggested you increse thw whole of the flat torque curve throughout the whole range then yes the resulting power curve would be steeper. But this seems to me to encapsulate the difficulty between theory and practice because in practice it seems almost impossible to increase the whole of the torque curve over the whole range by the same proportions everywhere. So while I must agree that in theory that is possible, to say in practice that a steeper power curve indicates more mid range or average torque is practically impossible and therefore could be misleading.

Zarcop – If we manage to fit the liners OK in a 2.5 engine then I think you could use a standard 2.5 head (with some modification) without too much detriment.

Finally, this whole exercise and the comments and replies has shown me that while many understand some if the issues, some do not still understand the important ones and even fewer have tried things to find out for themselves what is important and what is not.

Those that have eventually realise how to approach the improvement of an engine, firstly to establish what you want from it, secondly where you want it to perform (what top speed – what minimum speed, how many gears etc) and then after all that you can start to think about the best way to improve the acceleration while maintaining driveability (if that is also important).

More gears within the range of use will increase the average torque at the rear wheels and as long as the improvement in torque reduces the acceleration time more than the extra gear change time – then it is always worthwhile.

As tyres, suspension and engines have improved over the years – circuits have effectively become shorter and therefore more of an acceleration test than a top speed one and hence the ability to accelerate quickly - with control - using the relatively few gears you are stuck with (and therefore with relatively more wide ratios than are preferable for racing) makes the best choice of engine characteristics quite different to the purely theoretical route of more revs and peak bhp.

This exercise of ours is just a modest example of putting together a lot of these requirements, the coincidence of various parts from different engines fitting together quite easily and creating a whole new engine – that theoretically satisfies those requirements and then building it to test out the results.

Those who have built bigger and much faster 944 turbo engines and who have used new technology with variable vane turbos etc, etc and special fuels etc , developed new exhausts and waste gates etc etc have done much more work and with excellent results and are deserving of much more respect for it than us.

But most enthusiasts do not always have the resources to build or develop such incredible machines and we just thought that the results we got rather quickly and relatively easily and inexpensively (by comparison) were worth reporting on – so others may be encouraged to try something similar and out of it has come a commitment to manufacture a range of liners to potentially keep older engines with problems running or enable this conversion (or something similar) to be built by others.

I am sorry if in defending my teams work – I have disagreed with others and exposed their shortcomings – and I am even more sorry that this makes me seem a clever ba***rd – as I am quite happy to remain with comparative anonymity and get on with the work I enjoy here. It just seems to me that when these issues crop up with people obviously really interested in the whole technology of engine development and tuning and get so many basics wrong or misunderstand their relevance, and that when I can see their mistake, why they made it and how it will handicap their own understanding – that this is a part of the decline in our UK engineering and manufacturing in general and it is such a lot of wasted brain power that it would be helpful to them if I pointed out where I think they are going wrong and why.

Baz

[hartech]

If I seem a tad defensive it may be because I also receive private E-mails and telephone calls about my postings as well (arguing or telling me to do things differently) and they influence my replies - or it may be that I am getting too much like Victor Meldew or even that I pick up on more small mistakes people make in their replies and questions and it irks me that they are wondering about a technical issue while having a misconception about a basic rule and I just want to help put them right and stop them wasting time on a false premise.

Either way thanks so much for the positive comments and I will try and be more positive in future.

Baz

[hartech]

Yes a 6 speed box and coil overs at the rear would be great, (although if I was using one I would alter the power band characteristics to suit and lift the maximum torque through more boost, or a bigger turbo, and a larger exhaust and accept a reduction in bottom end torque as a result. But as someone else needs to take over the project I am just leaving it for now so they can do whatever they want. There is loads more potential because the boost is modest and much more to come from the engine.

We don’t really need a 6 speed box to prove our point because the wide spread of power/torque makes up for it – but if the inlet and exhaust was opened up – the cam changed (mainly to reduce overlap), and the boost pressure raised I feel sure it would reach the levels others have also achieved.

Regarding the torque – my emphasis has slightly exaggerated the advantages to compensate for the much held and misunderstood belief that all you need is to raise the revs and top end breathing to produce a higher bhp figure than anyone else and your car is perfect!

I am trying to encourage people see another issue to take into account as well such as track speeds, gear ratios, rear wheel torque etc.

Although torque and the right spread of power is absolutely right – there is an issue with power pulses and strain energy that is too complex to explain fully but basically means that more frequent power pulses (i.e. higher revs) with slightly less actual force on the pistons, can result in less power lost in the mass of the system and in strain energy and for the engine itself to accelerate itself just as quickly with slightly less torque at higher revs than at lower revs with slightly more torque. So the ideal power band is actually one with a slightly falling torque at the top end and the best place for maximum torque is probably nearer the bottom end of the power band (the range within the gear change points). You will find anyway that all engines have a falling torque graph but usually an increasing power graph at the top end of the rev range – which is about right. The mass of the system anyway has a damping effect on the power delivery making these changes feel less noticeable.

We have just made more torque than that at the bottom end to save costs and to improve driveability and it seems to have worked.

Driving experiences can be recalled by those driving the car in the next few days but the engine just feels to deliver equal acceleration whatever the revs above about 3000 rpm.

Anyway I don’t think I said that a falling torque curve would accelerate faster than before – what I said was “For those that quite rightly immediately recognised the achievement and the benefits of such a pair of graphs – can I add that although acceleration is proportional to torque (if resistance is constant or the same between compared cars) there is a phenomenon inside an engine and its transmission that relates to the stored energy between individual power pulses and strain energy lost as a result. It means that in actual fact there is a slight benefit from slightly reduced torque if the number of revs is increased and the strain energy lost is reduced. Put into more simple terms it means that a small increase in bhp at high revs can offer a small improvement in acceleration even if the torque is slightly lower and putting it simply it means that the best torque or power curve for the fastest acceleration, is one in which the torque is dropping off while the power is still increasing between the gear change revs”. Perhaps I expressed it wrongly because what I meant was that even when the torque is falling slightly and the revs are still increasing the actual acceleration still feels the same.

There are also usually benefits from slightly over revving an engine on reducing torque or power so that the revs you pick up on when you change up are higher – if for that particular engine the torque or power you pick up on at those higher revs is significantly higher (which it usually is) – but our engine has possibly too much torque at the lower end to make this relevant and the overall result should certainly be able to be improved upon – but probably with more expense and devotion.

Baz

[hartech]

My Report on the track test day.

A success, an experience and I learned a lot from it too.

Firstly a very strange coincidence – we rushed to grab a pit garage and there – sharing it with us – was the very person who I posted about recently - who originally owned the first 944 turbo we sold, tried to track it and found it too difficult to get out of the corners – that stimulated this project and there I was some 10 years later with the same model of car – having solved the problem – and he still has the 944 – spooky or what!

I was quite apprehensive about the engine working Ok and the conditions being too bad - when the track was damp, slippery (tree lined circuit in Autumn) and several red flags incidents, and here we were testing a bog standard car with huge torque on standard tyres! The whole idea was to invite others to test the car so the feedback was objective and impartial and so I decided not to drive myself. Although I would have been ok as long as nothing untoward happened – I was not confident that I would have had the right reactions if the torque spun the rear or a car in front spun in front of us in those conditions – and wanted the car to get through the test day in one piece.

Paul Follett was brilliant (thanks again Paul) he kind of took over the organisation of the day and calmly got things ready and set a professional stamp on the day which helped towards its success.

I did passenger and from that seat it was very clear that the engine did do what I wanted. From the hairpin at 2,700 rpm in third (around 40mph) the car would accelerate as quickly as the tyres would stand, under control, all the way to 6K. As I suspected we only used 3rd and 4th (except a brief snatch at second for the chicane). I was surprised how “short” the circuit had become with the chicane – virtually 6 short straights and tight bends and almost only using 2 speeds. With the box ratios allowing 4th to peak out just before most of the corners (so I think a lower pinion ratio may have not suited that particular circuit) although if it had 6 speeds and all that power band – you could always be in the perfect gear and the gear change is quicker in the newer boxes.

We had set the geometry as Paul wanted and it was spot on (tyres running even temperatures across the faces) and the lowering of the front he wanted - kept the tendency for the rear to squat under power - more under control.

Paul Smith drove and confirmed the power band was much as he expected from his own more powerful 3.2 version and I look forward to seeing the footage he took.

Later in the day the track dried enough for Paul Follett’s stickier club tyres and the car continued to take everything we could throw at it in a completely fuss free and relaxed way.

It helped me solve a problem that I didn’t fully appreciate before. Knowing that the std version with a bleed of valve fitted would spin up the rear wheels so easily, despite believing that a smoother torque delivery may be better or not spin at all (and reading the postings that others said the 50/50 weight distribution didn’t prevent the rear wheels gripping) I was still pondering on why all this extra torque didn’t spin up the wheels. If we wanted to spin them in a standard car we would snatch at the clutch because shock loading doubles the effective load - but we had double the torque so I kind of expected that to still be a small problem - but it wasn't at all. Then I realised that – just like a motorcycle doing a wheely – if the torque delivery is smooth enough to maintain traction (despite it being high torque) the torque reaction of the car is to transfer the weight to the rear wheels and lift the front – as given enough grip and torque – it too would eventually wheely like a motorcycle). Following one of the main Laws of physics (to every action there is an equal and opposite reaction) and providing the torque is not quite enough to spin the rear wheels, once the weight is being transferred - it becomes almost a self sustaining system where the extra weight then enables even more torque to be transferred with rear wheel spin – and so on.

I now totally believe those that told me so and have learned something valuable in the process – and am sure now that the front engine configuration could indeed handle more power and torque than we have available – providing the delivery is smooth.

Despite the huge torque dropping off as the revs rise – the engine didn’t feel like that and felt powerful right through to peak revs – although I now think that the engine would be faster if we could shift that curve up the revs by say 750 revs say pushing it more towards 340 bhp.

However that was never the initial intention which was to prove that the characteristics of putting together some standard parts, despite being high torque at lower revs, would produce good and controllable performance and still be ideal for a track car – at reasonable cost and within the capabilities of most home mechanics – which was proven.

The benefit has been that after giving up my previous involvement in racing and engine performance development 25 years ago (and like an alcoholic- having to maintain abstinence to keep off the drug since) – I found the day extremely lifting, exciting and it has left me feeling menatlly years younger. Although tired out last night I awoke early today early with a buz, an enthusiasm for the future and my head spinning with ideas – just like I had years ago. It feels like a great antidote to the big “R” and although we have forced ourselves to concentrate on providing excellent services for standard road cars and although I enjoyed the engineering side of solutions to improve and rebuild Boxster and 996 engines – I think the time has come for Hartech to embrace some involvement in some form of competitive Porsche motor sport activity in the future (more than the Boxster racing series that we are supporting with engine rebuilds next season).

I think the other shareholders and I will want to do this with a modern model more relevant to our daily income generating cars, and pass this Turbo on to someone else to enjoy - although I am intrigued by the fact that it seems that everyone who has built something similar has ended up with a similar power curve – whatever they have altered. It would be nice to solve that problem – what is the restricting issue – is it the turbo, the exhaust, the ports and valve sizes, the camshaft?

The fall in bmep with increasing revs is exactly in line with the torque graph (as they are proportional) so implies it is a mass airflow breathing problem rather than turbo delivery capacity – the same maximum air flow rates in and/or out of the engine being in use from 4000 rpm onwards if you reduce the volume by the reduction in valve open timing. I wonder if retaining the 16 valve head might have provided more scope?. I suspect from this 8 valve set up that it may be camshaft timing, and then possibly exhaust capacity – but any ideas or experiences from elsewhere will be gratefully received.

At the end of the day when we were reviewing the engine characteristics – I realised that apart from Paul Smith, no one else had driven a std 944 turbo – and although impressed with our car didn't know what all the fuss was about - but by luck another std one was being driven around – so a few words and the owner (who turned out to be a 944 specialist breaker and worth us knowing anyway - James Eaton of “only 9”) agreed to allow Stuart to drive his car and in return he had a couple of laps in ours. It is fair to say that Stuart was shocked at just how “laggy” the std car was (even with a waste gate bleed off Valve) and James really appreciated the big engined version and this brought home to Stuart just what a transformation we had managed to bring about (reported more on Porsche 968.uk).

Finally A big thank you to Rob (our technician), Paul Smith, Stuart Cookson, James Eaton and especially Paul Follett – not only for helping the day and the exercise becoming such a success but also for re-igniting my enthusiasm at a time when I am thinking of what I can do to stop getting bored in the future.

Baz

[hartech]

I think I have now managed to prove a few things to a few sceptics and also backed up the reasoning with scientific explanations but now need your help to decide on my next move.

I don’t know enough about tuning turbo charged engines as I would like to know and this project came about more from understanding a few key issues not really directly connected to turbo charging which was more the “vehicle” (if you will excuse the pun) for showing the difference a new/different type of power/torque characteristic can have.

(1) The time for an engine to breath reduces as the revs rise – so you have more chance of good breathing at slightly lower than peak revs – than at peak revs.
(2) Road gearboxes result in track cars having few gears and relatively long hauls in each gear before changing up again and so engines need relatively wide power bands.
(3) Acceleration is roughly proportional in some way to torque which is always higher before maximum revs.
(4) Smooth transition of torque prevents rear wheel spin up and allows torque reaction to increase effective rear wheel weight, enabling increasing grip.
(5) It is the average torque or power between the gear change revs that matters more than a high peak in one small area.
(6) Being unable to change gear ratios to suit circuits (as we did with our motorcycles) a power band that extends beyond the strict revs of the revs drops enables higher gears to be selected on first exiting a corner when the ratios do not exactly suit the circuit corners if the power band is wide enough to accommodate this.
(7) Increasing capacity while leaving remaining tuning as it was always increases the power band and torque and usually the bhp (but perhaps by a lesser amount).
(8) Smaller turbos than the engine capacity would ideally be run at for maximum peak bhp result in less snatch and a smoother and more torquey engine.

I thought all this could be proven by simply making the 944 turbo a 3 litre with standard running gear – and it has – but the problem is to decide where to go from here with my new found enthusiasm.

I agree 600 bhp is possibly obtainable but should I persevere with this car, or move on to a 986, 987, 996 or 997? Or should I do something in between.

In many ways the decision depends upon what it is that is holding back all the other versions like mine at the top end (as I think the engines could stand a shift upwards in the rev band and still retain the driveability and enough low down torque – or “grunt”) to keep the drive loading the rear tyres.

From my limited turbo knowledge it seems to me that the turbo exhaust must be running at a higher temperature and therefore increased volume and so evacuating it must be difficult. Perhaps therefore the exhaust cam should shut a little later than normal? Then it seems that too much overlap could result in back pressure limiting the intake charge (and we intend to test this shortly with a pressure gauge on inlet and exhaust and monitoring the pressures during a run). If this is right then the inlet must open even later than normal.

To test out different cam profiles with a single cam would be very expensive as there is no way to change the relative timing of the inlet to exhaust or overlap. However a twin cam could easily be modified to allow extensive test variations – so would it be better to let this old 944 turbo go to a new home and develop a 968 turbo with the 16 valve head using a cam sprocket that we can set in different positions to test out relative timings? We could even as a result retain the variocam use to broaden the resulting power band although I know we would have to then run with stand alone engine management but fortunately Rob (our technician with the 450bhp 1.6 litre drag Honda civic) is mastering that system on that car – so we should be able to set such a system up.

Is it worth considering this on what would still be an “old car” or to jump across to the Boxster or 996 and make a fresh start with one of those?

The air flow rates that 333pg333 stated for the std 8 valve and 16 valve head make me suspect the 8 valve head is the limiting factor and a 16 valve head may be the way to go (do you know what pressure differences that was measured at – because this would enable me to work out limiting boost pressures and estimate bmep etc etc)?

I also didn’t understand the reference to 293o inlet and 282o exhaust cam timing – perhaps someone could enlighten me.

If we did build something pretty special – using a 16 valve head we would have to increase the capacity again to continue with my preference for a torquey engine with wide power band but - is there any class we could then race it in?

In answer to the points about academia - I only ever use theory to try and understand what is going on and set objectives and a direction to go in but then always test out the idea by running with several practical set ups each side of what I expect to be the perfect answer to create a simple graph, from which I have always been able to see in a tangible way – what the benefit or downside was and where the best set up is.

My progress then is usually consider the technical issues, then create a couple of different set ups to test the theory and then – from the result – improve understanding and actual output. I don’t feel I know enough about this subject to know where to start. So if your contributions could simply state if they are from theory “T” or practical “P” experience – this would help me categorise the information better.

Whichever car and engine we decide to play with – I think it would be great fun to discuss the options and ideas on here, then build something some of us agree on and then all be a part of the results – wouldn’t it?

Well I need you ideas anyway – everything gratefully received – thanks!

Baz

[hartech]

There are more parallel threads on other web sites so various different issues under discussion. I don’t know about you but I am finding the more technical responses in recent postings - very interesting.

While I feel confident in my knowledge of naturally aspirated engines my experience of tuning turbos is relatively minimal and so now I am “hooked” by the experience I am interested in understanding the processes and problems and trying to work things out for myself while enquiring of others what they know – which has been most helpful.

I thought (for example) that the cam probably should have little or no overlap but bigger lift to compensate for the reduced timing and I thought that the exhaust on the 8 valve head would limit development and am now sure it will (even though I accept that it can be improved).

In considering what to do next I wanted to know what overlap (positive or negative) others were using and that is why I didn’t understand how the timing info without overlap or opening or closing degrees told me very much (sorry I didn’t explain myself very well).

I have got bitten by the bug and am in a position to make or try almost anything and so we look to the future considering what to do next,

I was extremely encouraged to see the results of my theory about torque etc working and also that the rear wheels can drive more high torque if the acceleration is smooth enough than I expected due to the torque reaction of the car and the extra weight loading can then sustain grip – and this has encouraged me to build a much more powerful version with a good torque spread. Acceleration is a strange concept and only ever an average between 2 speeds – the rate of change of acceleration being an interesting problem and almost certainly behind the lack of grip some turbos experience under sudden heavy acceleration - but it will be interesting to compare actual speed differences with the power or torque available and this engine allows for this – as it is possible to change up at different rev points and establish which part of the graph gave the best results (as Paul F thought it made little difference changing up @ 5K for example). So I am going to buy a more expensive road dyno to enable us to record full acceleration runs and use that to experiment with different power characteristics and finally see or prove once and for all whichever power band/torque delivery is quickest (to establish a curve we seek to replicate with more power/torque).


Using the old turbo we are going to test the inlet and outlet manifold pressures (already fitting the equipment to establish our view that the exhaust is choking the engine).

A single camshaft does not lend itself to allow easy adjustments to overlap and so I do think a DOHC engine would give more alternatives to optimise performance. I would also like to be able to change overlap with revs (which the 968 variocam could allow me to do) and so a 968 car (or a 968 head on the 944 turbo with external control) may be necessary to discover more – easily. As someone else said – this also would lend itself to eventually further develop more modern 6 cyl engines as they have the same system on board.

Developing the existing turbo would result in spiralling costs and eventual limitations so at the moment I think it is best to keep the existing 3 litre turbo as it is and offer it as a kit when I have worked out the costs.

We will also be making liners to repair all 944’s or alter their capacity and exploring the best pistons etc (none of which will be possible quickly).

I have had a few Eureka moments during my work with engines (perhaps the most relevant when I realised about the link between specific time areas and BMEP and that it was compression pressure that was the key and a variable compression ratio engine was the answer – normally aspirated). Of course a turbo is just that – by providing big changes to the pressure differentials it improves the compression pressure outside the scope of normally aspirated engines and as such varies the true C/R. The problem seems to me to be that while you can force more air into the engine by using a bigger turbo – you cannot force it out of the exhaust so easily and using a large exhaust and big turbo then results in it being top endy as the surge of acceleration makes the cars less driveable.

It makes no difference if I eventually discover that others have come up with the same idea (and it was in 1933 that I know of a variable compression engine first being built – but I didn’t know why or what the designer was trying to achieve until perhaps 30 years ago).

I have now had another Eureka moment of how to solve the problem of building a turbocharged engine to achieve the same bottom end torque as our turbo and a flatter torque curve to provide around 450bhp while keeping that lovely bottom end – but to experiment and build it I need to start with another engine. I will probably find that someone else has done it before as well – but for me it will be brand new territory and very exciting and satisfying.

As you can see I have well and truly got the bug again and am looking forward to testing this new theory in public.

I think absolutely that racing is the proof of the theory (and have had too many race wins with my ideas and engines to be excluded from that “club”) and I would intend eventually to put the results to the test on the track. I think I did that with this turbo – allowing others to drive it on the track after we had only completed about 20 miles of on road driving and with no alterations (apart from those suspension setting of P F).

I am not afraid of competition, discussions/arguments about all the technology or being wrong (as long as I learn why and grow) and find the journey both enlightening and stimulating.

I hope that by reading about all this – others will have learned something too.

Shame I also have to earn a living – oh well nothing worthwhile was ever that easy to achieve.


Baz