BIGGER IS BETTER
One day, the engine in your 911, Boxster or Cayman might need a rebuild. It’s a costly business, admits Chris Horton, but for perhaps an additional £1500–£2000 you could also increase the cylinder capacity for a worthwhile increase in power and crucially
Power from torque, or power from revs? Hartech reckon that torque talks and we’ve tested its latest 3.7-litre and 3.9litre ‘oversized’ engine conversions
There is, as we car enthusiasts like to say, no substitute for cubic inches. It’s a commonly quoted automotive ‘proverb’, one of those hoary old aphorisms that we use without considering their real meaning, but like most of its kind it is based upon easily observable fact.
American ‘muscle cars’ have for many years relied on the sheer size of their usually V8 engines to generate the power and specifically the torque needed to deliver the performance for which they became famous – albeit usually in a straight line only. And here in Europe, where smaller four-cylinder motors are the norm, both after-market and DIY tuners long ago realised that often the most effective way of significantly increasing a given unit’s output, its ability to propel the car satisfyingly quickly and easily from ‘A’ to ‘B’, is to increase its capacity, or in more scientific terms the swept volume of its cylinders.
This capacity increase, all things being equal, allows the engine to burn more fuel in the larger quantity of air which – if the induction and exhaust (and ignition) systems are up to the job – it can draw in to and then expel from its larger combustion chambers. That pushes the pistons with greater force, which in turn turns the crankshaft with more of that same energy. It’s much the same principle as throwing a stack of hefty logs on a sputtering campfire to create a bonfire, and thus generate more heat.
Porsche, too, has long been an enthusiastic proponent of ‘oversizing’ its engines for more power and torque; for upscaling, perhaps, to use a modern term. In standard production form the naturally aspirated air-cooled 911 expanded from its original modest 2.0 litres to 2.2, 2.4, 2.7 and 3.0 litres, and then finally to 3.2 and 3.6 litres. The water-cooled M96/M97 grew from 2.5 to 2.7, 3.2, 3.4 and 3.6 litres, and finally to 3.8. (The subsequent ‘shrinkage’ of the current gen 2 9A1 to 3.0 litres is another matter, prompted by the need to reduce overall exhaust emissions, but is arguably more than offset by turbocharging.)
Oversizing is not necessarily a simple process, though. At the very least it requires the engine to be stripped, and then the cylinder block to be machined to the size necessitated by the chosen larger-diameter pistons. Attention must also be paid to smaller but no less important details such as the resulting compression ratio (which will naturally increase), piston-to-bore tolerances, valve sizes, cylinder-head gasket(s) and, for optimum results, camshaft profiles and timing, ignition settings, and not least the intake and exhaust systems. And obviously you need to make sure that the rest of the engine’s structure is strong enough, too.
Arguably by far the simplest engine to modify in this manner is the air-cooled 911, with individual and easily replaceable cylinder barrels suspended from the external faces of a vertically split crankcase by through-bolts, once you have removed the camshafts and the individual cylinder heads. (Likewise the VW Beetle engine and its Porsche variants, of course.) That’s the way Porsche generally did it, and the reason why, even today, a good set of crankcases from even a 1965 2.0-litre can with the appropriate barrels and pistons become the basis of a no less than 4.0-litre motor. There can’t be many other designs with that kind of versatility.
It is, unsurprisingly, a different story for the water-cooled flat-sixes, designed for the easiest and cheapest possible initial assembly processes, and in effect as sealed-for-life units probably intended, like most massproduced modern engines, to be discarded when they wear out or go wrong. Few, if any, enthusiasts, then, would in this case routinely embark upon a tuning programme based first and foremost on capacity. A remap and an after-market exhaust system will deliver a
cost-effective improvement quite sufficient for most, and ultimately it is cheaper and easier to fit a larger second-hand unit – or perhaps just to buy another car.
But the times they are a-changing. History has shown many of the water-cooled engines to be disappointingly fragile (although some do manage to notch up impressive mileages; we know of several 2.5s that have happily exceeded 200,000), and a significant subindustry has sprung up to cater to the needs of enthusiasts who, understandably, don’t wish to consign to the scrap-heap a highvalue and certainly highly desirable sports car that they justifiably cherish.
Pre-eminent among those specialists here in the UK is Barry Hart, since 1985 the tireless engineering talent behind Bolton, Lancashirebased Hartech, and in terms of the molecularlevel metallurgy and thermodynamics of Porsche’s water-cooled flat-sixes probably the most knowledgeable – and certainly the most boldly and painstakingly innovative – man outside the factory itself. His long career in the automotive industry began way back in the late 1960s, and includes both the designing and manufacturing from scratch of a number of race-winning motorcycle engines and gearboxes. A beginner he is not.
Barry’s first reworking of the then contemporary M96 was as long ago as 2002, when it was becoming painfully apparent that the standard offering wasn’t quite as robust as we had all hoped, and since then he has gone on to develop solutions covering just about every aspect of the units. One of his earlier upgrades was to fit a purposedesigned bracing collar between the top of each cylinder and the surrounding coolant jacket, to prevent the former distorting and cracking, and it was surely no coincidence that Porsche itself later adopted this so-called closed-deck construction.
Thus far Hartech has in one way or another reclaimed at least 2000 of these power units for owners right around the world (there is at least one Hartech engine in Iran), and even now, with the potentially affected models rapidly ageing, and being replaced by more modern and almost certainly more reliable versions, there is no sign of the steady flow through the company’s busy machine shop diminishing. Like any genuine enthusiast, however, Barry couldn’t help thinking that for all its flaws – entirely fixable, thanks to those repair techniques and modifications – this was a design that had still more to give.
With time on his hands, then, now that the day-to-day running of the business is ably handled by his stepson, Grant Pritchard, and with a team of long-serving technicians and machinists behind him, Barry has spent the last three years devising a programme of capacity increases applicable to pretty much the entire range of M96 and M97 engines (but not the MA1/9A1 from the gen 2 997 and the gen 1 991). All are designed around the holy grail of the easily accessible and enjoyable performance that comes not just from their headline-grabbing power output, but primarily from the increased torque available. (See pages 84–87 for an explanation of the crucial differences between power and torque.)
Not for a moment is Barry suggesting that anyone in their right mind will spend the best part of £10,000 on an engine rebuild purely for the sake of a capacity increase somewhere between 200cc and 500cc, and increases in power and torque of around 15 per cent at the very most. His programme is based on the perfectly reasonable assumption that any engine receiving such an upgrade requires reconditioning anyway, either because of a failure, or because its owner wishes to avoid one. In which case, for the sake of perhaps just £1500–£2000 plus VAT out of that £10K bill, why would you not go the extra mile? Or in this case, perhaps, the extra cubic centimetre?
Central to each oversize conversion is a set of six Nikasil cylinder inserts and matching pistons, designed by Barry himself and manufactured under licence by Capricorn in Hampshire, and in principle no different to
those used in all of the company’s engine reclamation jobs. Thoroughly tried and tested, in other words – and things of undeniable beauty if you should be lucky enough to hold them in your own hands. Significantly, all have a nominal working diameter of 100.0mm, finished capacities (see chart below right) determined by the stroke of the crankshaft.
The original Porsche cylinders are machined out and the new ones pressed in, and crucially with a supporting ring not just at the top of each bore, but now also at the bottom. These are themselves drilled (see photo above), partly for lightness, but primarily in order to facilitate oil drainage back into the sump. There is an additional supporting interference fit between the block and roughly the mid-section of each cylinder; and the upper area of the tube, inside the coolant jacket, is ribbed to give additional surface area and thus more efficient heat transfer.
Cooling is further assisted by the opening up of the coolant pathways between the top of each block and the cylinder head – another proven Hartech modification to aid cylinderbore longevity. That said, Nikasil cylinders are inherently resistant to scoring, because the cylinder surface does not consist of hard silicon particles trapped in an aluminium matrix that gradually comes loose and scratches at the piston surface, but instead a homogeneous electroplated surface that remains permanently fully bonded, like a complete thin tube within the aluminium.
It would for similar reasons be the brave owner of an earlier engine who turned down the option of an intermediate shaft modified to take the later larger-diameter IMS bearing. This can be installed in either the earliest sprocket- or the slightly later gear-type shaft – and that first shaft is itself ‘updateable’ to the subsequent gear-style item if required, although naturally that requires the matching chain, and various other parts. Famously, that final iteration of the IMS bearing cannot later be replaced without splitting the crankcases again, but such appears to be its dramatically improved reliability and longevity that this should hardly be an issue.
It would be understandable if at this point you are imagining all sorts of exciting possibilities – turning your Boxster 2.5 into a 3.9, for instance; and remarkably every single one of these Porsche engines has the same overall external measurements – but not surprisingly there is a practical limit to what is possible, and Barry has no less sensibly set out just five options. (And, based upon the simple but often overlooked fact that today the 2.5 and 2.7 rarely, if ever fail, neither of those units is considered to be a viable basis for conversion.)
In order of original size, then, the range commences with the 3.2-litre Boxster ‘S’, which with those 100.0mm pistons in place of the original 93.0mm items gives a nominal 3.7 litres. The 3.4-litre 996 (96.0mm bore, again raised to 100.0mm) becomes a 3.7, and with the additional fitment of the appropriate longer-stroke crankshaft the 3.4-litre Cayman ‘S’ can be stretched to 3.9. The 3.6-litre engine (96.0mm) is another good candidate for the 3.9-litre upgrade, not least because it already has the same freebreathing cylinder heads as the 3.8, and likewise the 3.8 itself can easily be given those extra 100ccs, although obviously by this stage any benefits are beginning to be outweighed by the costs.
Precise increases in both power and torque depend on a number of factors, but generally, as we’ve said, amount to some 15 per cent, and with a no less useful lowering of the revs at which they are achieved. Both the 3.6 to 3.9 and the 3.4 to 3.7 will run with the original engine management system, we understand, although perhaps unsurprisingly the 3.4 to 3.9 will benefit from a remap – as, of course, will those smaller versions. There should also be an improvement in the engine’s overall efficiency and thus its fuel consumption, although since you would have to be as abstemious as a Franciscan friar not to use the additional performance at every available opportunity, that is unlikely to figure in any calculations. It is certainly not something that Barry Hart is championing.
For him – and, as you will see elsewhere in this story, for us, too – it is all about the torque, and the resulting improvements in not just the cars’ 0–62mph times, but also in realworld, mid-range acceleration. The ability not just to maintain a relaxed motorway cruising speed, but also to overtake swiftly and safely when needed. Sounds ideal to us. PW
Increases in both power and torque depend on a number of factors
One of Hartech’s most significant upgrades to M96 and M97 engines is to fit a supporting collar between the top of each cylinder ‘barrel’ and the inside of the surrounding coolant jacket, to prevent distortion of the bore and even cracking. Porsche itself eventually adopted this so-called closeddeck construction. Note, too, the subtly cut-back areas where coolant passes between the block and the head
From left to right: Hartech director and service manager Tobias Higgins; company founder Barry Hart; and Grant Pritchard – Barry’s stepson, and now MD. Barry began his career designing and manufacturing race-winning motorcycle engines and gearboxes, including one for the late Barry Sheene, although sadly he never competed with it. Barry (Hart) also designed the innovative squarefour motor for the Phoenix 4 bike that featured in the 1980 British movie Silver Dream Racer, starring David Essex
Belt and braces: it’s not just the middle and the top of each cylinder that is firmly braced against the crankcase. Latest Hartech innovation is to add these drilled rings at the base of each tube – holes are partly for lightness, but primarily to aid oil drainage. Pistons and cylinders (top) designed by Barry, and manufactured under licence here in the UK by Capricorn. Both first- and second-generation intermediate shafts, with small-diameter bearing, can be modified for the larger-diameter third-generation job (above, left of pic), and which has proved itself virtually unbreakable. Chart below gives a good overview of various capacity options