To scatter or not to scatter, that is the question.

This whole 'scattering' deal has befuddled many. The only real prose produced on the subject, and 'advice' handed out by many so-called 'specialists' tends to be somewhat tainted with large chunks of mysticism - inferring there's some kind of black art involved. The simple fact is the principle is very basic and easy; it's getting a cam with the right profile and figures applied to it that will actually work as a scatter profile that's the hard part. And this article does not mean I am all for scatter pattern cams. I will illuminate…

We are all pretty much aware that the A-series is a real oddity, largely because of the siamese (shared) port work - cylinders 1 & 2 share an inlet port, cylinders 3 & 4 share an inlet port, and cylinders 2 & 3 share an exhaust port. Very weird compared to the more familiar head designs found on almost every other engine type that has one inlet and one exhaust port per cylinder. You only have to observe how long the inlet valves are open together on one shared port to realise the potential problem this could cause. Next time you set the valve clearances, or now out of simple curiosity, turn the engine over by hand - or manually some other way - with the spark plugs removed for ease and rocker cover off. Watch one pair of inlet valves throughout their open and closing cycles. It's quite astonishing how long they are both open together isn't it? A situation worsened by sportier, longer duration camshaft. Consequently this shared port deal brings about it's own set of problems - cross-port interference is one, different flow potential another.

Cross-port interference is where one cylinder is effectively wreaking some kind of havoc on the other cylinder sharing the same port. On the inlets this amounts to some incoming charge robbing by one inlet valve from the other, on the exhausts a possibility of disturbing extraction and shock wave interference that may inhibit flow as well as ram-charging. Although cylinders 2 & 3 exhaust through the same port, they do so 360 degrees apart, so there is no cross-port interference of any sort going on there. Incidentally - that cylinder robbing is commonly expressed by the aforementioned 'specialists' as the sole reason for scatter pattern cam use. Just goes to show how 'specialist' they are as this simply isn't so.

Different flow potential in this case is concerned with the exhaust ports. The end exhaust ports not only have the luxury of one cylinder/valve per port, but also have a much better shape to them. The middle two cylinders not only share a port, but have a far less effective port shape too. An expertly modified cylinder head modified to give the best flow figures for any application (road or race) invariably has greater flow out of the end two ports than the centre port. It is obviously possible to limit the modifications to the end ports to keep them similar to the centre port - but that could be wasting power potential.

When considering the problems in hand, it is easier to cut the engine down into two two-cylinder engines rather than looking at it as one four cylinder engine, treating cylinders 3 & 4 as mirror images of 1 & 2. The basic principle then is to arrange the cam lobes to maximise performance by minimising cross-port interference and maximising port flow performance caused by the port sharing and mismatched ports problem. Told you the principle was easy. I suppose the simplest explanation of what a scatter pattern cam does is to say it reduces the shared time of the siamesed inlet ports, thereby reducing the negative interaction of the inner cylinders by the outer ones.

However, before any cam timing figures and phasing can be considered you need to appreciate a few things. The inner cylinder of each pair draws fresh charge first, followed 180 degrees later by the outer one - as observed by the little experiment outlined above, the valves are open together for a great deal of time. Hence the outer cylinder is robbing fresh charge from the inner one, further exaggerated by the scavenging action at the end of the induction stroke of the outer cylinder over-powering that of the inner one. And the inner exhaust port is fundamentally less effective in terms of airflow capability compared to the outer one; one side effect being different scavenging periods.

The theoretical solution to the problem then is to scatter (re-phase) the cam timing to optimise all four cylinders as follows -
- one set of timing figures for inlets 1 & 4
- a different set of timing figures for inlets 2 & 3
- a specific set of timing figures for exhausts 1 & 4
- a specific set of timing figures for exhausts 2 & 3
This is generally achieved by reducing the lobe centerline angle of cylinders 2 & 3, widening it on cylinders 1 & 4, exhaust timing similar or shorter of cylinders 1 & 4 than their equivalent inlets and exhaust timing longer than the inlets on 2 & 3. That's a proper scatter cam. Simpler forms are used where the exhaust timing is less exotic - largely because of cost and development work.

To grind a simple scatter cam is approximately twice the cost of a more standard single (inlet and exhaust timing the same) or dual (inlet has different timing to exhaust) phase camshaft because the cam has effectively to be ground twice. To grind a 'real' scatter cam would cost nearly four times as much. Then there's the development work necessary to evaluate exactly what is needed - you can't simply apply the scatter phasing to current off-the-shelf profiles. Some will work, some will be horrible. The development costs would be immense bearing in mind the time it takes to actually grind a real scatter cam, the all the dyno work to evaluate the profile, and how sensitive it is to engine build spec and components used. So the big question is - is it worth it?

In my humble opinion, the simple scatter cams that are already available that actually work are. But only on cams that have been specifically developed to be a scatter profile where the duration is less than 300 degrees or have been found to work by accident. This narrows the field down a lot - the Kent MD310SP works great. I have little experience with the scatter profiles offered by APT; however they have put far more research and testing into this cam phenomena than everybody else put together. I absolutely do not personally recommend using scatter profiles based on off-the-shelf profiles having durations less than 300 degrees. That means stuff like Kent's MD286 and 296 profiles. I have used the 296 extensively, and had very good results with it. I have used the 296-scatter pattern version four times, and all four engines were dogs - a situation I'm not at all used to. Others I have spoken to seem to think the 296 scatter is a fine cam…

From my limited experience of scatter pattern cam use, it would seem that the whole concept is seriously fine-tuning the A-series to the point of being an art where they will work handsomely if properly and specifically developed - but only when used with an exact component list and engine build that duplicates the test engine exactly. Either that or by complete accident a set of components and engine build accrues the environment where a previously abhorred scattered profile works.

Ultimately, providing a single phase cam has been correctly developed for the application – then a scatter cam is by and large redundant, as proven more recently (2006-onwards) by the range of cams derived at by Australian A-series guru Graham Russell. Whilst the MD310SP has been used for decades and accepted as possibly the best all-round race cam that improves torque and low-speed capability over other non-scatter race cams, Graham's 308 single phase cam gives it a right drubbing in the low-speed and torque area. True it may give one or two HP away at peak, but believe me – torque and low speed performance makes a car quicker on the track.