HOT AIR; About Branch Manifolds (Headers)

Developing branch manifolds in general and motorcycle header pipes in particular, requires endless hours on the dyno making comparitive tests.

We tested four into one headers with conventional four into one configuration collector, (i.e. not with a rotational 1-3-4-2 firing order collector), using various header diameters and lengths.

Once we were happy with primary lengths and diameters, as well as link pipe and silencer canister spec for maximum power, we tested the same header lengths and diameters converging into a rotational firing order collector. Each and every one of these dyno tests was done on the same bike using the same intermediate pipe and silencer which we had already developed previously.

These tests were conducted over a period of some three weeks under controlled conditions:

The carbs were jetted to perfection. The ignition timing was optimized, the coolant temperature was maintained at a constant level, the relative air density was taken into account, and so on.

The results were conclusive:

The conventional four into one i.e. 1 and 4 header pipes converging together on top of 2 and 3, made the most horse power, which was not surprising because up until that stage virtually all four into one bike headers had that configuration for track and road use.

However, the headers with the firing order collection made more  midrange power, ultimately making the bike easier to ride, resulting in better lap times even though there was no advantage on top end power. Not surprisingly there was even a slight increase in top speed down the straight, the reason being that because of smoother power delivery out of the corner into the straight, full throttle could be applied sooner, thereby exiting the corner faster, allowing for longer acceleration time.

The next step was to develop a four into one collector which would maximize flow from the header pipes into the intermediate pipe. We tried a merge mandrel – bent pipe collector:

These collectors were constructed using mandrel bends of various angles (we tried angles between 7° and 15°) which were cut accurately to join the four pipes and ‘gasflowed’ into the intermediate connecting pipe with the minimum flow restriction possible. A full day was spent in fabricating each collector and a lot of time was spent in the internal surface finish. The resulting pyramid point in the centre was ground to a polished finish.

The results were rather disappointing. Comparing the power and torque graphs we came to the conclusion that the extra time and effort were not worth the bother.

Some time later though, we developed a branch for a racing Golf and a special slip – on mandrel – bent merge collector was used initially. When tests were conducted with a conventional collector, this time the results showed conclusively that the special collector worked a treat!

There was an improvement in power all the way from 4000 rpm to 8500 rpm.

All this just proves that general theory is not enough when developing effective branch manifolds. Practical testing is absolutely essential to get the optimum result.

That’s all for now!

Abel dos Santos