Exhaust Manifold

The original plan was to take the car up to Leeds to Tony Law and fellow R1 Striker owner Marc Bell and myself had agreed to share the costs of transportation and use one of our cars as the basis for the work. Through Marc's involvement with Alastair Massaralla and Custom Autotech, Custom Fabrications in Norfolk were recommended as someone being good and also local to us both. Therefore, the car, well chassis and engine, were booked into Competition Fabrications in Attleborough as soon as a had ordered as the first window they had was the week beginning March 4.

Nick Paravani at Competition Fabrications explained that he would build a manifold to our specifications in mild-steel or stainless. It became apparent that Nick was an extremely talented fabricator, but did not know a great deal about how to design a good exhaust, so the search was on to find something to copy.

In discussions with Marc, we both decided that a 4-2-1 system would be the best bet in providing good peak power plus a good spread of torque lower down to compensate for the loss of the EXUP valve. Looking at numerous aftermarket full systems this seemed to be the way that everyone had gone too with the good systems providing a good increase in peak power and no drop off around 4500 where the EXUP valve did its stuff. Unfortunately it is impossible to glean any design details from looking at pictures and dyno charts and for obvious reasons there was no information available on exhaust header design.

I had previously made contact with All Bikes in Kent, who raced an R1equipped sidecar. When ordering a set of their sump baffles/windage plate I enquired about the spec of their exhaust. They were more than happy to share the spec with me and as the system was sleeved and held together by springs were able to easily dismantle it and obtain accurate measurements. These were relayed over the phone and then later by fax in order to be 100% certain what we were talking about. Big thank-you to All Bikes!

Also when I picked up the kit from Sylva, Jeremy mentioned that another fellow R1 Striker builder, Tom Louther, had just had a manifold fabricated by Tony Law. I called Tom and again Tom was more than happy to share the specification of his exhaust. Big thank-you to Tom!

Both these systems were virtually the same with identical primary lengths and pipe diameters and a 20mm difference in the length of the collectors, which would mean slightly different power characteristics at low-mid range, but not really significant. However there was one way in which they were very different.

A 4-2-1 set of headers, as with any tuned header, relies upon constructive interference of positive and negative gas pressure waves to assist in scavenging. This basically means that an exhaust port open to low pressure in the header which assists the flow of exhaust gases out of the cylinder and sucks in a greater charge for more power. With a 4-1 header the length of the primaries is such that a high pressure gas wave will flow to the collector and reflect a low pressure gas wave so that it arrives at the exhaust ports at exactly the moment when the next cylinder is opening its exhaust port at peak power rpm. A 4-2-1 header's primaries work in exactly the same manner, but the collectors are shared across pairs of cylinders and the primaries and collectors are tuned in length and diameter so that negative gas pressure waves coincide from the end of the primaries, the end of the collectors and the combined length of primary and collector to provide improved scavenging at lower engine speeds as well as at peak power rpm. As with a 4-1 set of headers the lengths and diameters of the pipes are critical in determining the resonant frequency of the system, but unique to 4-2-1 headers the pairing of cylinders is vital to generating the correct sequence of low pressure gas waves.

The problem was that the All Bikes system and every aftermarket set of headers I have seen pair cylinders 1&2, 3&4 and the Tony Law system paired 1&4, 2&3. Only one of them was correct.

The theory behind 4-2-1 systems is confusing. Even more so because there is a mixture of crank and distributor degrees used to describe the theory and there are 2 crank degrees for every distributor degree. Before this became clear statements like the pairs need to be 180 degrees and 360 degrees apart were obviously inconsistent.

Picking pairs of cylinders that are 180 crank degrees apart means that they are also 540 degrees apart too which seemed plausible given the theory of 180 degrees and as much time between gas pulses as possible. However the asymmetry of this pairing just didn't seem right as one cylinder only would be getting a pull from the other.

Picking pairs of cylinders that are 360 crank degrees (180 distributor degrees) apart seemed nicer because of the symmetry across all cylinders.

In order to be sure I did some sketches of the gas waves within the headers:

The numbers represent the  gas pulses from a cylinder. A positive value is a positive pressure pulse moving out of the system. A negative value is a reflection back to the ports. This assumes that the gas pulses are travelling from port to collector in 180 crank degrees and from collector to secondary and reflected back to the primary in 180 degrees, which seems plausible as the collectors are just shorter than half a primary length and are larger diameter, hence lower gas velocity.

First thing to know is the firing order. Looking at the relative TDC positions of the pistons it can either by 1-2-4-3 or 1-3-4-2. Turns out it is 1-2-4-3.

With the All Bikes system pairing 1&2, 3&4  yields a pair of cylinders 180 and 540 degrees apart. Sketching the gas pressure waves and low pressure reflections that result from this showed no constructive interference, or scavenging, at the exhaust ports other than that from the end of the collectors where all four ports reflect back to each other as with a 4-1 system.

However pairing 1&4, 2&3 produced low pressure gas waves exactly co-incident with the next exhaust port to open in the firing cycle. This just had to be correct. In both cases it was assumed that the engine was operating at the tuned rpm and that exhaust gases would travel to a join in the piping and reflect a low pressure wave back up the neighbour pipe to arrive at the port in exactly 360 degrees of crank rotation. Low pressure waves result from the merge from 4-2 into both collectors and from 2-1 from collector to secondary pipe. With the collectors being almost exactly half the length of the primaries this causes perfect alignment of low pressure waves at the correct port from the primary of the previous cylinder and the secondary of the one beforehand. I came to the conclusion that Tony Law's reputation of knowing what he was doing with bike exhausts was well deserved and that the aftermarket 4-2-1 headers I had seen that combined 1&2, 3&4 just did not work. Perhaps there are packaging limitations on the bike that prevent the combination of 1&4, 2&3. Perhaps not. Only Akrapovic have got it right and that's because they use a 4-1 system.

The final specs of the exhaust were as follows:

4-2-1 system with 550mm long 35mm internal diameter primaries, 280mm long 42mm internal diameter collectors between 1&4, 2&3 and a 2" external diameter secondary (outlet).

Here is the finished result and another rather tasty bit of fabrication work due for collection the same day.

Home Mail your comments