Sodium filled exhaust valves - not just a Norton issue

Recently I wanted to throw in a ‘curve ball’ type question to an interview for a young engineer to test their problem solving. The curve ball question I came up with related to the exhaust valve illustrated below.

Works Norton sodium filled valve

So this is no ordinary exhaust valve. 
It is a Works Norton Valve from one of their famous 500cc single cylinder machines. What is most notable is this is not a valve fitted to one of Norton’s manx production racers. The stem is the giveaway; and what a giveaway that is given that it is ½” in diameter!

Looking back now it is fair to say this is ghastly design. But this is where we need to put our problem solving hat on and consider why it has such a structure. The larger diameter is of course filled with sodium. Sodium is a group one metal which becomes liquid at relatively low temperatures. The liquid sodium can convect heat away from the valve head.

The exhaust valve is also noteworthy for its convex head. Once again this shape is chosen not to increase compression ratio (there are far easier ways to do this, e.g. material on the piston etc), but is instead incorporated to take more heat out of the combustion chamber.

So the convex head and large stem filled with sodium all points to increasing heat transfer away from the combustion chamber and most importantly away from the valve head itself. 
Why was this required?

Many far higher performance engines can operate perfectly satisfactorily with conventional solid exhaust valves.

Once again we need to look back at the period when the engine was racing; post WW2 when petrol supply was in short demand and resulted in low octane ‘pool’ type fuels. The lower octane fuels had a far greater tendency to result in engine ‘knock’. As we know there are a few mechanisms that cause uncontrolled combustion in a spark ignition engine (knock, pre-ignition), however one item that can be big influence is hot spots within the combustion chamber. It is no surprise that an exhaust valve is one of the hottest parts of any engine. 

So there we have it, on the very low quality petrol available post WW2 Norton tried all they could to prevent hot spots in the engine (hence heavily cooled exhaust valve) in order to prevent engine knock and allow them to use higher compression ratios and higher levels of ignition advance.

But the engineers of the time knew exactly this and were well aware that if different fuels were used, a sodium filled exhaust valve may not be required. In the words of Steve Lancefield when talking about Manx engines for 500cc racing cars which could run higher octane alcohol fuels having greater charge cooling characteristics:

“Yes, most of the Norton engines I prepare are fitted with ‘Sodium’ exhaust valves – it is open to question whether these are really necessary on such well-cooled engines using alcohol, particularly as very good results have been obtained with engines using non-sodium valves. However, for what it is worth, the sodium valve is a shade lighter in weight than its solid counterpart and the use of this type of valve is on this score alone worthy of consideration. When using hydrocarbon fuels, unhesitatingly – sodium-filled exhaust valves please!”

Sodium filled valves are however not just a remnant of engineering history.  Modern turbo gasoline direct injection (TGDI) engines with their very high specific outputs are also now requiring such technology once again to reduce valve head temperatures to prevent knock and allow higher compression ratios, higher boost pressures and more ignition advance. The images below details types of exhaust valves that can be used on engines and their influence on the valve surface temperature. The sodium filled valve (A) also utilise a hollow head filled with sodium to improve heat transfer – striking similarities to the issue and solution Norton came up with over 60 years previously! The sodium filled valve results in stronger valves (further from the fatigue line) with lower surface temperatures (better for knock) and also having a reduced weight compared to the solid equivalent.

Exhaust valve cross sections

Exhaust valve surface temperature and strength at peak power conditions in a modern TGDI engine

Edgar Franks - Norton Engineer

I was in a conversation with an old racer the other day about the development of Manx Nortons and the name Edgar Franks cropped up. I am sure many would have heard of Joe Craig who developed Norton’s Works racing machines, but the name Edgar Franks would be known by very few.

Edgar was a mechanical engineer at Norton who worked on many machines over the years. It was Edgar Franks who assisted Arthur Carroll and Joe Craig in the re-design of the Norton over-head-cam engine after Walter Moore left for NSU. Edgar also redesigned the Norton range for 1931. Changes he made included making new lower frames which made the rider adopt the knees fully bent riding position due to a saddle height of only 26.5”. Other changes included Norton making their own hubs, brakes and Webb-type forks, instead of buying in components.

In 1933 Franks designed the Norton model 50 350cc OHV motorcycle. It was also Edgar Franks that designed the Norton oil-bath primary chain case which was introduced in 1934 and used until the 1960’s. The first telescopic forks for a production Norton machine were designed by Edgar Franks for the 1940 Manx model. 

In January 1950 it was Edgar Franks who took Rex McCandless, the Featherbed frame and Rex’s jigs to Reynolds Tube Company. Reynolds went on to manufacture the Featherbed frames for the Manx Norton racing machines.

One of Edgar Franks’ biggest contributions was in the development of the Manx Norton production racing machines. For over two decades Franks continuously improved Manx to keep it competitive. It is a common mistake that Joe Craig developed the Manx machine – instead Joe Craig was in charge of the Works racers. So without Edgar Franks hundreds of racers would not have had such an effective racing weapon as the Manx Norton.

The Anatomy of a 1954 Works Norton Featherbed frame

I thought I would make this post to give an idea of the difference between a Works Norton racing machine and its Manx production racing cousin.

We are only looking at the frame here for a 1954 500cc Works Norton. Striking differences can be seen immediately between it and a Manx Featherbed frame. The radius of the curve in the main tube as it goes from the top tube to the gusset plates on the works frame is smaller than on a Manx Featherbed.

Modifications to the rear mudguard loop carried out at the 1954 TT

500 stamped on drive-side gusset

It can be seen that the rear mudguard loop has been chopped off the frame pictured here. This was done in a hurry at the TT in 1954. An extra bracing tube was added between the top tubes at the same time. This frame is for a 500cc Works machine, as indicated by the stamping to the gusset on the drive side.

Pannier tank mounts and bowed drive side main tube can be seen

The drive side bottom tube is bowed out on the Works frame in order to clear the outside flywheel that was used on the 1954 machines. The timing side gusset and bottom tube are also relieved slightly in order to allow the fitment of a 5 speed Burman gearbox.

Oil cooler mount

The frame incorporates many Works specific features. On the top tubes there are mountings which hold the pannier tanks on either side of the machine. The timing side front down tube has a bracket added which supports the oil cooler which was used on the Works machines.



Front fairing mount



Weight saving inside the steering head

Front fairing mountings are also different from that of a standard Manx. The 1954 Works bikes sometimes utilised the rather unusual looking Proboscis fairing. The steering head on the Works frame is also undercut inside in order to save weight.



Rear chain oiler



The Works frame is made so that it holds lubrication oil for both the primary and final drive chain. The 1954 Works machine used a floating rear brake, with the brake-plate being unusually on the timing side of the machine. In order to facilitate this, the frame had to have a mounting for the brake torque arm on the timing side.

Rear brake mounting and modifications to clear the Burman 5 speed gearbox

As well as incorporating all the special Works extras not seen on a conventional Manx, the frame is of course much lighter than that on the production racing machine.