Tuesday 23 March 2021

Exhaust Rings

This might not sound like a very weighty topic but the exhaust rings are pretty important for attaching the exhaust pipe to the exhaust port. And, here, there was a problem!

Exhaust rings for various Velocette models can be purchased from GroveClassic Motorcycles. The picture below shows one of these on a Mk IV KTT cylinder head in my workshop.


These are well made and work absolutely fine ….on a single cylinder Velocette engine installed in a Velocette frame. However, there was simply not sufficient clearance between the exhaust port and the frame tube on the front cylinder of the V-Twin to fit one of these.

Luckily, I also have a very early Velocette Model K flat-tank that is awaiting restoration and these earlier bikes use a much smaller finned exhaust ring, Velocette part number K66.


I tried this on the V-twin front cylinder exhaust and it fitted without fouling the frame tube …just! However, I am not aware of anyone that is making and selling these and, as I only have one good one that I would need for the Model K, I needed to make a couple more for the front and rear cylinders.

One might ask the question: “why bother with castings and rather have new ones CNC machined?” There are 2 reason for not going down this route: firstly, I do not have CNC machining capabilities and, in any case, the overhead in setting up for a maximum of 4 copies (I will use 2 on other projects) would be a relatively expensive exercise; secondly, the original was made in cast steel and the appearance of a fully machined finned exhaust ring would, quite simply, look out of place on a 90 year old motorcycle.

The starting point was to prepare the original for making a mould for casting resin copies to be used as foundry patterns. This involved filling in the threaded portion with an aluminium ring and the remaining part smoothed over with body filler. The aluminium ring is tapered so that the resin part that will become the pattern can be withdrawn easily from the sand mould at the foundry.

A couple of pieces of stainless steel locking wire were also fixed to the top surface with epoxy resin; these would be used to suspend the original part in the resin moulding box.

None of these additions do any damage to the original and can be easily removed later when the mould has been made.

The next step is make a small plywood box and to suspend the original part in the box

Which is then marked out and 2BA screws inserted a few threads into the wood around the periphery

The original can then be suspended in the box, surrounded by screws, about 1cm above the base

The small pieces of metal are to balance the ring so that it lies horizontally in the mould.

RTV silicone rubber is then mixed and poured up to the top of the part

and allowed to set.

After removing the supporting wire, 2 glossy paper cones are attached to the top of the exhaust ring to act as a feed sprue and riser for casting the top part of the mould. The silicone is also coated with Vaseline to avoid the upper silicone part adhering to the lower part.

Silicone is then poured onto the mould up to the base of the screw heads and, after setting, can be disassembled to remove the original part.

The mould can now be reassembled and used to cast one or more resin copies of the original

These resin copies can now be fettled (removing the sprue and riser and filling in any small holes with filler) and then sent to the foundry.

I made 2 good resin copies and sent these to New Pro Foundries. I have been using New Pro for many years and they have made many casting for me in AB2 Ni Al Bronze and heat-treated LM25 aluminium; rockers for Nortons, the AJcette timing cases and all the crankcase and timing case castings for this V-Twin. Having visited New Pro many times in the past and seen the quality of their work on extremely complex castings I wouldn’t go anywhere else.

In due course, a batch of new shiny castings arrived in the post

ready for finishing.

Machining these is fairly straightforward. The number of fins is divisible by 3, which means they can be easily held in a 3-jaw chuck, and after facing off, screwcutting the 16 tpi thread for the exhaust stub and putting them in the tumbler for a few hours they turned out pretty well.

The picture below shows 2 of the new completed exhaust rings on top of the original.

And, in comparison to the later K66/2 variant, the difference in size is obvious and with a lovely cast surface finish.


And, most importantly, it fits without fouling the frame tube

Shrinkage: These bronze alloys have a quoted volumetric shrinkage of 4% which gives a linear shrinkage of 1.3%. I measured the shrinkage of the whole process, ie original > resin > casting as 2% as best as I could determine for an irregularly shaped part; in practice, this isn't a problem for a component such as this.

You may ask: "why go to the trouble of making a resin pattern - why not just use the original part as the pattern?". Well, you could ...if it had been suitably modified with the aluminium ring and filler as I described. However, I have a golden rule of never sending a one-and-only original part to a foundry as a pattern ever since a good friend of mine sent an aluminium timing case to a local craft foundry  ...who mistook his original for a piece of scrap aluminium and tossed it into the melt ...never to be seen again.

Sunday 14 March 2021

The Positive Stop Gear Change

Gearboxes on vintage motorcycles evolved rapidly during the late 1920s and early 1930s. The first positive stop foot change mechanism is attributed to Harold Willis at Velocette and was first fitted to the 1928 TT winning KTT bikes. Up to that time, gear changing was by hand and, depending on the manufacturer, each gear had to be selected by either engaging the tank-mounted lever into slots on a gate or, in the case of the hand change mechanism fitted to Nortons using a Sturmey Archer gearbox, by feeling the engagement of a notch in a groove.

Many people believe that the advantage of a positive stop foot gear change is the speed of changing up, ie getting into the next gear faster and "getting the power on" – and, indeed, that is one of the advantages. However, from a riders perspective and as a regular rider of vintage bikes without a positive stop, there is another significant disadvantage of a non-positive stop hand change, namely that it is not possible to simultaneously change down a gear and apply the front brake and this results in the necessity of slowing the bike much earlier when entering a corner. I was not around during the late 1920s when bikes transitioned from non-positive to positive stop but I would imagine that it would have been clear to an observer at the TT at this time that the latter had a significant competitive advantage in their ability to out-brake the former.

During a period of less than 10 years, motorcycle gearboxes evolved from being non-positive stop hand change to positive stop foot change with the entire positive stop mechanism incorporated into the gearbox. However, there were a couple of important evolutionary steps along the way. The first was to incorporate gear indexing within the gearbox itself instead of allowing the gears to take any position on the layshaft and mainshaft (it is possible with a gated gear change to have partial engagement if the adjustment is set up incorrectly).  This consisted of 2 parts: the first is a series of notches, referred to as the part “LS 166 Plate” in the illustration below taken from the 1931 Sturmey Archer gearbox parts list and the second is a spring-loaded plunger that engages with the indents on the plate and has been oriented to show how it engages.

This, in itself, does not provide a positive stop as it would allow a single rotational movement of the selector to go from 1st gear through neutral and 2nd gear to top gear in one go, but it does provide the means of selecting unique and discrete positions of the gears on the shafts.

The second ingredient that is required is the positive stop mechanism itself whereby the movement of the selector (above) is restricted to only that required to move the gears into their next position – either shifting upwards or downwards.

The Sturmey Archer gearbox that I had chosen to use on the V-Twin already has gear indexing incorporated into its construction but there was no positive stop mechanism. I had 3 options of how to do this: 1) Make one from scratch; 2) Use a Velocette positive stop mechanism (I have 3 in my workshop) or 3) Use a positive stop mechanism from a Sturmey Archer/Norton Dolls Head gearbox (I had 2 available). I chose the last of these.

These Sturmey Archer positive stop mechanisms started as a “bolt on” extra located on top of the gearbox, as shown below,


but soon ended up in the “head” of the Dolls Head gearbox

As I had already used the entire gear cluster from this gearbox to refurbish another and not wishing to destroy the casing by decapitating the “head” from the dolls “body” I decided to make another casing to hold the mechanism. Firstly, the entire mechanism was extracted - the collection of parts is shown below:

And all the important dimensions for a new housing were obtained by reverse engineering the existing components.

Firstly, the inside of a 100mm diameter EN3B steel bar was bored.

This was then mounted on the milling machine and the internal details were machined

And the final machining operation on the main body was to form the boss for the spring holder on the rear

The other main part to be made was the bracket that supports the main body just above the gearbox and bolted to the right-side engine plate. This was machined from a piece of 75mm x 75mm x 6mm angle iron.

and the main body of the housing was machined at the bottom to “sit” neatly on the bracket


The 2 parts were then silver-soldered into a single component.

2 trunnions and a short length of threaded adjuster rod were machined and the complete housing with the mechanism inside could then be mounted on the bike for the first time

With an aluminium cover plate and a gear lever put on for testing purposes I found the positive stop worked perfectly, but only after removing a bit of material from the lower part of the engine plate to allow bottom gear to be selected.

Like everything else on the bike, it is a tight fit!

Tuesday 2 March 2021

The Clutch: Part 2

The second part of the clutch to be made is the clutch centre. This was made from a piece of EN8 round bar and the first machining operation was to make the taper for the gearbox mainshaft and the diameter for the bearing inner race. Narrow angle tapers are notoriously difficult to machine because the male and female tapers will not bind together if there is even the slightest difference between the male and female taper angles. The nominal taper angle, which was  4.50  here, can be determined by measuring the shaft and then set on the top slide on the lathe for machining but a deviation of even 0.10 will prevent proper contact. The only way that I have found to machine narrow angle tapers successfully in what is essentially a reverse engineering operation is to use engineers blue on the shaft/hole contact surface to determine if the angle needs increasing or reducing and to then make very small adjustments during the machining operation and try again until a good tight fit is achieved.

The axial location of the clutch centre on the shaft is obviously important as this is one of the determining dimensions for the chain alignment and because continual attempts at machining the taper would change that dimension the sequence of machining here is: 1) get the taper angle correct; 2) face off the end to get the correct axial location of the clutch centre on the shaft; 3) machine the diameter for the bearing inner race (0.001” interference fit).

The subsequent machining operations are relative straightforward, namely:

1)    Bore the inside diameters

 2)    Drill and tap the 6 holes for the spring holder retaining studs

 3)    Machine the grooves for the plane plates


and, finally, a trip to my spark eroder to make the keyway.

The comparison between the original clutch parts – chainwheel and centre is apparent.

New, longer, studs for the springs were made in EN24T

The spring-holding pressure plate is now at a much greater distance from the “mushroom” that is inserted into the end of the gearbox mainshaft and operates the clutch.

There are 2 possible solutions for this: 1) make a new mushroom that has a significantly longer shaft than the original or 2) modify the spring holder so that the original mushroom can make contact with an extension in the centre of the spring holder. I chose the latter because the shaft diameter of the mushroom is only ¼” and having over 1” of unsupported shaft pressing against the spring holder would be a poor design.

The original pressure plate, shown below in the lathe,

was bored and threaded to accept an insert that was silver soldered to the centre

A new threaded centre was made from silver steel and hardened and tempered at the contact end

And this completed the machining for the clutch centre adjuster/extender

All of the parts have now been made for the clutch and it was time to assemble it for the first time. The bearing was first pressed into the chainwheel and then the centre was inserted and pressed into place.

The change in appearance is a result of chemically blacking the centre. The backplate was now put on and the two main clutch components are now fixed together via the bearing.

After checking that they rotated freely and without interference, the assembly could be put onto the gearbox mainshaft and the plates and springs etc. inserted. There will certainly be some adjustments needed to spring pressure in due course and a locknut put on the centre adjuster but the chainwheel is, encouragingly, in exactly the right place for alignment with the crankshaft sprocket. I also noted that the pressure plate pulled off very evenly when the clutch was actuated so I am hoping that this will result in a very effective and relatively light clutch.

Last, but by no means least, a puller was made so that the clutch centre could be released from its taper on the gearbox mainshaft.