Friday, 28 February 2025

Rebuilding KTT 55s Cambox – Part 2: Assembling the Cambox

Although I have worked on 3 Mk 1 OHC Velo camboxes in the past none of these have been for a Velocette engine; these were for the AJcette and the AJS V-Twin, both of which were converted to chain-drive OHCs rather than bevel gears. Whilst many of the requirements are the same because the internals – the cam, rockers, rocker skids, tappets etc are the same, the Velo engines require the bevel gear meshing and large bronze bush, K-12, to be set up carefully.

Please bear in mind when reading this blog that this particular cambox has been assembled from a collection of parts that have not previously had the pleasure of each other’s company. I would expect (or at least hope) that a cambox that has been stripped for inspection and then reassembled would not require as much attention.

One fundamental requirement is that the cambox does not “rock” when placed onto the threaded cylinder head bolts and the 4 points of contact should contact simultaneously.


Why is this important? Because, if there is a problem with the cambox seating properly, then it will be stressed when it is bolted down and liable to break or crack at the weakest section. Damage invariably occurs at the lugs on the side opposite to the camshaft drive (the left side on the bike) and I have inherited a number of old camboxes with broken or cracked lugs that are a testament to a failure to address this.

There are 3 possible reasons why there might not be simultaneous contact – uneven spot facing of the cylinder head holes or machining of the mating surfaces on the cambox (resulting in non-coplanar contacts) or different lengths of the internally threaded section of the cylinder head bolts – or a combination of all 3! I am reticent to correct this by machining any material off either the cylinder head or the cambox and generally this can be addressed by adding shims – typically 0.005” to 0.010” maximum between one of the head bolts - to - cambox contact points.

There are various versions of the K-15/X camshaft. The earlier shaft has a larger diameter to support the K-17/2 cam whilst the later shaft has a smaller diameter for the K-17/5 cam. The picture below shows both variants.


Additionally, KTTs have a slot in the threaded section at the end into which slots an Oldham coupling that drives the cambox scavenge oil pump (and the shafts have a different part number – K-15/2 S or K-15/5 depending on which cam is used).

Although KTT 55 would originally have had a K-17/2 cam I decided that I would use a K-17/5 cam, especially as I had gone to the trouble of making one. However, I didn’t have any shafts (except on KTT 305) that had a slotted end and so the first step was to set the shaft up in the milling machine to add a slot using a 4mm end mill to engage a yet-to-be-made Oldham coupling.

This shaft also didn’t have any of the ¼” studs that hold on the K-18 top bevel gear and 4 new ones were made from EN16T.

The camshaft was put in the lathe after the studs had been fitted to machine off the small amounts of protruding studs to ensure the ends were all flush with the surface and then centre punched, as the originals, so that they will never shake loose.

The large bronze bush, K-12, needs to be “sandwiched” between the cambox and the bevel gear housing, ie it must be positively clamped. Why? Well, if it wasn’t then it would be free to rotate…. I am aware of 2 recommendations for the gap between the cambox and the bevel gear housing when assembled: 3 to 5 thou (0.003” – 0.005”) (Peter Miles -  Velo Club K series specialist – FT 141) and 8 thou (0.008”) used by Jeff Clew on a Mk 1 KSS (FT 93).

So, what do I have? I took 3 measurements: first, the thickness of the flange on the K-12 bush:


 and secondly the depth of the recess on the cambox

 

 and the bevel drive housing.

This indicated that I would end up with a gap of 0.252” – (0.122” + 0.112”) = 0.018” when assembled and which I also checked with a feeler gauge by assembling the bevel gear housing onto the cambox. This is too much and was cured by very carefully removing material from both faces of the flange

to give an acceptable ~0.003” gap between the cambox and bevel gear housing.

I do not know the history of these particular parts and so have no idea why the flange thickness was so much greater than the available gap in the cambox/bevel housing. I’ve measured the thickness of the flange on a couple of other K-12 bushes and found a ~0.004” variation between them. Maybe they were selectively assembled at the factory?

Finally, the K-17/5 cam was pressed onto the camshaft.

The 2 main requirements for meshing the bevel gears are that they are in contact over the whole width of the gear teeth and that the bevel drive/camshaft combination is free to rotate. Quite a bit has been written about this in the past so I won’t repeat it here – see, for example, the excellent article by Chris Benallick in FT 319 or the section on Camshaft Drives in Tuning for Speed by Phil Irving.

The simplest way of determining tooth engagement is by looking through the inspection cap

and if in any doubt as to their relative positions then the point of a scriber is useful to gently “feel” the engagement height of the teeth. If it needs adjusting then either the camshaft needs to be moved axially in or out and/or the bevel drive can be moved down by adding gaskets (it can’t easily be moved up! – at least, not without machining some material off one or other of the mating surfaces). It is unlikely that the camshaft needs moving outwards and can be moved inwards by adding shims behind the bearing supporting the camshaft at the other end. The OD of the bearing is 1 9/16” diameter and shims for the bearing are available from the Velo Spares Scheme – part number K128.

I ended up with good engagement with 4x 0.008” shims (which I would think is about the limit) behind the bearing and one 0.020” gasket between the K-14/3 bevel gear housing and K-46/2 bevel gear bush housing. As shimming the bearing changes the position of the cam relative to the rockers it is also important to check that the rocker skids contact the whole width of the cams. In this case, they were in the correct place so no further work was necessary   

New rocker skids, part number K-41/4, are available from the VOCs Spares scheme. However, it’s important to recognise that variants of the (original Velocette) skid exist with some differences to the section thickness at the “business end” – see picture below which shows an original on the left and a new yet-to-be-fitted skid on the right.

I believe the reason for this was to adapt the skid for different cams and for variants of the rocker as the latter was updated to give it greater strength during the evolution of the engine. A consequence of this is that the fitted length of a new K-41/4 skid needs to be checked and adjusted if necessary. There is also a question about how the load is transferred between the skid and rocker (at the bottom or the top of the skid) when using the K-41/4 skids on earlier rockers but I’ll leave that discussion to another day.

I have previously used the K-17/2 cam in both the AJcette and V-Twin. For comparison, I showed a picture of both the K-17/2 and K-17/5 cam on a recentblog. The K-17/2 cam has the same lift as the K-17/5 but a larger diameter base circle and the length of the skid needs to be reduced to position the rocker in the right place – it otherwise “sits” too high at the “cam end”.

Although I didn’t need to do this here, I have reduced the overall length of skids in the past by making a small aluminium sleeve into which the skid is a good tight fit and in which the sleeve has a 1mm slot from a slitting saw

so that it can be clamped tight in a chuck on the rotary indexer. This can then be set up on a milling machine, aligned with either the X or Y axis and then carefully machined using an end mill to retain the stepped feature (which positions the contact surface of the skid relative to the cam and prevents it from rotating in the rocker) but reduced in length. The skid is hard but can be cut easily with a carbide cutter.

(This is actually a picture from the V-Twin project which used K-17/2 cams).

The original adjusters are small fiddly little things to make (I made a batch of 6 of these in the past – see here) but there is an alternative that I discovered subsequently. The picture below shows, from left to right, an original Velocette split-collet adjuster (K-29 and K-40), one that I have just removed from KTT 305 in the centre and, on the far right, a new adjuster for a cammy Norton from Racing_Vincent

If I need any more of these then I will use the Norton ones.

Finally, I found it pretty well impossible to simply insert the felt washers that fit into the recess in the cambox at the end of rocker spindles and then assemble the rocker and spindle in place -the felt simply gets pushed out and damaged. I have therefore made a simple jig – just a short length of round aluminium bar that slides over the spindle

and can then be used to compress the felt washer in place.

I’ve also found that it helps if the felt is dipped in engine oil first and the whole cambox assembly put into the freezer (in a polythene bag!) overnight. The oil is then so thick that the felt does not decompress when the pressure is removed and the rocker can be slotted into position without a problem.

Finally, I have added a small aluminium disc that can be put inside the cam puller to remove the cam from the shaft

to avoid the pointed end of the screw on the extractor splaying the now-slotted end of the camshaft.

And…. I found the easiest way to remove and tighten the bearing locking ring is to use a small length of flat steel bar

and an adjustable spanner, being careful not to scour the aluminium thread with the end of the bar.

This cambox and upper bevel drive assembly (and the cylinder head) for KTT 55 are now complete.





There is one remaining job to do, namely making the Oldham coupling to drive the scavenge oil pump but I’ll postpone that until I’ve rebuilt KTT 305s cambox and then make them both at the same time.

Saturday, 8 February 2025

Rebuilding KTT 55s Cambox – Part 1: Repairing the Cambox Casting

The cambox that came with KTT 55 had been substantially modified because the bike came with a bronze head (which should be fitted to a later Mk IV) and, because a bronze head requires a different cambox casting, someone in the past had modified an earlier cambox to fit. Anyway, I had parted company with the bronze head and modified cambox to a buddy that has a Mk IV fitted with an iron head and can make better use of it. I had already rebuilt an iron head (see here) for KTT 55 and now it was time to rebuild the cambox.

There are a lot of details to pay attention to in a Velo cambox. The picture below shows the bits that need to be assembled to make a functioning cambox – but it’s not just a simple matter of bolting them together and expecting it to work, at least, not when assembling a cambox from a collection of parts.

 

Many of these early cammy Velo camboxes have structurally damaged castings, usually to either one or other of the lugs at the bottom where they bolt onto the cylinder head studs or the section between the holes where the rocker skids enter to contact the cam – the latter can be seen in the cambox in the above picture where the thin “bridge” has disappeared. The other damage that is often found is to the 2BA threads that fix the bevel drive casting or at the other bearing-end of the camshaft. The above cambox also has some damage to the corner when the exhaust rocker exits.

Anyway, I decided that I would repair another cambox that was structurally intact (regarding lugs and threads) except for a gaping and jagged hole where the rockers enter the cam chamber.

Apart from the obvious unsightliness of the damage, the question arises: why repair it – after all, it’s hidden inside the cambox? The reason, quite simply, is that oil “escapes” from the cam chamber – thrown out by the rotating cam - into (and over the top of) the gutters on both sides of the chamber ….and then to the outside world and invariably ends up covering the bike. For this reason, the gap that connects the cam chamber and the cambox through which oil can escape needs to be minimised – 2 small holes in the original design. For anyone that has read my previous blogs, I have had plenty of experience in trying to solve this problem on the AJcette and the AJS V-Twin, both of which use the same cambox.

The first step in making a repair was to machine a piece of solid Aluminium into an annulus,

clean up and remove the jagged edges from the casting and mark out, as best as possible, the shape of the hole on the new piece of aluminium (by inserting it inside),

chop off the top section of the annulus and cut out the shape to fit the hole

and with support from a piece of thin steel bent to fit inside the cam chamber

use oxy-acetylene and Lumiweld to fix the new piece in place.

This was then cleaned up on the milling machine

and finished by hand with a file and Dremel.

After vapour blasting

…..not perfect, but a reasonable repair to the damage.