Saturday 24 June 2023

Velocette KTTs 55 and 305: Wheels – Part 3 – Rear Wheel Hub

As I mentioned in a previous blog posting, whilst I had rear hubs for both KTT 55 and KTT 305 I was missing a hub for the Mk 1 OHC special . Specifically, one of these:

I’m not sure of the correct part number for this: the 1925 – 1931 parts book list this as K7 whilst the later 1932 to 1934 book has K7/4. Is there a difference between them? This hub is from KTT 55 and so mine is most probably the earlier component.

The mating brake drum part numbers are KS2 and KS2/3 respectively, the difference being the size of brake shoes that will fit. Grove Classics lists both of the brake drums (see eg here) and I had already purchased one of these

in the hope that I could find the hub. Unfortunately my efforts to find a hub were in vain -  I put a “Wanted” advert in Fishtail, scoured the autojumbles that I attend and put the word out among my buddies but there was nothing readily available. Not wanting to be held up in the project I decided that I would make one (note: if you need one look at the end of this blog....)

The first step was to make a drawing based on the hub that I had. However, there is one problem with this, namely that the taper roller bearings that were used originally are no longer available and the internal design would need to be changed.

By chance, I had a couple of new Timken taper roller bearings in my box of bearings, still in their original wrappings, that would be perfect for the job. The picture below shows one of the bearings together with an original (upper part of the picture - Velocette part # KS-18) from one of my other rear hubs.

I bought these over 50 years ago when still an engineering student with the intention of building an Aspin rotary valve Engine (I won’t digress into that here ….see eg but that project never happened ….but I still had the Timken bearings. These bearings are still available today and after checking their load-carrying capacity decided that these would be an excellent substitute for the originals.

Timken provide a lot of information about their bearings; the detailed specification of this particular bearing is summarized below:

Apart from the dimensions, one of the most important pieces of information here is C90, which is a measure of the radial load capacity of the bearing. The value of 1100 lbf means that a pair of bearings mounted opposite each other on the shaft will have a life of 90 million revolutions if subjected to a radial load of 1100 lbf. What does that mean in practice? Using 21” wheels (rolling diameter with a tyre is approximately 24”), 90 million revolutions equate to over 100,000 miles. How big a load is 1100 lbf? It’s half a tonne.

In simple terms, if the little Velocette and rider weigh one tonne (remembering that the weight of the bike + rider is distributed approximately 50/50 between the front and rear wheels) then they should be able to ride 4 times around the planet without worrying that the rear wheel bearing would fail. So, even going for a long ride after a heavy lunch shouldn’t be a problem.

The main dimensions of my existing hub were measured and modified internally to use the new bearing.

What material to use? I found that EN3B steel, which is quite adequate, could be obtained in 110mm diameter and a length of this was ordered.

In the above picture, the steel for the new hub is on the left, the piece of steel in the centre was also bought as I believed I would need to make a mandrel to hold the hub during machining and the existing hub is on the right. It turned out that I didn’t need a mandrel to support the machining.

The first step was to machine the larger (4.06”) and smaller (2.85”) spoke flange diameters.

This is a fairly large lump of metal to have spinning round and the surface speed of the cutter at these larger diameters is relatively high so the lathe is run at one of the slower speeds. In addition to producing large quantities of swarf (more of that later) the smoke from the cutting oil fills the workshop. In fact, it was so bad that I would initiate a cut and then leave the workshop whilst watching progress through the window from outside!

The next step was to bore the inner diameter for the bearing.

The workpiece was then turned around and, after carefully checking that it was on-centre with a dial gauge (I replaced the chuck on the lathe some years ago and it is pretty accurate) the larger diameter end was bored and the wheel and spoke flanges machined.

The workpiece was again turned around and the smaller diameter finished to size and the centre portion “carved out” using left and right hand tools alternately.

At both ends of the centre section, the corner was left unmachined

to leave sufficient material for a radius to be blended in.

The final operation on the lathe was to put a radius on the spoke hole flanges.

The lathe work is now completed and it was time to set up the hub in the dividing head on the milling machine.

The first operations were to drill and tap the 5/16" BSCY threads that secure the brake hub/sprocket and to drill 20x 0.15” holes for the spokes.

The hub was then turned round and the spoke holes on the smaller flange were drilled. It is important to note that these are staggered with respect to the spoke holes on the other end.

Apart from a ¼” BSF threaded hole for a grease nipple, machining of the hub is now complete.


A couple of statistics: the cost of the steel was 80 GBP; total machining time was 5 days in the workshop – around 28 hours.

And I have one full dustbin of steel swarf to dispose of!

Last, but my no means least, if you need one of these and either don't have the time, the equipment to make one or don't want to spend a week being asphyxiated by cutting oil smoke then please contact Peter Miles at as he is making a batch - get your order in now! Peter is the K Series Velocette specialist in the Velo Owners Club.

Saturday 10 June 2023

Velocette KTTs 55 and 305: Wheels – Part 2 – Cup-and-Cone Bearings

All the front wheels use cup-and-cone bearings and unfortunately, I was sadly lacking sufficient of these in good condition; I had one good cup and one damaged cup. The pictures below show pitting on the cup and typical brinelling damage on the cone.

The cone is from a different Velocette hub. There are a number of possible reasons as to the causes of such damage (interestingly, there is little agreement in the open literature - see eg here in the context of angular contact bearings) but, in this instance, it is probably the result of poor adjustment and/or inadequate lubrication.

The earlier hubs (part # W1/2), 2 of which are shown below with the bearing cups and the bolt-on brake drum removed,

use a really strange size of bearing. The hubs, where the bearing cup is located, is bored to 1.735” - that is 0.015” under 1 ¾” inches? Why did Velocette choose that size?

Now, the simplest solution would have been to bore the hub ID to 1.749” and fit off-the-shelf taper roller bearings. However, these bearings are not cheap; the best price I could find for the bearing that I would have liked to have fitted - Timken 05068/05175 TS was 182 GBP each – which adds up to 728 GBP for the 4 bearings required for 2 wheels! Whilst I don’t mind spending money on my restoration projects, I consider that a bit too much to spend on 4 wheel bearings.

Deep groove ball bearings are in plentiful supply and much cheaper but are not an optimum technical solution because they are designed to take radial loads but only minimal axial loads (they also need a spacer between the bearing inners to avoid the clamping force putting a large axial load on the bearings which would otherwise destroy them after a short period of time). Wheel bearings experience axial loads when cornering and huge axial loads if a sidecar is fitted.

On the other hand, angular contact bearings will take axial and radial loads, however, I could not find bearings of the size that I wanted and one that could possibly have been used was, again, horrendously expensive.

I therefore decided to make my own replacement cup-and-cone bearings.

The parts manual very usefully lists that these hubs use 9 balls of 3/8” diameter each side (part # W15).

I measured the radius of the cup using a radius gauge

and this was 0.25”.

Based on the dimensions that I had and not having an original cone to reverse engineer, I sketched out what I thought would be the optimum layout, shown below.


Here, I have used a 450 contact angle for the design (see picture below for the definition of this, shown for angular contact bearings) and the radii of the contact surfaces of the cup and cone are 0.25” and 0.315” respectively.

Picture courtesy of SKF

The overall plan would be:

1)    Use O1 tool steel to make both cups and cones

2)    Make blanks of the cups in the lathe and then use a ½” diameter carbide ball-nose cutter in the milling machine to machine the 0.25” radius bearing surface

3)    Use a 16mm diameter carbide cutter on the lathe to machine the cone bearing  surface profile. This cutter gives a radius of 0.315”

4)    Harden and temper all parts to ~ 60 HRC



The first step in making the cups was to machine 4 blanks to size, leaving sufficient material for the ¼” radius in the corner.

 Each of these was then mounted on the milling machine in the dividing head chuck

and, after a lot of handle-turning, there are 4 nearly completed cups. A damaged original is shown at the bottom of the picture for comparison.

The final operation was to chamfer the corners. The hubs have a well-defined radius of 3/32” at the corner where the bearing seats

and it is important that the bearing cup does not foul this corner. Another radius tool was therefore ground from a piece of square HSS bar to machine the corner of the cups.


The bearing surface profile was machined using a 16mm round carbide insert, as shown below. This has to be machined carefully because the tool is in contact over the whole length of the profile.

After tapping a ½” 26 TPI BSCY thread through the centre and milling spanner flats on 2 of the cones the machining of these is complete.

Before heat treating the cups and cones a few final checks were made on the assembly

and to ensure the balls are not in contact with each other when the cone is inserted.

All was good here so time to proceed to heat treatment.


Heat Treatment

Heat treatment – hardening and tempering, was carried out in exactly the same way as for gears described in a previous project (see here).

The first step is to make each component into a little package using stainless steel locking wire so that they can be easily removed from the furnace when hot and to coat them in Kasenit anti-scale compound.

They are then put into the high temperature furnace and heated to 815 0C

before quenching in oil (vegetable oil from the supermarket). The parts are then put immediately into another furnace for tempering at 200 0C and soaked for a couple of hours before removing the Kasenit coating with a wire brush and put in a tumbler overnight.

Although I can check the hardness approximately using a fine file it is much better to have a quantitative measurement if possible. Luckily, a local company has a hardness testing instrument and a couple of parts were selected for measurement. This came out exactly as I had hoped for, namely 60 HRC.


Last, but by no means least, new wheel spindles (EN16T steel), nuts and dust caps were needed.

All of the components for these 2 front wheels are now completed.

One Final Job

I have 3 sets of strutted KTT forks – 2 of these obviously belong to KTTs 55 and 305 but I plan to use the 3rd set on the KTP-framed special. However, the brake plate used on strutted forks differs from that on the standard forks by virtue of the brake cam boss being longer to avoid the brake lever and forks interfering. I only had 2 correct brake plates and a standard brake plate was therefore modified by making an extension piece that was held in place with a substantial phosphor-bronze bush.

The picture below shows the extension piece and bush before assembly.

After assembly, the bush was reamed to size, the extension piece was fettled to blend in with the brake plate casting and a new brake cam made to fit.

The work on the front hubs is now pretty well complete.

Thanks to Peter at Jayess Tools for grinding the carbide ball-end cutter and the HSS radius tool used to make the cups.