Velocette Mk 1 KTT Cambox Oil Pumps K-102/2 and K-152– Part 1

I decided to use an iron cylinder head for KTT 55 – and which it should have rather than the Mk IV KTT bronze head that came with the bike but was in need of a cambox oil pump. The oil pump for the bronze head, shown below in its component parts

is a gear pump (this is actually not an original pump but one that I made some time ago using a casting that I had acquired, modifying the internals of a spare oil pump and making an end plate) whereas the cambox oil pumps fitted to earlier KTTs were aluminium-bodied reciprocating pumps.

The earlier type of pump, shown below together with its Oldham coupling to connect it to the camshaft, is from KTT 305.

I bid for one of these on ebay but it sold for 361 GBP which was considerably higher than my bid. This is the first one that I have ever seen for sale (and this one had a weld repair to the valve lifter cable "arm") and simply waiting for another one to turn up didn’t seem like a very good plan so I decided to make one (actually more than one because of the effort involved) …but, how to make these?

As the aluminium body (K102/2) would be used to make a casting pattern, the first step was to replicate the missing valve lifter cable support arm. The pump body had been broken at the corner

and a "repair" was made with an appropriately shaped piece of wood attached with JB Weld and with reference to an undamaged non-pump “end” (K102/3) to ensure the correct orientation.

At the same time, all holes were filled with P38 body filler and the wooden arm blended in. Gasket cork has been added to the 2 machined surfaces and fixed with epoxy resin to ensure there is sufficient material on the casting to machine.

A ½” wooden dowel has been inserted into the hole for the reciprocating element and attached with epoxy resin.

This dowel will be used to support the pattern in position in the mould before pouring the silicone and, because it will be included in the casting, it will also assist setup for final machining.

So, there is now a part that can be used to generate a mould for casting resin patterns (those of you that have read my previous blogs about making castings will know that I never send original parts away to a foundry – they are too valuable), however there are 2 hurdles to jump to progress this project.

Firstly, making a simple 2-piece mould to cast resin copies is not straightforward for this particular piece because of the angle with which the valve-lifter arm protrudes. Using the 2-stage process that I described here for making resin patterns for some of the cycle parts will not work because the ”arm” would be totally submersed in silicone rubber and it would be impossible to extract the solid part from the mould without damaging either the mould or the component.

Secondly, I do not have machine tools that can produce the closed spiral groove on the ½” diameter steel pump plunger (K152) that provides the all-important reciprocation. I did figure out a way to cut an ellipse on a small bar with the equipment that I have but it would have been difficult to set up accurately and the cutter would have entered the bar at an angle which would produce a groove with slanting sides. The inside diameter of the groove would also be elliptical ...so not really a good solution!

 

Grooved Shafts

Luckily, we seem to be well endowed with small engineering companies around the area where I live. Realising that I would not be able to machine the groove on the pump plunger, I contacted Dave at Venture Precision Engineering Ltd and after a brief discussion on the phone and a quick sketch (from reverse engineering the existing component)

Dave was pretty confident that a 4-axis CNC machine that he has could do this.

A couple of days later he called me to come over to see progress – his premises is less than ½ hour away from my home.

The first step was to model the part in CAD – SolidWorks in this case and the screenshot below shows a sinusoidal plane superimposed upon and wrapped around the shaft.

The next step is to import this into the CAM software which generates the instructions to drive the CNC machine. It will also produce a simulation of the manufacturing process.

This is then used to drive the CNC machine – shown below making a test piece in EN24 steel (although a bit difficult to see behind the safety screen and copious coolant)

  

And the picture below shows the original and the test piece together.

There are 2 differences between the original and the new piece: using a sinusoidal variation of the motion is an approximation – the original has a steeper lift and fall and more “dwell”. However, the stroke is the same for both and I don’t believe this will make too much difference in practice. Secondly, a radius cutter has been used and this, by definition, produces a radius at the bottom of the groove. Again, this is not important as the inserted screw will be profiled to fit the groove. The radiused groove is, in one respect, a better design solution in eliminating the stress concentration from a sharp corner.

9 of these (from a 1m length) have now been made in 13mm diameter O1 tool steel and I will machine the final diameter, half-moon section pumping end and the slot for the Oldham coupling drive at the other end after the castings have been machined.

 

Aluminium Pump Bodies

The patterns for these have been made in resin using the original part and are now at the foundry to be cast in LM25 Aluminium and heat treated to the TF condition before final machining.

But first, how to make the silicone rubber mould to cast the resins? I decided that one way to do this would be to make the lower part of the mould in 2 stages. The first stage would be the main body and the second to incorporate the valve lifter arm. To accomplish this, a cardboard “dam” was made inside the moulding box to allow the liquid silicone to only come into contact with the main body of the pump.

The dam has been fixed to the wooden sides and base with epoxy resin before pouring the silicone. There are a few extraneous screw holes in the base of the box as it was previously used for making resin copies of a Norton Model 18 inlet rocker.

After the silicone has set, the cardboard dam is removed. As can be seen, the main part of the mould is filled (although a bit of silicone has leaked along the bottom – which is not important) and the part with the arm is ready for the second stage.

The next stage is to put Vaseline on all the wooden surfaces (to avoid the liquid silicone adhering to them), tilt the box so that the valve lifter arm is horizontal and pour silicone up to the level of the top of the arm.

The second pouring of silicone has a different colour because I ran out of the pink stuff and had to order some more.

Vaseline has not been applied to the pink silicone surface and so the 2 will adhere to each other to form one contiguous bottom-half of the mould.

The final step is to cover all surfaces with Vaseline and pour the silicone to form the top half of the mould

which can then be disassembled to remove the original part.

The mould can now be reassembled to start casting resin copies.

4 resin copies were made, shown below.

The different colours on the resin are both JB Weld and P38 body filler which were needed to tidy up the resin castings due to incomplete filling of part of the mould resulting from trapped air bubbles.

The resin patterns are now with a AJD Foundries in Brierley Hill near Birmingham (that's Birmingham in the UK not Birmingham Alabama) for casting and heat treatment. Interestingly, the new castings are being made only 16 miles from the original Velocette works in Hall Green where the original items were made over 90 years ago.