The DOHC 250 Velo Engine: Inside the Cambox

I have intentionally not disturbed the cambox up to this point in the engine build. Why? Well, because it appears to work just fine and there has been so much else to do but, as now the engine build is nearly complete, it was time to look inside.

This is what it looked like before stripping.

However, before dismantling the cambox I figured it would be a good idea to check the “as-received” valve timing. The relative valve timing, ie the relationship between the Inlet and Exhaust is determined by the position of the gears on the camshafts at the ends relative to each other and I have assumed that the setup reflects how it was run on the engine.

The absolute valve timing (ie the actual valve timing) is governed by the positioning of the bevel gears. Now, the large top bevel gear that is on the end of the shaft driving the other gears – part number K-18, has 44 teeth. One complete rotation of this shaft equates to 2 revolutions of the crankshaft and so the crank-angle change that would result from changing the engagement position of this gear by one tooth would be (360⁰ x 2) / 44 which equates to ~16⁰ crank angle. This would be a very significant change and this helps in determining the absolute timing by avoiding unfeasible extremes.

After a couple of changes to the relative positions of the bevel gears I came up with the following valve timing:

IVO 60⁰ BTDC

IVC 70⁰ ABDC

EVO 70⁰ BBDC

EVC 30⁰ ATDC

The 2 more reliable figures are EVO and IVC as the other readings can be affected by the loading of the other cam during valve overlap.

These can be compared with valve timings on other engines

and also indicates that changing the bevel gear by one tooth would put the valve timing outside the range used on any of the engines listed above.

After removing the outer steel plate supporting the intermediate gears all of the gears can be seen more clearly.

I have marked the tooth engagement of all gears (and also the bevel gear engagement) so that they go back in exactly the same place. The position of the intermediate gears is not actually critical as they have been set up with hunting teeth. The number of teeth on the gears from the camshaft on one side to the camshaft on the other side is:

30 – 31 – 32 – 30 – 32 – 31 – 30   

After removing the gears (which are all bushed with an oilway) the main structural support plate is revealed.

I initially wondered where there were 2x 2BA threaded holes in the plate that seemed to have no obvious purpose but it soon became clear; they were there to be able to insert 2 screws to lift out the steel plate from the aluminium casting.

After removing 4x ¼” BSW Allen screws and heating the entire structure, the plate plus everything attached to it was separated from the outer casing.

Both cams had minimal wear

but I replaced the balls (10x 7/32” diameter) in 2 of the Hoffmann bearings

which were showing a bit of play. I measured the base circle diameter at 1.008” and the lift + base circle came to 1.315” which gives a valve lift of ~ 5/16”.

All of the parts were in excellent condition, however, there was one conundrum, namely: what was the lubrication strategy? There are clearly 3 different ways that oil can flow in or out of the cambox. The oil drain is obvious – the pipe at one end of the cover, shown in the picture below.

An oil feed to the cams is also obvious, indicated by the arrows in the picture below.

There are quills of ~2mm diameter at the ends of pipes that would supply oil directly to the cams. However, it is questionable how much oil would be delivered using this method; the oil would have been provided from the upper bevel chamber via a slot in a K-12 phosphor-bronze bush in the same way as a standard “K” cambox oil feed and although there is an oil seal in the centre and immediately behind a bearing (to prevent the oil simply discharging into the cambox at the point of entry). I am sceptical that this approach would provide the amount of oil required to lubricate the cams effectively and, indeed, the entire gear train. This also indicates that the cambox was originally designed to work with an internally pressurized system rather than a targeted distribution approach.

I suspect that whoever designed the cambox must either have had the same concerns or, alternatively, the cambox started life on a pressurized “K” type system but transitioned to a strategic lubrication system (as on KTT 581) because that would explain why there are oil feeds that are positioned at the ends of the cambox. Alternatively, these could have been additional oil drains to lower the level in the cambox when the engine was stationary to prevent it leaking away via the pushers and covering the cylinder head in oil. The inside and outside (it is a 1/8” BSP thread) of one of these is shown in the picture below.

With the lubrication system that I have decided to use, the central feed system is redundant – the upper bevel chamber will not be pressurised and, instead, I will introduce a quill into both ends which will be fed under pressure directly from the oil pump.

Similarly, the upper bevel will also be provided with a direct oil feed ….more next time.

The DOHC 250 Velo Engine: The Final Build – Part 1

Before starting the final engine build there were a couple of jobs that I had postponed, simply because up to now there was always something more important to do.

The first of these was to replace the small end. The small end bush that had come with the MOV flywheels/connecting rod assembly had just a bit too much wear and so a new phosphor-bronze bush was made on the lathe and milling machine,

an oilway along the length of the bush put in with a long 1/8” end mill

and then pressed in and reamed to fit the gudgeon pin that had come with the engine.

The second outstanding task was to check the crankshaft balance. I would assume that my Ebay-acquired MOV crankshaft would have been balanced to an MOV piston and the only difference is that I am now using the steeply domed race piston that came with the DOHC engine. Nevertheless, the balance needed to be checked. The crankshaft must balance all of the rotational mass + a fraction (the Balance Factor) of the reciprocating mass.

The reciprocating mass is the mass of the complete piston assembly, m_p, plus the small end, m_se.

By mounting the crankshaft assembly so that it can rotate freely and adding weights to the small end the balance mass, m_bal, can be determined.

First, the small end was weighed

and then the piston + rings + pin + circlips, m_p, (which came to 347g) before the flywheel assembly was mounted in a pair of used main bearings on vee blocks on the milling machine table.

I have used this arrangement rather than the crankshaft balancing jig as the assembly can be rotated without the end of the conrod hitting the deck!

The first step was to ensure that the existing balancing puts the big end at the top, rather than one side, when rotated.

It does. And so, the next step was to add weights in a small polythene bag to the small end so that the crankshaft would stop in any random position, without any preference, when rotated.

The mass added to do this, m_bal, was 214g. The balance factor, BF, is then defined by:

BF x (m_se + m_p) - m_se = m_bal

Which gives a balance factor of 73%

Is the acceptable or not? It is not easy to find documented “hands-on” experience of balance factors for MOVs but I did come across one article titled “Making a MOV Go” on page 6 of Fishtail 18 (October 1959) which stated:

 “ ….to enable a balance factor of 70% to be achieved. No experiments were made with other balance factors, the motor proved smooth between 4000 RPM and peak revs of 8500-8700 RPM.”

and so the value of 73% in the DOHC engine would seem to be just about right.

I have seen extremely high values of 85% quoted for MOV balance factors but I can find no evidence to back this up. If I substitute the raw piston weight of 8 ¼ oz (235g) for a 6.5:1 compression ratio piston that is given in the Velo technical data here then this would give a balance factor of 80% and this value would be reduced if the slightly heavier 10:1 race piston data were used. For the record, the raw piston weight of the DOHC piston is 307g.

I had already sorted out the lower bevel gear meshing (see here) and now the solid steel dummy “bearings” that I had been using were removed, the oil pump was inserted into its hole, the new (modified) main bearings were fitted, the crankshaft was put in and the crankcases sealed – exactly the same procedure as described earlier for KTT 305. Prior to assembly, the modified Woodruff key that engages with the lower bevel gear was heat treated.

The K-45/2 timing gear cover plus all the internal bits and pieces were assembled. 2x 1.25mm thick steel discs were placed behind the oil pump driving piece to get good engagement of the K-34 and K-72 gears as described in a previous blog. There are 2 different gaskets (K-68 and K-68/2) depending upon the timing crankcase/timing gear cover combination – needless to say I had the wrong gaskets in stock and so had to make the other one.

I have also fitted a new spring and ball to the oil pressure adjuster.

With the changes that I have made to the lubrication system these will no longer be used to adjust the oil pressure but they do need to be there to block the hole that connects the chamber to the feed side of the oil pump.

The inner timing case then needed modifying

to enable it to fit over the boss that I had attached for the crankshaft quill oil feed.

If the magneto chain ran in an oil filled chamber then there would be a problem ….but it doesn’t.

I had checked the piston, piston ring and bore dimensions some time ago and found clearance of 0.008” on the thrust/antithrust faces at the bottom and the 2 compression rings had a 0.032” gap. Recommended values for these quantities are generally quoted “per inch of bore” and to be able to make comparisons my measurements equate to 0.003” and 0.012” per inch of bore respectively.

So, what values are recommended? That depends where you look…. For engines in standard trim (ie, non-racing) Hepolite/Heplex recommended a comparable piston/cylinder clearance of 0.0015” per inch of bore (see eg The Vintage Motorcyclists’ Workshop, Radco, page 53). On the other hand, Phil Irving (page 68 of Tuning for Speed) gives a value of 0.0027” – nearly double. The piston/bore clearance dimension on this engine is very close to Phil Irvings value (0.003" versus 0.0027”)

For the compression ring gap, Radco quotes a value of 0.012” (which, for comparison, equates to 0.0045” per inch of bore) for a Velocette KTT whereas Phil Irving suggests 0.010” per inch of bore – again, double.

I checked the ring thickness against published Velocette MOV data and they were the same and so I bought a set of new rings. These fit perfectly and gave an “off-the-shelf” value of 0.025” for the compression ring gap. This equates to 0.0093” per inch of bore which is pretty well exactly Phil Irvings recommendation.

So, it would seem that the clearances in this engine conform pretty closely to those recommended by Phil Irving.

Incidentally, I was amazed that I could simply phone Cox and Turner

and buy a set of NOS (New Old Stock) piston rings that arrived the next day for a fairly obscure British motorcycle that ceased production in 1948!

The next step was to put the piston + rings on the connecting rod and put on the barrel and cylinder head. The circlip grooves in this piston are flat-bottomed and so new Seeger circlips were used to hold in the gudgeon pin. There are not that many occasions when I need circlips but I bought a box of assorted imperial-sized circlips many years ago

and these do come in useful when needed.

The barrel and head had been given a coat of cylinder black before assembly. It was now time to set up the vertical drive.

The drive shaft (which would usually be part number K-49 on an early “K” engine) that came with engine is substantially longer, 7/8” longer in fact,

and although I don’t have a Mk 2 shaft in my workshop (except in the engine of the bike I recently purchased) my guess is that this is a shaft for the later engine. It also looks to be new with only minor witness marks of having been used.

This required a pair of Oldham couplings - the question is how thick should the flange on the couplings be? The shaft was put in place and a bunch of washers + shims + a feeler gauge placed between the shaft and the drive at the lower end

to find out. 2 new Oldham couplings were then made with a slightly thicker flange (0.218” here vs 0.188” for the thickest commercially available part (see here)) from O1 tool steel,

finished to size and heat treated

before cleaning up

and fitting. In the above picture, the 2 new ones are on the left and, for comparison, a standard “K” coupling on the right where the difference in flange thickness is apparent.

I have used a slightly higher tempering temperature than usual (260⁰C) for the heat treatment to trade off some hardness for additional toughness. The thicker flange should also help in that respect.

A check after assembly showed 0.008” vertical clearance

which should be fine.

Much of the engine build is now complete. The main outstanding tasks are: check the cambox internals, complete the lubrication system and fit the magneto.

And then it will be Christmas and I will be banned from the workshop for a few days.

The DOHC 250 Velo Engine: Securing the Cylinder Head and Cambox

Having completed the eccentric studs, the next step was to check the compression ratio – up to now I have had no means of bolting down the barrel and head properly.

The crankshaft assembly and piston + rings (the old rings – I have a new set but I’ll come on to that at a later date) were reassembled into the crankcases. The valves, without springs, were then inserted into the head and the head bolted down and checked for clearance at TDC. Plenty of clearance here - there would be no chance of the valves and piston making contact during the valve overlap period so the head was reassembled,

the piston was positioned at TDC

and with the engine held on its side with the spark plug hole uppermost, the combustion chamber space was filled with oil (R 40) from a burette.

The combustion chamber volume was measured to be 30.5cc which gives a compression ratio of slightly over 9:1. My guess is that this is probably around the value of when the engine was originally built and, although fairly high, should be OK with modern 98 RON pump fuel.

The next step was to determine the dimensions of the cylinder head studs, which also provide pillars for the cambox, and the cambox securing bolts. Like the eccentric studs for the cylinder, these were made in EN24 steel for strength and took quite a few hours of machining, for example, the 2 long bolts on the drive side of the cambox

and milling multiple hexagons.

Anyway, they all worked out well and the cylinder head and cambox fitted perfectly.

At this stage, I had intentionally left excess material on the studs/pillars so that there was a gap between the valve pushers in the cambox and the valves.

The height of the studs was then reduced to give a running clearance between the pushers and the valve stems with the cambox seating properly on its 4 contact points. I have set the clearances to 0.020” on the exhaust and 0.012” on the inlet. With care, I am hoping to build this part of the engine without resorting to shimming.

The final step was to chemically black all the bolts and studs ready for final assembly.

The DOHC 250 Velo Engine: A Clue to its History?

Historical Contingency is an expression used to describe the concept that the course of history is not predetermined but is the result of essentially random events. For example, if a well-placed bullet had killed Adolf Hitler during WW1, then European history during the 1920s to 1940s would have been somewhat different. Unfortunately, that didn’t happen ….and one could apply the same “what-if” thought process to today.

So, what is the relevance of this? It’s by chance that I was at Founders Day, stumbled across the DOHC engine and then bought it. It is also by chance that the person from whom I bought KTT 305 many years ago also saw the engine at Founders Day, took a few pictures and later sent them to James Robinson, Editor of The Motor Cycle. Sometime later, James was visiting another KTT enthusiast, Rob Drury (see Rob’s blog here), and showed him the pictures of the engine. After a bit of rummaging, Rob came up with this picture.

Acknowledgement to whoever owns the copyright of the above picture

The focus of attention here is not the Velo-engined bike in the foreground but the engine propped up against the van in the background

and which, I would say with 99% certainty, is the DOHC 250 engine now in my workshop. The cambox is characteristic and as far as I can tell unique to this engine - the picture below of the “as-received” engine in my workshop for comparison.

I do not know and have been unable to discover the origins of the race meeting picture. The bike in the foreground is a Beasley raced by Charlie Bruce but it is not known why the DOHC engine in the background is there. It is surmised that Charlie (referred to as “Chappie”) had bought it to race but found that it didn’t fit in the frame (I can attest that the engine is tall - see here). Apparently, Chappie had a reputation for not selling anything so it is unlikely that the engine was there “For Sale”.

For more information about Charlie Bruce and his racing career please visit TheYorkshire Ferret.

Unfortunately, this fleeting glimpse of the engine at a race meeting some 70+ years ago, whilst interesting, doesn’t answer the important questions of who designed/built the engine and when.

Maybe this will jog someone’s memory but unfortunately most of the people that would have had any recollection of these times are no longer with us …nevertheless I’m ever hopeful that something will surface.

Thank You to all involved in finding this picture, identifying the engine in the background and adding context to the imagery.

The DOHC 250 Velo Engine: Eccentric Studs

The distance between the centres of the threaded holes in the crankcase and those of the cylinder/cylinder head/cambox are different – the distance-between-centres on the DOHC engine is smaller.

As I mentioned in a previous blog, it seems that Velocette solved this by using eccentric studs to change the bolting pattern to allow different barrels to fit onto a standard set of crankcases. Clearly, this method could also be used to move the centres outwards rather than inwards. Ivan Rhodes sent me a sketch showing exactly this solution that was applied to one of the works DOHC KTT engines.

Ivan also mentioned that he had acquired a large box of eccentric studs when the Velo Race Shop closed down 50+ years ago. I have never seen any of these “in the flesh” but a reader of my blog kindly sent me a picture of a bundle of these (thank you Paul).

Although I do not yet know the origins of the DOHC engine that I’m rebuilding, the same solution was applied, evidenced by the “pear-shaped” holes in the base flange of the barrel which are required to accommodate the change in section.

I have carefully measured the distance between the centres of both the crankcase studs and the 5/16” diameter holes for the cylinder head retaining bolts. Allowing for some manufacturing errors of a few thou, the dimensions are shown in the picture below.

The resulting design for the studs to give this offset and with a 5/16” BSCY female thread in the top 1” for the cylinder head bolts (the same thread as used on the “K” engines) is as follows

and would be made in EN24 steel.

I chose to make these from a 1m length of 20mm diameter bar which allows the 4” long ½” diameter section to be turned on-centre

An additional 1” length has been added to the unmachined section for which the reason will soon become apparent.

The question now arises as to how to machine the off-centre section accurately? There are a number of different ways in which this could be done and I have chosen the following approach to set up the 0.18” offset in the lathe.

The part-machined stud was put in the rotary table chuck in the milling machine (that had been centred) and then moved off-centre by 0.18” and a 0.75” deep ¼” diameter hole drilled into the end.

This was then put into the 4-jaw chuck in the lathe and a close fitting ¼” silver steel dowel in the tailstock chuck was used to adjust the position of the stud until the hole was on-centre.

This can be “felt” (and seen when it is rotated) accurately and small adjustments made until the dowel slips in easily. Unfortunately, it was only possible to hold the workpiece in the chuck using the 20mm section rather than the much longer ½” section to avoid the jaws interfering with each other. Other checks were made to ensure the workpiece was parallel to the lathe bed. This is a very substantial chuck and although there is less than ½” in length on which the jaws are in contact there were no problems in carrying out the subsequent machining.

The entire section was turned to give 3/8” diameter

before sawing off (in-situ) the end section with the hole and putting a 3/8” BSF thread on the remainder.

Three more were made

and the 4 then inserted into the 3/8” BSF helicoiled threads on the top of the crankcase.

And then the acid test ….would the barrel slide over the studs?

Success! After positioning each stud in the right place, the barrel slipped down without a problem. It would not go any further at this stage and required the “pear-shaped” profile on the yet-unmachined round transition section to be completed before it would seat at the base.

The pear-shaped section is essentially the merging of 2 different diameters – the ½” and the 3/8” sections. Most of this was done on the milling machine

and then finished by hand.

After a few hours of machining and fettling there were 4 studs with the required section,

all of which fitted inside their holes in the bottom of the cylinder.

After juggling the position of each – small changes to their angular positions,

the cylinder could be dropped down all the way.

This, at least, showed that the calculations and the machining were OK but I could see that trying to simultaneously get all the studs to be screwed in fully “home” and the cylinder to seat on the base flange would not be easy.

I had, in the meantime, made a better estimate of the distance by which the cylinder barrel would need to be moved upward to avoid piston-to-head contact near TDC and decided that a 2mm thick steel spacer plus two 1mm paper gaskets either side should be sufficient.

The spacer was made from a square piece of stainless steel with the main hole for the barrel machined on the lathe

and the stud holes drilled on the mill.

A gasket was made and I then started to fit the barrel.

A couple of issue now arise. The first, and one with which I was aware of before I started, is that when each stud is screwed down onto the surface the orientation is unknown ahead of time – you can guarantee that it will be pointing in the wrong direction! This can be corrected by extending the threaded portion by an additional rotation with the knowledge that a 3/8” BSF thread has 20 TPI and therefore 360⁰ of rotation equates to 0.050”. For example, if another 90⁰ of rotation would put it in the right place, then extending the threaded portion by 0.0125” should do the trick.

I could no longer hold the stud in the 4-jaw chuck on the lathe and so each stud was carefully set up in the 4-jaw chuck in the dividing head on the milling machine, offset and with rotation about the axis of the threaded section to remove material from the contact face to be able to orient the stud correctly in the crankcase.

All this went according to plan, however, one thing that I had not thought of was that when the lower contact face of the stud was screwed down onto the spacer the stud was forced slightly off-centre – away from the centre of the cylinder - by the bending moment from the contact side and this prevented the barrel from sliding down. The effect is amplified over the 4 1/2” length of the stud and although the movement was less than 1/16” at the top of the stud this was more than enough to cause a problem. 

I therefore had to remove a few more “thou” from each to orient the studs correctly but with only sufficient contact to avoid this problem. After a few iterations of this I managed to get the 4 studs to seat in their correct orientations and with a cylinder barrel that could be fitted properly. During this sequence of fitting operations, the studs had been put in and taken out multiple times and I was glad that I had helicoiled the threads in the crankcase otherwise they would have been completely worn out.

At this stage, I also drilled and tapped the 5/16” BSCY threaded holes to attach the cylinder head.

There was still one more machining operation and that was to reduce the length of the pear-shaped transition section by 0.050” on all the studs to ensure there was clearance in the depth of the corresponding holes in the barrel. Compared to much of the previous work, this is straightforward and I have used a carbide insert with a small tip radius to reduce the sharpness of the corner.

With another 1mm thick gasket, between the spacer and the barrel,

the barrel could be slipped down and the cylinder head put on and bolted down for the first time.

There is still a bit of tidying up to complete this setup – rounding the corners of the spacer etc but the hard work on making these studs has been done.

I would imagine that Veloce would have experienced the same fun-and-games that I have had in making and fitting these some 90 years ago. Whoever was given the task of making these at Veloce must have groaned when the foreman came along wanting more. Maybe they developed some jigs and fixtures to help? Anyway, I am pleased with the outcome and can now get on with the upper parts of the engine.