The AJS V-Twin Crankshaft: Part 2 – Connecting Rods & Pistons

Design of the crankshaft and, in particular, the calculations for crankshaft balance, requires the weights of various reciprocating and rotating components. Part of the input to these calculations is the weight of the connecting rod small ends but before these could be measured some work was needed to make them fit the pistons. Apart from the small-end diameters being different the connecting rod was too wide to even fit between the small-end bosses of the piston, illustrated below.

As can be seen, the connecting rod small-end width is 0 .027” wider than the available space.

The connecting rods were reduced in width to 0.896” by removing equal amounts of metal from both sides to give a total clearance of 0.14”. The small-end bushes were pressed out before machining.

 

New phosphor-bronze small-end bushes were made to replace the HD bushes and to fit the Ducati piston gudgeon pin – the new ones and originals shown below

These were then pressed into the HD connecting rod small-ends and which now fitted properly to the pistons.

Weighing could now begin….

Weights are required for the piston(s) (which are identical), the crankpin and both small and big-ends of both connecting rods. A couple of pictures below illustrate this.

The question arises as to how accurate are the scales? I don’t have any precise weights to calibrate them and so I machined a short length of EN24T round bar as a test piece and measured the weight of that. The weight (more strictly, the mass) can also be calculated from the product of the volume and the density, which is quoted as 7.84 g/cm³. It turned out that the 2 values of mass – ie that measured on the scales and the calculated value, were identical. It could, of course, be possible that both are wrong and have compensated exactly to give the correct result but that is highly unlikely.

After making a number of repeat measurements, the following table gives a summary of the results. All masses are in grams (g)

	“Knife” Rod + Bearing     Big-Eng         Small-End			“Fork” Rod + Bearing       Big-End           Small-End
	396.6	205		557.8	213
TOTAL	601.6			770

Additionally, the connecting rods + bearings were weighed individually and the masses compared. These were: “Knife” Rod + Bearing = 603 g; “Fork” Rod + Bearing = 758 g. The correcting differences (+ and -) were distributed in the ratio 0.6 to big-end and 0.4 to small-end.

Other masses are:

Big-End pin = 515.7 g

Nuts (both) = 97.6 g

Piston + rings + circlips = 373.2 g

All of the information was now available to start detailed design of the crankshaft.

The AJS V-Twin Crankshaft: Part 1 – Determining the Stroke

The original plan was to build the engine minus the crankshaft and to then have Alpha Bearings make the entire crankshaft + connecting rods as with the AJcette engine. This turned out not to be an option and I had decided to use a pair of Harley Davidson EVO knife-and-fork connecting rods and big end assembly and to make the crankshaft myself. This decision had to be made early in the project to be able to machine the crankcases – in-line or staggered cylinders.

These con-rods/big-ends can be bought remarkably cheaply on ebay from the USA. The cost of the actual item was 60 USD (!!) and although shipping and import duty have to be paid when sent to the UK it is still remarkably good value.

 

These parts are manufactured in Japan; the quality is excellent and I really don’t know how they can make them for the price.

The top-down approach I decided on was the following:

 1)    Determine the stroke and calculate the compression ratio and swept volume

 2)    Weigh the connecting rod small ends/big end and pistons for balance calculations

 3)    Design the flywheels and mainshafts

 4)    Machine the flywheels

 5)    Machine the mainshafts

 6)    Assemble and check

The Harley Davidson HD connecting rod assembly that I had used in the design of the engine has a distance-between-centres of 7 7/16'' and the crankcase machining (distance between the crankshaft centre and the cylinder base) had been based on this.

In a previous blog I described how spacers had been added under the cylinder barrels of both cylinders to give the correct cam chain tension. This is really the only way in which the chain tension can be changed by any substantial amount as the alternatives would be either to remove metal from the cylinder barrel base flange or top face, neither of which is desirable, or to change the height of the cambox above the cylinder head, which is also not desirable as it would change the rocker-to-valve geometry. Shimming the cambox-to-cylinder head is an option for minor adjustment but not for gross changes.

The crankcases + cylinder barrels + spacers + cylinder heads + camboxes were therefore the starting point and the first step was to place the pistons at their TDC position in both cylinders and to determine the stroke that would result on each cylinder. From these 2 separate determinations of the stroke the minimum would then be chosen (to avoid the piston colliding with the cylinder head on the other cylinder!). This begs the question: “why would the calculated stroke be different on each cylinder?” Well, there are a number of possible reasons for this. Although the crankcase machining had been carried out accurately – the crankshaft centre – to – cylinder base flange was accurate to within a couple of thou and the 50⁰ V-angle was accurate to 0.1⁰, it transpired that there are differences between the cylinder head and barrel that were used on the twin-port Velocette KTP engine versus the single port engine.

With the piston (plus the top 2 rings) positioned at the top of the spigot on the cylinder, shown below.

 

The TDC volume was determined for both cylinder barrels/heads using a burette:

The volumes turned out to be quite different for the front and rear cylinders/cylinder heads with values of:

 V_rear = 87 cc       (KTP head)

 V_front = 74 cc      (single port head)

Why would the rear KTP cylinder head have a higher volume? I had always assumed that the only major difference between the single port and twin port Mk 1 OHC Velocette cylinder heads was the number of ports. But it turns out that there are at least 2 significant dimensional differences (there are some more detailed differences that I won’t go into here). The first is the height of the spigot on the cylinder head, 0.25” on the KTP head versus 0.3125” on the single port head, and the other is the diameter of the hemispherical combustion chamber – see pictures below

 

In which the KTP head is nearly 3.5mm greater diameter. No wonder there is a difference in volumes!  I have no idea why Velocette, in their wisdom 90 years ago, made these changes.

Before progressing any further, the major dimensions of the piston are required. In particular, the distance of the gudgeon pin centre to the top land, ie to the point where the piston would touch the cylinder head at the top of the spigot, which has a value of 1.116”  as seen in the sketch below of the important Ducati piston dimensions.

 All dimensions are in inches.

 Calculation of the stroke is then based on:

 L_{cr_to_spigot}  =  S/2  + L_conrod  + L_piston

 Shown schematically below

Where:

L_{cr_to_spigot}  is the distance from the crankshaft centreline to the top of the spigot on the cylinder and measurements are: rear cylinder  = 10.26''; front cylinder = 10.33''

S is the stroke - to be determined

L_conrod is the HD connecting rod length (7 7/16''  =  7.4375'')

L_piston is the height of the piston from the gudgeon pin centre to the top land (measured at 1.116'')

The measurements quoted above for L_{cr_to_spigot}  were made with the barrels set up to give the correct cam chain tension and with paper gaskets in place – this is the “as assembled” dimension. It can be seen that there is a difference of 0.070'' between the rear and front cylinders. This is very close to 1/16'', which is the difference in spigot heights between both cylinder barrels, and this is the reason for the difference in L_{cr_to_spigot} between the front and rear cylinders.

Inserting the values into the above equation for the stroke gives rear (S_rear/2) and front (S_front/2) half strokes of:

S/2_rear  =  1.7075''  = 43.345 mm

S/2_front  =  1.776''    = 45.125 mm

To avoid contact between piston and cylinder head a value lower than the lower value of Stroke/2 must be taken. The lower value of 43.345 mm is therefore rounded down to 43mm and this gives a stroke of:

S = 2 x 43 mm  = 86 mm

It is now possible to calculate the compression ratios of each cylinder and, of course, the swept volume.

To calculate the TDC combustion chamber volume there is now an additional volume that must be added to both rear and front cylinders resulting in the reduction of stroke from their maximum possible values to 86 mm. These are:

V_add,rear  =  1.57 cc

V_add,front  =  9.63 cc

And when these are added to the measured combustion chamber volumes for each cylinder we get the total TDC volumes as:

V_TDC,rear  =  87cc + 1.57cc  =  88.6 cc

V_TDC,front  =  74cc + 9.63cc  =  83.6 cc

It can be considered fortuitous that the difference between the maximum possible stroke for each cylinder and the actual stroke to be used has resulted in very similar values of combustion chamber volume. The original difference of 14 cc between the single port and KTP heads has been reduced to a difference of only 5cc at TDC.

With a bore of 76 mm the swept volume is calculated to be:

V_swept  =  390 cc

Giving a total engine capacity of 780 cc

And the compression ratios (CR) of each cylinder are:

CR_rear  = 5.4

CR_front  = 5.66

Whilst the relatively low value of compression ratio may at first seem disappointing the effect of this needs to be understood.

It originates from the use of Ducati pistons intended for road use. These have a lower dome height that a period Velocette piston. The picture below shows the Ducati piston (on the left) and one of the Velocette KTT pistons that I had made by JE Pistons and copied from an original Mk1 KTT piston. These have been positioned, as closely as possible, to align the top lands so that the dome sizes can be compared and it is seen that the Ducati piston is about 0.25'' less, thereby increasing the TDC clearance volume.

For the record, the AJcette, which uses an identical Velocette piston to that shown on the right above, had a (measured) compression ratio of 7.5:1 (see here and here).

I have also previously used one of these Ducati pistons in the K7 restoration. The picture below shows a comparison of the Ducati piston with the original K7 piston.

The compression height of the Ducati piston is 0.177'' lower than the original AJS piston. The compression ratio of a K7 fitted with a standard original piston, as delivered in 1928, is quoted as 6.25 (see instruction manual) (a race piston was available from the AJS works to give a compression ratio of 7.5). A reduced dome height of 0.177'' would increase the TDC volume by 19cc  (approximately, because the domes are not identical) and the compression ratio would be correspondingly reduced to 5.1. Although I unfortunately did not measure the compression ratio of the K7 at the time of restoration it would have been around this value (5.1); nevertheless, the performance of the bike was excellent.

So, what is the effect of using a compression ratio of ~ 5.5:1 ? In reality and for the purpose which the bike will be used (enjoyment in riding ….I don’t plan a re-run of the World Speed Record attempt!) it doesn’t really matter. I would not expect the power to be reduced substantially compared to a ~ 7.5:1 compression ratio; however I would expect a lower efficiency which, in turn, would translate into poorer fuel consumption and a slightly higher exhaust temperature.

These Ducati pistons are intended for “Road Use”. There are also Race pistons available which could be used.

Picture courtesy of Lacey Ducati

And maybe I’ll buy a pair of these at some time in the future for the “race kit”

 

The Mudguards and Oil Tank

Over the years, I have established what has become a well-trodden path for fitting mudguards; not so for making and fitting oil tanks, as will become clear.

Mudguards

The original AJS record attempt bike did not have any mudguards. That was fine back in 1930 when they were attempting to win the world speed record but not so good for riding around todays roads in the UK ….because it is illegal.  The bike will be run with open exhausts, which will possibly attract a bit of unwanted attention, and the absence of mudguards would only compound any consequential “issues”.

Vintage bikes, including the original AJS world record attempt bike, have a very “skinny” look – large diameter and narrow wheels, thin mudguards, narrow petrol tank etc which stems from the early days of motorcycling. I like this lithe appearance and wanted to preserve it in my recreation, even if the rest of the bike is dominated by a bloody great engine!. The wheels were already “skinny” – 21” front and rear and the mudguards were chosen to match these; specifically 4” across / 29” diameter for the front and 4 ½” across / 30” diameter for the rear. I have been getting my mudguards from Renovation Spares (contact Simon at renovationspares@mail.com) for many years – all my recent restorations including the AJS K7, the AJcette and 2 Model 18 Nortons use their mudguards. The mudguards come in freshly rolled raw steel which then needs to be drilled and cut to length accordingly.

 I’m sure everyone that fits mudguards has their own way of making brackets, stays etc. I make mudguard brackets by bending a piece of 3mm thick steel around an appropriately sized cylinder and then using a section of mudguard chopped off the original (there is plenty there) as a former to “push” the bend of the bracket into the corners. This particular bracket is for the rear mudguard that is bolted directly onto the frame tubes on either side.

This need quite a bit of localised heat with oxy-acetylene to get it red hot in the right place.

The front mudguard stays on vintage bikes up to ~1929, certainly AJS and Norton, often used simple ¼” diameter steel rod; in later years, thicker, usually ½” diameter tubing was used. I decided to use the thinner option here for the front stays which were made from ¼” diameter steel with an “eye” at the end to enable them to be screwed to the bracket; this approach is simple, cheap, light and effective and is identical to the earlier AJSs.

Again, quite a bit of concentrated heat in the right place is needed to form these ¼” ID “eyes” bending them around a ¼” steel rod.

On the rear mudguards, I use ½” diameter thick-walled tube. These are more substantial than the front because the rear mudguard needs to support the weight of a rider on the bump seat and the rear mudguard stay is often used by the rider to pull the bike onto the back stand, although I don’t plan on using one for this bike.

Some restorers flatten the end of the stay in a jig and then drill it to hold it to the bracket or frame. The thick-walled tube that I use is a bit too thick to flatten with a good appearance and I machine “ends” from mild steel that are inserted and silver soldered into the tube. The picture below shows the collection of “ends” for the rear mudguard before attaching them to the tubes.

The bottom “end” is made from 2 parts that have been TIG welded together.

The whole assembly of mudguard, brackets, tubes (appropriately bent to shape) and the “ends” are then assembled in exactly the right place and each tube and its “end” are tacked together with the smallest of TIG welds, shown below.

This is sufficient to hold the assembly together so that it can be removed and each of the “ends” and tubes are then silver soldered to make a permanent and strong fixture. After trimming each mudguard to length, the fabrication of these is now complete.

Oil Tank

The original AJS speed record bike has an oil tank clearly visible on the right side sandwiched between the upper and centre frame tubes.

and this is where I planned to put the oil tank. But how to make an oil tank?

As I have mentioned in previous posts on this blog, I am not a sheet metal worker. Yes, I can weld and braze a couple of pieces of sheet metal together but I have neither the experience nor the sheet metal specific tools for making the compound curves that are required here.

The alternative is to find something that already has the curves and about the right shape and size and to re-purpose it accordingly. I eventually decided that a Royal Enfield Bullet toolbox would be about right and bought this one

on ebay for £28. There are plenty of these around so one in good condition can be bought cheaply.

The plan was to use the nicely curved front in which would be the brass outlet and return fittings and the filler cap neck and to add a back and brackets.

The first step was to have the toolbox grit blasted and see if would fit in the available space.

Although a snug fit, it seemed that it would squeeze in nicely and so the next step was to start making fittings. The first was the neck for the filler cap. This was machined on the lathe from a solid piece of brass to fit the filler cap …another ebay find.

A cardboard template with the approximate curved shape of the filler neck/oil tank interface was then cut out of cardboard

and the neck was marked in white paint and cut out approximately on the milling machine.

It was then finished with a file and dremel to fit the tool box front.

Fittings were then made for the oil feed (which uses the original hole for the toolbox locking screw) and the oil return, into which is silver soldered a length of copper tube to bring the outlet up to the top of the filler, and a section of the sheet metal was removed from the top for the oil filler neck.

The next stage was to attach the back – which was the original tool box back reduced in depth, brackets to attach the oil tank to the frame and the filler neck using both braze and silver solder.

Unfortunately the heat and/or the attachment of the neck to the tinware distorted the filler neck so that it ended up oval  -  in fact about 1/8” out-of-round on the diameter and there was no way that the screw top would fit. It was either a case of throw it away and start again (not a good option as I had already spent about 3 days on this and there would be no guarantee that the same problem wouldn’t happen again) or try and fix it.

I squeezed it back to a round shape as best as I could in the vice but this is nowhere near accurate enough to screw in the filler top. I ended up making a 1.6” diameter tap using mild steel (it only has to work once – on brass) and managed to get this started on the existing thread to recut the thread. The “tap” started off as a taper tap and, as I was able to get it in further, was faced-off to eventually turn it into a plug tap.

Pretty crude but it worked and I was able to get the tap screwed right down

Such that the filler cap would screw into the thread properly

 

At this stage I could try it on the bike

Not too bad!