Wednesday 23 February 2022

The AJS 33/7 Trophy Model Restoration: INDEX PAGE

The links below are in chronological order of the AJS33/7 restoration to aid navigation of the blog.

First Assessment

Preparing the Gearbox and Clutch

The Camshaft Drive

Engine Plates

The Cambox

Finishing the Camshaft Drive 

The Primary Drive 

The Gearchange Positive Stop

The Petrol Tank, Battery Carrier, Oil Pipes, Gear Lever, Footrests, Carburettor and Wheels

Mudguards, Tank Panel, Kickstart, a day with Ron Langston and the 2022 Manx GP 

The AJS 33/7 Trophy Model Restoration: Headlight Bracket, Front Stand and Seat  

Cylinder Head Steady

Metal Finishing and a Trip to Ukraine 

The Crankshaft and Piston/Cylinder Assembly

Start of the Re-assembly ...and Motorcycling in Kenya  

Final Assembly and First Start





The AJS 33/7 Trophy Model Restoration: The Camshaft Drive


I am now fairly familiar with these, having made 3 – the first for the K7, the second for the AJcette and the last one for the V-Twin

The collection of bits that came with the bike is shown in the picture below.

Interestingly, this cassette-like arrangement is very similar in design to the AJcette camshaft drive that I built.

However, there were a number of problems to be solved before any of this could be used.

The first was that the total length of the gear + sprockets at 1.526”:

was 0.134” less than the inner distance between the ends of the housing.

Not only would this be an unacceptable level of end-float (!!) but there are also some critical dimensions that must be held to avoid the chains either interfering with each other or hitting the casing.

The second problem was that the hole for the bearing in the end-cap housing had nothing to prevent the bearing falling out of the end! There was no lip. As this is the first time I have ever worked on one of these later engines I have no idea if this is the original design (I can’t believe AJS would have done this) or whether someone has machined it with a larger diameter at some stage to be able to fit whatever bearing they had available; I suspect the latter.

And last, but by no means least, the gear appeared to have been heated at some stage and my FFT (in my working life this stood for Fast Fourier Transform ….but now it is the Fine File Test – applying a fine file to the surface to test the hardness) indicated that the gear had been annealed and was relatively soft. Not a good start.

The first step was to find a narrower bearing with an OD slightly larger than the existing diameter in the bearing cap and to machine the cap to give a supporting lip. After searching the various online bearing suppliers I was able to find a metric bearing that would allow a 2mm width lip and so the bearing housing was machined accordingly with 0.0015” interference fit on the ID

and the new bearing popped in after heating the housing.


The inner bearing housing, that locates into a large machined hole in the crankcase, had not been molested and after inserting a new bearing, which abuts against the crankcase casting on the inside, the distance between the bearing inners could be measured accurately by clamping the housing in the vice and using a telescopic gauge.

The next step was to sort out the heat treatment of the gear. My FFT test is pretty crude: if the material in question could be worked with a fine file – such as mild steel/EN1A, then it wouldn’t even register on the Rockwell C scale (it’s too soft); if the file glances off the surface and the material is impossible to file then it’s pretty hard, maybe around HRC = 60 (this is the hardness of my heat-treated O1 tool steel gears that I have made in the past).

It just so happens that NovaTransmissions is not far away and they have on-site hardness testing and, as I know the guys there, I dropped by to get an accurate measurement of the hardness. My guess was around HRC = 40 and it actually turned out to be 43. I must have used a reasonably accurate file!

However, there is a difficulty in heat treating steel if you don’t know its specification. Luckily, as this gear is original and would have been made in 1933, AJS would almost certainly have used a through-hardening carbon-based steel, not unlike the O1 tool steel that I use for these gears. I therefore used a similar heat-treatment, namely heat to 815C, oil quench and immediately temper at 220C.

After heat treatment, the FFT indicated the gear was significantly harder and a quantitative measurement showed it to be HRC = 53.

I would have liked it be closer to 60 but as there is a relatively light load on this gear and it has plenty of oil I don’t anticipate any problems.

The last problem to sort out was the spacing of the gear/sprockets to get the correct end-float and clearance for the chains.

Whilst it would have been possible to put shims or thin spacers between the gear/sprockets to correct this, I would consider this an inelegant solution and I therefore decided to make new sprockets with the correct dimensions. As with previous AJS camshaft drives, I acquired 2x 3 row 17 tooth sprockets (EN8 steel)

and machined these to give the boss dimensions that I wanted, retaining the middle row of teeth of the 3-row sprockets

and then spark-eroding the keyway and making a new shaft. Note that the keyway is aligned with a tooth on one of the sprockets and is staggered to be midway between teeth on the other to avoid the chain rivets on the adjacent camshaft and magneto chains from interfering.

The dimensions were checked carefully before pressing the sprockets onto the shaft

and then checked again in the housing. The square on the end is to drive the oil pump.

As with previous engines that I have built, the gear and sprockets were each laser welded to their immediate neighbour by EMP tooling. This technology did not exist when the bike was built back in 1933 and would have relied on transmission of the load between the gear and the sprockets by the interference and keyed fit onto the shaft. Laser welding them in this way allows the load to be transmitted directly between the gear and sprockets. The picture below shows the elegant 3600 weld between the gear and the inner sprocket and the filet weld between the two sprockets.

The dimension was checked again:

and gives end float = 0.004”. Perfect. And when inserted into the cassette and clamped up the assembly rotates freely, as expected.

One more little job done.

ps. I've just check that the movie plays OK does but I didn't realize that I had Classic FM playing in the background. I don't usually put any of my work to music.