But going back to that final year of development of the ultimate factory Works bikes - 1954, for that year the Works bikes were very special with obviously noticeable differences to the Production race bikes.
Most visually different, the engines that year employed an external flywheel on the drive side - easily identifying it from the 53 Works bikes. Other key differences were additional oil cooling to the head and lower vertical bevel housing - with the oil union from that housing leading to an external oil cooler located on the frame front down tube. 

But there were other less noticeable ancillary part details as well - if you look closely you will see in the picture above that the Lucas racing magneto - a Lucas SR-R at this time - was actually leaning back from the barrel slightly - while for the Production Manx models the SR-R (of which I have one fitted to my own 1955 DOHC Manx) the magneto sat vertically. Only a handful of these slanting magnetos were made by Lucas (specially for Norton's to my knowledge), and I assume this was to provide better air flow around the barrel and magneto.

And then there was the carburettor. Camshaft Norton's had been fitted with competition Amal carburettors since Amal was formed in the late 1920's (an amalgam of Amac, BandB and Binks) - and the race bike models had always used the latest competition versions - initially the TT carb through the early-mid 1930's, then the RN (Remote Needle) when that was developed in the mid/late 1930's, until by the early 1950's - Amal's were starting to develop experimental carbs for the Norton Works team initially for the first Featherbed framed Works machine, with what became the Amal GP carb - and which finally culminated in the carburettor that is shown here for the 1954 season - which was very different to any carb offered to the public - but I am guessing was the prototype to what later became the Amal GP3 (remembering this was for 1954 - the first year that Production Manx models were fitted with the 'Short Stroke' engine, and smaller Amal GP5 carbs - which my own 1955 Production manx is fitted with.

I confess, I know almost nothing about the development history of this carburettor - other than knowing it had been developed for the Norton Works team, and has an external copper fuel feed pipe from the base of the main carb body casting to the slow running bore - which I gather was done to improve slow running - but although on first glimpse, it is visually similar to the production Amal GP carb - on closer inspection, it is quite different with external (choke?) elements on both sides of the main carb slide casting.

As shown in the original 1954 Works bike photograph at the start of this section, it is clear to see that Norton were using a very distinctively shaped (velocity stack?) type bellmouth, with a large and very recognizable bell shaped profile. If you look at similar photographs of the earlier Featherbed Works bikes, even in 1950 the bellmouths looked slightly bigger than normal, but by 1951/52 they had developed this large bell shape profile - and 1954 (the last year of the special Works bikes, if not including the F Model) was much the same - the large bell shape, with a distinctive 'ring' lip at the open end.
Having looked at one of these Works carbs quite closely - it can be seen that this 'lip' at the open end was actually recessed on the outside of the lip/bellmouth - presumably to save weight.
As an aside, I remember many years ago seeing a similar bellmouth fitted to at least one bike built by that legendary Manx tuner - Francis Beart, but if memory serves me right - his bellmouths had a shallower external curve to the bellmouth lip, without the recess (I am sure someone will correct me if I have got this wrong!).

So, having talked a little bit about the original carburettor, and the Works Norton's it was fitted to originally - what about this carb, that you see photographed here and what is it for?
Well, first of all - it will be no surprise to anyone for me to say I am not in possession of an original 1954 Works Norton! . .. or even a complete engine, but like anyone interested in the history of the Racing Norton models of that period, I have always had a fascination for 'Works' parts . . . even though I know most of the proper Works parts are unlikely to fit a standard Production bike!

But I am fortunate to have been good friends for many years with a lovely gentleman who is also a serious collector of Norton camshaft models, and has owned quite a few of the Works bikes, including at one time a 1954 external flywheel Works model.

As many owners of old race bikes will know - race bikes and parts can have a hard life, and as part of restoring these rare Works bikes, my friend sometimes had to have made a new part from a worn original - and when doing so, as in the case of this carb, the original would have been used as a pattern, and maybe an additional casting made as a spare - this being the case of this carb body. But not for him to make something that was just 'close', he was fastidious to ensure the part was made as close to the possible specification as possible, and this one is no exception.
I am sure my friend would not mind me saying, he is now in his latter years and like so many of us, has realised he needs to cut down his collection - and with it, one or two of the special spare parts - hence, having spotted this lovely carb some years ago, I was fortunate to have been offered it (as an incomplete unit) some months ago. Needless to say - although not having a full Works bike to put the carb on - I did not want to turn down the opportunity of such an interesting trinket - and have already got a good idea of which bike project I am going to reserve it for - most probably my 1950 DOHC Gardengate Manx, but if not that - then possibly another slightly later engine project I have in mind for the future. – which may itself be the subject of a future article!

This Carb - As It Came To Me

As acquired, the main carb body was pretty much complete - but missing its throttle slide, its top cap, and most noticeably - its distinctive Norton Works type bellmouth - but I was fortunate that I was able to take close dimensions from an original bellmouth still in my friends' possession.
It was also missing its (remote) float chamber - and although in the first photograph of the carb earlier in this article (with it sat on a wooden cabinet top), it has an Amal 302 type remote float chamber next to it - that was just because I happened to have one to hand for that photo. If you go back to the original Works photograph Drive Side - shown on the right, you will see that the floatchamber fitted to the original 54 Works bikes looked slightly different - and I believe those floatchambers were referred to as 'Weir' type floatchambers, and I think (but am not entirely sure) were made specially for Norton Featherbed Manx models - I am assuming the Works carbs first (I believe from approx 1952), but also later Production models. And actually - the good news was I already had one of these available - I had acquired a very nice one of these many years ago, and was holding it to one side, with the 1950 Gardengate Manx build project in mind - so that was that sorted!  

Making a Norton Works Type Bellmouth

So, those that visit our online parts catalog (www.RacingNorton.co.uk - in case you haven’t been there before), you might know that as well as the Norton parts we manufacture – we also make a number of carb and magneto parts, as close as possible to the original patterns. Well one of the carb parts on my ‘To Do List’ when I got round to it – was the very early competition Amal 10TT bellmouths, of the pre-1934 type, where the bellmouth thread is internal and originally they were steel and a very different design to the alloy bellmouth and lockring that followed on all Amal comp carbs from 1934 onwards. Many years ago I had supplied these in steel, but my manufacturer at that time was no longer doing them – so I had decided to make a batch from solid alloy billet, and had already had the wide diameter alloy billets already cut. To get these rather large billets onto my CNC lathe, I knew I would need to cut a big ‘soft’ 5c Collet, to hold the billets while they were machined. 

Finding a Suitable Billet

It occurred to me – having taken measurements of the Works 54 type bellmouth – if I could find a billet of alloy of similar outer diameter, maybe I could hold it in the same ‘soft’ collet I had already made up for the pre-34 type bellmouths. Again, to cut a long story short – I was eventually able to find a large billet of alloy on E-bay . . . which was very close to the required dimensions – just a couple of millimetre short of the original Norton diameter (at its widest part), but unfortunately – about 8mm short of the required length to ensure that it could be held in my custom lathe collet and leave enough metal for ‘Parting off’.
However, with the price of custom cut alloy billets being high . . . and this billet being reasonably priced – I had no issues with just amending my outer diameter by a couple of millimetre, and deciding I will get the length as close as possible, and not part it off, just craft the outer lip on the manual lathe and a 2nd op . . . after all – who would know the difference if it was a fraction shorter!!  

Holding The Billet In The CNC

This allowed me to load it into my CNC machine, held by the big (soft) 5c collet - previously custom made for holding the smaller pre-34 billets, and then move in the (pneumatic) tailstock revolving-centre I have on the CNC lathe. This would allow me to machine the main outer profile, as well as the outer diameter thread, in relative safety, without worrying that the side thrust would throw the billet out the collet and turn it into a garage missile! As a bit of background to programming a CNC lathe at this point – I write all my CNC lathe programs in raw G-Code, i.e. I do not use a CAD-CAM application, which is the normal modern method I gather (although I do use CAD-CAM for the occasional milling programs I write for my small CNC vertical mill). However, I do have an emulator for my CNC lathe – so I can run a CNC program in ‘virtual’ mode first, check what it looks like on the emulator, and make sure I haven’t got any silly errors.
However, I normally expect that when I come to setup the actual machine . . . and set the correct offsets on all the induvial tools, that I might end up wasting a couple of pieces of material, until I have got everything setup correctly. That is fine when I am working on bar stock (most of which comes in 3 metre lengths) . . . but obviously with this job – where I only have a single large alloy billet (which still cost a reasonable amount of money) . . . my approach had to be a bit different – I couldn’t afford to screw things up!
So, I had to go much slower in my approach than I normally would – and on a couple of the turning tools – make sure the offsets (i.e. where the actual tip of the tool sits in relation to where the software on the CNC thinks it is) were generous – I would do some trial cuts for each tool, knowing they should be a few fractions of a millimetre greater than the actual dimension should be – and then take a micrometer reading – and adjust the offsets from there.
The result of this was that the outer profile of the bellmouth went relatively smoothly – although taking much longer than it would do for a production program – knowing this was a one-off job, and it was my ‘own’ weekend time I was wasting, after all – it was still much faster than having to make hundreds of individual passes on my old Smart and Brown manual lathe! Likewise, I only had one shot at the thread. I had already measured the thread diameter on the carb body – this of course being a female thread, and figured out that the pitch was the normal pitch used by Amal’s in that period for their competition carburettors.
I already had a threading sub-program written for RN type bellmouths (which we manufacture) – so took the format for that as the starting point for the much larger Works type (GP3) thread dimensions – but just like the turning tools – I only had one shot at the thread. So, without going into all the complexities of single point thread turning on a CNC (using solid carbide thread turning inserts) – the result was a nice looking thread, but actually if I am honest – it was a few fractions of a millimeter down on perfect – but actually, being such a big diameter thread, when I trial fitted it into the body of the carb, it went in smoothly, with only a minimum looseness – and when the shoulder of the bellmouth locked up against the carb body, and the lockring was then tightened . .. the thread did its job just fine for a one-off . … after all – once the bellmouth was in place, I wasn’t intending to remove it again!
There was quite a lot more to the outer profile of the bellmouth – not least being the special machined groove that was added to the large outer bell profile by Norton’s on the original version (presumably to reduce weight by removing as much excess weight as possible) – but as the CNC program progressed, and the outer profile was looking quite good – I could see that cutting the groove on the outer profile would be best done after having done the inner bore – and actually . . .having so little room for error – it may be more prudent to switch to my manual lathe for the next step – boring the inner inlet tract of the bellmouth – if nothing else, just looking at the amount of billet protruding, and knowing how little of that billet was being held in the collet, I was afraid that once the pneumatic tail revolving Centre was removed – there was too big a risk my bellmouth may become a projectile! So at that point I took the billet – which was already starting to look a bit like a bellmouth – into my 1950’s Smart and Brown manual lathe and started the long and laborious task of boring out the inner profile. Although this lathe is not small – because it does not have a clutch, and is single phase, I don’t have the biggest motor on this lathe . . . and with bigger jobs, I just have to take more time than is ideal and take smaller cuts. Eventually I did have it roughed out (probably 2 or 3 separate afternoon sessions, with Classic FM on the radio to stop me falling asleep!) – and although I considered putting it back on the CNC for the final profile (which I had already written and programed) . . . I realized that as the final bell profile would be quite thin . . . and I would only get one shot at it on the CNC . . . it might be better to just bite the bullet and resign myself to continue to machine the inner profile manually by hand, and use my fingers to test the final thickness . . . after all – who needs a CNC, Norton’s didn’t have them – and there is nothing better to test thickness than the feeling you get by touching it with your own fingers!!

Swop Over To Smart And Brown For Internal Bore

First task once I had mounted the bellmouth in my old Smart and Brown toolroom lathe was to rough out the recessed lip on the outer profile – as per the photograph on the left.
For this I used the same lip I turned on the billet to hold it in the special 5c collet - but now in my old fashioned 3 jaw chuck(!), but again, also applied a tails stock centre to hold. It. That allowed me to safely rough out the lip, but I knew I would need to come back to it, once I had bored the internal bellmouth profile.

Having done that, it was then a case of removing the centre and running numerous twist drills through the billet, until I got my largest machine drill, somewhere just over an inch. After that, I fitted a boring bar, and then it was just a case of taking long and boring (no pun intended) passes, incrementally opening out the bore on each pass.

Because this lathe was originally built for a 3-phase motor . . . but many years ago I fitted a large, but less powerful single phase motor, and because it does not have a clutch (I never understood why – it was built as a high quality industrial lathe in the ‘50’s when Clutch’s were fitted to bigger lathes) - I decided to play safe
. . . I knew I could not take as big a cut as you would expect with a lathe this big – I had adjusted it so the belt would slip at a certain point . . . a rudimentary safety device of sorts . . . but this meant that large boring jobs take longer than I would like – so might as well just turn the music up on my I-Pod headphones and stare into space as I flick in the auto-feed on the saddle for each pass!

Therefore, I know with these larger jobs, I just have to take smaller cuts as the bore gets larger - and accept . . . it is what it is - and it will take longer than I would like, but at least if the worst happens I will hopefully not lose my fingers!

Boring Multiple Passes - Slow and Boring

Boring the main bore with multiple drill and boring bar passes. Not very exciting, but a necessary evil - with no clutch (as explained above), I set the belt so it will slip if too bige a cut is used.

The photos that follow show the bellmouth mounted in the Smart and Brown Toolroom lathe, first with a series of machine twist drills – up to just over an inch, and then after that a long boring bar, that went almost, but not quite the full length of the inner bore.
After this point, and much time spent looking into space, I had the inner bore on the thread end, just shy of the finished bore. And at this, I then turned the bellmouth round, so I could duplicate the bore on the bellmouth end and then opend out the inner bell profile. Most importantly, to do this and ensure the billet was well supported for this final shaping – I was able to mount the bellmouth mid way along its length, but then fit small alloy spacer pieces, so the jaws of the chuck did not mark the outer profile, you can just see these in one of the photographs.
Having first opened out the parallel bore so it matched the other end, I then set the crosslide top slide with a taper, so that it was able to rough out the major taper of the bellmouth. This roughing process of the taper was relatively straightforward, but took quite a while, but as I started getting closer towards the finished bore – I was faced with a dilemma – do I put the bellmouth back on the CNC lathe, and trust the CNC program I had previously written to automatically machine the final profile (and had tested on an emulator), but knowing if I had got my offsets wrong – or the indented lip I had manually machined was in any way different to the CNC program – I could bugger it up and waste all the time I had already taken, and would have to do it all again.
So having made the decision to continue with opening out the bellmouth profile on the manual lathe, I spent a fruitful hour or two gradually opening up the bellmouth and changing the taper angle on the topslide and gradually opening up the 'bell' until it reasonably closely resembled the general profile of the original. As this profile started to get closer to a finished version - although at this stage still a series of joined up straight lines, I figured I might as well just stay as I was on the Smart and Brown, and not bother putting the billet back on the CNC.

So as this was just a one-off bellmouth, I decided it would be safer to keep it mounted on the manual lathe and continue to do the final inner bellmouth profile by hand – and trust that by continually using my finger and thumb to gauge the thickness of the wall by ‘feel’ alone, I would not break through and screw it up.
So, by continually changing the taper angle on the lathe topslide, I was able to rough out the final bell as much as I could – but when it got to the final finishing shape, as I have done many times in the past, I swopped over to a lathe ‘scraper’ tool.

Now to Shaping Out The Inner Bell

To anyone not familiar with using centre lathes, this most useful tool may seem out of place – but actually, I have two or three hand scrapers (normally with a particular shaped and ground blade head – as per the one in the photo), that are used primarily for chamfering and removing burrs on internal bores – and have never been without one, in fact I still use one that came from my father’s workshop, who was a toolroom engineer himself.
But the other thing that hand scrapers can be used for, if you are careful, is for internal bore shaping on softer materials like aluminium – very much like using a long handled chisel on a wood lathe. But unlike a wood lathe, when using a hand scraper on a metal lathe – I definitely don’t use a steady bar to hold the scraper against, instead I move the saddle out of the way, and use my left hand to hold the scraper and my other hand to help support that hand – then if the worst happens and the scraper digs in or ‘catches’, it is less likely to get ripped out of my hand and turn into a flying dagger! And you can actually be surprisingly accurate with one of these tools with a little practice, and in the case of the final bellmouth profile this was the ideal tool, it allowed me to nicely shape the final profile of the bellmouth, but also gave me a better feel of the amount of metal I was removing.

As the wall thickness between inner and outer profiles was reducing, I also had to shape the outer bell curvature to take into account the lip I had machined in the end of the billet to mount the billet in the CNC collet earlier and there was very little wriggle room with this – in fact, I was still worried I would not have quite enough metal to allow the outer curve to be formed without leaving a small groove because the lip was too deep, or the final profile just looking wrong.

Final Shaping and Finishing - Would It Look OK?

But as it happens, I just about had enough meat left on the billet, and all turned out ok, using a combination of hand scraper, light cuts with a finishing tool, then the final finish and polishing with a combination of cloth emery tapes of varying grades, until I had achieved a pleasing blending of all the different tapers into one single continuous flowing inner bellmouth profile - which did not look that different (well visually anyway), from the original Norton bellmouth profile.
The final result looked pretty much like I had hoped it would and I was quite pleased with the result.

Final task before removing the bellmouth from the lathe was to pull out my favourite alloy polishing compound - Brasso!, and having first used it poured onto 600 grade wet and dry paper, then finallly some cloth soaked with Brasso - providing a lovely polished internal profile that looked really nice.

I was also able to blend in the outer profile and achieve the correct profile, carefully removing metal until I had 'just' reached the indent for holding it in the collet - which I had been most worried about removing. But very carefully blending back (again, almost entirely with a hand scraper), eventually I just caught the inner point of the lip, and removed it, without affecting the overall profile.
This outer profile, including the recess behind the bellmouth lip was then given the same emery cloth and Brasso polishing treatment - until eventually I could step back and breathe a sigh of relief - Job Done!

The Finished Works Carb - For Now

As you can see from the final photographs, with the bellmouth now fitted to the carburettor – the carb looks far more like the original Works Norton carb as fitted to the 1954 bikes. I will need to come back to it at some point in the future and address a few other small parts and jetting – but for the moment it makes a very pretty shelf ornament and talking point – and as I mentioned earlier, there is a very good chance I might try and fit it to my 1950 Gardengate Manx or a possible short stroke engine project at some time in the future.