The TR25T crankshaft that I am busy liberating from its crankcases has had a rough life - possibly since it was half dismantled by a bike breaker and then thrown about, or perhaps by previous strip downs involving hammers. Both ends of the crankshaft are damaged: the oil feed end is very slightly lipped - so will require some lathe work. And the drive side end threads are smeared. Seems that they have had a wrong thread nut on them at some point, and possibly some other damage.
I think there is a remote chance of saving this thread. But I want to move the metal back where it should be - so not by cutting or filing (so not a die or thread file). As the threads at the flywheel end are undamaged, I need to start with a nut at that end (which is impossible...), and then undo it. With plenty of oil. And repeat. At the moment my thought is to split one side of a nut and open it sufficiently to slide it up the shaft to the good threads, and then close it back up into a spanner and unwind it.... Has anyone got any better suggestions? Plan B is taking it to be built up with weld and rethreaded....
You could completely split the nut (through the peaks of the hexagon rather than the flats) then use mole grips or similar to clamp it onto the good portion of the thread. Might work......
I have now found various threads on other forums debating the thread form on A65/A50 cranks from around 1970. Seems that 20TPI cycle thread may or may not have been changed to 20TPI UN (not the same as UNF) and may or may not have been changed to a 60 degree 20 tpi thread. Meanwhile I have found two dealers selling 5/8" 18TPI rotor nuts...
I will split one of my 5/8 UNF nuts fully as you suggest. And before I do anything els I'll try it on the good threads to check the match...
I'm not aware of UN as opposed to UNF (what does it stand for?). I accept that theoretically UNF 20 tpi is different to Cycle 20 tpi but both are 60 degree threads and in my opinion the minor differences in geometry are not practically important. In your case I'd suggest not an issue.
I'd never heard of a UN thread either but this screen shot shoes 5/8" threads in Unified Coarse (UNC) 11tpi, Unified Fine (UNF) 18tpi and Unified Extra Fine (UNEF) 24tpi forms and Unified (UN) with 12tpi, 16tpi, 28tpi and 32tpi forms.
You can read the whole Unified specs here Unified Inch Screw Thread (UN UNC UNF UNEF) (https://www.threadspecification.com/un/)
Interesting web site. It drove me to look at my Machinery screw thread book (18th edition 1961). This has a comprehensive section on Unified and American Series and UN doesn't appear anywhere in it. On the website, UN seem to be in between UNC and UNF TPI. There isn't any reference to standards that I can see however.
Quote from: iansoady on 28 Oct, 2026, 12:04 You could completely split the nut (through the peaks of the hexagon rather than the flats) then use mole grips or similar to clamp it onto the good portion of the thread. Might work......
I would be tempted to try only cutting one side and then wedging that cut to get the nut on whilst slack and then clamping for the removal . But sounds like a good idea from Ian .
So - the 18TPI rotor nut arrived from one dealer. I checked the threads - and they are 20TPI.... (I haven't worked out what thread form.) The other dealer is in the US and doesn't ship to the UK. I note that he lists his 18TPI nut as interchangeable with his 20TPI nut. So now I think I have been led up the garden path...
My experiments with a split 18TPI UNF nut were succesful in parts... Yes it sort of worked in that the reconstituted nut, held in a box spanner, would eventually run off the shaft. But I am now less convinced that the original thread was 18TPI.
I think I am back where I started - a trip to an engineer to have it built up and rethreaded. Probably to 20TPI cycle thread - which is what BSA used before the alleged change to UNF (UN? whatever). I shall make certain I have a nice clean rotor nut with a known thread to hand!
Have you actually checked the thread on the shaft with a thread gauge? 18 and 20 tpi are different enough to be fairly obvious in my experience.
Just a quote from the 1969 parts book:
"NEW PARTS HAVING THREADED FIXINGS INTRODUCED IN THIS CATALOGUE HAVE THE UNIFIED
FORM OF SCREW THREADS. THIS CHANGEOVER WILL BE CONTINUED AS AND WHEN NEW PARTS
ARE INTRODUCED. EVEN WHERE PARTS ARE SIMILAR IN APPEARANCE NEW PART NUMBERS
WILL BE ALLOCATED TO IDENTIFY THE UNIFIED THREADS. "
It would be worth asking one of the specialised dealers eg CD Autos.
Have you tried a thread roller? This reshapes the thread without removing metal.
A 60 deg and knife edge needle files work very well, its a slow careful job but you can work on one area of damage at a time.
Have you actually checked the thread on the shaft with a thread gauge? 18 and 20 tpi are different enough to be fairly obvious in my experience.
Yes. But a lot of the threads were smeared over - which made working out where the crest originally was a bit of a guessing game.
If you look closely, some of the 'crests' look like two crests that have been pushed together.
I'm also intrigued by the marks on the shaft shoulder.
OK its bad but not terminal. The thread roots look OK and true to pitch but it looks to me that a wrong pitch nut has been forced on and damaged the crowns. Using a roller guided by the roots would work perhaps even a tread file but first be 100 % sure of the correct pitch. That 18 tpi does not look like its correct, the parts book shows 20 tpi.
Try with a steel rule and count the troughs over 1/2". A lot of BSA / Triumph cam shaft nuts are coarse CEI which is 20 tpi. 37-1798 is 20tpi CEI
The nuts is holding a gear locked by a key so I don't think the nut toque needs to be that high which means a less than perfect thread should be OK.
Buy the correct part number nut and confirm the thread in that first.
A useful guide to some thread sizes can be found in the factory torque figures listings, in workshop manual or service bulletins.
Below are the torque settings for the 1969 B25 Starfire.
The cranks are sometimes found abused due to the difficulty in dismantling the cases with ball bearings, maybe a previous owner using a hammer on the shaft end to get it out of the case.
The marks on the end of the plain section could be an attempt to fix a loose fitting rotor?
That is a very intersting table!
Look at the:
crankshaft pinion nut - 0.625 x 20 TPI BSF .919" AF
rotor mounting nut - 0.625 x 20 TPI BSC 1.01"AF
Same major dia (5/8), same TPI 20, one (CEI) 60 deg and one 55 deg (BSF) on the same crank
And one taking a 1/2 BSW spanner and the other a 9/16.
Sometimes I think we are being set up to fail ;D
Either way I think your crank is 20 tpi but get the correct nut then the correct die since both crank threads appear to be the same TPI but different threads.
It is indeed interesting, especially the mixture of Cycle, BSF and UNF! I've also never seen torque wrench figures for an British bikes.
I now have sets of BSF, BSW, BSC, NF and UNC taps and dies. All cheap off e-bay, but its the only way I can be sure of identifying a thread and even then there are bear traps like 1/2 x 20 CEI and 1/2 x 20 UNF.
I still don't think there are appreciable differences between 20 tpi CEI and 20 tpi UNF. Both are 60 degree thread angle; the only differences are in the rounding / truncating of the thread peaks, which in real world use disappear once the threads have been assembled a couple of times.
I have noted that PES do a 20tpi rotor nut with the note: 5/8 BS Hex 5/8" x 20 tpi late cycle.
This nut will not fit early cycle threads.
https://pesltd.uk/?product=29-2053
I wonder if that is UNF?
I'm not sure if I mentioned that I have three crankshafts - all variously distressed. The best one (from a 1971 engine) has the worst threads....as previously pictured. This week I bought two rotor nut NOS from Vale Onslow. On the crank with the best threads (which I already have a nut for that goes on and off perfectly) one of the NOS nuts goes on perfectly. And the other locks after 1/4 of a turn... My suspicion is that V.O. probably muddled different thread form rotor nuts in the same parts bin.... I'm trying to think of a way of working out what thread form is in the naughty nut. But of course it could just be a badly made nut?
Have you tried a thread roller?
Is this a thread roller? I have one of these but due to the shoulder there isn't room to use it.
That does the same job. It "may" be possible to do some careful reshaping with a hammer and edge pushing the thread over a bit at a time. I've had to do this on an axle.
Re 20 tpi CEI and 20 tpi UNF, yes they are very similar and I'm certain some UNF nuts are sold as CEI. I can't remember but one is a slightly tighter fit which may actually be a help in your situation.
I note from that link: "This nut will not fit early cycle threads.". Cycle threads whether early or late should all conform to the relevant standard (BS 811:1950). So I don't understand the statement. The whole point of standards is that they describe - well - standards.
Agreed, doesn't make sense. I wonder if they have mixed up the CEI pitches? 7/16", 1/2", 9/16" and 5/8" CEI come in both 26 and 20 tpi which has confused me in the past. BSA used the 1/2" x 20 CEI for the A10 camshaft but 1/2" x 26 everywhere else.
https://www.triple-mregister.org/uploads/retro/thread_data.pdf