Chuck location on a Vertex Dividing head

[Brian WoodParticipant @brianwood45127]

A puzzle which seems illogical

[Brian WoodParticipant @brianwood45127]

I have been cutting some gears recently. The blanks were mounted on a purpose made mandrel to which they were keyed to maintain register. My dividing head is a Vertex BS-O

I observed that those cut so far have teeth that are not all to the same depth, varying cyclically around the gear and on checking everything through I found that the mandel has about 10-15 thou run out at the gear blank end. I am using a tailstock too, but these measuiements were taken with the end free of the tailstock.

The chuck and mandrel ran perfectly true on the lathe. All the mounting threads are clean, register faces likewise. I checked the mandrel out with 3 other good quality self centreing chucks, the same some of errors were repeated, despite seeing each combination running true on the lathe before transfer to the dividing head. A good quality test bar gave the same errors.

A lathe catch plate fitted to the nose of the Vertex head gave a 1/2 thou runout on the face at the outer edge, not enough in my view to translate to the sort of errors I have measured at 5 inches from the chuck.

The BS-O register is ground and correct on diameter, there was no detectable movement on that or the ground vertical face of the register with a sensitive DTI.

It is though only a short register, not as long as that on the lathe.Just for good measure I checked the truth of the ground Brown and Sharp centre supplied with the head, there was no detectable movement on the DTI

My question is; has anyone else observed this rather bizarre behaviour and if so what did they think it was caused by? My current thinking concerns the fit of the BS-O thread in the chuck backplates before they are butted up to the register, they are a bit snug which might hold things off a trifle.

I am seeing the job through with a 4J chuck holding the gear blank mandrel, it is though a bit annoying that I can't use any of my 3J chucks instead

Brian Wood

[Stephen BensonParticipant @stephenbenson75261]

Yes I had that problem and I made an adjustable backplate and clocked the gear/wheel bland on the mill a bit over engineered but I do make a lot of clock wheels got some pictures on my web site

**LINK**

[_Paul_Participant @_paul_]

I use the MT2 taper inside the Heads spindle to mount mandrel/s when gearcutting never a chuck.

Paul

[Brian WoodParticipant @brianwood45127]

Hello Stephen and Paul,

Thank you both for your input, I am more irritated by the apparent fault than anything else. Current use of the 4J chuck gets over the immediate problem. The taper in my example is Brown and Sharp which is not compatible with my MT tooling.

When the job is done and I have the time I'll strip out the mandrel and follow the nose thread on it in the lathe, I suspect it is slighty drunken but can't yet prove it. Depending on what I find I'll either machine it off and make a new one on a shrunk on sleeve, or make a new mandrel entirely with an MT taper so that I can use that as well.

I'll report again on this in the future.

Brian Wood

[jason udallParticipant @jasonudall57142]

“….5 inches from the chuck”
10 thou at 5 inch is a tiny angular error…

” “

[Brian WoodParticipant @brianwood45127]

Hello Jason and John,

Thank you both for your replies.

Jason The runout error was negligible with things mounted on the lathe, even at 5 inches from the chuck, but present on every chuck when fitted to the BS-O head; all the chucks are otherwise spot on. The error was sufficient to give variable cutting depths on the teeth of the gears I was making, visibly wrong.

John I ran a die over the BS-O thread and got a whisker of metal away, it didn't correct matters though. When I have finished this batch of gears I'll transplant the 4J chuck with the holding mandrel still gripped in it back to the lathe to observe the situation in reverse.

I don't like silly mysteries of this kind and like a terrier I'll plug away until I've found out and corrected whatever is wrong, it would remain as a niggle until then.

Brian.

[blowlampParticipant @blowlamp]

Try some Engineer's Blue on the register faces.

Martin.

[Brian WoodParticipant @brianwood45127]

Hello Martin,

A good idea, I hadn't thought of it; it would reveal one sided contact for example. Many thanks

Brian

[Author]

Posts