Re: [multimachine] Re: Romig style lathes
- Right but if you were turning very short brass say near the h/s, what would you say the repeatable tolerance is?
On Sat, Aug 31, 2013 5:58 PM PDT Dennis Shelgren wrote:
>I should clarify, the entire top of the bed is worn nearest the headstock.
>The front and rear edges of the bed are still o.k. In a way the front to
>back accuracy is o.k. but the toolpost effectively "sinks" as it hits the
>worn area. The tool, therefore, is cutting lower on the work piece and also
>farther away. It isn't something I measured other than to lay a test piece
>on a straight edge and and I can see light on both sides of a short piece.
>It is so slight that I still used the lathe until the headstock itself
>If I wanted to use this machine for high accuracy work I would just buy a
- As long as I measure as a I work and don't expect perfection, I imagine my time fixing it will be well spent.I have a very old(1890) Sheppard lathe and it's bed is probably wavy, twisted, and worn. I'm still going to clean it up and put a motor on it, and it'll be used for larger parts that don't go on it's smaller cousins.Dave is right in that it's more a matter of the man and not the machine.As long as you know where the "bad" spots are you can compensate to a degree.The repeatability is very high, all the parts will have the same error. The accuracy is off by an unacceptable amount on small parts.This is probably why it's viable to build a lathe bed with components of less than ideal accuracy.
I only detected the problem when a smaller shaft I cut fit it's hole oddly, and I realized(and mic'd) it had to be barrel shaped. I finished the part with a file on the lathe to correct it and moved on with my life. For thin parts or tubes I could cheat with a follower rest which bends the part in line with the worn bed.
Even on a much worse worn lathe the bed wear/error is divided by the geometry of the part and tool setup. In other words on larger parts the height change of the tool is less with relationship to the part.