Making a lathe test bar

[graham-xrf]

You'd be amazed at what gravity will do with just a little bit of help.

Though I was taught about stress vs strain diagrams, yield point, plastic deformation, etc. roughly at a time when when gasoline was $0.40/gallon, it seems the same stuff is still the prevailing wisdom.

Pure cast iron has no plastic deformation, and come to that, not much stretch either. It is brittle, and weak in tension, and just breaks. Lathe beds like mine (South Bend circa 1942) are made of "semi-steel", a loose term which can cover many grades, but apparently heavy duty, durable, and lower cost.

If it is stressed from being "un-level", so OK, it will deform. Take the stress off, and it should return back to where it started. Provided the yield point was not reached, that should happen right away. If it did yield, but not break, then it will return to a new place, permanently deformed. I did not know there was a "get there slowly" time delay thing involved.

 

[graham-xrf]

Please look at post #13 here:

Test bar

Hello. I recently-ish ordered a mt2 test bar, along with some other stuff. Unfortunately the test bar went out of stock before my order was packed. The company offers another test bar that is a combination mt2/mt3, and I told them they could send me that one instead. I had considered that one...

www.hobby-machinist.com

Bob Korves gives an excellent approach.
Robert

Not only the (very detailed) explanation from Bob Korves, but the whole thread even. It is all excellent!

They make the point point that the lathe must first be aligned to not turn a taper even without the tailstock.
This could be done either with a ready-made test bar, or any old bar (Rollies Dad), or by the two-collar cut.
Otherwise, you have to make allowances if you then use a tailstock. If the bed has a twist, the tailstock alignment would only be good at the tailstock position where the alignment was done. Even then, a perfect 2-collar test bar at that position becomes not-so-good if the tailstock is in any way turned so the quill is not extending parallel to the ways.

It seems to me that step #1 is to check the headstock end with a known test bar, or use any old bar, and get up the Rollies Dad Method. Likely it will be OK, but you at least have to look.

Step #2 is to make sure the tailstock is at least at the right place. The center's points should touch exactly. If the tailstock is already aligned, the centers should stay on the line as the tailstock is moved. You use the two-collar-cut test bar to reveal what is happening. Maybe a good thing would be to put a ground test bar on the tailstock, and use a dial indicator to become sure of what happens that end first, even checking extending the quill. Then finally, a two-collar cut should confirm things. A contrary measurement here would be a disappointment, but also a huge diagnostic!

 

[nnam]

There are lots of good info.
If I may add, is this a question of making a test bar, or how to test the set up with a test bar?

With a good test bar, it can tell if a lathe has issue. Making one to check a lathe is more about check anything that is wrong as pointed out already, meaning it is hard to know from the cut to know which caused the issue.

I saw Adom on YouTube (abom79), created a bar so he can align the tailstock. He was making a same diameter with the tail, then trace with an indicator between the two to see any offset.

I wonder why he didn't just put the indicator on the chuck and rotate to indicate around the tail.

Also, 1.25" sags greatly. So a good tailstock center position needs to support it, or using a lathe rest.

Once that is done, the carriage can be dragged left/right on the rod to check the bed for twist/wear.

A machinist level would be used to check for twist.

Head and tail can also need check for alignment.

 

[graham-xrf]

There are lots of good info.
If I may add, is this a question of making a test bar, or how to test the set up with a test bar?

Hi nnam
Definitely it is about both kinds of test bar, and making clear the differences between them.
We throw in Rollie's Dad's Method, which uses any old bar as a circular cam, and uses the averages of the highs and lows as it wobbles it's end. Spectacularly accurate outcome, just by being clever - but time consuming.

With the right amount of thought, and making test cuts on just about anything, one can figure out what to adjust to make it straight. These methods are a more focused set of steps to check, confirm, and diagnose.

At some points, they do overlap. You could argue that with full application of (say) a ground reference test bar, you don't need to ever make and try a two-collar test bar. Conversely, making a test bar and measuring it, is what you do to check you are not turning a taper. This is instead of discovering the taper is on your favorite project! From what you find, you can infer things about the lathe alignment, and figure out what to adjust. All these methods compliment each other.

 

[lagweezle]

The guys who take this to an art form will tell you that there needs to be a slight offset to account for tool pressure, and that some lathes were made that way so that over their working life the "taper" would wear off, and then swing the "other" way. I don't know how true this notion is, but is was certainly believed by enough folk.

To be clear, I'm incredibly inexperienced. With that out of the way, I thought the entire point of doing one or more "spring pass" cuts was to mitigate the effect of tool pressure deflecting the work surface. Also, given that the cut, speed, and materials in use change how much deflection there is of the work surface, I don't see how an intentional offset of the tailstock makes sense. I'm inclined to consider this one of the many, many bits of knowledge that are passed around as true but ... really aren't. There is likely some truth to it, but the original truth, etc. has been lost to time and only if we're lucky can we find it again.

 

[Dan_S]

To be clear, I'm incredibly inexperienced. With that out of the way, I thought the entire point of doing one or more "spring pass" cuts was to mitigate the effect of tool pressure deflecting the work surface. Also, given that the cut, speed, and materials in use change how much deflection there is of the work surface, I don't see how an intentional offset of the tailstock makes sense. I'm inclined to consider this one of the many, many bits of knowledge that are passed around as true but ... really aren't. There is likely some truth to it, but the original truth, etc. has been lost to time and only if we're lucky can we find it again.

He isn't talking about the tailstock. back in the day when manual machines were still used in a production environment, the head stock would have an ever so slight angular offset for Chucked work. Time is money, so you don't want to make extra finishing passes when you're making hundreds or thousands of parts.