Making a lathe test bar

[tjb]

If I turn a long piece of stock (aluminum or mild steel) using the three-jaw chuck and the tail stock, I get a slight but consistent taper, so I'm pretty confident my lathe needs to be aligned. I have a 36" by 1 1/4" piece of tool steel that I want to use to make a test bar, but before butchering anything, I have a few questions:

1. I'm assuming I should align between the three-jaw chuck and the tail stock. Right/wrong? (The chuck is a Bison that I bought new. It's in very good condition.)
2. My lathe is an OTMT 13 x 40 (looks almost identical to the PM1236 but larger capacity). How long should I make the test bar?
3. What's the best way to cut it?
a. Lathe? I have an Aloris BXA tool post with a Kennemetal 'EVSCTR120326C' cutoff toolholder that can easily handle the diameter but I am uncertain about the hardness.
b. Band saw? I have a horizontal band saw. Same issue regarding hardness.
c. A better way?
3. Is it necessary to cut it at all? 36" is pretty long. Should I be nervous about trying to face and center drill with that much length hanging out the back - even if supported? It'd be nice to keep it intact, but I don't mind cutting it if necessary.
4. Any other tips I should know before wrecking a nice piece of tool steel?

Thanks for any advice.

Regards,
Terry

 

[graham-xrf]

Hi tjb

First let me say that I have not done this yet for myself, because I am still in the middle of restoring my South Bend, but I know it is something I will ultimately have to do, so I have been researching. I believe that using a test bar has to be done between centres (I think).

Certainly a test bar I bought for headstock alignment came with a MT3 taper to fit up the spindle without chuck, and a MT2 on the other end to fit up the tailstock. This, of course, is a different kind of measurement, done with a dial gauge onto the bar, instead of micrometer measurements of the diameter at each end of the test bar.

I have seen a test bar belonging to a friend who is into model engines. He made his from a piece of 40mm (about 1.57 inch) aluminum pipe which had "end bungs" welded into them, and faced off. He tried a few "centres" by hand using only a punch mark until he got least run-out, then committed to proper centres. He said the tedious bit was to cut a relief out of the middle of the pipe length, to leave the slightly raised dumbell end bits for the proper measure cut. It looked about 400mm (about 15 or 16 inches) long.

If you do want to use the chuck, maybe true it up first. Check out YouTube for what to do to grind the chuck jaws. One guy used a Dremmel clamped down onto the compound. Tubalcain also has a video.

 

[tjb]

Hi tjb

First let me say that I have not done this yet for myself, because I am still in the middle of restoring my South Bend, but I know it is something I will ultimately have to do, so I have been researching. I believe that using a test bar has to be done between centres (I think).

Certainly a test bar I bought for headstock alignment came with a MT3 taper to fit up the spindle without chuck, and a MT2 on the other end to fit up the tailstock. This, of course, is a different kind of measurement, done with a dial gauge onto the bar, instead of micrometer measurements of the diameter at each end of the test bar.

I have seen a test bar belonging to a friend who is into model engines. He made his from a piece of 40mm (about 1.57 inch) aluminum pipe which had "end bungs" welded into them, and faced off. He tried a few "centres" by hand using only a punch mark until he got least run-out, then committed to proper centres. He said the tedious bit was to cut a relief out of the middle of the pipe length, to leave the slightly raised dumbell end bits for the proper measure cut. It looked about 400mm (about 15 or 16 inches) long.

If you do want to use the chuck, maybe true it up first. Check out YouTube for what to do to grind the chuck jaws. One guy used a Dremmel clamped down onto the compound. Tubalcain also has a video.

Thanks, Graham.

Sounds like I will certainly need to cut this piece of tool steel. Thanks for the information, I'll definitely want to study it before proceeding.

Regards,
Terry

 

[benmychree]

I would just find a piece of soft steel, perhaps about 1- 1/4" diameter and 9" long, turn one end down to a bit over 1" diameter and 2- 1/2 long to chuck onto, then relieve the remaining length down to about the same diameter, leaving a collar about 3/8" wide at each end; chuck it up and take very light cuts on each collar, and adjust leveling until the cut the same size.
Trying to make a fancy test bar run true enough to be useful for alignment purposes is an exercise in futility.

 

[graham-xrf]

I would just find a piece of soft steel, perhaps about 1- 1/4" diameter and 9" long, turn one end down to a bit over 1" diameter and 2- 1/2 long to chuck onto, then relieve the remaining length down to about the same diameter, leaving a collar about 3/8" wide at each end; chuck it up and take very light cuts on each collar, and adjust leveling until the cut the same size.
Trying to make a fancy test bar run true enough to be useful for alignment purposes is an exercise in futility.

John is right, and you can even do it in aluminum. Tweaking the tailstock to get it to run true to the point you get a near-the-same measurement on each end of the bar removes the "turning tapers" symptom. 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.

The levelling of the lathe is one cause, by introducing twist between the ways, and this does account for some of the cause of getting unequal diameters on the test bar, after the light cut. One tip I have seen is that one should not trust the setting that advances the tool. You take the light cut on one end, then move keep moving carriage to the other end, and continue the cut. Get the levelling right first, to remove twist stresses. This game can somewhat compensate for wear anyway. You do not have to knock yourself out getting absolute about it when a homebrew test bar can let you set up and get going, working around the effects of wear and age.

 

[tjb]

I would just find a piece of soft steel, perhaps about 1- 1/4" diameter and 9" long, turn one end down to a bit over 1" diameter and 2- 1/2 long to chuck onto, then relieve the remaining length down to about the same diameter, leaving a collar about 3/8" wide at each end; chuck it up and take very light cuts on each collar, and adjust leveling until the cut the same size.
Trying to make a fancy test bar run true enough to be useful for alignment purposes is an exercise in futility.

Questions:
1. The way I'm reading your approach is about 2-1/2" is turned to @1" and chucked, so the other 6-1/2" sticks out. Should it be faced and center drilled for tail stock?

2. On the 6-1/2" length, leave a collar @ 3/8" wide on the 'chuck' end, another @3/8" wide on the 'tail stock' end, and turn the center down to something approximating the 2-1/2" section in the chuck. Right/wrong?

3. Adjust 'leveling' or 'alignment'? My understanding is that they are two different things. I suppose it could be a leveling issue, but I've been assuming it's an alignment issue. Is there a way to identify which it is before proceeding?

Thanks for responding.

Regards,
Terry

 

[Cadillac]

I wouldn’t waste your length on a test bar. 16” bar would be fine 1 1/4” something cheap. Chuck close to chuck and cut centers in both ends. Make or get yourself some aluminum donuts 1.250 I.d. Maybe 2” o.d. About .750 wide. You could set screw or press these onto your bar about 2-3” from the ends. This will be the test bar, chuck one end and use a good center in the tailstock. Take a light cut on both donuts with a sharp tool without moving the cross slide. measure and adjust tailstock half the amount.

 

[Cadillac]

Even though leveled one would still have to perform this test to align the tailstock to the chuck. Perfect level technically means nothing in this case cutting cylindrical parts is the goal

 

[tjb]

I wouldn’t waste your length on a test bar. 16” bar would be fine 1 1/4” something cheap. Chuck close to chuck and cut centers in both ends. Make or get yourself some aluminum donuts 1.250 I.d. Maybe 2” o.d. About .750 wide. You could set screw or press these onto your bar about 2-3” from the ends. This will be the test bar, chuck one end and use a good center in the tailstock. Take a light cut on both donuts with a sharp tool without moving the cross slide. measure and adjust tailstock half the amount.

Okay, got it. That makes sense. Thanks for responding with this and your second post.

Regards,
Terry

 

[graham-xrf]

There are two ways to use test bars. One type of test bar is a precision ground bar, with ready-made centers, and perhaps also, precision ground Morse taper ends. You only ever measure up to it, relatve to parts of your lathe, and you take care of it. The other is a test bar you repeatedly cut on the lathe itself, and you use a micrometer on it.

The alignments they make possible are different. the reference test bar is used for very fundamental stuff, like aligning the headstock to the ways. Other kind of test bar, the kind you make and repeatedly use for test cuts is the other sort, used for tailstock alignment, aligning centers, and the like. It reveals the "taper turning" tendency.

You have a cylindrical bit at each end, wide enough to comfortably use a micrometer on, and a void bit between that is cut deeper, so that you can move the tool between the end cylinder parts that you will measure. If the tail end ends up a smaller diameter, then it may be the tailstock is misaligned closer to the tool side than it should be. It can also mean that the headstock is pointing slightly towards the tool side.

If you start by putting a center in the chuck, and also the tailstock, and push them together. Hold a steel rule or metal strip between the points, and gently press the points together. The metal will madly tip if there is the slightest mis-alignment, and the direction will be obvious. If things need a shim, this is where you see it.

Even though you get the centers together at the chuck end, it does not mean they will still be so when the tailstock is moved back. If the ways are not parallel to the headstock spindle, it may not matter how much you adjust the tailstock. It will drive you mad!

Assuming such fundamental stuff is not what ails it, and all it needs is the taper effect taken out, you can go ahead and align the tailstock.

There is a variation to this. One measure is to mount up an indicator to measure up to the cylinders. The other measure is to use a micrometer on the diameter of the trial cut cylinders. This trick takes time, but is superbly accurate.

Tubalcain, in detail, explains how, and he shows both measures being done.

There are LOTS of videos on this kind of alignment, as well as using a precision bar, which is agreed to be the faster method, though it does involve the purchase of a test bar you do not cut on. I found a 16mm solid carbide rod about 300mm (about a foot) long for £13 (about $17) that would be good for this - if I could figure out how to get some centers drilled into it!

"Rollies Dad's Method". I don't know who Rollie's Dad was, but after you read it, you can quickly start inventing your own test variations. I think, on this forum, there has to quite extensive discussions already, maybe in older threads.