Test bar

[Shootymacshootface]

I checked my tailstock alignment on my mine by using 2 mt3 dead centers and making sure the tips meet. Is this not sufficient?

 

[rwm]

I checked my tailstock alignment on my mine by using 2 mt3 dead centers and making sure the tips meet. Is this not sufficient?

It is pretty hard to be sure the ends meet to within .001"

Bob K - Why would it be hard to drill new centers if you could fit the bar into the spindle. You could use a 4 jaw and then center drill? No? I guess if they are really hard that could be a practical issue.
Robert

 

[Bob Korves]

OK, here is the way to make an accurate test bar for testing between centers. An accurate tail stock is not needed. In fact, this is how you get your tail stock properly aligned:

Use a piece of relatively thin wall tubing. Something like 1/16" to 1/8" wall thickness. The O.D. should perhaps be somewhere between 3/4" and 1.5", length 8" to 16" (for a 12" swing lathe.) The dimensions are not critical at all, but the O.D. of the tubing must be about 1/16" or more smaller than the spindle bore. With the tube mounted deep into the spindle bore, so the end just sticks out far enough to machine the end, face the tube, lightly chamfer the O.D. and I.D., reverse the tube and repeat at the other end. Turn 2 inserts with a larger O.D. than the tube (but smaller than the spindle bore) to a light to medium press fit for the tube bore, chamfer the ends, and then press them into both ends of the tube. The shoulder needs to have a larger O.D. than the tube, with a small allowance for turning it concentric with the centers. Press the end plugs into the tube. Mount the tube deep into the spindle, deep enough for just enough to stick out to turn the O.D. of the plug. Turn the plug O.D. It does not need a super finish, but should be pretty good. Then use a center drill and do a really good job of placing the center. I also relieved the outside corner of the center a bit with a larger angle chamfer tool, and then touched up the center again, to help protect the center from any dings at the plug surface. Remove the tube from the chuck, reverse the part, and repeat for the other end. Install a piece of round stock in the chuck, and then turn a 60 degree center on it. Do a good job of it. Mount a 'known to be good' dead center in the tail stock, and get it as closely aligned as possible by sliding the tail stock to meet the spindle and tail stock points together. Adjust until they are in line side to side, ignore any vertical mismatch. It is not critical to success, but do your best work anyway. Clean and lubricate the test bar centers with oil and mount the test bar between centers, snug, but so it can still be rotated against the centers with little effort. Lock the tail stock spindle. Now, get a good, sharp!, turning tool and take a light but real cut on the plug O.D. at the tail stock end, doing a really nice job of it. Do not touch the cross slide or compound rest feed screws after making the cut. Remove the bar from the centers, reverse it, and repeat the cut at the other end of the test bar, still cutting at the tail stock end.

The bar is now finished, and will be highly accurate if done correctly. With mine, a .0001" indicator does not move when rotating and indicating the collars, and the collars have identical feel when testing for diameter with a locked micrometer at both ends. Reversing the test bar gives identical readings.

To use, turn in place a dead center mounted in the spindle, mount the test bar between dead centers, and mount an indicator, preferably .0001" resolution, to the carriage. Set it up to read the side of the test collar at one end of the bar, zero the reading, and then move the carriage with the hand wheel to indicate the collar at the other end of the test bar. The readings need to be the same for the lathe centers to be properly aligned side to side. Repeat the test with the indicator reading the top of both collars to test for equal center heights at both spindle and tail stock. The tail stock will likely need adjustment to get identical readings. The screws on both sides of the tail stock can be used to correct side to side errors, and shims and/or scraping can repair vertical errors. Consult your lathe manual, the desired settings are usually not preferred to be .0000".

 

[rwm]

Excellent instruction Bob! I will be making one of those. My old one is junk!
Bob is also correct about sag in that long test bar with heavy ends. After considering cutting it in half, I don't see a need to put centers on the cut ends. You would primarily use this to align the lathe spindle. You would not want the tailstock involved for that. If you tried to use it to align the tailstock you would get alignment to the bed ways but that would not guarantee concentricity to the spindle. Bob's technique is the standard method.
Robert

Edit: By this- "Mount the tube deep into the spindle, deep enough for just enough to stick out to turn the O.D. of the plug. " I assume you mean using a 4 jaw chuck and dialing it in to the highest accuracy? Or does it not need to be that accurate since the final cuts are done at the tailstock end?
R

 

[Bob Korves]

Edit: By this- "Mount the tube deep into the spindle, deep enough for just enough to stick out to turn the O.D. of the plug. " I assume you mean using a 4 jaw chuck and dialing it in to the highest accuracy? Or does it not need to be that accurate since the final cuts are done at the tailstock end?
R

Accuracy of the tube in a runout sense is of no consequence. All that is important in the end is that the drilled centers and the cut collars are cut concentrically and with minimum taper for easier measuring. The tube is only a structural element, there are no important dimensions for its purpose. The sequence of operations uses the length of the test bar to reduce any possibility of the test collar faces being tapered. The long bar reduces that small error by a large magnitude. Still, I dialed them in with the 4 jaw as good as I could. I do not know if using a 3 jaw with a few thou runout would make any noticeable difference in the end product. Try it and report back!

 

[Splat]

I've done Lyle's (Tubal Cain) method using two drill chuck arbors to get my TS aligned. Highly recommended. Unlike being restricted to test bar length, you can position your TS anywhere along the bed and measure alignment.

 

[rwm]

Accuracy of the tube in a runout sense is of no consequence. All that is important in the end is that the drilled centers and the cut collars are cut concentrically and with minimum taper for easier measuring. The tube is only a structural element, there are no important dimensions for its purpose. The sequence of operations uses the length of the test bar to reduce any possibility of the test collar faces being tapered. The long bar reduces that small error by a large magnitude. Still, I dialed them in with the 4 jaw as good as I could. I do not know if using a 3 jaw with a few thou runout would make any noticeable difference in the end product. Try it and report back!

Yes Sir I will!
R

 

[rwm]

I've done Lyle's (Tubal Cain) method using two drill chuck arbors to get my TS aligned. Highly recommended. Unlike being restricted to test bar length, you can position your TS anywhere along the bed and measure alignment.

Tubal Cains method looks quite simple but I am totally confused about how he did this? He first machines the MT3 part in the spindle. Then he removes it and places the MT2 arbor in the spindle? How is this possible? Did I miss something?
Robert

 

[darkzero]

Tubal Cains method looks quite simple but I am totally confused about how he did this? He first machines the MT3 part in the spindle. Then he removes it and places the MT2 arbor in the spindle? How is this possible? Did I miss something?
Robert

He machined the MT2 first using a sleeve & machined the MT3 second. You can see the sleeve better when he removes it at 5:32.

 

[rwm]

He machined the MT2 first using a sleeve & machined the MT3 second. You can see the sleeve better when he removes it at 5:32.

Thank you. I missed that. That would introduce the small error of the sleeve I suppose. Still seems like a good method.
Robert