View Full Version : Improving my lathe's compound's screw.

Tony Ennis
02-19-2011, 04:27 PM

This is a lovely drawing of my Atlas/Craftsman 12" lathe's compound screw and parts.

Now, I intend to replace the 1" dial with a 2" dial so I can see it easier. While I was taking it apart to see how it worked and to remove some backlash, I decided perhaps it could be improved with a bearing.

I'll describe the drawing. You see the compound's screw at the top. One end is acme threaded, the other has normal threads.

The bushing slides over the end of the shaft and is bolted to the compound casting. Note where I've put the black arrow - this shoulder butts up against the bushing. This is what prevents the screw from backing out of the compound. The 3/8" hole in the bushing is a good fit on the compound screw. This bushing prevents the screw from flopping around.

Next, the dial goes on. Note the set screw. I suppose this is how I'm supposed to 'zero' the dial.

Now the 'thin nut' goes on. It's so thin none of my wrenches can tighten it once the handle's on. I need to make a wrench from bar stock I suppose. I believe backlash is removed by tightening this nut up against the shoulder in the screw.

Next the handle and its woodruff key go on.

Finally, the normal nut goes on. I believe tightening this down actually tightens the handle against the thin nut, and thus locks the thin nut.

Now, I'm already going to replace the 1" dial with a 2" dial. While I'm hacking away, is there any reason not to add some real bearings?

1. The shoulder just begs for a bearing of some sort. It rubs against the bushing, and I have to remake the bushing anyway; the dial's 'zero mark' is cast into it. And it will be too small for the larger dial. I could drill the hole in the new bushing a little bigger and install this bronze sleeve bearing (http://www.mcmaster.com/#bronze-sleeve-bearings/=b3rv2j). The bearing's flange would rub against the screw's shoulder.

2. The dial rubs directly against the bushing. Since I am making a new dial, and it's 1/2" thick, is there any reason not to sneak a ball thrust bearing (http://www.mcmaster.com/#thrust-bearings/=b3rsha) in there?

3. I really dislike the setscrew that's on the dial. I'd prefer something without a set screw. My friend's 9x20's dials move easily enough but stay put due to friction. I believe the dial has an inner ring and an outer ring, and an o-ring between them. Anyone know a different snazzy way to accomplish this?

Any other ideas?

02-19-2011, 04:53 PM
The normal assembly sequence is thick nut, handle and thin nut outboard of the handle. I see no reason not to add bearings if space permits and that is your desire.

A small thumbscrew on the indexing collar will make tightening simpler. I prefer positive locking of the collars to friction as there is less of a chance to accidently move them when feeding.

A plug for Tall Grass Tools offering of kits for larger feed dials among other neat things is in order.


Tony Ennis
02-19-2011, 05:29 PM
I tried that assembly order. The thick nut encroaches upon the woodruff key slot making it impossible to assemble. Perhaps I should go try again.

02-19-2011, 06:22 PM
Thin the nut or shorten the key. It has been a while since I have had an Atlas apart, but things can be adjusted fit the need without making any major changes to the operation of the machine. You can machine a bit off the shoulder of the leadscrew to provide more room for bearings and/or thicker graduated collar.

Tony Ennis
02-19-2011, 06:30 PM
Normally in these situations I'm just being stupid. I bet if I reassemble it carefully, it will work. Or I'll find it's not the original nut. 2mm off the key and it will fit.

I had some major drama with my thread dial the other day. I had never mounted it to the lathe before. It simply didn't fit right, at all. Only later did I realize I had mounted it backwards (flipped right-to-left.) "Hey, why are these numbers upside down... DOH." lol.

02-19-2011, 07:43 PM
IIRC the compound on my Atlas has had the key filed off on one end so a thicker nut will fit.

I suppose you can add all the bearings you want but the application is a slow speed and there will be some nominal drag from the ways if they're adjusted to be snug so I"m not convinced reducing friction there will make a noticeable improvement in operations. If doing anything I'd wonder about a teflon washer with a couple .002 shimstock washers adjacent so the nut adjustment could be tightened up to reduce backlash without introducing much additional friction. OTOH, backlash is a fact of life with mechanical parts and developing a mindset and habitual mode to account for it in precision moves will always stand you in good stead.

Thanks for the plug, Jim. ;)

02-19-2011, 09:14 PM
I think I'd avoid roller element bearings as unless the assembly is designed for them there is a good chance that crap will get into them and make them feel lumpy. TGTool noted that it's a low speed operation and bushes do fine in this role.
Having said that, I do have ball thrust bearing on my lathe but they are buried deep in housings and shielded from garbage by large washers on the open side of the bearing.
You might want to do something like put a lubrication point on the new bush so that you are running on an oil film rather than something dried or gummy.
From experience with a previous lathe, if you are wanting to improve the feel of the lathe I would suggest having a look at the gibs and dove tails. It's amazing how much better a slide will feel if it is correctly adjusted with scraped in (that is flat) surfaces.


02-20-2011, 11:36 AM
I just reassembled the compound on my Logan 820. I had similar problems and I made a pair of teflon washers from a bit of scrap that I had. I made the washers about 40 thou and cut a 30 thou recess in the face of dial and the face of the bushing where it bears against the retainer. seems to work just fine, so I did the same thing on the cross feed dial. Larger dials would have been nice, but they required other work.

02-20-2011, 11:53 AM
I replaced the screws with new screws from MCmaster. Then I went back and installed ball bearings for the screws.

The new screw: http://www.machinistweb.com/forum/showthread.php?t=1015&page=3

The bearings: http://www.machinistweb.com/forum/showthread.php?t=1073&highlight=cross+slide+bearing

J Tiers
02-20-2011, 12:15 PM
An opinion from way out in left field.......

I'd not do ANYthing to the compound. If it is tight, so be it. I don't want it moving around anyway, unless I am turning the screw. It tends to be a set-and-forget item.

If I AM using it, once in a while for tapers, etc, I can put up with some minor issues. I find no advantage to focusing on that particular screw, unless everything else is perfect, which since you are discussing Atlas, it is NOT. (same with some other low-end brands)

If I am going to do anything to "upgrade" the Logan, it will be to the crosslide, which gets used all the time. And, if I put a thrust bearing in, I will also put in a LOCKING means, because I don't want it turning from vibration, etc.

In fact, I already upgraded it with an added bushing arrangement to have a settable dial, and to minimize screw-bearing-based backlash.

Mike Burdick
02-20-2011, 02:10 PM
3. I really dislike the setscrew that's on the dial. I'd prefer something without a set screw. My friend's 9x20's dials move easily enough but stay put due to friction. I believe the dial has an inner ring and an outer ring, and an o-ring between them. Anyone know a different snazzy way to accomplish this?...


One can make a "friction" type dial that uses three small internal ball bearings on springs. These work great! A general plan is shown below:


Also one can make graduated dials like shown without using special tools...


02-20-2011, 03:29 PM
An alternative way to give the dial friction is firstly to replace the inner nut by a shouldered sleeve, the sleeve end going on first, and the shoulder acting as the old inner nut. The dial gets bored out - half of its thickness is bored out so there is an annular gap between the sleeve and the inner surface of the dial, and the other half of its thickness is bored out so it is a working fit on the sleeve.

Into the annular gap between sleeve and dial goes a bit of spring steel bent into a cross between a 'U' shape and a 'V' shape. The ends and the tightest radius lock the spring against the inner surface of the dial, and the sleeve is squeezed between the 'arms' of the spring. Adjust with pliers and filing till the friction is just right, and the sleeve is a difficult hand fit in the dial and spring.

The spring needs to be 1/4" or more wide, so if the dial is 1/2" wide you've got enough meat there to do it.

I wouldn't do anything else. unless the main bush is worn so badly you notice the slop in normal use. Replacing the inner nut with the sleeve, with the extra length of thread, will reduce the tendency for the inner nut to move on the threads when you lock up the handle, and make fine adjustment of the backlash that much easier and more positive.

Tony Ennis
02-21-2011, 11:50 AM
@JTiers - good point about rebuilding the compound before the cross-slide is putting the cart before the horse. I intend to make a new dial for each. I'm pretty happy with the cross-slide otherwise, though I do wish it locked. I'm going to be embarrassed if there's a big honkin' bolt on the side for just such a purpose :D

@rohart - You're saying there doesn't need to be an inner nut? I wonder how Atlas missed that cost-saving idea! :p With your idea, does the 'v' of the spring press against the dial or against the sleeve, or does it matter?