[RotarySMP]

On the exploded view, you start at the back with a seal. This can be a simple turned part out of delrin (I use Ikea chopping board plastic).

Next is the bearing block. You size this around you bearings, 7200 bearings are good.
2 Bearing 7200B 10mm Diameter 30mm Angular Contact Ball - eBay (item 360189223748 end time Nov-07-10 18:27:15 PST)
You want that bore to be a sliding fit within about 0.0002". Nice sharp HSS boring bar, with lots of rake, and one just hanging out long enough to reach the bottom of the bore.

This is how you learn precision.

The bearings are shown wrong, as they are both pointing the same way. The need to be inside face to face. A thin shim is used between them to preload them.

The cover of the bearing block should have a boss sticking out on the back the diameter of the bearing, so that it pushes the two bearing outer races hard together.

On set of holes holds the cover onto the bearing block, and the second send mount the block to carrier which mount it to the base casting.

[RotarySMP]

Since we have chosen you 7200 bearings, and they have a 10mm bore, 9mm wide, you know that the end machining on your ball screw needs to be

1/ A short section with two flats to get a wrench on. say 1/2" diameter, with 3/8" flats 1/4" wide 1/2" long.

2/ 10mm diameter 18mm long for the bearings.

3/ A UNC 3/8" thread 5/16" long to mount a spacer (washer) and locking nut.

4/ the bit you drive coupling goes on. If you are using Keling Nema 23 motors with a 1/4" shaft, then specify this as 1/4" diameter 3/8" long

[nateman_doo]

Reading all your info. The motor has a 3/8" shaft.

http://www.kelinginc.net/KL23H2100-50-4B.pdf

I only drew the bearing on the right because I was going to mirror the other side of what I did. Just was lazy drawing it in paint.

[RotarySMP]

It is very difficult to measure and manufacture the screws and bearing blocks to put the correct preload onto the bearings using the ball screw as a spacer. That is why you preload the AC bearings back to back, and capture the screw axially on one end. Then the other end is not critical. It could be left unsupported, or just use a single ball bearing floating axially. It's function is then only to dampen vibration and prevent whipping.

You can redraw the ball screw end for a 3/8" to 3/8" coupling. You will probably have to go up to 12mm ID bearings, as you can't thread a 3/8" nut onto a thread which starts half way down a 3/8" shaft.

[nateman_doo]

A Better more detailed picture

[RotarySMP]

How much backlash do you want?

[nateman_doo]

as little as possible. Thats why the upgrade to ballscrews correct? Is there an inherent design flaw with what I doodled? I don't see the problem. If I make the bearing housing for both sides like the drawing, it will clamp each end of the ballscrew.

I do see a slight slope on the back end of the base, but I have no way of accurately measuring it.


Edit: The base is 11", and the X axis table is 18 5/8"

[RotarySMP]

The goal is good. Zero backlash should be your target, as backlash is never a good thing in a machine tool.

Ballscrews can be made with zero backlash (preload) in the nut, but you need to find ground screws for that (read up on the 5bears site).

Since you are planning to use Chinese rolled screws, the nuts will have backlash. If you want to get rid of it, you will need two nuts arranged with preload (normally done with springs).

Normal ball bearings have free play, which causes backlash. If you have no way of accurately measuring the length of you base, how are you going to specify the length of the screw between the bearing journals to the accuracy needed to preload those bearings? How are you going to machine the ends of the base to be parallel planes? How will you ensure that the bearing bores on each end of your table are aligned? Different thermal expansion of the base and screw would mean the preload will constantly change.

The way you have drawn it, you could hold a screw so that it won't fall out, and it will turn. It will not hold the screw rigidly in the Y axis, which is needed if you want to prevent backlash.

The simplest way is to use AC bearings back to back like NSK, THK, etc offer. At first this looks complicated, but it is easier than your method because:

1/ you only need to precisely machine one bore
2/ Both bearings are installed in the same bore, and will therefore automatically be precisely aligned.
3/ The bearing cap locks the bearings outer races axially very rigidly.
4/ A simple nut locks the bearings inner races onto a thin shim, thus locating the screw rigidly axially, removing the preload to remove backlash.
5/ Since your Y screw is short, you could leave the other end free. Then you don't even need to specify ball screw length. You could just order it a couple of inches long, and them chop it off to length with an angle grinder.

On X you will need a second bearing block, but it does not need to be axially aligned. You just making the bearing inner race a sliding fit on that journal.

[RBIEZE]

Nate, I think you should avoid the X-Y ballscrew conversion for now.

go with the Z axis conversion, but just use the Acme's on X-Y for now..

Once you have the motors-drives-breakout board-and pc all setup and working, you can then attempt the ballscrew conversion with a functional (albeit sloppy)Cnc mill..

This will allow you to have a working machine to make parts on.

The backlash in the Acme's can be minimized by gently overtightening the nuts for a short duration, or making some mogul nuts for use while making the ballscrew mounts etc...

You may find that .002 -.005 backlash is acceptable for the parts your making anyway...
Especially if you compensate for it when coding your programs.

I havent installed the ballscrews I purchased on my G3102 yet cause Im ok with the slight backlash I have , but... yes one day I will get the screws installed...

My 2 cents.

[nateman_doo]

I have been reading the 5 bears sight all morning, from start to finish. I think I can understand about 60% of what he is saying, and 90% of the pictures I am looking at. I may not know the lingo, but I understand what I see.

He actually used my table in part of his project, but he got rid of the dovetails and replaced with rails. Really looks sweet & efficient. Far better then dovetails. After looking at that I am half tempted to build a table from scratch instead of getting this table to work, but that will cost a bit more.

The ballscrew end machining is starting to click, but it still a bit fuzzy.

[nateman_doo]

Nate, I think you should avoid the X-Y ballscrew conversion for now.

go with the Z axis conversion, but just use the Acme's on X-Y for now..

Once you have the motors-drives-breakout board-and pc all setup and working, you can then attempt the ballscrew conversion with a functional (albeit sloppy)Cnc mill..

This will allow you to have a working machine to make parts on.

Is this what you are referring to:
http://www.cnczone.com/forums/bencht...preciated.html

[RBIEZE]

I was thinking that starting out with the basic conversion using the existing Acme screws is a good first step for you.



Then later when you feel more confident and have a gameplan, you could do the ballscrew conversion separately.

If you already have the motor mounts etc done, the only thing you'd need besides the ballscrews and bearing mounts would be a new set of couplers.

[nateman_doo]

Thanks, not a bad idea but I am not ready to throw in the towel today. I had a few visions lately & have a rotary table on the way. I am going to work on that, but in the meantime anyone have any ideas for a nice 12x18 tooling plate? Where is a good place to buy such a thing.

[RotarySMP]

Hi Nate,

I have the same mini lathe as you, so I if I can make the precise bits for my tools, you can also.

I remember when I started, some of the "best" ways of doing things seems too complicated or beyond my ablity. Sometimes I did things the "easy" way. As you grow better at machining, these things all seem to get done over.

[nateman_doo]

just bought the lathe, and I have never used one in my life. I feel more comfortable making parts on the mill. I don't even know what type of bits to buy for it, raking, slope, bore-all of it is another language at the moment. All I have is the lathe out of the box and I bought a drill chuck which like you read in the other thread is too long.

I agree I would rather do things right the first time, but the right way comes with time and experience. I have neither.

Im just sitting here lapping my X2 head spacers at the moment.

[RotarySMP]

Hi Nate,

For me it is the opposite. I had a little lathe experience as my Dad had a Boxford when I was young, and did some turning in shop class, and again in my apprenticeship. I had no milling experience till I bought an old Deckel G1L engraver frame from the scrap dealer for €20, and put a sherline spindle on it.

wrathall.com

The right way comes either through getting sick of doing it wrong and having to redo it, or though asking for help, and taking it.

There is nothing complicated in what you are planning, just break it down into small acheivable parts.

[nateman_doo]

Very sage advise. Break it down into manageable parts. I planed to tackle this in several phases.

Phase 1: Purchase ballscrews & bearings based on measurements from others who have done this already to this same table.

Phase 2: Once those are on hand, make & install the end blocks that secure to the Y axis.

Phase 3: X Axis. I will cross that bridge when I get to it.

On the bright side, I found a sight that has an incredible amount of info on the lathe: http://www.mini-lathe.com/Default.htm

[Hans_G]

The MIT machine shop video series is another good "getting started" resource.

MIT TechTV – Machine Shop 1

[nateman_doo]

Very helpful videos! watched a few of them and learned a few things. Still have more to watch.

I got the green light for the ball screws. No idea how long they will take, but the order was placed.

On a side note, I just bought a 6" rotary table. How and what do I get to the chuck of the 7X lathe in it? Any adapter plates?

[nateman_doo]

The rotary table.


Guess I need a center/edge finder too.