An X Y Z R Table Part 1

An ongoing project

(still in process, so it's not quite done yet)

Actually, as of October in 2004, it's pretty much done except for some adjustments here and there and adding limit and home switches.

 

I'm going to try something different, here.  I really don't know how this will turn out, since I'm designing it a bit as I go along.  However, if you check back every week, I should have some more stuff added.  Normally, I just put up a project as done.  This time, I'm going to let you look over my shoulder as I make mistak... design changes.

 

There's several ways to do a project like this.  The mechanical design has various options, and the electronics can be done with either servo motors or stepper motors.  I have lots of stepper motors, and they're more available than servo motors, so steppers are a way to go for me.

 
One option is to have a baseplate, and two rails.  A carriage runs on those rails, which runs back and forth.  On top of that, there's another set of rails crosswise, and a second carriage runs on that.  Now for Z, you put a tool in the middle of the travel range, and control the height of that with the z axis controls.  For R, you put a rotary table on the second table.  That gives you good radius stuff and the other degrees of freedom.  You also might be able to put  the tool on a rotary joint so it can come in from the side.  These are all possibilities.  The method of drive here is to use a screw supported on both ends with a permanent half nut on the driven table.

 
Another option is to have the same two rails on the bottom, then a very narrow set of rails riding on the first set.  The tool is put on a very small platform and moved over the work.  This is the way a normal XY plotter works.  Now there's nothing wrong with that, but if you want to do certain types of work, where the tool comes in from the side, you've got a problem.  The other difficulty is that the Z axis travel is severely limited in this design.  Drive is either by screw drive, or by captive cables.  The cables need to be thin woven steel, so that's another interesting point, find the cables....

 The problem with the first design is that the table is 1/4 the base size.  Depending on what you do for the tables, the amount of mass needed to move can get rather high, since the table and bars of the second table need to be driven by the bottom table drive.

The problem with the second design is the cable drive and the limited height of the tool.  On the other hand, it can be a good design for drilling PC boards or routing the copper off them to make tracks.

 

Table without Y axis,  This picture is done in Carrara, and if I had enough space (it costs half a megabyte so far of space) I'd put a movie in here to show how it moves.  Suffice it to say, the table moves left and right on the rails.  I'm going to base the design on this model.

 

Well, all things considered, I decided that the movable table design was the one I wanted to build.  So the first item was to try to figure out what I wanted to do, and how I wanted to do it.  My first try at a design used a 3/16 inch thick piece of aluminum, and square rails for the design.  I very soon ran out of metal without even cranking up a single tool.  A trip to the local metal supply store found a number of things. 

First is that there are no precision ground squares that they had.  The precision ground flats that were available were only 1/8 inch thick, and while 36 inches long, their price for the piece of metal was well over 50 dollars.  Time to try something else. 

A quick change of design decided me on two 1/2 inch diameter pieces of drill rod.  They were about 9 dollars for a 36 inch piece, which was considerably more reasonable, by my standards.  So that, and some 1/4 inch plate for the base and table were bought, along with a lot of 1/2 inch thick by 1 1/2 inch aluminum bar. 

 

With the model a bit more developed, you can see the Y table perched on top of the x table.
Now the piece they gave me wasn't square, although it did seem to have some of the angles close to 90 degrees.  I'll probably put it on the table saw and trim it a bit, it needs a little straightening.  I'm also going to give the cross panel a bit of a trim.  Not my favorite thing, but at that size, I have little else to trim things with.  Time for body armor....

The design has two parallel bars on the sides.   I bolted them together to keep them in the same  orientation.  The first thing to do is to put everything in a vise and drill a pilot hole for the tap.  The next thing is to drill through 1/2 of the bar width (that's to say, one of the 1/2 inch bars) with a body drill for the screw, then tap the hole.  All of this is done without removing the bars from the vise and changing the orientation.  The exact position of the holes isn't critical , they just need to be out of the way.
Nothing spectacular here, but none of these corners are quite square, and each side is a different length.  You just can't get anything good for scrap any more.  However, it's roughly 16 inches on a side, plus or minus about a sixteenth.  We won't even mention that the price I paid for it wasn't a scrap price. 
Now there's nothing particularly special about about drilling a hole, but the neat thing here is that the indicator is mounted on a magnetic mount, and is set so that with the drill in the hole, just ready to drill, the indicator is zeroed. 

The hole is drilled until the dial reads 1/2 inch displacement, which is just through the top piece of aluminum.  The whole is tapped using the drill press axis as a reference, there's a small spud used to align the tap.

Before the vise is loosened, the bolt is tightened.  That way, the alignment of the rails will be exact. and they should be parallel if the two bars are parallel.
This is the drill rod support milled out.  I used a 1/2 inch  ball mill, which does not cut deeply, but does cut a nice 1/2 inch radius..  Those should be in close alignment.
I had a small piece of drill rod left over, so I used it as a precision spacer.  I then measured from the bottom of the bars to the top of the flat to set the depth of the cut.  I got it within 0.001 inch, I think.  That ought to be close enough.  I do expect to have to shim things a little, but no big deal.

 
Each piece of drill rod has a flat milled on the end, which will allow the head of the cap screw to seat well.  The dimension isn't critical, just enough to give full support for the head of the screw.  The spin indexer is used hold the rod.

Next, the center of the flat is found, and a through hole is drilled of the same size that will be used to tap the support base.

However, I did one without thinking of the other, and had to align the flat so that I would bore a hole on the diameter.  How to do that?

 
Here's the answer.  You use the same indicator and run it across the flat.  You change the spin indexer degree by degree (it was already close) for best alignment, and then you've got it level.  You can see the lathe in the background and the 4 inch mill vise.  I really like the new toolbox.
So here's the main base plate with the left hand support mounted.  I very carefully measured a spot for the screw in the support, and a similar one on the base plate.  I drilled both, and tapped the support for 1/4-20.  The base plate was drilled to accommodate the cap head screw.  That was the first hole.  Then, I aligned the support square with the front edge, clamped it, then drilled through the base plate into the support.  I tapped the support and bored out the base for the screw.  This made certain that the two holes would be aligned.  Details next.
When I drilled the base plate through, I did not unclamp it at all.  I put a depth indicator on the mill head, and set it to zero with the drill just touching the plate.  I drilled down about .300 inches, since the plate was .250 thick.  This would go through the plate, and leave me plenty of material in the support.  What I didn't want was the 1/4 inch diameter (and deep) hole to go 1/2 inch into the support.  This worked fine.   I think that's Trading Spaces in the background.
I did the same thing for the second support bar, the one on the right.  I squared it with the front plate edge.

Next, I attached the drill rods at the other end, then clamped them down as you see.  The clamp on the far right clamped directly to the base.  I had to be sure that the middle clamp holding the rod was not too tight, it would cause the rod to buckle.  I ought to have gotten a clue here, but not quite yet.
Here's the base clamped.  I knew that the plate was not quite square, so the underhang didn't worry me.  It probably ought to have....
The rod mounts in the slot and is held in place with a 1/4-20 cap screw. 
To mill the other end of the rod, I found it to be easiest to cut a very small flat on the top of the rod.  The setup I used wasn't secure enough to be able to hold the rod while it was being milled.  I guess it had something to do with a 16 inch piece of aluminum and a 4 inch milling table.  That small flat, though, was enough to allow me to put it in the spin indexer, align it flat (already covered the method), then mill it the rest of the way down. 
Well, if you're wondering what this little gem is, it came about like this:  I mounted both the rods, front and back, and rather congratulated myself.  However I made the fatal error of measuring the distance between the rod supports front and back.  Hmm, I said, off by 1/8 an inch.  Now, that's house carpentry tolerance, not machine shop.  I had not thought of the effect of the non-square piece of metal.  Disassembled everything, took the metal base out and squared it up on the table saw with an aluminum cutting blade.  Ground the edges a little to clean them up, and then went back.  I slotted the right rear hole in the base plate, and then slotted the support rod after aligning it again.
You can see how far I had to move the support over.  However, I have some more adjustment room, so I can get the two support bars very parallel now. A consequence of this minor adjustment in the plate is that there's an overhang on the back of the plate.  The bar is too long by  a bit.  I'll trim it off later, for all I know, it needs to be like that.
Note how far the bar has moved, but it can still go some more, if needed.  I see some way of extending the dial calipers in my future.
The bearing supports, (you'll see how they go in a bit) are cut from 1/2 inch 304 stainless steel.  I both love and hate stainless.  It looks very pretty, but, well, we won't go into how much fun it is to thread.
This bearing assembly (there are 8 needed like this  for the lower table/bar assembly, is nothing more than a 1 inch piece of bar, with the first 1/4 inch or so turned down to 8 mm, which is the diameter of the ball bearing assembly axel.  Those bearings were made for in-line skates, and are about 13 dollars for 16.  Wonderful price, but only one size ever.  The shaft end is drilled and tapped for 8-32 hardware, and that holds the bearing in place, along with a stainless washer.  in this assembly, everything is probably stainless.
The next assemblies are almost identical, but provide an cam adjusting action.  To get this, I put the blank in a 4 jaw chuck, and adjusted the chuck so the offset was about 50 thousandths.  The dial of the indicator I used showed a total movement of a bit more than 0.1 inch.  I put the indicator in a holder and magnetically mounted it to the headstock of the lathe.  This is what such a piece starts to look like as you turn it.  The lathe tool alternately cuts, and is free.  This puts stress on the lathe, so I only removed about .020 maximum at a time.  That's diameter, not radius, the cuts were no more than .010 deep.
Measuring the cut with the digital calipers.  Using the knife edges allows me to measure a diameter without worrying if the calipers are perpendicular to the lathe axis.  Otherwise, I'd get different readings as the calipers wobbled from side to side because the jaws would pry open a bit.  At .3865 diameter, I want to take three cuts of .010 depth, which removes 0.060 in radius.  This would get me down to about .3265, which when I remove (and measure first) about .005 material, would take me down to about .3165, and I'd sneak in from there.  Basic thing is that I don't want the bearings too loose.  I also want the shaft part to be a little short of the bearing thickness so the washer will hold the bearing in.
While still in the lathe, I drilled a hole in the center of the offset part completely through.  Since this is a cam action, I need something to be able to grab the cam and twist it.  I settled on using another cap head screw from the back.  This means that I tap the stainless rod for the bearing hold down, and again for the adjustment screw. 

I used the collet fixture to hold the bearing assembly, then a spud (pointed rod) to help align the tap with the work piece.  This worked rather well, except that I found stainless to be a rather difficult metal to tap well.  For this design, I have to tap the top and bottom of the bearing support.  The top hole, shown here, is for the bearing.  The bottom hole is for the offset adjust.  You'll see what I mean soon.
I got four of them done, and broke two taps in the process.  Needless to say, they broke off without enough sticking out to dig them out again.  I was not pleased.  I'd kept them vertical, and was not tapping deeply, but it was enough.  This is the bearing assembly as first designed: side view.  You can see the offset in the picture.  The rear screw, on the right, is how you adjust the cam in the hole, you just twist.  I could have done the same thing by milling a pair of flats on the end of the bar.

To eliminate the tap problem, I've redesigned the thing a little, and I will not need to do any tapping of stainless.  It did require hardware I don't have, but it was worth it.  I also have taps on order.  I'll show the redesigned one in the next picture.  After the redesigned bearing shows up, I'll go into the table support blocks, which these bearing supports mount on.  The only critical dimension here is that the bearing fit the shaft without any play at all.
The bearings are mounted on a block.  There are four vertical bearing assemblies on the X table.  While the mounting holes need to be accurately placed, the real critical dimension is the distance between the fixed bearing and the top.  This is the top view without the bearing holes bored.
This is the bearing block with the top (adjustable) bearing ready to be bored.  I start off with a center drill, then move up to a 15/64/ths drill size.  The last little bit is to ream it out with a half inch reamer.  Now fortunately, this is a press fit, unfortunately, the adjustable bearing needs to turn in the mount (the shaft, not just the bearing).  So for the top one, the hole has to be bored out a little more than 1/2 inch.  Note that in the other pictures, I've added a stop to accurately position the other blocks as they are machined.
You can see the boring head doing it's little thing here.  I'm rather glad I bought it, since I knew I would have to make some odd sized holes in plate.  I had mostly bearings in mind here, and they're metric anyway.  note that there's support under the aluminum, but enough space so I don't drill into the table.  The boring head, by the way, moves outward by 0.001 each division in rdius, so that's 0.002 for each division in diameter.

In drilling the next hole for the fixed bearing, I supported both sides of the block, used a stop to fix the position of the block, and still use the 123 block.  That block is just there to space the block out from the tall fence.

 
This is a look at the bearing assembly.  The block above is adjustable, the one below is fixed.  I would have reversed them, but there's no neat way to adjust the lower bearing lock (which there is none of).

There will be one at each corner of the X table.  The bearing is reversed as shown.  The bearings will be inside, not outside as shown.  Note that there is nothing that keeps the table from sliding off the rails.  That's the job of the horizontal bearings and the horizontal bearing assembly.
This is the horizontal bearing assembly.  You'll note that the supports are longer, and the adjustable support sticks up through the X plate.  The screw on the right holds the fixed bearing assembly in place.    the screw on the left (the horizontal one) locks the adjustable bearing support in place.
Here's the picture of the bearing assembly, disassembled.  There's an extra adjustable bearing shown assembled to give you an idea of how the assembly goes together.  The supports are counterbored for 1.25 inch long screws, so the head is actually inside the support.  You reach in with an allen wrench to adjust it.
Detail of the adjustable assembly.  note the groove cut in the shaft, the screw fits in that and keeps it from dropping out.
Here's the bearing assemblies mounted....  It was at the end of the day and it was a long one.  See a problem?  Well, Ididn't until I flipped it over and tried to put it on the rails.

Unfortunately, it's a good 1/4 inch too far towards the edge, so the rail will go right through the vertical bearing assembly supports.  Not too good. 

The simple solution is to slot the holes.  However, before I did that, I just bored these out and tried mounting the plate. 

That was not really a good thing to do, mostly because I found that to tighten the bearing assemblies enough to reduce slop, I ran up the friction so high that the table almost wouldn't move.  This is not what I wanted.
Well, I tried another approach.  I had some 1/2 inch square tubing, which I substituted for the rod.  I've got a larger bearing surface, and that made the whole thing seem to move better.  I find less friction this way.  So what I had to do is to cut two pieces of 1/2 inch steel tubing and put them in place.  That meant that I had to modify the supports for the rails, and also had to clean up the holes for the "adjustments".
It was easier to mill out the bottom of this square than it it was to make it round.  So much for the ball mills right now.
Here's the bottom view of the completed X table bearing assembly.  The rod is just threaded thoughthe bearings to show what it's going to look like.  This picture isn't possible to take from this angle with the X plate mounted.  I think that it even looks rather impressive.  Now only if it works well.

A side view of the vertical bearing assembly, viewed one last time.  The adjustable bearing assembly is on the top.  The trick here is that I added a fender washer to the adjustment screw.  This keeps the bearing assembly from popping through, and makes it look nice.  Hey, nice is good.

Ok, so if all this works as well as I want it to do, then the rail assembly is finished.  I don't need to do any more documentation on it.  The Y table mechanism will be similar.  Now the next part is the lead screw assembly and the worm drive.  In the picture below (simulated), you can see dht ewlrm gear drive and the stepper motor.  The stepper is the simulated little box on the very front in this view.
The first step will be to put in the X lead screw supports in, which will need the boring head.  Then the next item will be the lead screw and the threaded drive block.  Next, the worm gear and drive gear will be made, and the stepper motor mounted at the last.  I'll have to be careful there, if the stepper does not have enough power to move the table under load, and well, I'll have to get a more powerful stepper.
I decided on the 10 TPI acme screw.  I don't have an acme tap, so there's a little problem here.  However, there's a workaround.  I turned down an acme thread nut to just get off the flats.  Then, I bored a hole in the table block that fits the circular part of the nut.  I will use something like Locktite to hold it in.  The block will be mounted to the table bottom by screws in oversize holes to allow alignment.
The far end of the lead screw is turned down to fit the 8 mm inside diameter of the bearing.  I further cut down the end of the shaft to 0.25 inches, just in case.  The end was a little too loose, so I knurled the part under the bearing, and now it fits very snugly.  You can't take the bearing on and off too much, though, the knurl will wear down.

As a further note, I'm not sure what this is made of, might be drill rod, but it's not fun to machine.  Not nearly as nice as stainless.  I think it might be work hardening.  Not a nice thing to deal with, but more or less acceptable.
The bearing supports are bored almost completely though at the bearing diameter, and deeply enough to support the entire bearing.  However, the back is bored out enough to clear the inner race of the bearing, since it's a bit wider than the outer race.  Now the next trick is to align everything.
The bearing supports are adjustable vertically, this shows how the holes are elongated vertically.  The table can be adjusted horizontally.  This takes care of most of the possibilities in terms of the adjustments.
I have no idea of the metaphysical significance of this, but we're going to get a tower of babel here when milling out the slots.  I plunged milled the holes, which actually worked better.  I only offset a 1/4 inch hole by about 1/16 of an inch, which worked well enough when I used a very slow vertical feed.
A side view of the bearing and table, showing the drive end of the bearing assembly.  You can see the washers for the table adjust on the top of the table.
Here's the business end of the drive.  There's a small worm gear on the drive shaft, but it's too small given the way that this is built.  The vertical plate at right angles that supports the worm gear is adjustable over about about 1/4 inch or so.  There's a small collar and insert in the bearing inner race that adapts the 0.312 or so bearing inner diameter to the slightly trimmed down shaft of the 1/4 by 20 thread of the worm.  It's made of stainless.  
The back view of the worm screw bracket, but the screw heads cover up the adjusting slots. 
Top view of the worm gear assembly.  I'll make a larger worm gear here to take care of the adjustments.  There will be some words on the drive system in a few.  There are some tradeoffs that I'm still trying to make.  They involve the size of the stepper, and slewing speed.

The bottom line seems to be be, if you want speed, then you need a big stepper.  If slow is OK, then a smaller stepper will work.
Here's the overall view, but without the stepper and the motor mount. 

Stepper Motor Considerations

Well, the question now is not so much whether or not to use a stepper, but what kind to use and how to build it in.  Let's look at a direct drive situation first.  The leadscrew is 10 turns/inch, so one stepper turn is automatically 0.100 inches.  If I want to have a step resolution of 0.0005 or so.  Now that requires a stepper of at least 100 steps, which is relatively easy to get.  However, the problem here is torque.  Steppers in floppy drives are about 10 in/oz.  That's not enough to move the table.  If I gear it down by a 10 to 1 ratio, I'll get at least 100 in/oz.  That'll work.  However, let's assume that the stepper can run at a 100 pulse/second rate.  That gives me 1 rev/second.  Now that (assuming a 10/1 ratio), means that I get 6 rev/minute of the main screw, which is a maximum feed rate of 0.6 in/sec.  That might not be fast enough.  On the other hand, it's doable, in terms of the mechanics and the availability of the motor.

Now if I can find a 100 in/oz motor, I might be able to do a direct drive.  The bigger steppers run off something more than 12 volts, so that means that I need a larger power supply.  No big deal here, I do have one.  However, the question is how to drive the leadscrew.  A belt drive will work well, but I'll need to find a good belt, as well as some decent pulleys.  I'll probably end up making the cogged belt pulleys.  For right now, I'm going to try the smaller stepper unless I can find a few larger steppers.  I'm not going to have a fast slew rate, but if it's all under computer control, then I suppose it doesn't matter how long it takes.

If I want to do some stuff to the lathe and to the mill, I'm going to need those bigger steppers, I think.  Have to see what I can find.  Needless to say, I want them at low prices....

As it is, I'll need to build a stepper motor tester so I can try different motors, and so I can even just run the motors I've got.  More on that as another project. 

Well, I've done some of the last stuff on the X table.  First, I decided to put in the limit sensors.  I found some at 75 cents each, with wire leads.  I've used them before on the barn door and lathe tachometer.  Since I wanted them to be all exactly the same, I bought them rather than salvaged them.  There's no actuator yet.
This is a simple actuator, made from a piece of 1/4 inch aluminum.  Two screws hold the flat plate in position, and the single screw holding the actuator in place allows me to adjust both the angle and sense point.  There's utterly no stress on the actuator, so all it needs to do is to stay in place.
An overall view of the two actuators with the carriage at the full right position.  Those who have been paying attention will notice that the stepper seems much different.  It is.  I replaced the 12 volt 300 ma. stepper with a 5 volt 1.4 amp stepper.  The original stepper had perhaps 10 oz-in of torque, this one has 130 oz-in of torque.  It's fast enough to move the carriage directly, and will run at about 400 steps/second maximum. However, it did need a new mounting system.
I used a 4 inch wide piece of aluminum plate 1/4 inch in thickness.  I'd probably want something thicker, but that's what I had and it seems to work well.   Since this is direct drive, it doesn't get a bearing support other than the motor.  The replaced bearing assembly is going to be used, as is, on the Y table.  However, the motor has a slightly raised circular plate that needs either a 3 or so inch hole in the plate, or an inset.  I decided for the inset since the plate would be stronger that way.  Since it was too big for the lathe, time to use the (brand new....) rotary table.
In the previous picture, I started at the center.  (center the plate on the table, then center the table.  Offset it only in X for the diameter, milling as you go.  Once you get the right radius out, lock the table and start to crank on the rotary table.  You'll get the result you see.  Next, I move the table back in, and just crank again to mill a smaller diameter circle.  I like this.... I really do
With the result, that here's the newest stepper.  I'll get pictures of the mounted motor and wired bracket in a few, but for now, go to part 2 which covers the Y table so far...