PDA

View Full Version : Second MDF machine



spalm
07-25-2005, 12:48 AM
Well I thought it was about time to post what I have been playing with for the last couple of months. I had a lot of CAD to learn. It seemed to pay off. It was slow, but I now know how to model a machine. Thanks to all who helped.

I’ll post a bunch of pictures of my dabbling. I was just prototyping in MDF and planned to switch to Baltic Birch ply for the real machine (you can see the sheets of ply in the background in some of the pictures) but I am so pleased with the results in MDF that I might just go with what I got. The results really rock!

It is basically a torsion box design of the JRGO machine (thank you JGRO for your design). Same dimensions, but I just turned the machine inside out and put the bearings in compression pressing in against the black pipe. (This is nothing new, see Ger21’s machine and several others) The bearing adjustment blocks work great. I did some tests for parallelism by marking the four extreme corners with a pen on the Z axis and then measuring the diagonals and it is dead on. Flex is not detectable with my initial testing. I still need to add the second skin to the bottom of the X (long) axis, but I want to drill and install a bunch of T-nuts for hold-downs first.

The design is held together with just 8 T-nuts on the gantry: 4 on the upper torsion box and 4 on the bottom box of the gantry. 4 T-nuts for each of the bearing blocks supply adjustment and 2 T-nuts lock each bearing block in place. (No tapping!) This is the first time I have used T-nuts (they seem kind of cheesy, but I must admit they work well in this application).

I had planned on using Igus round slides for the Z until I read some of the recent posts. So I don’t quite know what to do. Anyone have suggestions?

I have accumulated ½ inch by 8 two start Acme rod, Dumpster’s lead nuts, thrust and standard bearings, lock collars, Belleville washers, and motor couplers. (Plan to re-use the Hobbycnc drivers and 200 oz steppers) But I have not arrived at a solution for the drives yet. Now that I have the XY done, I plan to work on this and the Z.

I am so happy with the design so far I can’t decide on whether to use the dancing banana or the sheep.

Thank you all for your help (this place is the best),
Steve.

zoltan
07-25-2005, 05:13 AM
Hi Steve,

Do you have any plans of your machine?

Thank you.

Zoltan

DieGuy
07-25-2005, 06:49 AM
very nice indeed! ;)

spalm
07-25-2005, 09:55 AM
Hey Zoltan,
Yes I do have plans, but the Z axis and the drive screws are not complete. So not sure if you want them in this state.

zoltan
07-25-2005, 11:44 AM
Hi Steve,

It is OK even if they are not complete. When you have the Z axis and leadscrews you can share. In the meantime I can build the other parts.

Thank you,

Zoltan

mvaughn
07-25-2005, 12:48 PM
Steve,

Your machine is incredible. I was just thinking about how to design a machine that uses torsion boxes to support for the X axis bearings.

I would like to know your opinion on how much stiffness and rigidity has been gained by supporting your X axis pipes compared to the unsupported pipe on the original JGRO design?

Keep those pics coming.

spalm
07-25-2005, 01:28 PM
Zoltan, I plan to supply plans but some things changed in the shop compared to the original design. Things like the bottom of the gantry does not line up with the bottom of the box, and the torsion supports changed in shape. The leadscrew mounting determines the holes so I think I should wait. I am using Inventor for CAD and any change in one of the small parts ripples into the total design and drives the constraint manager crazy (or maybe it’s just my CAD skills).

Mvaughn, support was the one of the main reason for the redesign. Also it is much easier to align. Those four screws really work. This puppy is solid, still need to do more testing and if it doesn’t perform, I’ll redesign it again. This project has totally consumed me.

Steve

spalm
07-25-2005, 11:19 PM
Ok, I did some more testing. I started by tightening the bearings to make sure that all the slop was taken into account. Then I clamped a square to the Y axis and moved it around to test for level. The right side was over a 1/8 inch shorter than the left. I think this is because the right gantry wall is flexed from the pressure. By tightening the top set of the bearing adjusters and loosening the bottom, I was able to completely correct for this by causing the bearings to ride more on top of the pipe (see picture). It is now dead on across the width of the table.

The same tests on the long length of the table was never a problem. Flex is not noticeable at all (even though the long pipe is also half inch).

I wanted to take a movie of the ruler just skimming the entire top but thought it would be too nerdy.

I am really liking this…

Steve

lurch
07-25-2005, 11:48 PM
Great rework of a classic starter design. Simple, well thought out improvements...nice, great for beginners! I like it...tell us how it performs, and more pictures please.....

mvaughn
07-26-2005, 12:35 AM
I wanted to take a movie of the ruler just skimming the entire top but thought it would be too nerdy.


I don't think it would be nerdy at all! Everyone on this site has a little nerd in them anyway, we build CNC machines as a hobby. :wave:

spalm
07-26-2005, 10:08 AM
Just some random thoughts. Anything to keep me from documenting this design.

To make the X axis pipe supports for the large torsion box, I started by cutting a sheet to 19 ½ inches and then ripping it to make the eleven 2 ½ inch strips. This made sure that they were all exactly the same size. Then I used a jig on my router table to add the curve for the pipe. By using a pattern cutter bit (with a bearing on the bottom of the bit) I inserted a strip in the jig, routed it, flipped it over, routed it again, flipped it end for end, routed it, flipped it again and routed it. This way I was assured that cutout was in the center and all the pieces were the same.

Lurch, thanks. That was my intention.

I am now thinking about using two rods and delrin slides for the Z using a jig to drill the holes. Any thoughts or a source for these rods or smooth pipes. Maybe half inch by 12 inches?

Steve

mvaughn
07-26-2005, 11:24 AM
I was under the impression that you cut all your parts with your first machine. It's impressive to see that you made it so accurately by hand.

Here is a link to some 1/2 inch drill rod that you can use for your z-axis. I've used drill rod like this for all the axis's on my machine and it works very well. A little 600 grit wet sanding will polish the drill rod up very easily.

1/2 inch drill rod (http://www.use-enco.com/CGI/INSRIT?PARTPG=INSRAR2&PMAKA=408-0029&PMPXNO=943418)

spalm
07-27-2005, 12:52 PM
I cut the top torsion box supports and the gantry walls by CNC. The rest of the parts on the table saw. My current CNC is still a little too sloppy.

Just a note about my table saw. It is a 1954 Sears model. My Dad bought it for me on the day that I was born (I guess my Mom was preoccupied). He remodeled two houses with it as did I. I started using it when I was 10 and took it with me when I left the house. So if any of you need an excuse to buy more tools use the one my Dad did and buy it “for the children”.

Mvaughn, thanks for the link. It is cheaper than I thought. Any links on bearings or bushings for the slide?

Steve

ger21
07-27-2005, 01:00 PM
They made Beismeyer fences in 1954? :)

damae
07-27-2005, 01:03 PM
... So if any of you need an excuse to buy more tools use the one my Dad did and buy it “for the children”.


Your dad is a very wise man! It's actually brilliant! I'm currently trying a variant of that tactic -- the CNC project at my house is a "father and son" project. My son is only 7, and if I'm going to be spending all that time in the garage (instead of playing basketball with my son), at least I can involve him.

My wife buys it so far =)

mvaughn
07-27-2005, 03:23 PM
Mvaughn, thanks for the link. It is cheaper than I thought. Any links on bearings or bushings for the slide?



Do a search at http://mcmaster.com for "UHMW Polyethylene Plain Bearing". That should give you a page that lets you pick the dimensions for your required bushing. They are really inexpensive!

I purchased oil impregnated bronze bushings for my machine but scrapped that idea because I couldn't keep them from binding. Here is a link to victorbl's thread where he used UHMW bushings. http://www.cnczone.com/forums/showthread.php?t=7614

Mark

spalm
07-29-2005, 11:35 PM
Like my new spindle motor? I was doing testing for flex and needed some way of applying stress. It weighs 30 pounds. Again with the square clamped to the Y carriage, I detected absolutely no difference with or without the added weight. I drove it around the table for over ten minutes with just pinky pressure. It was amazing.

The carriage did tilt forward by about a sixteenth of an inch. I thought what the heck and tried to adjust the screw thingy’s to bring it back to vertical but the ½ inch MDF was just not up to it. I also realized that I had to re-think how I had the dados arraigned so I could maximize strength. It’s probably time to switch to plywood.

I made some other decisions:

The gantry walls flexed with the increased pressure to bring it into level. They need to be stronger. I’m going to try doubling up to 1 ½ inch to see if it helps.

The T-nuts inside the torsion boxes work better than I thought, but just seem cheesy. I plan to switch to threaded rod running all the way through (left to right). This will require a new box on the bottom of the gantry. On the upper gantry I will run it through the pipes. This will spread the pressure applied from a three inch spread to six inches.

I am going to try to mount the long lead screw where it is in the picture. The table seems like the most stable surface and the left side has no adjustment. I believe that the first several inches over there are wasted anyway. Comments?


Please comment on what you think so far. I am trying to design a beginner’s machine 2.0 and would appreciate any suggestions on what you see as any weakness. She who must be obeyed will be preoccupied this weekend so I plan to be able to fiddle with some of these changes.

I am having a blast,
Steve

mvaughn
07-30-2005, 01:52 AM
IMHO I would put the lead screw in the center of the gantry this will cut down the max distance that the z-axis can get away from the lead screw by half, thus reducing the amount of effective torque. Others have mounted lead screws to the side and wished they hadn't.

spalm
07-30-2005, 11:24 AM
Hmmm. Interesting point. I am so new at this, never thought about that. Is there some easy way of measuring the needed torque? I was trying to detect the amount of pinky pressure in the center and on the sides, and it seemed to be the same. I’ll mount the spindle motor again and try it with the carriage at the opposite side.

I was thinking about pulling it back and forth with a fish scale and then noting any difference at different situations. Wouldn’t be calibrated at all but at least could see if any changes happen.

I am not looking forward to working under the table and adjusting the lead screw.

Steve

ger21
07-30-2005, 11:29 AM
Can you rack the gantry by pushing one side and pulling the other. If so, you could put 2 leadscrews, one on each side, connected by a belt to a single motor.

spalm
07-30-2005, 01:28 PM
Well something had to go. Machine #1 back in kit form. Kinda sad. I needed my table saw back.

Gerry, I can not rack the gantry at all (that I can feel or see). Maybe I should try to lock one side with clamps and then see if I can move the other. I remounted my 30 pound spindle and could not detect any difference whether the carriage was close to me or far away. I sure would like to measure it somehow.

Think I might mount the screw on the left and see what happens. This is just supposed to be a prototype.

Steve

spalm
07-30-2005, 11:14 PM
I don’t have a metal lathe so am trying to come up with bearing block for the motor end. This is ½ inch 8 turn 2 start, so regular nuts don’t work.

The plan:

MDF holders at either end of the table. These have dados and alignment bars glued to their bottoms. They will be much stiffer than in the photos (in fact little torsion boxes), I was just working out dimensions.

Motor mounted to holder.
¼ inch to ½ inch shaft coupler.
(Small gap)
½ inch lock collar (locked)
½ inch thrust bearing.
MDF thrust plate.
½ inch thrust bearing.
½ inch lock collar (loose)
(Big gap)
Dumpster’s anti backlash lead nut
(Big gap)
½ inch radial bearing in MDF holder

Before attaching the lead nut, back it into the bearing block to apply tension and tighten the second lock collar.

We will see,
Steve

CNCRob
07-31-2005, 05:59 PM
Great looking machine Steve. I like that torsion box design.

spalm
08-01-2005, 12:06 AM
Thanks CNCRob, I think it has real potential.

After moving it off the table saw and onto the 27 inch tall table inherited from machine #1 (which had already bumped my wood lathe into the growing junk corner) I was able to study it much further.

The gantry walls really flex from left to right. Double thickness is easiest. Double ply would be better. Torsion boxes would work but add quite a bit of width. There is another thread with gantry stiffening going on where sandwiching aluminum pipes or square channel inside the double layer is talked about. I am trying to decide on a reasonable compromise for this size of machine. Any suggestions? Something in the back of my mind wants to not over-design this and also keep it simple enough for others to build if they want. I promise to document when done. (I used to teach and it has never left me)

One other problem that I detected now that the machine can be easily manhandled is that the long pipes can move up and down. Not by much and only it a lot of force is applied. This is caused by the fact that the curves in the torsion brackets are larger than the pipe and the pipe is not attached in any way other than gantry pressure. The curves are created from a 7/8 drill bit’ed jig that is then double routed to get an exact center (see post #11). This is larger than the .84 inch pipe. Two solutions that I have thought of;

#1-Route a V instead of a half circle thus having the pipe touch two points.
#2-Insert threaded rod though the pipe and use a turnbuckle to pull the pipes together.

Thanks for help and sorry if I ramble,
Steve

acondit
08-01-2005, 12:54 AM
Steve,

I am not an expert, but I would lean towards the V-solution even if you used something to fasten the pipes in. It would give you two points of contact that could be shimmed if the rail varies in thickness along its length (and black pipe often does). Unless you were to use something like the ground linear rails you are probably going to have some variance in thickness of the rails. So a semi-circular notch large enough to allow the rail to set in is probably only going to have one point of contact (except at some lucky spot where it is a precise fit.

Alan

strat
08-01-2005, 07:26 AM
how wide is your gantry?? mine is done with mdf i notice alot of ppl talking about motion side to side when their gantry is straight cut mine is like 12" at the bottom and 4" at the top it ( more of a triangle shape) eats alot of my motion i know but seems more stable i dont' have the flex

OCNC
08-01-2005, 10:28 AM
The gantry walls really flex from left to right. Double thickness is easiest. Double ply would be better. Torsion boxes would work but add quite a bit of width. There is another thread with gantry stiffening going on where sandwiching aluminum pipes or square channel inside the double layer is talked about. I am trying to decide on a reasonable compromise for this size of machine. Any suggestions? Something in the back of my mind wants to not over-design this and also keep it simple enough for others to build if they want. I promise to document when done. (I used to teach and it has never left me)

Any geometry change that increases the depth of the vertical gantry wall cross section will improve its stiffness. The stiffness increases with the cube of the depth. Twice the depth is eight times stiffer for the same material (and geometry). The maximum bending stress occurs at the outer surface of the material so this is where it is best to have the highest modulous (E) material in a composite panel. For the same thickness a 2-sided melamine skinned panel would be stiffer than a plain MDF panel. You could double skin an MDF panel with plastic laminate using epoxy or possibly even high quality carpenters glue for the adhesive. ( I would stay away from contact cement). You might also be able to bond aluminum skins to the MDF. I believe that in this case it could be done well enough with epoxy. The stiffness only changes linearly with a change in the modulous of a material. If the E for two materials of the same geometry is different by a factor of two then the stiffness varies by two. MDF has an E of about 400,000 while Douglas fir has an E of around 1,800,000. Aluminum (6061-T6) has an E of around 10,000,000. Adding a couple of 1x2 D-fir strips to the outside of the vertical MDF panel will give an increase in stiffness by 'the cube of the depth effect' and by 'the higher modulous at the surface effect'. Using an aluminum shape could be even more productive. Whichever material you use, it's important that the mating surfaces be thorougly adhered to each other to maximally distribute the shear stress at the join. Otherwise complete or partial failure of the join can occur and the ultimate potential stiffness will be degraded.



One other problem that I detected now that the machine can be easily manhandled is that the long pipes can move up and down. Not by much and only it a lot of force is applied. This is caused by the fact that the curves in the torsion brackets are larger than the pipe and the pipe is not attached in any way other than gantry pressure. The curves are created from a 7/8 drill bit’ed jig that is then double routed to get an exact center (see post #11). This is larger than the .84 inch pipe.


I would mix up some thickened epoxy using colloidal silica, which makes a thixiotropic mixture (doesn't run), and apply it to the curved pipe seats. You could apply the mixture without removing the rails using a small plastic syringe to inject it into the gaps or you could remove the rails, apply the mixture to the seats with a small stick and then replace the rails bedding them into this. One method or the other will be more appropriate depending on what is involved in aligning the rails. When the epoxy cures the rails will be firmly seated at each location and you should have no flexing at all.

spalm
08-01-2005, 10:38 AM
The gantry is 23 inches wide (and 8 inches deep). I just picked this number because it allows the supports to be cut from a 2 foot board. The table top (X) is 41 inches deep.

I kind of like the V groove also. Good idea about the shims, much easier to do this on flat surfaces.

The right gantry wall is in a constant state of curvature. I didn’t care because I was able to correct for it by changing the bearing alignment. I now feel that I need to take care of this or I will never be able to get the Y lead screw straight as the motor is attached to this wall. Again pointing to the need for stronger gantry sides.

Wow, thanks OCNC. Just caught your post as I was replying. Going to take some time and absorb it.

Steve

spalm
08-01-2005, 01:52 PM
OCNC thank you. You a knowledgeable man. Thixiotropic, sweet!

I didn’t think epoxy would work here. Not sure why I thought that. I still have some left over silica from when I built my plywood canoe. Need to get some fresh epoxy and a syringe. That will be easy; I don’t have to rebuild it.

There are two dados on the inside of each gantry wall used to achieve an easy transition from horizontal to vertical. I believe that these would mess up any skin I apply to the inside. I have some ¾ inch fir handy panels that I could rip into 2 inch strips and glue one on either side of the pipe holes. Again, easy, I like it.

Is there a modulus (E) chart somewhere? Do you know what Baltic birch would be rated?

Steve

OCNC
08-01-2005, 05:51 PM
I have some ¾ inch fir handy panels that I could rip into 2 inch strips and glue one on either side of the pipe holes.

You would be making a double-Tee section. Two strips of solid fir lumber would be stiffer for the same nominal thickness because you can only count half of the thickness of the plywood to run in the long direction. If you're familiar with using epoxy and silica besides simply gluing the 2" strip to the upright, run a fillet on each side of the strip and lay in a length of fiberglass cloth. (The 3/4" plywood strip without the fillets and cloth is probably sufficient however).


Is there a modulus (E) chart somewhere? Do you know what Baltic birch would be rated?


Paper birch is around 1,600,000 psi and yellow birch is 2,000,000 psi. With plywood you need to consider that you have only half the nominal thickness in either orthogonal direction. Wood being anisotropic (not the same in all directions) has three different elastic moduli, one in the direction along the length of the grain, another perpendicular to the length of the grain in the direction of a radius towards the center of the tree and the third perpendicular to the length of the grain in the direction tangent to the circumference of the tree. The elastic modulus most commonly sited is the one along the length of the grain (it has the largest value of the three moduli). You can find info about wood as an engineering material here. (http://www.fpl.fs.fed.us/documnts/fplgtr/fplgtr113/fplgtr113.htm) The values will be metric though. For additional info Google 'engineering properties of wood' or 'properties' of the specific material you're interested in.

spalm
08-01-2005, 11:06 PM
Well I took apart the gantry and added the fir stiffeners. I think a bit overkill (and butt ugly), but boy do they work. When I reassembled and leveled the Y axis to the table, the adjustment plate thingy is vertical, just like planned. Yeah! No flex or bow that I can detect, and I am able to run the bearings with a lot less pressure against the pipe, which seems like a good thing to me. I installed the pipes from machine #1 which were already cut to size. Also grabbed the newer bearings from machine #1. The ones on this test machine were actually stolen from one of my kid’s roller blades and were in nasty shape.

Good stuff MrCNC, thanks for the class.
Steve

spalm
08-02-2005, 12:51 AM
There are two major things in life that I have never mastered; backing a trailer and drilling a hole where I want it.

I spent Sunday afternoon trying to come up with a predictable assembly. This ½ inch rod is not a consistent ½ inch outside diameter (I can’t blame them; their tolerances were for the threads). My ½ inch radial bearing is exactly ½ inch inside diameter and it gets terriblely stuck trying to slide it up and down the rod. A piece of MDF with a ½ inch hole does the same thing.

Not sure where I am going with this or if it will work. I really don’t want to have to turn the ends to except a normal bearing arraignment, but am getting discouraged. I want to stay with this 4 TPI screw in order to increase speed and reduce RPMs (read whip).

The table top is the most stable thing here. That is one reason why I want to reference it and drive the screw on the left side. If I can’t make it work here, working under the table would be a disaster. One problem is that it can chew up table space. Maybe with my new gantry stiffeners, I can move it to the outside of the gantry, but I can not figure out how.

I spent a bit of time with a small hole exactly 1 ¾ inches above the main rod holes that I shined a laser pointer through for alignment. Other than proving that I need new glasses, it was a flop. Anybody else try this kind of alignment technique?

On the bright side, I got to use my special 1 ½ inch speed-bore bit. It rocks. I got it about a bazzilion years ago, but I just love it. It is a cross between a Forstner and a hole saw. Just look at the MDF swarf that it produces. I would love to get a set of these, but have never been able to find them. I think they were intended for plumbers as they are short and just rip through a 2x4.

Steve

OCNC
08-02-2005, 07:40 AM
Well I took apart the gantry and added the fir stiffeners. I think a bit overkill (and butt ugly), but boy do they work.


Nice job with the stiffeners. The maximum bending moment is in the area of the x-axis bearing carriage and again in the area where the lower edge of the y-axis beam intersects the vertical support. If curves are less ugly here's an alternate shape for the stiffener.


Did you cut the dados with a router bit or a dado blade? From the photos they look like router cuts. In either case that seems to be a good system for maintaining alignments and accuracy during assembly.

OCNC
08-02-2005, 07:51 AM
Not sure where I am going with this or if it will work. I really don’t want to have to turn the ends to except a normal bearing arraignment, but am getting discouraged. I want to stay with this 4 TPI screw in order to increase speed and reduce RPMs (read whip).

The table top is the most stable thing here. That is one reason why I want to reference it and drive the screw on the left side. If I can’t make it work here, working under the table would be a disaster. One problem is that it can chew up table space. Maybe with my new gantry stiffeners, I can move it to the outside of the gantry, but I can not figure out how.

I spent a bit of time with a small hole exactly 1 ¾ inches above the main rod holes that I shined a laser pointer through for alignment. Other than proving that I need new glasses, it was a flop. Anybody else try this kind of alignment technique?


It's not clear to me what you're trying to accomplish. It seems like you had this worked out in an earlier post. Can you clarify?

spalm
08-02-2005, 10:36 AM
I guess I just got frustrated trying to align 4 holes of different sizes in multiple dimensions, especially when I just suck at putting a bit where I want it. Just need to be more careful. I also was worrying about using up table space, making it repeatable and strong, and not having it look like a big kludge. It seemed so easy at first.

Not being able to slide the radial bearing up and down the rod is also big headache. It can really get stuck. I guess I could try to find some with a slightly larger inside diameter. Maybe some wet or dry on the rod will solve it. It really only needs to slide over the last several inches.

I’ll cheer up and get back to it.

The dados were cut on the router table. My table saw dado blade is a wobble type and I don’t like it too much. I leave my big router mounted under the extension table on the saw so I can use the same fence for both.

So finding out that fir is 4 times stronger than MDF, seems like it would make a good choice for the interior supports for the two gantry torsion boxes. Thoughts?

Steve

ger21
08-02-2005, 10:50 AM
So finding out that fir is 4 times stronger than MDF, seems like it would make a good choice for the interior supports for the two gantry torsion boxes. Thoughts?

Steve

Since you're using so little, I'd buy a piece of maple, which should be quite a bit stronger. Just guessing, though.

spalm
08-02-2005, 11:59 AM
I just received my ¾ inch drill rod and UHMW bearings from MSC for the Z axis. The rod looks great for the price but the bearings are way oversized. Part number 35443340. Is this the way they are supposed to work? The inside dimension is .775 and the OD is 1.137

Do they shrink when you insert them into a 1.125 hole?

OCNC
08-02-2005, 01:13 PM
Since you're using so little, I'd buy a piece of maple, which should be quite a bit stronger. Just guessing, though.

'Strength' is a catch-all word and often a bit of a misnomer. What we're really talking about here is resistance to bending. Shagbark hickory has the highest E at 2,160,000 psi while Black Locust is second at 2,050,000 psi. Sugar maple is at 1,830,000 psi (less than D-fir). Black Locust has a compression strength parallel to the grain of 10,180 psi while for D-fir it's 7,430 psi and Sugar Maple is at 7,830 psi. The highest shear strength parallel to the grain is found in Black Locust at 2,480 psi, Shagbark Hickory is at 2,430 psi, Sugar Maple is at 2,330 psi and D-fir is at 1,160 psi. Shear strength parallel to the grain indicates the ability of the wood to hold fasteners against end grain pullout.

E is the ratio of the stress to the strain and indicates the inherent stiffness or resistance to deflection of a material. If we were working with forces near the limits of a material's strength the modulus of rupture would come into play which deals with maximum stress in a material at rupture. Shagbark Hickory has a MOR of 20,200 psi while D-fir is 12,200 psi and Sugar Maple is at 15,800 psi.

I took these numbers from 'Understanding Wood' by Hoadley. Of the 54 woods listed none have an E lower than 1,040,000 psi (Incense Cedar) which is still more than twice the E of MDF. Also Western Larch (1,960,00 psi) is right up there with D-fir (1,950,000 psi) for those who might have acces to it.

spalm
08-03-2005, 12:11 AM
So if I made the gantry walls out of fir, it sounds like it would be much “better” than MDF. The problem I see it warping and seasonal instability. Seems like a compromise is MDF or plywood for panels and fir for support. (?)

I drilled a 1 1/8 inch hole and inserted (pounded) one of my UHMW bearings. It did go in and stay nice and tight, but the inside diameter is still .77 (as read on my calipers, I do not own a micrometer) 20 thousandths sounds like a pretty close fit, but I can wiggle the bearing a lot as I slide it on the ¾ drill rod. Did I buy the wrong bearings? Is this to be expected and by the time I have 4 of then mounted to a plate it will not be noticeable?

Steve

damae
08-03-2005, 03:04 AM
... 20 thousandths sounds like a pretty close fit, but I can wiggle the bearing a lot as I slide it on the ¾ drill rod....
Steve

20 thousands (.020") may sound close, but when you're talking metal on metal it is actually a lot. That's half a millimeter. It doesn't seem like the play should just go away...

OCNC
08-03-2005, 08:47 AM
So if I made the gantry walls out of fir, it sounds like it would be much “better” than MDF. The problem I see it warping and seasonal instability. Seems like a compromise is MDF or plywood for panels and fir for support. (?)


If you used a quality D-fir stair tread (measures 1 1/16" usually) for stock to make the uprights you should have all the stiffness you need and then some for the size machine you're building without needing to add the webs. Good stair tread material (if your lumber yards don't have it go to a local stair building company) is usually vertical grain and very flat and uniform in thickness (equal to or better than MDF). For stability against movement from varying moisture levels apply a couple of coats of epoxy.

OCNC
08-03-2005, 08:52 AM
20 thousandths sounds like a pretty close fit, but I can wiggle the bearing a lot as I slide it on the ¾ drill rod. Did I buy the wrong bearings? Is this to be expected and by the time I have 4 of then mounted to a plate it will not be noticeable?


A sliding fit should be just a couple of thousandths. I believe that the tolerances for the bearings you have are too great. Look at 2570K13. The tolerance is -0.000 to +0.0015. For info on tolerance look at page 971.

spalm
08-03-2005, 10:33 AM
Stair tread, excellent idea.

Yup, the bearings that I bought from MSC must be for a different use. They come without any specs but they are “good for the food handling industry”. Oh well, need to find some other ones.

I assume that you were looking at McMaster Carr’s catalogue, as MSC’s page 971 is for Tweco Mig Guns. :)

The 2570K13 looks like it is a bit to precise for my application: “Bearings are lightweight, corrosion resistant, and self-lubricating. All are designed for applications where shaft misalignment isn't likely to occur. Use them with end-supported shafts with a Rockwell hardness of C25-C30 and a 4-12 micron smoothness finish.” I have oil-hardened drill rod (MSC 06000483) and will probably have some misalignment.

How about 57785K27? I really don’t have any idea what tolerances I should look for. I guess I am worried about binding with “some” shaft misalignment.

OCNC
08-03-2005, 12:45 PM
I assume that you were looking at McMaster Carr’s catalogue, as MSC’s page 971 is for Tweco Mig Guns. :)
Sorry about the catalog confusion. I did mean McMaster and not MSC.


How about 57785K27? I really don’t have any idea what tolerances I should look for. I guess I am worried about binding with “some” shaft misalignment.Look at the shaft tolerances on page 1012. Stay away from the UHMW as the tolerances are wrong. Try 6362K166 RulonLR for hard shafts. The tolerance is +.002 to -0.000. That's more like it.

spalm
08-03-2005, 01:04 PM
Thanks.

Stopped by Home Depot over lunch and looked at their stair treads. They just called them “pine”. Pretty nasty looking, but they only had 5 left in the bin. I’ll wait until they restock and also look elsewhere. $10.50 for 4 feet.

Actually the solution I already have works pretty well also. Maybe if I cut them down to 1 ½ inches and round over the edges they will look a little less clumsy.

mvaughn
08-03-2005, 04:45 PM
What if you used the 1 inch square steel or aluminum tubing that Home Depot sells. That would definitely add some rigidness to your gantry sides. Angle cut the ends and paint them a nice accent color and it would look awesome.

OCNC
08-03-2005, 05:53 PM
Stopped by Home Depot over lunch and looked at their stair treads. They just called them “pine”. Pretty nasty looking, but they only had 5 left in the bin. I’ll wait until they restock and also look elsewhere. $10.50 for 4 feet.These might be yellow pine treads. D-fir would be better. Ideally you would want a D-fir tread that has vertical grain.

ger21
08-03-2005, 07:23 PM
I know you won't find D-fir at any Home Depots near me.

pminmo
08-03-2005, 08:11 PM
Stiffness, try a sandwich MDF exteriors with an OSB interior. Wood strengthens as grains are mixed. You can take five 1/16" sheets of a hardwood veneer, press laminate them and will be shocked as to how strong and stiff the result is.

spalm
08-03-2005, 11:38 PM
OCNC and Gerry, I suspected yellow pine. That’s why I called it out in quotes as an indirect way of asking. I guess my handy panel that I ripped for the existing supports are pine also. Sure is strong enough for this job although, and cheap and easy to obtain. Until now any non-deciduous tree was just pine to me.

Pimino, I am not familiar with OSB. I have seen many laminations that seem stronger than they should be. I was at Ikea with my daughter last night and they had several examples. (Did you know that just tapping the ESC key stops your typewriter? :) I am old and get easily distracted.)

Family duties keep getting in the way, but I was able to steal an hour to finalize my main lead screw. Got the T-nuts installed in the main table and the holder’s dadoes cut. Now back to drilling.

I think why I am so disappointed in my drilling is I am spoiled by my table saw, router, biscuit cutter, and Leigh dovetail jig. These all have really good fences, indexers, or guides. This is where I live when I build. When I drill it all seems free hand and not at all like the other operations. I use my little square to draw two pencil lines and form a center point. Center-punch it sometimes. Take it over to the drill press and try to clamp a fence and a stop so that I will hit center of my pencil marks. I am at least three degrees of movement from the original objective. Does anyone understand my problem? Do I just need a better fence for the drill? Take more time? That’s life?

Thank you all for your help.
Steve

OCNC
08-04-2005, 08:04 AM
I am not familiar with OSB. I have seen many laminations that seem stronger than they should be. OSB id oriented strand board and is a sheathing material used in building construction. It's made of what look like flakes of wood fiber that are glued together under pressure. It's very dense. It has a poor surface finish (you'd have to sand it before gluing it into a lamination) and it's generally not flat like MDF (residual curves and twists from the manufacturing process). It is cheap but I can't see that it has any place in a CNC machine (except mabe as a shipping container). Also the suggestion to stiffen the neutral axis of a lamination is flawed. It's the outer surfaces that need to be made of a stiff material. The inner portion only needs to be strong enough to carry the shear forces between the two outside surfaces. Think of a hollow box section. It has no middle. Where the edges join top and bottom the area of contact is quite small compared to what it would be if it were a solid section. This illustrates how little contact area is really necessary to resist the shear forces.

Does anyone understand my problem? Do I just need a better fence for the drill? Take more time? That’s life?
You need good sharp drill bits. You might want to use a spotting drill to get the punch prick lined up with the quill. Then clamp the work to the table. Change out the spotting drill for the hole drill and you should be ok. It's not likely that you'll get a very accurate hole if the work isn't firmly clamped in position or if the tooing is dull. Also use wood specific bits when possible.

ger21
08-04-2005, 08:38 AM
Use good brad point bits, or forstners, but they are expensive. clamp or use double face tape to hold parts together when holes need to line up.

Jimmy Southern
08-04-2005, 09:41 AM
Hello everyone,

I was just reading the posts and noticed someone talking about OSB. They mentioned having to sand it. Well if you don't need the extra thickness then go with vinyl floor underlayment. It is also an OSB product but is only 1/4" thick. It is sanded very smooth because it is used to level humps and ridges in the floor so they will not show through the vinyl.

Hope this helps
Jimmy (group)

ger21
08-04-2005, 09:50 AM
I've always seen 1/4" Luan plywood was used under vinyl floors?

Jimmy Southern
08-04-2005, 10:17 AM
Hi Ger21,

When I relayed the floor in my house I had the choice of Luan or the OSB. I chose to use the OSB. It is a much higher quality material than the sheathing version. Very flat & stable.

Jimmy :o

spalm
08-04-2005, 10:55 AM
Anybody have experience with the Incra T-Rule12? Seems like it might ease my pain when marking out. Amazon has it for $32. Little holes every 1/32” for a mechanical pencil to poke through.

I was just talking to the shop that does our sheet metal. They are getting a new waterjet CNC machine. I asked him for a link and here is his reply. Good site, lots of info.

The machine is manufactured by Flow International. That company has produced 70% of the installed waterjet cutting machines and has been selling machines since about 1990. The more we learn about the capabilities of the machine, the more impressed we are. This is an emerging technology, but it has been developed to the point that the machines cut with precision (within a few thousandths of an inch), the pumps are reliable and deliver consistent pressures. Because the cutting method is essentially erosion, there is not high heat affecting the base material, and there is little or no burr. Following is a link to their web site:
FlowCorp (http://www.flowcorp.com/waterjet-products.cfm?id=112)

ger21
08-04-2005, 11:00 AM
Anybody have experience with the Incra T-Rule12? Seems like it might ease my pain when marking out. Amazon has it for $32. Little holes every 1/32” for a mechanical pencil to poke through.


I've got the protractor version, but never use it. I use an Incra gauge for layout. http://www.woodcraft.com/family.aspx?DeptID=2095&FamilyID=4216

spalm
08-04-2005, 11:07 AM
I have one of those in a box somewhere. Used it as a router fence long time ago. Never thought to use it for layout. Hmmm

OCNC
08-04-2005, 11:22 AM
Little holes every 1/32” for a mechanical pencil to poke through.I use a tungsten carbide tipped scribe and a machinists sliding square and machinsts rule for laying out (along with a headband magnifier). A pencil is too thick and 1/32" graduations are too coarse for good hole drilling layout.

spalm
08-04-2005, 11:40 AM
Somehow I knew you were going to say that.

spalm
08-05-2005, 12:22 AM
Well I got the long axis lead screw mounted. The free end bearing got stuck badly, so I just clamped it’s holder to the table. Have not mounted the lead screw nut yet as I ran out of time.

I fired up Mach2 and entered 6400 as my pulses per inch (200 x 8 micro steps x 4 tpi). It spins quite nicely but there is some whip. At max speed (180 ipm) or a more reasonable speed (60 ipm) it seems to be the same. I never checked to see if this cheap screw is bent, but even out of the machine it has quite a bit of flex. Should I really expect it to run true at 40 inches in length? Has anyone ever spun a 40 inch half inch ACME with no lead nut at a decent rpm and seen it run without whip? This is defiantly better that machine #1, but I would like to improve what I saw tonight.

I want to make some whip measurement. Basically just raise a block until it touches the spinning screw in the middle of the table at different RPMs, and compare that with the expected distance measured at the bearing blocks. Then I can tell if anything I do later helps.

I suspect the motor couplers since I saw whip at low RPMs (or that this cheap rod can not support it self). They are a single set screw type and are attached to just the threads. Luckily this would be the easiest section to turn down with a grinder thingy to something like 3/8 and give it something more to bite on.

Another thing that I want to play with is that the hole in the thrust plate is ½ inch. The rod is a piston fit through it. Maybe open this up a bit.

Steve

OCNC
08-05-2005, 07:17 AM
I never checked to see if this cheap screw is bent, but even out of the machine it has quite a bit of flex. Take the screw out and roll it on your table saw top for a quick check to see how much it's bent. The natural frequency of the rod as you've specified it is around 860 rpm with both ends 'fixed' and about 550 rpm with one end 'fixed' and the other end 'supported' ('fixed ' meaning it can take moment loads as well as axial and radial loads and 'supported' meaning it can take only axial and radial loads). Although at first it looks like the driven end is 'fixed' I think it's more of a supported end and even then I'm not sure that the bearings are actually supporting the radial loads (it seems like the MDF is doing this). I noticed that your thrust bearings are just sitting on the surface of the MDF. I think that you should conterbore the mounting surface from both sides and set each bearing into it's repective counterbore for true rolling support of the axial loads. This should contribute to reducing the whip and also take the coupler out of the equation. Do this at both ends for best results. If your thrust loads are taken care of at the driven end you only need radial bearings at non-driven end . (You might want to double or triple up the MDF between the two thrust bearings so that they are further apart and will be better able to resist a moment load. Don't think of the motor bearing as being part of the 'fixed' end condition. The lead screw should be properly supported without regard to the motor. The motor's sole function should be to rotate the screw.)

spalm
08-05-2005, 11:34 AM
OCNC, thank you so much for watching.

I guess if the rod was in fact fixed; the motor couplers wouldn’t cause my whip. I will check the rod for straight, maybe also try a second plate and counter sink the thrust bearings (I was trying to cheat). These thrust bearings seemed to have a lot of slop in them so I was thinking that the ½” hole would help. I now think otherwise.

Thrust bearings: MSC 03379831
Radial bearings: MSC 35433390

Could you please explain natural frequency? Is this the max rpm?

Could you please explain moment load? Is this a start/stop property?

Back to thixiotropic epoxy, I think I used up my silica, but did find my Wood Flour. I had used this on the canoe for fillets with good results. Is the difference between these mostly cosmetic?

nikolatesla20
08-05-2005, 12:31 PM
The thing is about whip is I am thinking that as long as your linear axis is firm and stable, whip shouldn't throw off your movement since the "lever" or "screw" is still being turned. Which means it will force movement.

For example, say you had a lead screw facing upwards. Say that lead screw moved side to side a bit while turning. Thing is, the screw still turns, and therefore pushes the coupler nut upwards, even if the screw moves to the side a bit. As long as the side to side motion is allowed and not carried into the linear slide motion, the linear slide will still move the same distance up as if the screw wasn't moving side to side. As long as the whip isn't a severe angle which causes more motion in X direction than in Z direction then I'm thinking it doesn't really hurt accuracy, it just makes everything louder and wear faster.

That's the benefit of using a screw, is it will always force motion in the preferred direction.

-niko

OCNC
08-05-2005, 12:47 PM
Could you please explain natural frequency? Is this the max rpm?The screw has a natural resonant frequency for each length (like a guitar string). There's a certain speed of rotation that will 'excite' the screw into vibrating at that resonant frequency. If you operate at or near that rotational speed (the critical speed) the screw will start vibrating dramatically. The numbers I gave earlier were the critical speeds for a 1/2" x 40" ACME lead screw. The rpm goes up as the length goes down and the diameter goes up and is also influenced by how the ends are supported. If you want to read more about this check out the Roton website. (http://www.roton.com)
Could you please explain moment load? Is this a start/stop property?The whipping action of a shaft is applying a bending moment to the shaft that the bearing support must be able to resist. Using the double bearings separated by a suitable space (somtimes called the 'bury') provides a reaction to the bending moment caused by the whip. A shaft supported only at its bitter ends might as well be supported by ball joints and offers little resistance to any bending moment. If the shaft were infinitely stiff this wouldn't be an issue and even ball joints would work.

Back to thixiotropic epoxy, I think I used up my silica, but did find my Wood Flour. I had used this on the canoe for fillets with good results. Is the difference between these mostly cosmetic?Besides being thixiotropic silica is harder than the wood flour but in this case the epoxy itself is hard enough. Wood flour is not thixiotropic so your mixture will be prone to running while it cures which you should to be able to control well enough by using as thick a mixture as possible (consistent with being able to place it where you want it), positioning the work to let gravity help and perhaps making little troughs around the seats out of masking tape. Don't bother with the silica if you don't have at hand.

OCNC
08-05-2005, 01:09 PM
Thrust bearings: MSC 03379831
Radial bearings: MSC 35433390Without being able to actually handle the thrust bearings I would be inclined to go with the radial bearings exclusively. I suspect that they will be able to take the thrust loads that this machine will produce while I'm not sure that the thrust bearings will do a good job with the radial loads. Without getting into angular contact bearings, better bearings for the driven side might be a set of double row ball bearings.

spalm
08-05-2005, 01:27 PM
Using the radial bearings throughout is tempting. They are less than ½ the price and seem to have a lot more quality than the thrust bearings. I did play with this concept but found two problems. The shaft collars rub on the outside ring (race?) of the bearings. I suppose this could be solved with a washer. The other is that their I.D. is .5 inches and the rod’s O.D. is about .5 inches causing them to stick. In sections of the rod where I could slide them, tightening a shaft collar next to them caused enough deformation of the threads to render them virtually non-removable.

OCNC
08-05-2005, 03:44 PM
The other is that their I.D. is .5 inches and the rod’s O.D. is about .5 inches causing them to stick. In sections of the rod where I could slide them, tightening a shaft collar next to them caused enough deformation of the threads to render them virtually non-removable.Take 3 MDF squares and drill a 1/2" hole similarly laid out in each of them. Attach the squares to a flat surface in such a way that you can run the lead screw through all three holes, placing one square near each end and the third in the middle. Get a helper. Attach a 1/2'' variable electric hand drill to one end and slowly rotate the shaft while laying a sharp fine cut flat file on the screw end where you want to place the bearings (Poor man's lathe). Check the bearing for fit periodically. If you're careful you should be able to get the bearings to slide on and off without difficulties. Don't be spinning the screw for this next part. Determine where you want to place the shaft collars and then use the edge of the file to make a flat. Go down nearly to the root diameter. You might need to get a longer set screw. Now when you tighten the set screw you shouldn't be deforming any threads.

spalm
08-06-2005, 12:56 AM
Well, half an hour of carefully filing the rod spun by my drill seems to have done it. I picked up some nylon spacers to be used between the shaft collars and the bearings. I had to bevel the edges on them to be sure that they would not touch the outer ring of the bearings. Washers don’t work as they are thin enough that they fall into the threads.

I think I finally have settled all the components for my linear drive.

Lovejoy couplers. $5 x 3
Clamp type shaft collars $1.50 x 3 x 3
Nylon spacers $.50 x 3 x 3
Sealed radial bearings $4 x 3 x 3
Dumpster’s lead nut $15 x 3
12' ½” 8 tpi 2 start Acme $52

Total $166 for all three axis

I also received my new RulonLR bearings for the ¾” Z rods. They are sweet, slide really smoothly with no noticeable play. Can’t wait to start building it.

Steve

OCNC
08-06-2005, 07:08 AM
I think I finally have settled all the components for my linear drive.

Lovejoy couplers. $5 x 3
Clamp type shaft collars $1.50 x 3 x 3
Nylon spacers $.50 x 3 x 3
Sealed radial bearings $4 x 3 x 3
Dumpster’s lead nut $15 x 3
6' ½” 8 tpi 2 start Acme $52

Total $166 for all three axis

I also received my new RulonLR bearings for the ¾” Z rods. They are sweet, slide really smoothly with no noticeable play. ...and your Z-rods are oil-hardening drill rods?

It's looking good! Did you file flats for the collar set screws?

ger21
08-06-2005, 10:30 AM
You can get 1/2-8 2 start acme from MSC for ~$26 for 6 ft. http://www1.mscdirect.com/CGI/NNSRIT?PMPXNO=4454587&PMT4NO=598320

spalm
08-06-2005, 01:17 PM
Oops. That was 12 feet of oil hardened rod from MSC (26 * 2 = $52). I edited that post. It only comes in 6 foot sections. Don’t want to say much until I really beat on it, but it spins fantastically. Just like predicted, spins at 480 rpm (120 ipm), and starts to break up at the next step up. I’ll take it. If I can get 120 ipm out of a MDF and gas pipe machine, I’m loving life. (Probably will be lower, but hey)

Need to attach the lead screw. Initial tests with clamping it to the gantry did not seem to change the above outcome. Want to throw a 15 pound weight on the gantry and try some G code for repeatability.

Yabba dabba doo
Steve

ger21
08-06-2005, 03:01 PM
Just like predicted, spins at 480 rpm (120 ipm), and starts to break up at the next step up. I’ll take it. If I can get 120 ipm out of a MDF and gas pipe machine, I’m loving life. (Probably will be lower, but hey)

Steve

What voltage are you running at? 100+ipm was my target, so it's looking good.

spalm
08-06-2005, 09:32 PM
Well I clamped my 30 pound motor to the Y carriage (is there another name for this) and drove it back and forth. It did not seem to make a difference if the carriage was all the way to the right or to the left. It just did what it was told. Not sure what more to test.

OCNC, yes that is ¾ inch oil hardened rod for the Z. It is on sale for $9 for 3 feet at Enco #408-0045. Feels really nice to me. The shaft collars are clamp type, no set screw, MSC #87856779. Again, I really like them. Only set screw is on the motor coupler and yes, I did file a flat there.

Gerry, I am using a PWM HobbyCnc controller and their 200 oz steppers. I am supplying 3 amps per coil from my 39 volt DC supply. Is this what you were asking? Isn’t your machine 8 feet long? How you going to spin a rod that long, that fast?

Steve

dwwright
08-06-2005, 11:12 PM
I built the jgro machine a while back and noticed that some of the new ones are being built with some added support between the pipes for the Y axis'. Is there a reason for the multiple struts? I was thinking about adding a support that was simply two 1/2" plywood or MDF panels with the long edges cut at a 45 degree angle, then laminate them with one on each side of the pipe. Would allow room for the threaded rod of course. Anyone see any issues with this or able to give a reason for the multiple struts?

Thanks,

spalm
08-07-2005, 12:37 AM
Darren, by adding the struts between the two sheets of wood, you are building a torsion box. These are much stronger and resist flex and twist. They are not really that hard to build and the rewards are worth it. A good torsion box example is the modern hollow core door. It is large, light, and really quite strong. Search this site for the word torsion and you should find a few examples.

Steve

mvaughn
08-07-2005, 01:48 PM
Darren, by laminating 2 sheets of plywood between your rails you will gain stiffness for verticle forces but I would suspect that you will still have problems with twisting. As Steve suggested, do a search for torsion here on the zone and you will find several pictorial examples of torsion boxes.

ger21
08-07-2005, 02:37 PM
Is this what you were asking? Isn’t your machine 8 feet long? How you going to spin a rod that long, that fast?

Steve

Yeah, that's what I was wondering. It's only 5 ft. I'm going to try to put the screws under tension a little to see if that helps.

spalm
08-07-2005, 08:49 PM
The carriage that slides back and forth across the Y axis needed strengthening. The one I had been using for bearing and flex testing was two skinny (6 inches) to hold my new Z, and would almost fall apart when I flex tested it. But before throwing it out, I decided to try to beef it up and see what happened.

I added two 2” vertical pine strips (that were the cut offs from gantry supports) to both the front and back sides. I filled in the lower top with MDF to accommodate. I then skinned the sides with ¼ inch ply. I had predrilled access holes to be able to route out the lead screw and torsion box holes. After the glue dried, I chucked a laminate trimmer bit (with the bearing on the top) and routed the outside and the holes. Looks pretty and is one strong little box.

Remounted the carriage and tried out my 30 pound motor. Success. Not sure if you can tell by the pictures, but the torsion hole is parallel to the box. The carriage actually seems to move back and forth easier with the weight attached.

Steve

spalm
08-12-2005, 04:57 PM
I am working on CAD drawings and one thing is driving me nuts. What is the hole spacing for a NEMA 23 motor based on? Is it really a 1.5 inch hole surrounded by 1.856 inch spaced mounting holes? Where did this number come from? Tried converting to metric and also as the hypotenuse of a triangle, but nothing seems to make sense.

Steve

rweatherly
08-12-2005, 05:02 PM
Might be a 2 5/8" bolt circle.

spalm
08-13-2005, 12:20 AM
2 5/8 is it. I was blind. I kept looking for a even decimal number. I wasn’t thinking eighths.

There are several things going on in the photo.

1) Got a Forstner bit set. Yeah. No more spade bits.

2) Cut my first aluminum on the power miter saw to make the motor standoffs. Piece of cake. I believe aluminum has now entered my lexicon of materials.

3) Cut some rings from a copper pipe for bearing spacers. I like them more than the nylon ones that I was using. Still looking for brass spacers.

4) Finally found 5 inch 10-24 screws for the motor mounts. Had to buy a box of 100.

5) Cut my drill rod and did a bunch of testing with the RulonLR bearings. These things are fussy, but when aligned, they slide like silk. Because of their sensitivity to misalignment, I made the decision to make the entire Z axis a stand-alone unit that I can fine tune on the bench. It will then fit into a ply-skinned box as in the previous post. (People seeing this box either really liked it or referred to it as a bird house or cuckoo clock.) Hopefully the Forstners will help in drilling. If I still run into alignment problems, I plan to split the upper set into individual pillow blocks, and let the rods align them while the glue dries.

Side note: If I google RulonLR, I get a hit on the first page to this build log. Cool.

Thought I’d throw in a dxf drawing of my Z assembly.

Steve

spalm
08-14-2005, 11:06 PM
Well, designs change in the shop. I first built the new Y carriage to allow a removable Z axis. It ended up to be 8 inches across. When I finally got to see it, it was way too wide. So I went back to the drawing board and shrunk everything that I dared to. The new width will be just a hair over 6 inches. I built the new removable Z. The Forstener bits are way cool. It slides effortlessly with absolutely no binding. I love it. I ran out of time to build a new carriage, but thought I’d post pics anyway. You can see how the Z will fit inside and hang from the carriage, but I will add some side bolts to secure it.

Looking for a new router. Anyone have thoughts? PC892, PC690LRVS? I want quiet, variable speed and ½ inch collet.

Steve

spalm
08-17-2005, 11:13 PM
I completed my carriage for the reduced size Z. I’d post a photo, but it would just be more of the same, only smaller. I routed dados into the side of ½ inch ply for the sides to lock into rather that building with supports and skinning with ¼ inch ply. Saved a bit of space and parts, it is rock solid.

The boss was out today, so I got to do some extra CAD work. I saved the top level assembly as a DWF file and zipped it up. You need a viewer that you can get for free from www.autodesk.com (http://www.autodesk.com). It is quite an impressive viewer. You can of course rotate (select the Saturn tool), and zoom into the most detailed views. But what is also cool is that by selecting a part, right clicking, and then select either Hide or Transparent, that part vanishes allowing you to see what is behind it. Or click Isolate to just see that part. You also get to see all the sub-assemblies and all the stupid names that I gave every part.

Just ordered a Bosch 1617EVS. Yeah!

Hope to get some more build time,
Steve

acondit
08-18-2005, 12:07 AM
Steve,

Can you post it as a dxf or a dwg?

Thanks,
Alan

spalm
08-18-2005, 10:36 AM
Alan,

Do you really want DXF versions, or is PDF OK? Here take a look at these. These go together with the DWF file; you can see the same part names and placement. I feel kind of funny releasing data of something that I have not totally completed, but I have tried and tested all the components individually. Hope to be cutting in the next week or two.

This set of drawings is not finished, but pretty close. Three known omissions/problems are the X axis leadscrew mounting as I have mine temporarily mounted on top of the table for speed/whip testing (and yes, it does go 120ipm with minimal whip), Z axis router mounting and attachment, and bill of material work (things like callouts, bolt lengths, and manufacture’s numbers). I also cheated by drawing round nuts, goofy bearings, etc. This is my first real CAD project, so bear with me.

Let me know,
Steve

acondit
08-18-2005, 12:03 PM
Steve,

The pdf is fine. I use VectorWorks on my Mac and I can import DWG, DXF and a bunch of other formats but not DWF. In this case I just wanted to see a bit more detail than shown on the photos and the pdfs worked fine.

Thanks,
Alan

eqreservoir
08-18-2005, 12:51 PM
Great looking CNC Router table. What is the cutting area expected to be when you are finished.

Thanks. Herman

spalm
08-18-2005, 01:27 PM
Alan, I think you’re right; they don’t have a Mac version. Boo. If you get a chance to get on a Windows machine, try to give this 3D viewer a look. In my limited mechanical CAD world, I have never seen one this good.

Herman, good question; I’m not sure. I had no requirements when I started other than beefing up the JGRO design. I used those measurements as a guide. The table is 40 inches long and the gantry walls are 23 inches apart and 8 inches wide. The Z flies about 5 inches high, is 6 inches wide and 3 inches deep. Then there is the diameter of the router itself. If one can keep the leadscrew components from bonking into each other….

Side note about the router that I just ordered. I bought the Bosch 1627EVS factory reconditioned from toolking via eBay for $120. If you go directly to toolking.com the price is $160 for the same thing. Same thing happens with the new version: ebay -> toolking = $163, straight to toolking.com = $200. Shopper beware!

Steve

spalm
08-20-2005, 10:30 PM
Well, it’s alive. I had to use an extra set of MDF bearing holders on the gantry sides as the existing holes have been moved. One of the first things that I want to machine is a set of the new gantry walls, and then these holders will go away. I attached the router mounts from machine #1.

Showing a photo of the removable Z. It’s kind of neat, but I’m not sure it is really needed. I was worried about sticking linear bearings and working in such a tight space and getting the router mounted, etc. All in all, this was not a problem.

Hope to do some test cutting tomorrow.
Steve

OCNC
08-21-2005, 10:56 PM
Well, it’s alive. Did you epoxy the gas pipe seats and if so how did that work out?

spalm
08-21-2005, 11:46 PM
Hey, OCNC, glad to see you’re still watching. Thanks again for all your help. Confidence and diligence solved my drilling problems.

Nope, I have not epoxied them yet. Plan to. I have not found a local supplier for fresh epoxy. Home Depot only had Bondo style. Guess I will have to go the internet route. Thought I might try a test with Gorilla Glue or even Liquid Nails, but with the pressures involved, I bet it won’t work. Being able to remove the pipes on the shorter axis for disassembly is a real plus, but the long axis defiantly needs some sort of glue.

DieGuy
08-22-2005, 12:22 AM
How about setting up a router to mill out the oval shapes and replacing them with an aluminum v-blocks. You could contain the aluminum on 3 sides and the pipe would rest nicely in the v shape. You might only have to do it every 2 or 3 ribs.

OCNC
08-22-2005, 01:07 AM
Nope, I have not epoxied them yet. Plan to. I have not found a local supplier for fresh epoxy. Home Depot only had Bondo style. Guess I will have to go the internet route. Raka (http://www.raka.com/EpoxyPricing.html) has a sample kit that includes 12oz. of resin + hardeners and some silica (along with other items) for $18. Talk to Larry and he may be able to set up a special package that contains more silica and less of the other sample items which you won't need.
Thought I might try a test with Gorilla Glue or even Liquid Nails, but with the pressures involved, I bet it won’t work. Being able to remove the pipes on the shorter axis for disassembly is a real plus, but the long axis defiantly needs some sort of glue.I'd stay away from both the Gorrila Glue and the Liquid Nails. Gorilla Glue is gap filling but also porous and Liquid Nails takes too long to set up really hard (if it ever does). I still favor the epoxy with the semi-circular seats.

Kudos for your consistent progress and reports posted.

pminmo
08-22-2005, 09:01 AM
What about drilling and tapping the gas pipe. would allow for tweaking and removal

Jimmy Southern
08-22-2005, 10:14 AM
spalm,

I have an idea for you! If you have read many of my posts you will know that I do alot of fiberglass work in my business. So here goes take the wife or sgnificant others saran or plastic wrap. Then take one strip and coat your pipe in it. then take the bondo and liberally coat the pockets with a good bed work very quickly! Clamp the pipes in place, let dry.

Don't underestimate the strength of bondo, it just hates to be unsupported. Really good stuff! :banana:

Hope this helps!
Jimmy Southern

DON'T BE AFRAID TO THROW OUT CRAZY IDEAS SOMETIMES THEY WORK!

rweatherly
08-22-2005, 02:54 PM
What is the reason for the epoxy -- is it to create a surface that conforms better to the pipe?

truman
08-22-2005, 06:26 PM
I love your work this is very inspiring and I can't wait to get started. Thanks Truman.

spalm
08-22-2005, 10:45 PM
My 12 year old son had shoulder surgery today. All is OK. I am just exhausted. The emotional part of being a parent can be so draining.

The problem with the pipes riding in half round cutouts is twofold. Neither is a show stopper.

Problem #1 is that the diameter of the cutout is larger than the pipe. The pipe rides nicely in the bottom of the grove because of the pressure exerted by the gantry. But if enough force is applied (by wrestling the gantry), I can get it to ride up and down. V groves would probably help here, but the same problem might still be noticeable.

Problem #2 is that the pipe flexes (again ever so slightly) as the gantry moves up and down the track. With the gantry in the center, both the outside edges move out of the bottom of their cutouts. Or with the gantry at one end, the other end is loose. This is self correcting because the only spot that I really care about is where the gantry is located, and at this spot it is tight. But I still don’t like it.

So attaching the pipe to the torsion box seems necessary. I can’t figure out how to drill, tap and screw it to the MDF in an easy and strong fashion. This leads me to glue. I am really leaning towards epoxy. I used a boatload (pun) of it on a canoe that I built and really got to like it. Mixed with silica it forms a paste that can be shaped with a Popsicle stick and even sanded. You can make it look quite professional.

Thanks guys, suggestions always welcome.
Steve

(Do you think this thing would be too difficult for most people to build?)

phillby
08-22-2005, 11:53 PM
Spalm
What about good old tapping the Edges of the torsion boxes & strengthening the threads with CA glue. Then use 1/4" counter sunk screws which should be out of the way of the Bearings and holders. All this of course after conforming the rails using the cling wrap and epoxy suggested before. If you don't like the idea of threads in wood (championed by Jgro) maybe epoxy nuts into the wood. These methods should be stronger than epoxying the rail to the wood. Surely if the rail was attached in many places the movement would be stopped.

OCNC
08-23-2005, 08:32 AM
Spalm
These methods should be stronger than epoxying the rail to the wood. Surely if the rail was attached in many places the movement would be stopped.
My suggestion to use epoxy is not rooted so much in actually generating a glue bond to the gas pipe but rather to create a fully conformed seat for the pipe. That in itself should eliminate the noted problems. Using the epoxy-silica mixture could be a fifteen minute fix assuming you're familiar with it. Drilling and fastening the pipe with screws involves more work and may be over kill for this size machine. It comes down to a judgment call for Steve.

ger21
08-23-2005, 08:45 AM
One thing to be concerned about if you glue them in place, is keeping them straight while the glue cures. You say that it now only stays seated where the gantry is. How will you keep it seated evenly and straight while the glue cures?

spalm
08-23-2005, 09:34 AM
In some ways we are making a mountain out a mole hill. You literally have to bear hug the gantry to move it up and down. The sideways (or inward) flex is also minimal. Imagine the pipe bending in the middle when supported with so many ribs, it is not much. It is audible when the leadscrew is not attached and the gantry is slid up an down, there is kind of a creaking sound as the pipe readjusts.

That said my thought was to use the gantry as the clamp. Slide it to one end, epoxy the other end, and then slide it back until the glue dries. Repeat.

Steve

Bent
08-23-2005, 03:11 PM
One way to fix the rails to the ribs would be to drill a hole from the end of the rib and at the end of the hole you cut a slot in the rib and put at nut in the slot. This way you don´t have to rely on screws holding in MDF. This way is much used in cheap furniture.

Hope you understand my explanation

Bent

spalm
08-23-2005, 04:10 PM
How 'bout this? Seems to fix most of the problem. I drilled a 3/8" hole through the ends of the pipes and ran a 1/4" threaded rod through them. When tightened, I can not move the pipes at all.

Steve

spalm
09-05-2005, 10:57 PM
Isn’t that cute? The little router is routing a bracket for his big brother. I used three1/2” pieces arranged to form a mortise and tenon to lock them together. It seems quite strong. Have to wait for the glue to dry to try it out.

I noticed quite a bit of slop in the Z bearings. I never expected this. I was so worried about binding. I can not believe that it is from the bearings. It must be the drill rod is not really3/4”. I plan to order some precision rod to see if it fits a bit tighter.

Steve

ViperTX
09-05-2005, 11:06 PM
Why don't you measure it first?

spalm
09-05-2005, 11:15 PM
I know, that sounds dumb doesn’t it? I only have a caliper, and it reads just barely less than 3/4” on the rod. The drill rod comes without specs, and the bearings are rated at -0.000, +0.002 so I tend to trust the bearings more. They are loose, and I didn’t think that .002 would really be detectable with just a wiggle test. Do you have any suggestions on what else it might be?

miljnor
09-06-2005, 12:38 AM
.002 is very noticable on a wiggle test on most parts! Depending on your length of contack, ie: how long are the bearings.

I don't know anything about your bearings but -000 sounds good, but, I have never heard of a bearing with a +002 thou tol. usually they are tighter like in the .0005 range. Of course I haven't read the whole post and Iam assuming your not metric.

If your metric....never mind!

spalm
09-06-2005, 11:25 AM
The more I learn, the more I find out I don’t know. I was so worried that I would have binding that I purposely asked for bearings with a bit of slop. Never thought 2 or 3 thou would be noticeable. Evidently I was able to drill holes where I wanted them. The rods themselves do not seem to move.

Original combination was 3/4” drill rod (no spec, but I measured one today with a digital caliper at .748”) with 3/4” long McMaster RoulonLR 6362K165 plain bearings -.000 to +.002. (3’ drill rod $12.50, bearings $4.70 each = $31)

New planned combination is 3/4” precision shaft McMaster 6061K34 -.0005 to -.001 and bronze bearings McMaster 6391K263 -.001 to +.000. (12” rod $12.50, bearings $.95 each = $29)

I should be able to easily drop in both of these with no modifications. Any thoughts about this? Is bronze OK here (they sure are cheaper)? I really do not know what I am doing.

Steve

Jason Marsha
09-06-2005, 12:47 PM
If you use bronze bearings (bushings) then your two shafts must be precisely parallel. The bronze bearings will bind with for the smallest misalignment. There are UHMW bearings available from mcmaster which are more forgiving with misalignment problems, thay are however much more expensive.
Check out this thread
http://www.cnczone.com/forums/showthread.php?t=7614

Jason

miljnor
09-06-2005, 01:21 PM
I don't know how your luck is but if you try to mate a part with a -.0005 to -.001 part with a -.001 to +000 part you most likely are going to get a non-fit situation. I would personaly go with a better bushing. The bushing I have been able to get are usually +.000 to +.0005 and thats for an ID application usually you can make them a little bit tighter by putting them in an undersized hole. But at least you would have some "adjustement" however small.

Or go with the UHMV bushings and then you can play witht the fit but makeing the hole tight and loosen it up to make your tolerences wherer you want them.

Phule
09-06-2005, 01:42 PM
I realize this is a little late in the game, but has no one tried using alum. angle for the tracks instead of pipe. It seems if the angle was mounted with the leg tips down and resting on a flat surface and the 90 degree angle at the top it would be less likley to deflect than pipe. It would also be easier to mount the axis bearing carriers.

By attaching(building in) a horizontal shelf along the length of the vertical base edge, one edge of the angle would be restrained by the vertical edge of the base, and another restraining strip could be bolted(screwed) alongside the outside tip of the angle's leg to restrain it there. The longitudinal restraints would deny movement laterally, and the narrow shelf the angle rests upon could be constructed in such a manner as to eliminate deflection. I think the section modulus of 2x2x1/4 angle is probably much higher than that of 3/4" dia iron pipe.

This is the method I intend to use for the long axis on my equipment. It would be inexpensive, and less prone to deflection/movement than the pipe. Why go to the trouble of building a strong torsion box base, then using a less than optimum support system for axis travel.

The alum. angle stock can be found in many hardware stores, lumber supply stores, and sometimes craft or hobby shops. Easier to install, easier to support, easier to restrain, stronger and less prone to deflection.

This does not apply to vertical axii.

just my $.02 worth anyway.

jcb

spalm
09-06-2005, 05:04 PM
When I look at McMaster’s UHMW bearings, they are +.004 to +.015. Isn’t this going in the wrong direction? I want to get tighter, not looser. The only non-metallic bearing that McMaster carries that has a tighter tolerance than the ones I am using are the Vespel ones, but they are $67 a pop.

I don’t know how to make a slightly undersized hole with my available tools.

Miljnor, do you have another source for the types/specs you mentioned?

Should I be looking a Linear Plain Bearings (not just Plain Bearings)?

Thanks guys.
Steve

acondit
09-06-2005, 05:11 PM
jcb (Phule),

I may be missing something, but how do you propose to control the lift if you are riding on the vee lying horizontal. I believe that the reason many people are using pipe is because it offers some control over up, down, and sideways movement.

Alan

OCNC
09-06-2005, 05:52 PM
The more I learn, the more I find out I don’t know. I was so worried that I would have binding that I purposely asked for bearings with a bit of slop. Never thought 2 or 3 thou would be noticeable.According to your caliper the slop is more likely 4 thou.
Evidently I was able to drill holes where I wanted them. The rods themselves do not seem to move.

Original combination was 3/4” drill rod (no spec, but I measured one today with a digital caliper at .748”) with 3/4” long McMaster RoulonLR 6362K165 plain bearings -.000 to +.002. (3’ drill rod $12.50, bearings $4.70 each = $31)McMaster's 8888k109 is 3/4" drill rod +or- .0005. If you take the Rulon bearing and cut it along its length with a fine tooth hacksaw you should then be able to insert it into your mounting hole and use some fine thread machine screws to compress it from the side. It shouldn't take much.

New planned combination is 3/4” precision shaft McMaster 6061K34 -.0005 to -.001 and bronze bearings McMaster 6391K263 -.001 to +.000. (12” rod $12.50, bearings $.95 each = $29)If the bronze bushing is -.001 and the shaft is -.0005 it's not going to work.

Check out this page (http://www.misumiusa.com/PDFViewer.aspx?Metric=false&Page=81) on the MISUMI website regarding fits for plain bushings.There is more on page 82, 83, and 84.

spalm
09-07-2005, 11:14 PM
Well, before I was informed that worst case specs would not work, I had already ordered parts. I felt pretty stupid. But I got real lucky and the best case scenario occurred. I received my Thomson linear shafts and bronze bearings today. Wow, they are nice. I dropped them into my Z and they slide as smooth as silk. I can hardly believe it. Not a bit of wiggle now.

Thanks guys,
Steve

miljnor
09-08-2005, 11:49 AM
Thats great news, You get lucky once in awhile and murphy dosn't join the frey!. But as far as the plastic bushings go, You are suposed to make the hole that they fit into compress the plastic bushing to give the fit required by you application.

I believe when Mcmaster carr gives you a large +.010 to +.014 this is without any compression on the bushing. When you compress the bushing the hole tolerence gets alot tighter (of course you have to play with the hole size to get it just right).

This is how I used to make some fitness equipement with IGUS busihings. And they lasted forever using this method and were smooth as greased pig in a mudhole! sorry my roots are showing again! :D They were smooth! ;)

Lionclaw
09-14-2005, 02:25 PM
Steve,
Looks like you've got an awesome machine going there. I'm looking for something to replace my JGRO as well. I'd like to try to make one fashioned after your design, but I'm a bit concerned with adjustability. If I were to assemble the machine and find that the x-axis wasn't perfectly level, would there be a way to adjust it?

-Andy

spalm
09-14-2005, 02:48 PM
Hey Andy,
Hmmm, X axis not purely level…. Not exactly sure what you are asking.

I really did not want to have too many adjustments. I was never able to get my JGRO aligned even though it had a ton of adjustments. So this version basically has three.

1)XY perpendicular. This is accomplished by over-sized bolt holes at the bottom of the gantry where it attaches to the bottom torsion box. Not a lot of movement allowed, but for example by sliding the right gantry wall back and the left gantry wall forward before tightening.

2)XY (or Z) level. Not a lot of movement but I am easily able to get 1/8” inch of height adjustment with the gantry bearing adjusters on the right side. You can always skim cut the top to make it perfectly level.

3)YZ perpendicular. Again with the top bearing adjusters.

If your rails end up to not be parallel, not sure how I can help.

I can run a new set of pdf files and email them to you if you (or anyone) pm’s me with an email address. I have not totally kept them up to date, but I think they are a little closer than the set that is posted.

Steve

ger21
09-14-2005, 08:01 PM
2)XY (or Z) level. Not a lot of movement but I am easily able to get 1/8” inch of height adjustment with the gantry bearing adjusters on the right side. You can always skim cut the top to make it perfectly level.


Perhaps flat is a better term than level. You want the x axis rails perfectly parallel. You also need to have your gantry the exact same height on both sides, so Z=0 on one side is the same as Z=0 on the other side.

If any of the above are not correct, skim cutting the table will not make it flat. It will just match the table to the machine

Adler
09-15-2005, 04:17 PM
where is a good place to buy Acme lead screws and nuts?

Rance
09-16-2005, 05:51 PM
Steve,

I've finished reading your entire thread through last night. I REALLY like your torsion boxes. I also like the fir strips on the gantry verticals as well as the mortise and tenon for the router mount. However, I have the same concerns that Andy brought up. I have imagined building this machine and could probably get PDC (pretty darn close) on getting it straight but I can't see how I could 'guarantee' that it would be +-0.00? Using your jig for routing the slots for the rails is about as good as can be had. If for some reason it doesn't come out exactly right, then I don't see a good way of fine-tuning it. I'm not saying YOU need to change it, I'm saying that I'm trying to come up with a solution but just havn't found it (yet). :)

How have you measured the rails against each other to make sure they are parallel with each other (vertically; both parallel with the table)? Thanks.

Rance

scott wiggins
03-12-2006, 12:58 PM
Did you ever get a finalized set of plan's completed?
Scott

esmiller
03-12-2006, 09:57 PM
Perhaps flat is a better term than level. You want the x axis rails perfectly parallel. You also need to have your gantry the exact same height on both sides, so Z=0 on one side is the same as Z=0 on the other side.

If any of the above are not correct, skim cutting the table will not make it flat. It will just match the table to the machine


What is the difference, as long as the endmill is always the same height above the table? Am i missing something obvious, and don't see it ??

Eddie

ger21
03-12-2006, 10:18 PM
Because if the table is not flat, but the workpiece is, the cutting depth will vary, depending on table position. Unless the workpiece is flexible and can conform to the table. :)

spalm
03-13-2006, 11:26 PM
Scott,
The short answer is no, I don’t have updated plans available. I’ve moved on to build another style machine, and I think it would be in bad taste to publish my last set of drawings without a machine to back it up. This thread was a real learning experience and with a lot of help (I am still really a newbie here) I was able to discover some solutions to a few ideas that others have continued to pursue. It did perform pretty well. I cut quite a few things with it, and it got me through my spirograph project.

A few machines that are like this (and maybe better documented) are Joe’s, LionClaw’s, MadClicker’s, and of course Gerry’s.

I still fancy returning to this style of machine, and come up with a simple, straight forward design for the beginning builder that has limited tools and has no existing CNC to cut it out with.

Steve

zachjowi
09-25-2006, 05:57 PM
does that machine work well?

I really think it looks like it wouldn't be that hard for me to reproduce one like that. I want one that is big enough to cut out ski templates.

spalm
09-25-2006, 10:28 PM
Wow, blast from the past.

Yes you can build a machine like this. It really isn’t that hard. I would highly recommend a machine such as this. Several others have taken it to completion:

Joe’s 3rd
http://www.cnczone.com/forums/showthread.php?t=15139

LionClaw’s
http://www.cnczone.com/forums/showthread.php?t=14331

MadClicker’s
http://www.cnczone.com/forums/showthread.php?t=14863

And the guy who started it all ( but never did finish it :) )
http://www.cnczone.com/forums/showthread.php?t=1356

Steve

ger21
09-25-2006, 10:47 PM
And the guy who started it all ( but never did finish it :) )
http://www.cnczone.com/forums/showthread.php?t=1356



Soon. ;)

SirDave144
01-10-2007, 03:21 AM
where is a good place to buy Acme lead screws and nuts?

www.use-enco.com

If you are looking for inexpensive, that is.

blackbeard52
12-27-2007, 07:44 PM
Yep they have the 1/2-10 acme at 6 ft for only 7 bucks right now....also I found a 12 foot screw at this site for $24.00 for us dreaming of an 8 foot table. And antibacklash nuts for $15.... gettin pretty interesting.

http://www.dumpstercnc.com/leadnuts_acme.html
http://www.gocav.com/Storefront/Rod/Acmelstore_rod.htm
http://www.use-enco.com/CGI/INSRIT?PARTPG=INSRAR2&PMAKA=408-0220&PMPXNO=942694

just another newbie

Bob

pminmo
12-27-2007, 10:24 PM
From a motor perspective for long travels, I'd skip the 1/2-10 single start acme and goto 1/2-10 2 start. You get twice the travel rate for the same rpm and efficiency improves. Steppers loose power as they go up in rpm. You spend a few more bucks in the mechanical, but will easily save in the motor and electronics to match the performance.

posthocfallacy
10-10-2015, 11:30 AM
294626 (Picture taken from the original post.)

Great thread. I learned a lot from this. Great input OCNC! Thanks for the references to wood properties.

One thing that I think has been left out of the discussion is that the lead screw that is applying the moving force would ideally be placed at the Center of Gravity of the moving assembly. The X direction uses two leadscrews so that that effective force is near the middle of the gantry, but the other motions have one leadscrew. The Z lead screw is not too far off where I think the CG would be. If you don't have a CAD program like SolidWorks (I'm lucky enough to have a Home Use Licence Agreement (HULA) that allows me to legally use SW at home using the license that we have at work) the center of gravity is pretty difficult to find exactly. On the other hand, we know the gantry CG is (horizontally) somewhere between the rails and the motor. The motor moves the CG down some, so we know that the center of gravity is somewhat lower (vertically) than midway between the two rails and on the motor side by some amount.. The lead screw mounted on the opposite side puts a lot of twist into the Y when it moves. Seems that it should have been mounted on the motor side of the Z instead of opposite. On the other hand, we know from reading to the end of the thread that the thing works. You can't argue with success. The Y gantry must indeed be very stiff, and the gantry bearing spacing must be large.

Side point: I've forgotten who suggested the aluminum angle as rails, it is one to be considered only if steel angle is used. Aluminum is too soft to last very long. We also know from ger21's experience that we don't want any zinc plated steel.

One other thing that we have learned from this thread is that lead screws become unstable with length, regardless of diameter. I had an experience building a robot in 1984. The engineer in charge wanted to use a huge lead screw, 1½" in diameter and ball screws with many leads (maybe as high as 5 starts). The problem was that the length of travel needed was five feet, so the lead screw had to be that long as well. They also wanted some fast motion. I have forgotten how many leads it had, but many more than two. It was my job to build motions to test the mechanisms to find the failure points. The whip in the lead screw, even with the 1½" diameter beat the hardened balls in the ball screw into hexagon shaped affairs in a couple of hours. OCNC can probably tell us how long a lead screw can go without destroying itself and the nut(s) it rides on. The final product replaced the lead screws with kevlar reinforced toothed timing belts, the same as those used on the newer Harley-Davidson Motorcycle drives. Almost zero stretch and stronger than most chains.

I know this thread is 10 years old right now, but it is still a good one for a learning experience. Thanks CNCZone for keeping this thread around.