Fixed Gantry - First Build

[Schweeb]

So, since I'm new to these forums, I'll begin by introducing myself. My nickname is Schweeb, which is what I like to go by most of the time. I currently draw, program, set up, and run parts in a VM1 Hurco for a small JCI supplier named Alex Products. About the only thing I actually don't do is design the parts. I only have one year of work experience under my belt, but I'm a quick learner.
I am familiar with at least five CAD/CAM programs and proficient in at least two. I have formal schooling in Rhinoceros and Autocad, but I've probably spent the most time in SolidWorks, my personal favorite. My CAM program of choice is Mastercam since that is what I use every day at work.
Despite my profession, I'm actually not very mechanically inclined. I have failed several home projects already due to lack of planning and failed attempts at improvisation. My skills mostly revolve around math and physics; although, I'm capable of quickly learning anything I might need.

With the introduction out of the way, I'll get down to the machine. This mill is designed to be extremely cheap and an introductory course into home built machines. I do not expect to be able to mill steel of any kind, nor do I expect to do much with aluminum other than engraving. My target materials are engraving Lexon or milling plastics.

My current design has a work area of around 30"x36"x14", and it will consist primarily of 1" square tubing welded together. I currently have the rails as 1/2" round stock, but it's well within my budget to go to 3/4". The threads will be 3/8-16 UNC B7 threaded rod. This was virtually my only choice due to cost concerns. I will be using nuts and clamps from DumpsterCNC and Oilite bearings. I have already received my driving kit from Probotix so that's the one component that's not up for changes. I'm using 180 oz in. stepper motors, and I will more than likely using a dremel or a really cheap router with a simple rheostat. The dremel will be thrown in there first to see what I can get away with.

My primary areas of concern are the rails, Oilite bearings, and the size of the machine compared to the motors.
The rails are a huge concern due to their length. The X and Y rails are currently ~4' in length so it might come down to just reducing the size of the machine to relieve the flex in them.
I had never heard of Oilite bearings before stumbling across a thread here. I was struggling in trying to bring the price down when I came across microcarve's design. I honestly have no idea if it's even viable to use them on larger machines. My big question here is what size I could get away with using these things, as linear bearings are quite expensive.
The motors are already bought so they are my absolute largest limiting factor. I honestly doubt they will move such a massive machine which is my primary reason for asking for advice here. Is it possible to drive a machine with over a 2'x2' work area using these tiny motors?

Any comments or concerns are welcome. I have never done anything like this before so it has been quite overwhelming. My biggest fear is that I'll be forced to limit my work area to a 1'x1' area. I'd like as big of a work area as possible.

Eventually, I will build a larger machine, but for now I don't have enough spare money to indulge in this. Most of this is being paid for by my tax return
Some other notes: I'm not interested in buying machine. My total budget is under $1k. I will not use acme threads or ball linear bearings for this machine due to cost. My minimum work area for this to be worth it is 10"x10"x4" with a table around 16"x16".

[louieatienza]

I think your motor choice is fine; you probably could have even gone with a smaller stepper, since your thread pitch gives you a mechanical advantage.

I'm not sure if 1/2" or even 3/4" unsupported rods over that span; if you want 24" cutting area you have to add the spread of your bearings, plus "limit" area on each side. On your table, cantilevering the table over the bearings that far could cause your cut quality to suffer in those areas. It might be a good idea to use some MDF also to "skin" the frame, which should stiffen it up some.

[DonFrambach]

I think you are likely to have problems with the relatively long unsupported rails in your planned build. Can you get by with a smaller router? Have you looked at MicroCarve's thread? He is successfully using unsupported rails but on a much smaller router.

Here's a linke to MicroCarve's thread: http://www.cnczone.com/forums/diy-cn...p_machine.html

[Phife]

My suggestion.. do a lot more reading..

You machine is probably not going to be of much use in machining plastics, you may be able to do some light woodwork with it.. but 1" square tube frame is not rigid enough and the 1/2" or even 3/4" unsupported rods are going to be extremely flexible..

The whole idea of going fixed gantry is so that you can beef up the gantry and make it more rigid, you are doing the opposite.

Use atleast 1.5" 1/8" wall square tube or larger, and if you must use unsupported rails you should go with 1" or larger, although if it was me I wouldnt even dream of using unsupported rails.

Machining plastic is actually harder and machining wood.

Also, the 3/8" screws are probably not going to work well either, you want at least 1/2" for those spans.

Oh and 14" Z is just crazy for this machine..

If you are trying to keep it under $1000 then I suggest making a much much smaller machine. 12x12 or less. I hope the $1000 doesnt include your electronics and software!

Put off spending any more money.. you will just be wasting it, and spend the next several weeks reading this forum and seeing what works and what doesnt..

Sorry, but I have to say, machines like this hurt my head..

[ger21]

I've seen a machine with 1/2" round shafts deflect nearly 1" with pressure from 1 finger.

[Schweeb]

Software is not an issue. I have access to a multitude of CAD/CAM programs through work and friends. EMC2 is free. Electronics are already purchased and cost less than $300.

I figured the Z was a bit overkill, but I was attempting to get a work area similar to my VM1. I can easily knock the Z down to ~6" or less, which would decrease the weight on the rails considerably.

What would you say the longest span I could get away with using 1" unsupported rails? They will be pre-hardened 4140 ground and polished round stock. I've worked with 12' pieces of cold rolled so I do know how much they bend. I'm just not sure how much I can get away with on rails with 4140.

If I drop the work area down to ~16x24x4, it would reduce the span to 2.5-3', and it would allow me more money to spend beefing up the rails. I'm not actually sure how to beef up the gantry. Most of the fixed gantry machines I've found are too small to worry about it. Reducing the Z, however, should help eliminate that problem.

As for reading the forums, I've been doing that for many months. I started doing research over 2 years ago when I was in college. I think it's time for action.
I have actually read microcarve's thread. All of it(that's a lot of pages >.<). It's actually where I got a lot of my ideas from. I'm just pushing the boundaries to see what I can get away with. The bigger the work area, the more things I'll be able to use it for. A small machine doesn't have that many uses for me.

Sorry if there are any grammatical errors or incomplete thougths in this post. My posting time was cut short.

[microcarve]

I have actually read microcarve's thread. All of it(that's a lot of pages >.<). It's actually where I got a lot of my ideas from. I'm just pushing the boundaries to see what I can get away with. The bigger the work area, the more things I'll be able to use it for. A small machine doesn't have that many uses for me.

Yep, that's a lot of pages. Thanks for reading...

I've been pushing the boundaries of what works with plain round
rails for a long time. With lots of different designs...as can
be seen in the long thread.

I like round rails for a number of reasons....but IMO...they need to
be something like Thomson or other high quality case hardened shafts.
Plain/tool/drill rod steel, rather or not it's ground, hasn't been as
reliable as the better quality case hardened shafting.

It has on occasion, but it's not usually as round as it looks. It's been
too hit & miss to justify taking a chance for what it costs to ship it back
or replace it all over again. Precision shafts are pretty inexpensive for
what they are. Precision. Really!...

Though the math may say otherwise, I think the case hardening adds to the
stiffness of them. They, at least, seem to me to be a bit less flexible than
plain steel. Maybe not....?

Matched with the inexpensive Oilite bearings, that's a fully qualified precision
linear solution to the problem of what to use for a very cost effective DIY
machine.

There are problems though....

They shouldn't be too long without being a certain thickness for the length.
And, it has ev-er-y-thing to do with what the intended use of the machine
is. With a Dremel cutting foam....(no kidding...there are lots of machines
that, that's the intended use) you can get away with fairly long spans of
relatively thin shafts. But for wood and plastics, there's some considerable
resistance to cutting, so the rails can flex too much if they're not a good
choice of size.

So, how to tell.

I'll use a 24" stretch of 1/2" drill rod for this example.
Prop the rail/shaft on ends of two pieces of, say, 3/4" MDF.





Now, I was quite surprised here. I weighed the chunk of steel fully expecting
over 16 lbs to easily flex the drill rod by at least 1/4". The piece of MDF
in the center is 1/2" so with the drill rod propped on 3/4"...the difference
is the 1/4".


http://www.microcarve.com/zone17/deflection/


But it didn't. The weight scale has a 70 lb capacity. I can easily
press that down to it's max. But to force the drill rod to flex the 1/4" and
touch the 1/2" piece of MDF took a lot of force....so I couldn't hold the
camera easily enough to make a video. I'd guess something around 90-100
lbs to flex it that far.

I was very surprised. I thought it'd surely be more flexible. So for foam
cutting that'd work fine. For wood/plastics...no, as it's a long way to
1/4". Even 1/16" (.0625) flex is too much for a cnc machine.


Now, here's an interesting side note.

Lots of people use anti-backlash nuts. Those are generally only rated for
about 10 lbs load capacity. They go up to 350...but those aren't anti-backlash.
Most of us will end up with 25 lb max.....unless someone uses 5/8" screws and pays
about $60 for a nut.

So, how's all that figured...I have no clue. But it'd seem to me that the forces
of driving the router around are somewhere in the 20-40 lbs of force range.
Based on what the people who make these things they sell us use for whatever
ratings they have.

So, 20-40 lbs of force on a shaft seems fair...and from lots of practical
experience with them...it does seem about right. If it flexes an UN-accepteble
amount with that force, then the shaft size needs to go up...or the length
go down.

My thread shows a lot of playing around with sizing up and down in very small
amounts to arrive at what works best for what the machine is intended to do.
A little bigger is fine, but for the long term, stability, reliability, and
mistakes and some outright abuse, I went to a bit smaller lengths and they're
much less prone to flexing, even with use of considerable force.

Mine are 18" and 16" 5/8 shafts. But they are supported correctly. And that
matters a whole lot. And the bearings are spaced closer to ideal.

I'll end it here for now, but I hope there may be something useful in that.
Round shafts can make making an affordable and very high quality machine simple.
It can be very high precision, and very reliable. The thing to be considered
most is, what's an acceptable amount of flexing? They will flex, but the force
to cause that may be very well within acceptable limits for thousands of uses.

You'll need much thicker rails, or a much smaller machine...unless you want
to mill strictly foam...



John

[jsheerin]

John, (lots of John's here it seems...), to address some of the design questions you raise, there are generally two issues to consider. One is the load rating of components. This usually relates to component life. In other words, if you run a nut rated for a 10lb load at 100lb, it's going to wear out pretty fast. Linear bearings are the same way - THK has a lot of data in their catalog to let you calculate the expected life of a bearing given various loads on it. The other issue is stiffness. You can look at stiffness on the component or machine level. The components all contribute to the overall stiffness of the machine which is what ultimately matters. It's really the stiffness between the material you're cutting and the cutting tool. Stiffness is force divided by deflection, so in your example with the bar, you're measuring the information needed to calculate that. There are ranges of stiffness that let a machine cut certain materials well, and it's related to the modulus of elasticity of the material. Higher modulus of elasticity materials (like steel, for example) require a higher stiffness machine to cut them well. I've posted specific numbers on the forum previously if you're interested.

[Schweeb]

Thank you very much, microcarve. That post actually helps me a ton.

I actually used a similar method on a cold rolled bar last night. I found a bar of 1" and pressed down on it to see how much it'd flex under various loads. Since 4140 is about twice as hard as 1018, I can assume it will have around 3/4 the flex. Based on that, I know that 4ft would just have a little too much flex for milling harder plastics. I might be able to get away with 3' of 1".
If I can find cheap case hardened rod, I would love to go with it, but I was under the assumption that's it's moderately expensive. I can get 4140 HT TGP for dirt cheap from a few places. I've worked with 4140 a lot so I'm pretty familiar with it.

I'll try a more in depth test when I actually get my rails. I'll actually use a height gage when that comes around. That will give me definite confirmation of what I can get away with.

My planned max load force was ~50lb (including table weight). Generally, it will be much less. A peak of 75 on plunges would be pretty reasonable with the slow movement of the threads I chose. My co-workers gave an overall thumbs up for 3' of 1" for cases like these. Of course, you always want to build it for 150% of what you expect, so I'm still a little bit concerned.

I'll probably drop the X travel as small as I can without limiting work area too much. I'll get back to this thread after I've redesigned it.

[microcarve]

My planned max load force was ~50lb (including table weight). Generally, it will be much less. A peak of 75 on plunges would be pretty reasonable with the slow movement of the threads I chose. My co-workers gave an overall thumbs up for 3' of 1" for cases like these. Of course, you always want to build it for 150% of what you expect, so I'm still a little bit concerned.

I'll probably drop the X travel as small as I can without limiting work area too much. I'll get back to this thread after I've redesigned it.

I use 36" lengths of 1" Thomson shaft on this quick-built machine
that I use for cutting all my other MDF machine parts...

My Newest Desktop machine

That's what it's intended purpose is and it cuts as fine as any machine
I've ever built. Really, I doubt the quality of the cut could be any
better. Basically no sanding more than a quick brush with a sanding
sponge.

I run it at about 90IPM with a 1/2" 2 start screw. 270 oz motors at
36 VDC with something similar to a Hobbycnc board...(Pminmo's Halo board
...which I like a lot...

It cuts about 15 x 24

It's not pretty, but it's not supposed to be. It's just supposed to do
the job and it does it very well....

What I don't do ever on any of my machines is plunge so fast that that
alone causes flexing and bending. And that's pretty easy on machines that
use powerful leveraging screws on the Z axis.

Just from the experience of using these round rails...I may limit the
1" shaft to 30" length max if it were used on the gantry for cutting wood.
2-3" of that would be end supports. I've gotten a feel of what works well
from using thousands of the shafts over time. Just a few inches can matter
pretty often.

They do work great and are highly reliable when used within their
limits.


John

[microcarve]

John, (lots of John's here it seems...), to address some of the design questions you raise, there are generally two issues to consider. One is the load rating of components. This usually relates to component life. In other words, if you run a nut rated for a 10lb load at 100lb, it's going to wear out pretty fast. Linear bearings are the same way - THK has a lot of data in their catalog to let you calculate the expected life of a bearing given various loads on it. The other issue is stiffness. You can look at stiffness on the component or machine level. The components all contribute to the overall stiffness of the machine which is what ultimately matters. It's really the stiffness between the material you're cutting and the cutting tool. Stiffness is force divided by deflection, so in your example with the bar, you're measuring the information needed to calculate that. There are ranges of stiffness that let a machine cut certain materials well, and it's related to the modulus of elasticity of the material. Higher modulus of elasticity materials (like steel, for example) require a higher stiffness machine to cut them well. I've posted specific numbers on the forum previously if you're interested.

Thanks!

Yes, I do follow your posts...

I'm not great with the calculation side of things, but I try. And the
stuff you post is some of the best explained and easiest to follow along
with.


John

[microcarve]

I actually used a similar method on a cold rolled bar last night. I found a bar of 1" and pressed down on it to see how much it'd flex under various loads. Since 4140 is about twice as hard as 1018, I can assume it will have around 3/4 the flex. Based on that, I know that 4ft would just have a little too much flex for milling harder plastics. I might be able to get away with 3' of 1".
If I can find cheap case hardened rod, I would love to go with it, but I was under the assumption that's it's moderately expensive. I can get 4140 HT TGP for dirt cheap from a few places. I've worked with 4140 a lot so I'm pretty familiar with it.

Ah, there's another calculation I have no clue about....

But.... When I've used other than case hardened shafting in the past.
If...I didn't have to sand it for hours to get the oilites to slide perfectly
enough....and that happened very often....

I noticed that a fair change in temp would be the cause of binding
sometimes. I can't attribute it to anything else but temp changes.
No other reason I could imagine.

So, I don't know if the crystalline structure of the case hardening
makes it less likely to do that. So far it has been much more
reliable to the point where I've rarely noticed any difference in
travel due to the shop temp variances....(From say, 45 to 85 F)

The oil in the bearings can thicken a bit, but a few manual jogs
before cutting anything seems to distribute the oil and things proceed
from then on to work.

A 36" shaft 1" of Thomson is about $40. And is nearly guaranteed to
be problem free. I just had too many problems with other than the
Thomson shafts.

(That, of course, does depend on how it's worked into a design....


John