Need help - Fixed gantry design

[boltzzz]

Hi guys.

I'm in the process of conceptual design of a fixed gantry CNC router and would like your feedback.

Firstly, my goals. What I want to machine capable of. Some are optimistic I know.
+/-0.005" machining capability.
Machine up to and including aluminium.
To be used as a desktop machine in an apartment.... (will try fully enclosing machine with soundproofing).

It will be an all aluminium design using 5083 plate and 6061 T6 angles/channels bolts together and machined surfaces to mount rails.
Ideally I would've liked to have used rectangular hollow sections, but the maximum wall thickness I could find locally was 0.25" which seemed a bit thin for my liking.
The gantry is made of 2 plates bolted to 2 c channels to create a box section. I've hidden some parts in one of the views so you can see.

All will be bolted together with interference fit bolt and nut (no thread tapping).

Hiwin HGR20 rails and HGH20 bearings on all axis.
1605 ballscrews on x and y, 1205 ballscrew on z.
2.2KW water cooled router spindle

Cutting area: 12" x 24"
Cutting height: 3.5"

Overall machine dimensions: 25" (L) x 33" (W) x 22" (H)

At a guess, this machine will cost me around $5k AUD.
I will have to find a local shop to do the machining for me which might cost a bit.

Would love any feedback from previous experiences.

Similar Threads:

• Newbie- Fixed Gantry Design
• Build Thread- Hybrid Design - Fixed Gantry Router for High Speed Aluminum Milling - Metal Nibbler 5
• Fixed Gantry Design Question
• Newbie- Small Fixed Gantry Design
• New Machine Build- 8020 Fixed Gantry Design

[handlewanker]

Hi....looks good........I like moving table designs over moving gantry types.

Couple of variations I'd have........mount the Y axis rails on strips on top of the base plate not under it as you show....this makes it simpler to build and cuts down on machining.......also you don't want a swarf trap with the slotted holes etc.

You could also make the X axis saddle longer (higher) to get a longer slide as opposed to a wider one and also make the Z axis slide narrower and longer to give you more support when the Z axis is all the way down......you can go above the X axis crossbeam with the Z axis slide length as it's in the open air.

Before you start scaling the design prior to a build, project the spindle assembly package onto the design and Z axis slide to make sure it won't run out of slide way.

I would also make the base as a separate box with one channel cross member underneath the middle and have the shortened sides bolted to the box sides and have the X axis crossbeam as a rectangular box bolted between the side plates at the top..

By having separate modules bolted together you can align and adjust them in the final assembly and correct any out of squareness you might get..

If you bolt the linear rails onto strips on the base and cross beam you can correct any flatness by careful filing where necessary.....also this will give you height clearance for the ball screw nut under the table and under the X axis saddle.

I would also make the table from a slab of aluminium and have tapped holes instead of the Tee slot extrusions.....just my preference.

BTW....you don't need to machine the linear rail mounting strips....just use the linear rails themselves as straight edges to reveal the high spots and work on them with a flat file etc.....this will cut down any outside machining costs as you can DIY it quite easily.
Ian.

[boltzzz]

Thanks for your input.

So when you say mount the Y rails (moving table?) on the base plate, you mean like wade o's design below?
DIY CNC Router Build; Fixed Gantry, Steel - Wade'O Design

The concept I was using was from the link below. The plate with slots in it just below the moving table is only thin, to catch the swarf and stop it falling on the rails and ballscrew. With Wade O's design, it was all open and I see he later made improvements to cover the rails/ballscrew from swarf but I wasn't a big fan of it. I was going to install brushes or a thin rubber cover in the slot to try and prevent sward dropping in on the rails.
MAKERDREAMS |HEAVY DUTY DESKTOP CNC MILL AND HIGH QUALITY 3D PRINTER EVO-ONE Desktop CNC milling machine

Thanks for the input on the Z plate. I'll make it slightly narrower and taller to increase the distance between the bearings in the z direction.

I was going to go a slab of aluminium for the table. Never really been a big fan of the T slot extrusions.

I see you're from Dandenong. Can you recommend any local places for aluminium and machining? I've been to Aluminium Trade Centre, Ullrich and Capral.

Know of any pros/cons for ATP-5 plate vs 5083 plate? Know of any Melbourne suppliers of ATP-5 plate?

[handlewanker]

Hi......sorry, can't help on the supply.......those you mention probably would have stock of various stuff, but I'm only messing with bit stuff and not large amounts of ally for projects or production etc.

I'm not really a big fan of anything made from aluminium as it's a difficult metal to work with....that is, joining it up to other bits and pieces along with it's relative softness etc.

I'm mainly orientated to steel for a build.....one that actually I started but put on the back burner due to the arrival at last of a mill I ordered from China....long story.

Steel is extremely forgiving in that you can weld it, bend it, heat it, grind it and cut it with Plasma without too much drama........just my opinion.

Aluminium has the property that if you go to a weld option the expansion capability will throw all semblance of accuracy and squareness in the too hard basket, so the only option is to bolt together and doing a 90 degree join is not easy with bolts or screws.

The cost of steel, either tubing or plate, is far less than any aluminium sheet or extrusion, so that is how I think when machinery design comes up.

The design I started was for a 3025 CNC moving table router to be able to mill steel, and the steel tubing for the complete frame components cost $150....that was for 2 machines as it's cheaper to buy complete lengths and cut them to your sizes at home.

This is my opinion pure and simple, doing a build will be what you see as an ideal machine with the tools, money and skill available, and not all roads lead to Rome.....LOL.
Ian.

[handlewanker]

Thanks for your input.

So when you say mount the Y rails (moving table?) on the base plate, you mean like wade o's design below?
DIY CNC Router Build; Fixed Gantry, Steel - Wade'O Design

The concept I was using was from the link below. The plate with slots in it just below the moving table is only thin, to catch the swarf and stop it falling on the rails and ballscrew. With Wade O's design, it was all open and I see he later made improvements to cover the rails/ballscrew from swarf but I wasn't a big fan of it. I was going to install brushes or a thin rubber cover in the slot to try and prevent sward dropping in on the rails.
MAKERDREAMS |HEAVY DUTY DESKTOP CNC MILL AND HIGH QUALITY 3D PRINTER EVO-ONE Desktop CNC milling machine

Thanks for the input on the Z plate. I'll make it slightly narrower and taller to increase the distance between the bearings in the z direction.

I was going to go a slab of aluminium for the table. Never really been a big fan of the T slot extrusions.

I see you're from Dandenong. Can you recommend any local places for aluminium and machining? I've been to Aluminium Trade Centre, Ullrich and Capral.

Know of any pros/cons for ATP-5 plate vs 5083 plate? Know of any Melbourne suppliers of ATP-5 plate?

Hi, something like that, but without the base plate being attached to a bulky table.

I envisage a box base with the side frames bolted to it and the rails bolted to the top of the box on strips.

There are umpteen different configurations to apply 3 axes to one another....you just have to apply your desire to how you want to build it and what you think will work.....no two people agree on the absolute ideal configuration......not even with a commercial product.

One word of advice........don't over engineer the specifications for something that will only ever be a hobby project, it would be another matter if it was to be a money earner......then you'd be a fool to go to war with a home made cannon.
Ian.

[NIC 77]

Cutting height: 3.5".

I'm not sure what that means. Does that mean gantry clearance or Z movement? Are you going to put a vice on this at any time? You mentioned you want to cut aluminum. How you will do your work holding is important to consider in the design stage.

You'll want to match your motors for the pitch of ballscrew. 10mm lead may be a better choice depending on your motors and the speeds you want. Why use 12mm diameter on the Z? Use 16mm minimum for everything, I say.

So that is a 1.5" bit cutting a 1.5" thick piece with 0.5" left over? So with a 2" bit you couldn't cut a 2" thick piece that is flat on the table?

Like I said, I'm not sure what what cutting height means in this instance, but if that is the case I think you'd want more. With a fixed gantry, I have to think you could make the height of the gantry adjustable (with a couple different bolt locations or spacers) and add some extra inches of Z movement. If you put the Z rails on the part that goes up and down, you can keep the two lower bearings at the lowest spots. That's what I'd do. Also, flat plates bend easily. At least add some angle iron to the sides of your Z (the part that goes up and down)

I do like the concept of what you've come up with. I think many of the design options you are exploring have merit. At the end of the day, it will be easier to use a solid gantry tube. If you are concerned about the wall thickness, you can just weld some flat bar where you will mount the rails. If you can't weld, find someone who can, not alot of work to put a couple of 1.5" by 0.5" pieces of flatbar onto your gantry tube, either steel or aluminum.

For a fixed gantry, your weight limit is what you and your friends can move into your apartment, and what doesn't pose a risk of falling through the floor. So for a small machine like this, I don't think a steel 6" x 6" x 3/8" square gantry tube is out of the question, you can probably find something at the scrap yard for almost nothing. If you're handy with an angle grinder, and have alot of patience you can even use a sandpaper flapper disk to make your mounting surfaces flat. Of course, that makes a big mess, so, do you have somewhere you can work on this? Barring that, there's epoxy. Do you have a drill press, welder, angle grinder?

BTW....you don't need to machine the linear rail mounting strips....just use the linear rails themselves as straight edges to reveal the high spots and work on them with a flat file etc.....this will cut down any outside machining costs as you can DIY it quite easily.
Ian.

I'm going to agree and disagree with this. You may be able to use a file if you are working with aluminum. It could take you until doomsday if working with steel, depending on how far you are off. I used an angle grinder with a flapper disk on my build and a piece of extremely flat 2" x 3" solid aluminum that I had handy and some cosmetics to find the high spots. I disagree that you should use the linear rails as straight edges. Use a straight edge or a piece of metal you know to be straight by measuring with a straight edge. The rails become flat when you mount them to a flat surface and the flatness can be off a small amount prior to mounting.

By having separate modules bolted together you can align and adjust them in the final assembly and correct any out of squareness you might get..
.

I totally agree with this statement.


It sounds like you want something that you can get cut out of a sheet and assemble easily in your apartment with minimal tools. Not a bad idea, I just don't know the best way to accomplish it. I can only think of ways that I would do it, and I would use square steel tube, probably 6" wide for the whole thing with 1/4 to 3/8 wall, and it would have welds and bolts holding it together. I know I could get that for 30 cents / lb at the scrap yard near where I live, and build it with the tools I have. Not saying your idea is bad though, and after a couple design revisions you may come up with something that is easy to assembly and doesn't take forever to build. My stuff usually takes a long time.

[handlewanker]

Hi....on the topic of a straight edge.....two flat faces placed on each other will reveal any bowing in one or the other.....but one can be concave and the other convex and both appear to be flat ....it takes a third one to indicate the other two's irregularities.

I would use the flat way on a lathe bed or a mill table to test a DIY straight edge......or the back side of the linear rail to see if it's good enough for one.....when the Devil drives then needs must be.

I used a 600mm Starret combination set ruler as a straight edge because when I tested it on a granite inspection table there was less than a smidgeon (.001") over the total length.

Rendering a steel plate flat is not an art.....going in from the mill scaled surface you use an angle grinder to lightly grind the surface in a cross hatch pattern until you get bare steel, using chalk to indicate the high spots.

It worked for me over 60 years ago when I was an apprentice on the mines and far away from the workshop.

The question is what design and what material.......one goes with the other in various combinations, and it doesn't pay to be completely dominated by other people's preferences.
Ian.

[NIC 77]

I would use the flat way on a lathe bed or a mill table to test a DIY straight edge......or the back side of the linear rail to see if it's good enough for one.....when the Devil drives then needs must be.

I used an actual long straight edge to check the piece of aluminum I used. I've used the straight edge and aluminum against enough things to know that they are dead on. You can't really rub cosmetics against something using a straight edge, which is why something else is needed.

In my experience, you can't use the back side of a linear rail because they aren't flat enough until they are bolted down, and if you are using it to rub for high spots, they can flex.

using chalk to indicate the high spots.

I never tried chalk. I used cosmetics and wax paper. If chalk works, that would be a better option as you wouldn't get as much debris sticking to it as you do with the cosmetics, which have to be wiped off in between applications for that reason. If I do it again, I'll have to try some chalk.

The question is what design and what material.......one goes with the other in various combinations, and it doesn't pay to be completely dominated by other people's preferences.
Ian.

Agreed. Hopefully boltzzz will come up with something awesome that I would never think of. Not sure how he's going to bolt things together, and I think this will take a fair amount of refinement to the design to get it right, but it's a good start. Nothing wrong with considering other options and opinions too.

I have nothing more to add to this. Good luck boltzzz.

[Bubba]

Hi,
A few years ago, I also built an aluminum framed fixed gantry machine. Like you, I wanted bolted connections (main reason was that I had no TIG!) so instead of just bolting, I also "glued" it with Loctite product E-20NS.

I finished fabrication back in 2011 and it is still glued together to this day! I use it mainly for etching PCBs, but have also done some plastics, Aluminum, and brass on it.

Good luck to you in this project.

[handlewanker]

Hi, using the chalk to reveal high spots is just a rough and ready way to get steel plate with mill scale on it to a reasonably flat surface so that you can bolt something down on it.....the chalk stays white in the hollows and turns shiny black when a high spot is rubbed.

If you're careful and don't dig the angle grinder disc in you can get really flat as the disc only wipes the surface when presented lightly at an angle.......don't use a new disk....older partly worn discs work better or round the corner off on a brick.

I used this method mainly for outside in the field fabrication and construction.

I also used it on the rough back of my lathe headstock casting to give a flat surface to mount the belt drive countershaft where there previously wasn't one.
Ian.

[boltzzz]

Thanks guys,

So that is a 1.5" bit cutting a 1.5" thick piece with 0.5" left over? So with a 2" bit you couldn't cut a 2" thick piece that is flat on the table?

Z movement of 3.5". As the router can be moved up and down on the z plate, this can allow for different length cutters and still get the full 3.5" height in Z.
A question, what is the standard length of a cutter?

If you put the Z rails on the part that goes up and down, you can keep the two lower bearings at the lowest spots

It's the simple things that sometimes can make a big difference. I can't believe I couldn't think of it!
A lot of other designs I see still have the rails mounted on the Y plate instead of the Z plate. Do you know why that is?

I do have to have a bit more of a think of how to join the aluminium together.

Looking at various different designs, sometimes I think it will be easier to make a strong moving gantry design. I'll keep on improving my design.

[RCaffin]

pros/cons for ATP-5 plate vs 5083 plate
ATP-5 is slightly modified 5083. There is not a lot of difference.

Cheers
Roger

[boltzzz]

Okay. All great input!

So there were a few reasons why I was leaning towards aluminium. It's what I'm comfortable working with and lighter. I have very little experience with steel and no experience welding.

Even though mass is a good thing for a CNC router, I had to keep in mind that this machine will have to be moved a few times.

When refining my aluminium design, I ran into a few issues.
No large RHS profiles available locally meant built up structures.
Built-up structures meant increased build time, complexity and probably increased built in errors.
No easy way to join parts together.

I did a few build comparisons for the gantry structure between aluminium and steel. I knew aluminium was going to be more, but from locally sourced suppliers, the cost of the aluminium design was 3-6 times more expensive!

So I've decided to leave the aluminium design and go with steel. I know long term it's the right decision.

Here is more first iteration of my steel design.
You'll notice it's very similar to Wade'O Design, which has been of great help.

I've tried to keep weight down while improving rigidity and simplicity.
There are 3 basic parts.
The gantry which is 200x100x6x900mm RHS and weighs 24kg.
The two uprights weighing 3kg each.
The base which is made up of 125x75x6mmx900mm RHS which the uprights sit on, then 2x 100x100x6x600mm SHS notched out and welded to the 125x75 for the X rails, then 2x 65x35x4x300mm RHS cross members between the 100x100. All this comes to 35kg

Having them built in 3 sub-assemblies will make it easier to move around. It will also enable the parts to be adjusted for alignment.

I would have like to use a 250x150x9mm RHS for the gantry, similar to Wade'O Design to increase the distance between the Y axis rails but it comes in at 47kg and realistically, I'm only going to be milling aluminium on the odd occasion. I could go a 250x150x6mm RHS which will bring the gantry weight up to 32kg, but by the time you add all the rails, bearing blocks, Y & Z plates and router, it will be north of 40kg.
Using the 200x100x6mm for the gantry spaces the y rails 130mm apart at the centerline.

As I have no experience or tools for welding, I was looking at getting the base welded and stress relieved but I am yet to get this priced.

All the interfaces between the sub-assemblies will be machined as well as the surfaces for the rails. I'll get holes drilled at 150mm increments or so dowels can be installed to assist with mounting the rails straight.

What do you guys think?

[handlewanker]

OK, so you're going to cut out 2 sections from the 2 base tubes.....best of luck as you're working with steel........and weld in a length of tubing to mount the uprights on......make the cut outs shallow as opposed to almost all the way in, about 1/3 rd the tube width and do the same for the tube at the bottom end......this will give you a level tube base each end.

I would also add another cross piece between the inside tubes half way down the base frame to stiffen up the base a bit.

That will give you a strong base, but make the uprights from the same base tube section and bolt them onto the back of the X axis cross beam and the base cross tube.....you'll need to weld steel plates to increase the thickness to tap hole in at the bolting points

Being an all steel build now, the forces that affect the uprights will not be large enough to cause deflection, unless you get ambitious with a large cutter on a piece of steel etc.

I say this, as it would be normal to make the side supports with a larger section longitudinally to give some stability in the fore and aft direction, but mostly overkill with a modest build.

This design would need to be bolted together in 4 pieces to allow machining or hand working the various surfaces that need to be attached together, but it does allow you to make the structure firm and square and correct any misalignment that ca occur.

DON'T consider stress relieving the base part once it's completed....mostly it won't be necessary as you won't be cutting into any of the sections once welded, and drilling and tapping won't cause the structure to distort or re-align.

Bear in mind that this is an alternative to an aluminium construction and so won't need to be super strong to resist the forces that you would apply to an aluminium framework......thinner tubing can therefor be used....2mm wall will be quite OK, easy to work with and as strong as you'll need.

Design with the same tube section size all round so that you can buy one section length and get it cut to your requirements......you can work with a hand hack saw mostly if the tubing is thinner walled.......2mm wall will be easy to gasless Mig weld.

You could consider making the X axis cross beam as a composite from two smaller tubes welded to a flat steel plate to give a channel section......this will give you 2 tube surfaces to bolt the rails onto and the gap between to give a gully to clear the X axis ball nut.

With all the frame sections welded, the next stage will be the surface finishing for the mounting of each section.
Ian. .

[RCaffin]

Gantry uprights look extremely thin to me. Can you double them?
Angle inserts needed between X and Y axes (so to speak).

Cheers
Roger

[boltzzz]

and weld in a length of tubing to mount the uprights on.... .

I'm not sure what you mean by this. Can you please explain?

.make the cut outs shallow as opposed to almost all the way in, about 1/3 rd the tube width and do the same for the tube at the bottom end......this will give you a level tube base each end. .

I agree that cut outs almost all the way through is not ideal. I thought it was move important to have an entire base cross member intact than to cut into it. Maybe I could cut half into the base cross member and half into the x rail supportd.

.this will give you a level tube base each end. .

The base of the machine is level. The top of the base cross member sits 25mm lower than the top of the x rails supports to allow for the ballscrew.

DON'T consider stress relieving the base part once it's completed....mostly it won't be necessary as you won't be cutting into any of the sections once welded, and drilling and tapping won't cause the structure to distort or re-align. .

My knowledge in welding is very limited. I thought that when you weld and it cools down, it would contract and warp the material. I was looking a stress relieving to eliminate this preload so the parts would be straight.

I would also add another cross piece between the inside tubes half way down the base frame to stiffen up the base a bit. .

The maximum gap between the base cross member and the end cross members is 300mm, so I thought additional cross members weren't needed.
I was thinking of adding 45° tubing from the base cross member to the x rail supports as shown in the photo.

That will give you a strong base, but make the uprights from the same base tube section and bolt them onto the back of the X axis cross beam and the base cross tube. .

Like the photo below yeah?

This design would need to be bolted together in 4 pieces to allow machining or hand working the various surfaces that need to be attached together, but it does allow you to make the structure firm and square and correct any misalignment that ca occur. .

I'm not sure what you mean. I was going to build these as sub-assemblies and get them machined as sub-assemblies. Any minor variations can be shimmed out when bolting the parts together.

You could consider making the X axis cross beam as a composite from two smaller tubes welded to a flat steel plate to give a channel section......this will give you 2 tube surfaces to bolt the rails onto and the gap between to give a gully to clear the X axis ball nut. .

I was actually thinking about that to increase the y rail spacing, but it would increase the complexity and part count. I've also read that channels aren't great with torsion, but having a box section at each end of the web might change those properties. I'll give it more of a think.

[handlewanker]

Hi, consider the assembly and what you have when it comes to machining the various faces to mount the rails on and the interfaces for the sides etc.

I previously mentioned 4 sections....that is the base tube assy.....2 side uprights and the cross beam.

Making the base as a complete flat entity allows you to machine or hand work the top rail mounts for the Y axis and the faces for the side uprights..

The side frames would then be separate items bolted to the base.....the cross beam would bolt to the faces of the side uprights.....each can be machined true and square without anything getting in the way

The design has changed a bit now.......it's more complicated than the previous one.....also getting too massive to rate it as a build in steel as opposed to aluminium.....that is over engineering for no good reason.......to make it as a replacement for a 6040 all you need to do is make it as strong as the ally model, as you won't want to cut steel with a high speed spindle .....ever.

BTW....the large box section crossbeam puts the stick out of the spindle much further forward.....that will increase the cantilevering effect on the Z axis slides.....that is one reason I suggested a flat steel plate with 2 steel tubes......probably 50mm each square with a 2mm wall and spaced at least 150 -200mm apart.....with rail mounts welded on them this will give you enough clearance for the ball nut.
Ian.

[handlewanker]

BTW.....on the stress relieving....it only causes a problem if you significantly machine the work piece as this will release the stress randomly.

I weld progressively and in small runs to keep the heat from building up.....as soon as it gets hot I douse it in water to keep it cold.....that way it's always cold when you weld the next bit and not expanded to lock in the stress......if you weld when it's hot it will expand and then pull the structure when it cools down.....law of physics.
Ian.

[NIC 77]

So I've decided to leave the aluminium design and go with steel. I know long term it's the right decision.

OK, the party is on now!

I would have like to use a 250x150x9mm RHS for the gantry, similar to Wade'O Design.

Good choice. The drawings I've done for you are 6" x 8" x 3/8". The 250 x 150 works a bit better.

I could go a 250x150x6mm RHS

Yep, that would also work. If you decide to build it so it bolts together, the weight might not be so important. But either should be fine. You're not dealing with very large spans here. Deflection is a factor of the length cubed, so you should be fine with the 6 mil also.

I'll get holes drilled at 150mm increments or so dowels can be installed to assist with mounting the rails straight.

Not necessary. You can put the bearings on your table and slide your table back and forth to align the rails, then you clamp the rails down at the ends, and make sure the movement is smooth. Then you can use a hand drill with a bit as close to the hole size in the rail as you can get to drill little divits that you will later use to guide you when you use a drill press or mill to drill the holes for tapping. It's not that hard to do. Don't sweat it, you're over thinking it.

DON'T consider stress relieving the base part once it's completed....mostly it won't be necessary as you won't be cutting into any of the sections once welded, and drilling and tapping won't cause the structure to distort or re-align.

BTW.....on the stress relieving....it only causes a problem if you significantly machine the work piece as this will release the stress randomly.

What he said.

Don't worry if you have to grind off a bit of a surface to make it flat. you won't be removing enough to be a problem, plus you would be doing more than one pass, so like he said, don't waste your money.

Lots of people on here have big routers, all welded frames, never stress relieved, never had a problem. The small amount of drilling you will do after welding should not be a problem.

So, a picture is worth a thousand words. Instead of trying to describe what I want to in writing, I did up some drawings.







I used
6" x 8" x 3/8" square tube
6" x 1/2" flat bar
1.5" x 1/2" flat bar

A 1" flat plate for the table. In reality, you'd want to add some ribs underneath, so you'd want to lower the ball screw down a bit and add an extension to the ball nut bracket. I assume you can make the table from aluminum. Perhaps you want to add a protective sock to the ball screw. What I designed here is very preliminary, I just wanted to express my opinion in a way that is easy to understand.

The design is 4' x 3'. The cutting area is a bit larger than what you wanted.

This is the way I would do it, but it requires some welding. I would use my skill with an angle grinder to make the mounting surfaces flat. There's not much work in welding the flat bar on once everything has been cut and drilled (for the big holes, not the rails, you do that after). A day perhaps. If you lived near me we could make a deal for some welding rods, and beer.

You could also do it this way without the 6" x 1/2" or 1.5" x 1/2" flat bar and just bolt the tubes directly to each other. This may require you to use some stainless steel shim tape and or conformal shimming (epoxy) on the surfaces that bolt together. There's plenty of space to get tools into the tubes to bolt them together without using the flat bar. You could make it this way without doing any welding at all.

You could drill the holes with a 1/2" bit and use M10 bolts. This would give you some room for adjustability. The separate pieces mean that you can move this into a basement or apartment. Obviously, you need at least a good drill press to do this.

Feel free to use this design for your machine or modify it and then use it, or do something else entirely, whatever you feel like. It's just a suggestion. Take what you will, and leave the rest. Good luck.

[RCaffin]

The vertical 'columns' are bits of big box tubing side on. They MUST have a flat plate welded across one end - I think that is shown. Otherwise there could be some ringing or vibration.
I suggest that the Y axis box section should be capped at the ends to also prevent deflection. Trouble is, that makes it hard to get inside for bolts.
The vertical movement allowed is probably fine for a router. For a full CNC you would need more ht.

However, all in all, that is a far more robust and reliable-looking design.

cheers
Roger