Cost to rigidity balance... where is it? Need help planning large gantry.

[Jeff0000]

When I started looking here I was set on building a large mpcnc, 4'x8'. I quickly learned that size was outside it's ability for the speeds I was wanting to cut, 3/4" plywood in 2 passes at 65 ipm.

I am handy with steel, but only have basic tools, nothing to make really precision cuts or grinds. I am sure I can get it close enough though.

I am thinking that rack and pinion is going to be my best bet. But I have to keep it to nema 23 motors for cost reasons. And even then I am not sure what driver and control board to use now since my knowledge topped out with what the mpcnc had.

I did find rack and pinion drive, which I think gives me 3x the power of the stepper via gearing? PRO Rack and Pinion Drive, NEMA 23 | CNCRouterParts I assume I would need 3 of these? one for each side of the x axis, and one for y (gantry). Do I need these drives even?

So now with costs in mind can I get away with one rack on the gantry? Or is two needed?

How heavy can I make the the gantry with nema 23?

Would I be able to make this rigid enough to cut 3/4 in one pass at a reasonable rate? 2.2kw spindle I assume, or would 1.5kw be large enough?


Planning to make the table from 2x4 1/8 wall tubing with 1x2 1/8 wall cross bars, Might need to go thicker on the parts that hold the rack.
Planning on making the gantry from 2x6 1/4 wall tubing.
Planning on making the 'legs' for the gantry from 5/8 plate (because I have it on hand).
Planning on welding what I can.

Any advice would be great. I figure I am going to be 100% over budget, so anywhere I can cut costs would be awesome. My goal of 1000 probably isn't realistic, but I still want to try.

Thanks in advance.

Similar Threads:

• Need Help!- Planning stages, gantry router for aluminum cutting.
• Need Help!- Need advise/opinion for 8020 based gantry (planning my first build)
• LOW COST BEST QUALITY LARGE RELIEF MODELS
• Newbie- Gantry rigidity question
• Gantry: planning to design moving open gantry

[subsoniq]

Nice! I will follow this as i am thinking of the same build? Do you ever plan on cutting aluminium as well? What is the z height?


Sent from my iPhone using Tapatalk

[Eldon_Joh]

this may not be realistic but to buy 48 plus 96 inches of the gear rack from that website will cost you at least 211 dollars, based on .68 inches per dollar. maybe 250 and then you have to pay for shipping.

If you can find three identical flywheels at a junk yard, you can take the ring gear off of them, heat them up red hot and stretch them out flat, maybe even put just a little bit of a backwards curve in them. depending on what steel they are, they should be fully annealed by the time you're done with straightening.

I would use 6 by 6 1/8th wall tube for your gantry. put diagonal stiffeners internally and some kind of vibration dampening material.

[Jeff0000]

Nice! I will follow this as i am thinking of the same build? Do you ever plan on cutting aluminium as well? What is the z height?


Sent from my iPhone using Tapatalk

I figure it should be rigid enough to cut aluminium, isn't a specific goal, but will for sure give it a go. I've kinda hoping it will cut thin stainless, like thin for a little sign letters or something.

this may not be realistic but to buy 48 plus 96 inches of the gear rack from that website will cost you at least 211 dollars, based on .68 inches per dollar. maybe 250 and then you have to pay for shipping.

If you can find three identical flywheels at a junk yard, you can take the ring gear off of them, heat them up red hot and stretch them out flat, maybe even put just a little bit of a backwards curve in them. depending on what steel they are, they should be fully annealed by the time you're done with straightening.

I didn't realize that was even a thing. Looks like I can get new ring gears for like 15 bucks each, and each ring should give me 50", being a much coarser thread, would that cause much of an issue?

[Eldon_Joh]

I didn't realize that was even a thing. Looks like I can get new ring gears for like 15 bucks each, and each ring should give me 50", being a much coarser thread, would that cause much of an issue?

you can get a new one for 15$?

so, wikipedia says the teeth are hardened but the core is soft, mine is definitely through hardened; 12 DP, 192 teeth, about 16 inch diameter. yours may not be 12 diametrical pitch but rather some other pitch.
So of course you will need to find a gear to drive it. the coarseness of the gear teeth is not an issue theoretically, it may be an issue once the gears start to wear, you're going to get a cogging type effect where the rotary motion is not perfectly translated into linear motion, and its going to repeat at the pitch of the gear.

The tooth form will not be perfect but it will be extremely close to a proper gear rack, the most difficult part will be getting it straight enough to lay down flat against whatever you are using. i don't think that will be particularly difficult, but i haven't tried yet.. typically the ring gears are not exceptionally thick, about as thick as they are wide.

[hanermo]

I would recommend just buying the straight gear racks.
It will cost you maybe 10x more at minimum wage to adapt something like a ring gear, and it will never be remotely accurate.

Likewise, for anything *practical * on a router running full sheets...
Figure on upgrading stuff 1-2 steps.

Servos. Nema 23 is fine.
Real spindle with VFD.

If you actually want to make money with it ...
You need a toolchanger, and now the spindle is a lot more expensive than you proposed budget.

The best, nearest commercial level router is a mechmate.
Costs about 8000 $ to make, with no toolchanger.
Add real liner rails, toolchanger, servos, good to go.

Google mechmate for samples, ideas.

A 6x6 gantry will work ok, for light use, with small/tiny ? bit, IF AND BECAUSE you hope to get a 1.2 m free span for 1.2 x 2.4 m sheets.
Strength will not be the issue, vibration over the long span.
6 x 6 / 1/4 is very flexy.

I ran a 200 kg table fine, with tiny nema 23 steppers, on a mill.
Mass is a total non-issue.

My new bridge is about 700-800 kg in mass, goes up and down, on 32 mm ballscrew.
Tiny nema 23 servos are plenty (== 600 kgf push force), gives 1/2 g acceleration.

Think of it like this.
A 2-3 kW spindle probably needs about 100 kgf push force to run well (ie you are using 3 kW at spindle).

GO stand on the 6x6 beam, in the middle. Beam on nuts at ends, so its free floating.
Does it bend ? Measure with dti.

I will guess it bends 0.1 mm or more, ...
about 5x too much.

The 6x will work fine .. but it wont get you full-sheet router capacity at 3 kW spindle power.

Example:
Look at commercial machines for full sheet use (chinese ones in the 8000-10.000 range).
Cast iron or steel beams, maybe 20x more rigid than a 6x6.
Linear guides, in the 20-25 mm range.

There Is A Reason.

I have a free beam of 1600 mm, which is my table width.
Scratch built mill.
Just the uprights are 300 kg each, 2 cm thick steel on z axis, 2 cm thick steel plates, 300 bolts on each.
120 cm tall, or 4 feet.
The box section is 20x400 mm, on each end.

There Is A Reason.
The reason is that a wide span bends, a lot.

[ger21]

With your budget, even at $2000, for a 4x8 machine, everything you choose will be a compromise.

Here are the bare minimum components that I would use:

Spindle - $325
FOUR BEARING ER20 2.2KW WATER COOLED MOTOR SPINDLE AND DRIVE INVERTER VFD CNC | eBay

Rack and pinion drives - $390
PRO Rack and Pinion Drive, NEMA 23 | CNCRouterParts

Gear rack - $360 (You can get cheaper gear rack. Check with Moore Gear)
PRO Rack and Pinion Drive, NEMA 23 | CNCRouterParts

Gecko G540 Motor/Drive kit - $560
http://www.automationtechnologiesinc...in-48v7-3a-psu


This puts you at ~$1600, with no frame, linear bearings, wiring, and all the other miscellaneous items that can cost a few hundred more $$.

Keep in mind that these are minimum components. This is why a decent 4x8 router will usually cost between $4000-$8000.

These components can allow you to cut at 300+ipm. I'd recommend 3 passes at 300ipm vs 1-2 @ 65ipm. Your bits will last longer, and you'll still get done a lot faster.

[Jeff0000]

I would recommend just buying the straight gear racks.
It will cost you maybe 10x more at minimum wage to adapt something like a ring gear, and it will never be remotely accurate.

Likewise, for anything *practical * on a router running full sheets...
Figure on upgrading stuff 1-2 steps.

Servos. Nema 23 is fine.
Real spindle with VFD.

If you actually want to make money with it ...
You need a toolchanger, and now the spindle is a lot more expensive than you proposed budget.

Google mechmate for samples, ideas.

A 6x6 gantry will work ok, for light use, with small/tiny ? bit, IF AND BECAUSE you hope to get a 1.2 m free span for 1.2 x 2.4 m sheets.
Strength will not be the issue, vibration over the long span.
6 x 6 / 1/4 is very flexy.

I ran a 200 kg table fine, with tiny nema 23 steppers, on a mill.
Mass is a total non-issue.

My new bridge is about 700-800 kg in mass, goes up and down, on 32 mm ballscrew.
Tiny nema 23 servos are plenty (== 600 kgf push force), gives 1/2 g acceleration.

Think of it like this.
A 2-3 kW spindle probably needs about 100 kgf push force to run well (ie you are using 3 kW at spindle).

GO stand on the 6x6 beam, in the middle. Beam on nuts at ends, so its free floating.
Does it bend ? Measure with dti.

I will guess it bends 0.1 mm or more, ...
about 5x too much.

The 6x will work fine .. but it wont get you full-sheet router capacity at 3 kW spindle power.

Example:
Look at commercial machines for full sheet use (chinese ones in the 8000-10.000 range).
Cast iron or steel beams, maybe 20x more rigid than a 6x6.
Linear guides, in the 20-25 mm range.

There Is A Reason.

I have a free beam of 1600 mm, which is my table width.
Scratch built mill.
Just the uprights are 300 kg each, 2 cm thick steel on z axis, 2 cm thick steel plates, 300 bolts on each.
120 cm tall, or 4 feet.
The box section is 20x400 mm, on each end.

There Is A Reason.
The reason is that a wide span bends, a lot.

6x6 1/4 wall square isn't very rigid either though. Would a 6x6 i beam that is gusseted every 6" be better? I know it's the twisting force that is going to be the greatest. Or actually maybe a double beam 2x6 with the spindle in the middle? I could go 3/8 wall too I guess.

I was thinking about a tool changer, and came across this Hackaday Prize Entry: DIY Automatic Tool Changer | Hackaday

Mechmate is the general design I am following, although several ideas have come to mind reading threads on here.

With your budget, even at $2000, for a 4x8 machine, everything you choose will be a compromise.

Here are the bare minimum components that I would use:

Spindle - $325
FOUR BEARING ER20 2.2KW WATER COOLED MOTOR SPINDLE AND DRIVE INVERTER VFD CNC | eBay

Rack and pinion drives - $390
PRO Rack and Pinion Drive, NEMA 23 | CNCRouterParts

Gear rack - $360 (You can get cheaper gear rack. Check with Moore Gear)
PRO Rack and Pinion Drive, NEMA 23 | CNCRouterParts

Gecko G540 Motor/Drive kit - $560
G540 4-Axis kit Stepper Motor | Stepper Motor Driver | CNC Router | Laser Machine | 3D Printers For Sale


This puts you at ~$1600, with no frame, linear bearings, wiring, and all the other miscellaneous items that can cost a few hundred more $$.

Keep in mind that these are minimum components. This is why a decent 4x8 router will usually cost between $4000-$8000.

These components can allow you to cut at 300+ipm. I'd recommend 3 passes at 300ipm vs 1-2 @ 65ipm. Your bits will last longer, and you'll still get done a lot faster.

Thanks! I know my budget is weak, it's pretty loose, just might take longer than originally thought if things start to get too expensive.
I actually started looking into building a router sled to plane live edge slabs that are too big for my planner, that led me to the mpcnc, to basically just set a program to plane for me. But was convinced that the mpcnc is simply not rigid enough for that large. And now I am worried I am stepping well outside of hobby territory and will need a way for this machine to bring in a little bit of income.

For the gear rack there is a local to me old guy that makes and sells it for about the same price as I can find anywhere online, but no shipping. I should actually ask him about helical rack.

And linear bearings I was thinking last night just to build a skate with bearings, but after sleeping on it I think a double profile rail for the width of the gantry, and then double round rails for the track the gantry rides on. I know profile is better, but it's just so much more expensive for the long way.
But again now I feel like i am building this machine to be a monster.

For the feed rate and passes, 3 passes at 300 ipm would be quicker, I still have a lot to learn on what is best for feed rates and how many passes. I didn't realize it would be able to handle 300 ipm, that is moving pretty quick, lol.

[ger21]

6x6 1/4 wall square isn't very rigid either

6x6 1/8" wall is about 8x stiffer than 1/4" 2x6

And now I am worried I am stepping well outside of hobby territory

I think that you'd be surprised how many hobbyists have spent $5000-$10,000 on CNC machines.

And linear bearings I was thinking last night just to build a skate with bearings, but after sleeping on it I think a double profile rail for the width of the gantry, and then double round rails for the track the gantry rides on. I know profile is better, but it's just so much more expensive for the long way.

If you're linear bearings are a weak point, then you can't take advantage of all the other components.
The weakest part of the machine will always be exposed, and if you improve that part, then the next weakest becomes the weakest link

But again now I feel like i am building this machine to be a monster.

You're a long way from a monster. What I'd consider a DIY monster gets into the $10,000-$15,000 range.

[Jeff0000]

6x6 1/8" wall is about 8x stiffer than 1/4" 2x6





I think that you'd be surprised how many hobbyists have spent $5000-$10,000 on CNC machines.





If you're linear bearings are a weak point, then you can't take advantage of all the other components.
The weakest part of the machine will always be exposed, and if you improve that part, then the next weakest becomes the weakest link



You're a long way from a monster. What I'd consider a DIY monster gets into the $10,000-$15,000 range.

Looking around it seems like the 6*6 still suffers from torsional flexing. Would putting it between two 2*6 be best? Or just go 3/8 wall on the 6*6

I guess it's just a monster for me, 5 weeks ago I had a budget of 70$ for a good bit for my router and some lumber to make a router sled.

Will it tough to program to act as a planner?

[ger21]

Just reinforce the beam like this:
http://www.cnczone.com/forums/cnc-wo...ml#post1689518

[Jeff0000]

Just reinforce the beam like this:
http://www.cnczone.com/forums/cnc-wo...ml#post1689518

I read that thread already actually. It's interesting using rebar and epoxy, I'd probably cut a few slots in it to drop a plate in and weld on 2 sides, maybe a couple plug welds on the other two sides. Or even just do plug welds all the way around.

[SteveWill]

This is above my knowledge, but I think you need to think about how accurate this need to be. ( 1 inch per foot or 0.001 per foot. ) Right now I am using a 1.5 by 6 inch wood 30 inch long for my X axes and would think that a 2x6 steel tube for 4 foot would work, for 0.01 accurate cuts. 0.01 is not very accurate but it work for me.
Not sure how a 8 ft drive would be done with out rack and pinion, but your speed will now be more than 70 ipm.

[Eldon_Joh]

If a 6 by 6 1/8th is too floppy, and maybe it is, go with 7 by 7 or 8 by 8. the difference scales to the fourth power with the outside perimeter of the element under torsion or flexing. increasing the wall thickness (within limits) is a linear relationship.

6x6x0.125 is 16.9 in^4, 1/4th inch wall thickness is 31.75
7x7x0.125 is 27 in^4 1/4th wall thickness is 51.32
8x8x0.125 is 40 in^4 1/4th wall thickness is 77.66


as for flattening a flywheel ring gear, turns out mine is not through hardened. by flattening 5 inch wide segments at a time in my vise, I now have a radius left in the "gear" of about 26.5 inches. it would be pointless for me to try and give you the radius needed to bend it backwards, since every gear will be different. but i bent it backwards against a leaf spring of perhaps 100 inch radius and now i have a rack segment with a radius of about 50 inches.

as far as a flywheel ring gear's accuracy goes, it would surprise me if that was a problem. it appears the gears are broached these days, all the teeth cut at the same time, then induction hardened.

Btw, after cutting the gear, it sprang "closed" by about a 2 inch overlap. so be careful to clamp it while cutting it with an abrasive disk. i anticipated this and cut it 90% of the way through the gear, then broke it apart.

[Jeff0000]

This is above my knowledge, but I think you need to think about how accurate this need to be. ( 1 inch per foot or 0.001 per foot. ) Right now I am using a 1.5 by 6 inch wood 30 inch long for my X axes and would think that a 2x6 steel tube for 4 foot would work, for 0.01 accurate cuts. 0.01 is not very accurate but it work for me.
Not sure how a 8 ft drive would be done with out rack and pinion, but your speed will now be more than 70 ipm.

Realistically 0.1mm accuracy per foot is probably adequate. I mean this cnc will be replacing me with a cordless skill saw and or jig saw. 1mm over the 8' would be fine. It would be nice to have 0.05mm per foot accuracy even with aluminum.

If a 6 by 6 1/8th is too floppy, and maybe it is, go with 7 by 7 or 8 by 8. the difference scales to the fourth power with the outside perimeter of the element under torsion or flexing. increasing the wall thickness (within limits) is a linear relationship.

6x6x0.125 is 16.9 in^4, 1/4th inch wall thickness is 31.75
7x7x0.125 is 27 in^4 1/4th wall thickness is 51.32
8x8x0.125 is 40 in^4 1/4th wall thickness is 77.66


as for flattening a flywheel ring gear, turns out mine is not through hardened. by flattening 5 inch wide segments at a time in my vise, I now have a radius left in the "gear" of about 26.5 inches. it would be pointless for me to try and give you the radius needed to bend it backwards, since every gear will be different. but i bent it backwards against a leaf spring of perhaps 100 inch radius and now i have a rack segment with a radius of about 50 inches.

as far as a flywheel ring gear's accuracy goes, it would surprise me if that was a problem. it appears the gears are broached these days, all the teeth cut at the same time, then induction hardened.

Btw, after cutting the gear, it sprang "closed" by about a 2 inch overlap. so be careful to clamp it while cutting it with an abrasive disk. i anticipated this and cut it 90% of the way through the gear, then broke it apart.

Hmm that makes sense actually.
I really need to keep this build realistic I don't need 0.001 accuracy so 6*6 is probably lots.

Actually thinking of down sizing to 36 wide and 44 long but having a way to extend it for the handful of times a year I'd need longer.

[hanermo]

As others said .. it is the 4th power of section size (of half height, iirc).
So a 7x7 will be very rigid vs a 6x6 for example.

Accuracy is NOT what you are concerned with here.
At All.
Its the vibration you need to control, and that comes from rigidity - not strength.
A typical machine is only loaded at 2-3% of its "strength".

So if you cut a finish cut with a tiny bit, you will get excellent local accuracy, somewhere around 0.02-0.04 mm.

No way around it .. the machine will either not work at all, or be pretty "accurate", in local terms, and fairly accurate in global terms, probably about 0.1 - 0.2 mm with (weak, not good, yuk) round rails.

But if the bit vibrates, when cutting with say a 10-12 mm bit, its liable to snap.
On a vastly heavier and more rigid industrial machine (Haas SR100, Sheet Router, 6.6 kW, steel, 35 mm linear rails iirc) the same bit lasts hours and hours wizzing away at 500-700 ipm (1000 ipm cut speed max, 2000 ipm rapids).
The SR actually look like a pretty nice model.

And I think the gantry beam seems to be about 20 cm wide, perhaps 20-25 cm tall.
So 8-10 " in section size.
Another good think, imho, about the SR is that it has only 6" of clearance. This makes it rigid.
And it is only 1270 kg in mass.

The last thing you should be concerned about at this time is speed.
The easiest thing is to just upgrade to servos, at some point in time, if you wish to or need to.

The nr 1 concern, the major concern, is rigidity.
And as mechmates show, even hand built linear guides on v-rollers work fine- they just need to be straight and well aligned.

And on that happy thought, you can always glue another 6x6 next to the current one for added rigidity if you need to.
And yes, glue is actually plenty strong enough.
Industrial epoxy, as rigid ie non-elastic as you can find.

Hope this helps.
I think the Haas routers are discontinued, and in any case not sold in the EU.
However, pics and video exist on the net and youtube.

[Jeff0000]

As others said .. it is the 4th power of section size (of half height, iirc).
So a 7x7 will be very rigid vs a 6x6 for example.

Accuracy is NOT what you are concerned with here.
At All.
Its the vibration you need to control, and that comes from rigidity - not strength.
A typical machine is only loaded at 2-3% of its "strength".

So if you cut a finish cut with a tiny bit, you will get excellent local accuracy, somewhere around 0.02-0.04 mm.

No way around it .. the machine will either not work at all, or be pretty "accurate", in local terms, and fairly accurate in global terms, probably about 0.1 - 0.2 mm with (weak, not good, yuk) round rails.

But if the bit vibrates, when cutting with say a 10-12 mm bit, its liable to snap.
On a vastly heavier and more rigid industrial machine (Haas SR100, Sheet Router, 6.6 kW, steel, 35 mm linear rails iirc) the same bit lasts hours and hours wizzing away at 500-700 ipm (1000 ipm cut speed max, 2000 ipm rapids).
The SR actually look like a pretty nice model.

And I think the gantry beam seems to be about 20 cm wide, perhaps 20-25 cm tall.
So 8-10 " in section size.
Another good think, imho, about the SR is that it has only 6" of clearance. This makes it rigid.
And it is only 1270 kg in mass.

The last thing you should be concerned about at this time is speed.
The easiest thing is to just upgrade to servos, at some point in time, if you wish to or need to.

The nr 1 concern, the major concern, is rigidity.
And as mechmates show, even hand built linear guides on v-rollers work fine- they just need to be straight and well aligned.

And on that happy thought, you can always glue another 6x6 next to the current one for added rigidity if you need to.
And yes, glue is actually plenty strong enough.
Industrial epoxy, as rigid ie non-elastic as you can find.

Hope this helps.
I think the Haas routers are discontinued, and in any case not sold in the EU.
However, pics and video exist on the net and youtube.

So making it rigid and heavy is more about reducing harmonics and vibration.
I did see a video where they filled everything with sand, figure it's gotta help.
Wonder if water filled would be better?

You say not to worry about speed. Does that mean if i wanted to cut something like steel on a wood machine it would be possible but would just have to go slow? Or is there limits?

[Eldon_Joh]

You say not to worry about speed. Does that mean if i wanted to cut something like steel on a wood machine it would be possible but would just have to go slow? Or is there limits?

water won't work to dampen anything but the added mass reduces the frequencies of the resonant frequencies, and that is often enough; you could heat up tar and mix it with sand, maybe a 10:1 sand to tar ratio. then pour it into the beam. what a mess though. i would try sand alone first. then experiment later.


yes if you want to cut steel on your machine you will need to experiment. i offered some speculation here: http://www.cnczone.com/forums/chines...21470-cnc.html post #14

also: http://www.cnczone.com/forums/epoxy-...ml#post1953422

[hanermo]

Yes, and no.
1. Rigid works for everything. If it is rigid enough, it will cut steel.
A mill is rigid enough.
Routers are not.

With a tiny cutter, say 1-2-3-4 mm, a heavy router will cut steel fine.
Industrially, molds are finished with tiny cutters.
The finish cuts can easily last 100 hours (10.000 rpm, 1-2 mm cutter, 2.5 m mold, auto stuff) running 24x7.

2. You want mass, and vibration dampening materials.
Water is useless - liquids dont compress.

Cast Iron aka CI is great. It is 10x better than steel re:vibration.
Granite == CI.
Concrete == CI / granite re vibration.

Concrete, granite and cast iron are all quite flexible.

I meant re: speed that speed is the least important aspect of a machine.

If one wants/needs fast speeds, these are easily, and not too expensively, obtained with brushless servos.
400W ac brushless servo == 300€ per axis.
It is about 5x faster than a stepper in acc, 3-5x faster in speed, and 5x more accurate.

The speed is mostly not desirable.
For example, I gear my machines down, and don´t use top speeds.
At 1:2, with a 3000 rpm servo, the 1500 rpm at screw would make it wear really fast, and heavily stress the machine.

Industrial machines can mitigate this by using ground screws with less slop - but this is really expensive.
A 32/4 screw, C7, high quality, rolled, is == 200€.
A 32/4 ground screw, C0, == 2000 €.

So making it rigid and heavy is more about reducing harmonics and vibration.
I did see a video where they filled everything with sand, figure it's gotta help.

You say not to worry about speed. Does that mean if i wanted to cut something like steel on a wood machine it would be possible but would just have to go slow? Or is there limits?

[Jeff0000]

water won't work to dampen anything but the added mass reduces the frequencies of the resonant frequencies, and that is often enough; you could heat up tar and mix it with sand, maybe a 10:1 sand to tar ratio. then pour it into the beam. what a mess though. i would try sand alone first. then experiment later.


yes if you want to cut steel on your machine you will need to experiment. i offered some speculation here: http://www.cnczone.com/forums/chines...21470-cnc.html post #14

also: http://www.cnczone.com/forums/epoxy-...ml#post1953422

Tar and sand sounds way to messy!
I think it is already going to be rigid enough for what I am really wanting, but i think I will fill anything I can with sand, and have seen people adding oil, although not sure what the oil would do at that point.


This build here I a have been watching.

And the results he gets are good enough for what I will be doing really. Because I like to tinker I am still going to try and built it up as much as I can realistically push it, but I think I have been worrying about something that isn't worth worrying about. Even a 2x6 steel gantry has to be better than that aluminium he has. And I did find some 6x6 that I can use for almost free, it is 3/8 wall even.

Yes, and no.
1. Rigid works for everything. If it is rigid enough, it will cut steel.
A mill is rigid enough.
Routers are not.

With a tiny cutter, say 1-2-3-4 mm, a heavy router will cut steel fine.
Industrially, molds are finished with tiny cutters.
The finish cuts can easily last 100 hours (10.000 rpm, 1-2 mm cutter, 2.5 m mold, auto stuff) running 24x7.

2. You want mass, and vibration dampening materials.
Water is useless - liquids dont compress.

Cast Iron aka CI is great. It is 10x better than steel re:vibration.
Granite == CI.
Concrete == CI / granite re vibration.

Concrete, granite and cast iron are all quite flexible.

I meant re: speed that speed is the least important aspect of a machine.

If one wants/needs fast speeds, these are easily, and not too expensively, obtained with brushless servos.
400W ac brushless servo == 300€ per axis.
It is about 5x faster than a stepper in acc, 3-5x faster in speed, and 5x more accurate.

The speed is mostly not desirable.
For example, I gear my machines down, and don´t use top speeds.
At 1:2, with a 3000 rpm servo, the 1500 rpm at screw would make it wear really fast, and heavily stress the machine.

Industrial machines can mitigate this by using ground screws with less slop - but this is really expensive.
A 32/4 screw, C7, high quality, rolled, is == 200€.
A 32/4 ground screw, C0, == 2000 €.

I think I might have confused a bit with the speed question, I mean more couldn't a plywood built machine cut steel if it just went really slow at it?

While I actually looked at how to cast iron, it is definitely well outside my ability in my garage, lol. But man do I want to try and sand cast something. If filling with concrete would be better than sand I can do that, concrete is cheap enough.

I was looking at servo's instead of steppers, but their price just keeps it out of reach.