[Buzzard88]

Hi all

I have a load of rectangular hollow section 250x150x12mm, or US approx 10x6x7/16". I mean I have a lot of this stuff.

So I am wondering if / how I can use this to build a fixed gantry style machine base. Target machining materials would be mostly aluminium but some steel. Ideally the machine would have a base size of approx 1000x750mm. So three pieces of the section joined side by side and then maybe EG filled.

I would use a cast iron surface plate as a base but they are hard to come by in this kind of size in Oz.

1. Is this a stupid idea?
2. If not quite so stupid, what is the best method to join these sections? Welding along the lengths top and bottom?

TIA

[joeavaerage]

Hi,
the only problem with welding is that there will be a huge amount of residual stress build up in the finished structure and that as soon as you machine it, it will move. The solution is ideally
Thermal Stress Relief (TSR) or a lesser known alternative Vibratory Stress Relief (VSR).

Most hobbyists throw up their hands and say 'It can't be done'....which is rubbish, it not only can be done but is a widely known and practiced process
industrially....its just not cheap. I priced it (TSR) recently here in New Zealand and it came in at $6.70NZD/kg + GST which works out to $4.70USD/kg.

If you plan on welding then you need to find a heat treater and find out what size oven they have. No point in building something that wont fit it!
Another alternative is to use a pottery kiln. TSR requires taking the structure up to about 650C and then slow cooling....well within the capability
of a pottery kiln.

I would not bother filling the sections.

Vibration damping comes from a material dissipating energy as a result of flexure. If you fill an internal void with concrete or EG, both of which are a small fraction
of the stiffness of steel, you are relying on the very small amount of flexure of the steel shell transmitting that flexure into the filler and thence dissipate energy. This relies also
on perfect adherence of the filler to the inner surface of the steel which is in itself unrealistic.

If the structure is stiff enough vibration will be vanishingly small. You would be better increasing the wall thickness of the steel rather than filling. Make the structure as stiff as possible,
damping is nice but way WAY WAY down the list whereas stiffness or rigidity occupies the top three places in the list.

Craig

[catahoula]

Read through my thread: https://www.cnczone.com/forums/cnc-d...132-forum.html Machine footprint is roughly 900x900mm, work area is 610x460x210mm. FEA stiffness theoretically 100+ N/um at spindle nose (tests in my thread show ER20 collet deflection too).

Steel is an excellent material. The most rigid commonly available material by far. And 250x150x12 is an excellent size. Welding is fast and effective. As Craig mentions you need thermal stress relief after welding and before having it machined but this is not a big deal. I drove my frame 3 hours to the nearest industrial heat treater, it cost $350 or so (roughly $1 USD per kg), which compared to any other expense or time effort involved in building a CNC mill is absolutely trivial. When the frame was machined flat, the machinist commented that there was zero measurable distortion after the operation.

I wouldn't join them in the arrangement you have shown in your cad model. Look at Onsrud or Thermwood fixed gantry machines for more inspiration on the optimal arrangement. in fact, tubular structure instead of a flat, solid base allows easy draining of coolant and chips to a collection tray.

I agree with Craig that EG fill is probably more hassle than it's worth for a properly rigid machine. With steel you have the advantage of drilling and tapping holes very easily, but EG fill adds a significant difficulty in your order of operations, as you can no longer drill and tap once it's filled. Of course it can be done, but I've done a lot less EG filling as I've built mine than I initially planned because of this issue. I will still pour EG fill in the base, but that's because mine has an open bottom. I'm simply filling the tubes with dry sand as I go. I can't speak to the vibration damping of loose sand vs EG from an engineering/theoretical perspective, but it adds mass and certainly some amount of damping (based on sound from tapping the frame), and it's cheap and easy to do. And the frame is designed to be plenty rigid on its own.

The major changes I would make to my machine if I did it again are 1. to use larger section tubes for the base frame and 2. simplify the overall design. In a moment of weakness I reduced the size of the main longitudinal tubes on the base, thinking I was overbuilding. Then during fabrication, I added a bunch more material to the base and gantry to make up for it, so the design was needlessly complicated. Now, if I support the base by three points and put my body weight in the center of the base, I can measure about .04mm vertical deflection. I am going to epoxy bond some additional material under the base to reduce this before final assembly. 250x150x12 tubes, situated with the larger dimension vertically, would be ideal. With that size tube you can make a very simple structure a la Onsrud or Thermwood and have it plenty rigid (for a machine similar size to mine or slightly larger).

The other change I would make is to be very deliberate about rail vs ballscrew height. I welded flat bars to the tubes, which were machined down for the rail surfaces. That is the right concept, but if I did it again I would make sure that during the machining process, the height differences between rail and ballscrew are fully compensated. As mine is, I will still have to use 5-10mm thick spacers plus fine tuning shims to match the two, which is fine but annoying.

Edit to add: also think a lot early on about designing around an ATC and way covers/ballscrew covers. These things are really hard to add in later.

[joeavaerage]

Hi,
I'm not a fan of a big flat base either. In order to hold a workpiece you have to either drill and tap holes to mount a vice or have tee slots.

I had three axis beds (700mm x 250mm x 130mm) cast in grey iron at 115kg each, then VSR and machined. The X axis bed has tee slots milled into it, and there is a good reason
that they are so widely used, they are just so damned good!.

Have you considered getting a piece of 32mm or maybe 50mm steel cut as a base?. You might be surprised how cheap it can be. If you chose the size carefully you could mill tee slots
and surface it in one setup. Depending on the size you could surface grind it. Provided you do not weld to it it will be pretty damned stable as is.

Craig

[peteeng]

Hi Buzz - The section size is useful but you need to figure out your workflow especially final machining and thermal stress relief as you are going to weld the structures. Read through some of the steel mill build threads and glean as much info on the errors they make as possible. Stay away from EG filling, not necessary with the big thick tubes you have (plus epoxy is costly better put the $$$ somewhere else in the machine). Sort out whos going to final machine the bits and make sure they know what a zero stress set up is as people have come to grief at the final machining stage and correcting that is an issue at the end of the project. Stitch weld say 50% vs full welds in symmetric sequence and you'll be close to the money. Also sort the spindle as this is a big topic and if you start designing at the base you will run out of real estate when you get to the spindle. So design from the spindle outwards. The Z axis is a place that lets a lot of people down. Your size machine is at the edge of portal design. Unless you really need a large square bed make the 750mm a bit smaller and the 1000mm a bit longer and it will be a better machine if built as a moving column design. Keeping two drives synchronised in an accurate mill sometimes is difficult. A moving column design only has 3 drives (XYZ) so is easier to keep in square & tram... That size machine implies it will take large heavy chunks of billets. That means you have to design the drive systems to move heavy masses. If you use a moving column design in which the job stays still you can optimise the drive systems better as they only need to move the machine which is a known mass/inertia. Plus the drives are up out of the much. Keep at it. Peter

[joeavaerage]

Hi,
you need to research what you want for a spindle.

Many people go for a high speed/low torque spindle, say 24000rpm and 2.2kW....and they are great, but poor, even very poor for cutting steel. Steel requires low rpm and high torque.
That is not to say you cannot cut steel with a 24000rpm spindle, but is very slow by virtue of taking such very fine cuts and it wrecks tools like they are going out of fashion.

I myself have a 24000rpm spindle, but only 800W. I can and do use it on steel with 1/8th coated, expensive tools. About the best tool life I can get is four hours cutting, and that is with 1mm DOC and less,
and that is with continuous flood cooling. Lesser cooling solutions will result in lesser tool life. The same spindle has been excellent in aluminum, brass, plastics, and I've used it for thousands upon
thousands of hours for over ten years....so don't get me wrong I love my little spindle...but its a very poor choice for steel.

So much so; that for steel I made a second spindle out of a 1.8kW Allen Bradley servo, 3500rpm, 6.1Nm (cont), 18Nm (peak). It great for steel and stainless, but too slow really for aluminum
and small tools with many of the small parts I make.

Its a bit of a dilemma really. A Chinese high speed spindle and VFD are vey well priced and very appealing....just they are poor at low and part speeds.You can of course buy a spindle
that will do a good job of aluminum, brass etc and engraving with small tools, and do steel....but the cost!!!! Thousands upon thousands!!

Do some research and make a decision early on about the spindle. Even if you get a small spindle to start with....with the plan to get the 'proper and decent one' in a year or twos time....you need to have
planned for having that 'you beauty' 20kg spindle from the start. Among the choices is ATC or not. If you want ATC then you may require an extra 100mm or even 150mm extra Z axis clearance so you can get
the tool holder out of its taper. Too bloody late once you've built your machine to discover that cockup.

Craig

[peteeng]

Hi Buzz - There is a thing called designer attachment or designer inertia. In the past I have been involved with design projects teams and managed design offices and it's a real issue for some people. If you start designing your machine from the base you will put a lot of effort into the base, bridge and various bits then you will hit the saddle and Z axis. These will be then compromised to squeeze all the bearings and bits into the shrinking space. Then you hit the spindle and yes if you go ATC you'll need more space and you will be reluctant to change all the work you have done on the base and bridge (your attachment and inertia) and it will become frustrating, So start at the spindle and make sure its correctly sized to the Z axis and then enlarge that to the saddle and then the bridge and base. Sure you can pencil in the base and bridge etc but use block geometry as a visual aid to get a feel for how things are going. Once the blocked out design is looking good detail out the spindle area then the Z axis then the saddle. Then there will be no back tracking and angst over where all the space went... Been there a few times... Peter

[joeavaerage]

Hi,
contrary to peteeng's suggestion I went for a variation of a column design. The axis beds are bolted to a steel frame. This I found gave me the best flexibility
to change my mind at a later date and shift the axis beds around on the frame, or even build a new frame. The axis beds themselves were very expensive, and they will perforce last
the life of the machine, but the frame is not sacrosanct.

Peteeng's idea about starting from the spindle is however entirely justified whether that be a moving column design or a fixed column design or any other design. Start with the expensive and noisy bit....
and work outwards.

Craig

[joeavaerage]

Hi,
the last thing you might like to think about is whether you want a fourth and/or fifth axis. Many a CNCer dreams of doing so one day.

If you do then you must allow for the height of the fourth and/or fifth axis, it will be substantial, often meaning that the Z axis does not have clearance under it.
If there is no easy means to increase either the Z axis height or its travel then a fourth/fifth axis may be precluded.

I have built a trunnion table and fifth axis platter, and I kept the build at the lowest height I could within the bounds of budget and rigidity, but the platter height is still
236mm above the table. Fortunately the column design means I can unbolt the Z axis bed and shift it upwards to recover the lost Z axis travel. I deliberately made the column higher
that I needed initially anticipating the day that I would need to shift it upwards.

There are means to have the fourth axis in a well for instance that obviate the need to raise the Z axis.......but the principle still applies that you give good thought to your initial design
that it might accommodate future developments. Four/Five axis and ATC are two of the most common and desirable amongst hobbyists.

Craig

[Buzzard88]

Wow, thanks guys for all the awesome info. I had alerts to thread enabled but never got any alerts for some reason so I just checked in today and found this little treasure trove.

Seriously can't thank you guys enough. I will go thru in detail but just scanning quickly over things here, you've all provided such useful stuff.