Bench-top or welded frame

[Doddy]

All good points

[Neale]

Thinking about it, I cannot remember seeing a single aluminium extrusion-based machine using diagonal bracing under the bed (although that might be a faulty memory!). And some of those machines have certainly been capable of machining aluminium. Additionally, the biggest "shaking" loads are involved in accelerating and decelerating the gantry, followed by accelerating and decelerating the Z assembly along the gantry. That is much less than the first, obviously - the mass of the gantry is the biggest load. I've reduced the effect of the second load by raising the side rails so that the Z assembly is not as high above the X rails - moving mass is more in line with the support points - although that means that the rails need good support. I think I've achieved that. I instinctively feel concern about the designs that use a single tall plate each end of the gantry, which I can imagine bending under dynamic loads. Clearly I'm wrong there as there are plenty of successful machines that are built like that although they need pretty heavy plates to take the bending loads. But engineering is all swings and roundabouts and there are always trade-offs to be made. My gantry design is more fiddly to make, for example, and probably needs a vertical mill as a minimum (which I have available, hence the design).

I do have a single diagonal corner to corner across the end (shorter) sides of the main frame. I have two more-or-less diagonals across each long side, but these are arranged to take the load of the intermediate X rail tube supports down to the bottom of the legs and do not run corner-to-corner. That does leave space to get the control cabinet in (although to be honest I had forgotten that when I built the frame, and I was lucky that the cabinet fitted the space available with a few millimetres to spare). Occasionally the gods smile on us.

I don't have a support leg in the middle of the bed, and I happily crawled over the bed structure when I was building the machine. My bed is all 50x50x3; if you used your 100x50x4 as cross-rails and braced it with intermediate 50x50 segments between them, I'm sure that would be stiff enough.

I would advise having a suitable device for cutting the box section to length accurately and squarely; I used an angle grinder in a small pivoting stand and struggled to achieve accuracy. This was a bad decision on my part, born out of lack of experience. One of those cut-off saws, maybe, with a decent size blade? Personally I would go for a metal-cutting bandsaw but that's because I have other work I do that would make use of it.

[Doddy]

Analogy is always a good reference. Yes, you're right, plenty of aluminium framed designs without so much bracing. As I said, part of this was to provide a strong, rigid base for the bed (to avoid sag), but I've finally clicked why you used 100x50 horizontally - to provide a base for the bed down each side of the X axis (for info, my very first router was a 2nd-hand MD affair and I could measure 0.7mm deflection on rails over a 400mm span... I don't want to go back there). I've got to be careful not to go back to my first post here - I could design and build a desk-top router and simply bolt this to the existing timber bench surface which would provide some mass if only limited rigidity. For me space is the constraining factor here, more so than cost, and the existing timber bench offers me storage and a little work area.

But, I do like the portable table idea, and steel is cheap (argon, on the other hand...). I might revise/simplify the table base (which may start to look even more like AVOR with the 100x50 on the flat). There is a lot of deflection across the 1200mm span with 50x50x4 (and 50x100x4) which was the intent with the support leg to transfer some of the load into the second, lower 50x50, though I don't know if that is better implemented with a deeper beam at the bed (elsewhere on here I did ponder whether transferring a load into a lower, unsupported rail is sensible or not)... but yes, some diagonals to transfer load into the legs.

Yes, I have a metal-cutting chop saw. It was one of those machine-mart moments when you go in with £30 for a hand-held angle grinder and walk out £180 poorer with a Makita chop saw. I don't fancy balancing 6m lengths of steel box on a band-saw table :) (partly tongue in cheek - I'm hoping to source from somewhere that will cut down to a standardised "usable" length - most places do)

[JAZZCNC]

Again can't be arsed to read all previous posts but I get the gist you want to aluminum but there are other areas of design that could be better worked on than building a base which could hold up blackpool tower without flexing.
Also, do you realize just how much heat will go into those joints so closely connected and how much distortion will occur from that heat.?

The frame needs to be strong but doesn't have to be massively built because the cutting forces are mostly lateral. If Was using cutting 250KG blocks of Aluminium then yes go big to stop defelction but doubt you will be.

The best thing you could do to give best cutting capability in hard materials is to build strong gantry and think about how can keep cutter deflection to a minimum.?
Cutter deflection comes from a mixture of Z-axis/spindle and Gantry vibration/deflection. Gantry deflection is easy dealt with by building strong. Z axis/Spindle is not so easy because this impacts on other areas of machine like travels, tool lengths, clearences etc.

If want high clearance for thicker material or long tools then Long z-axis required which deflects more.
Likewise if only cutting thin material then a lower clearance allowing shorter Z axis is possible, but still got to account for tool lengths, clamps etc.
One option is to lift material to the tool and keep Z extension to a minimum but this isn't practical if wanting to cut range of materials and soon gets pain in arse lifting and surfacing bed each time moved.

But with little out the box thinking, there is a better way which allows Minimum tool deflection of any machine setup while still having high clearance. In fact, can have very high clearence and still only have delection equal to tool stickout.?

This is done by Getting rid of the Z axis altogether and raise/lower the whole Y axis. This way the only deflection is from the tool stickout and any deflection in the gantry beam and supporting frame.
Build the support frame and Gantry beam strong enough and then the only deflection can come from tool it's self regardless of clearence and material thickness.
Because your lifting the Gantry beam which spindle runs along fixed to solid back plate you remove the Z axis from the equation altogether so then can have very high clearence and still have low deflection because there isn't a Z axis to deflect only the length of tool sticking out the spindle.

[Doddy]

This is done by Getting rid of the Z axis altogether and raise/lower the whole Y axis. This way the only deflection is from the tool stickout and any deflection in the gantry beam and supporting frame.
Build the support frame and Gantry beam strong enough and then the only deflection can come from tool it's self regardless of clearence and material thickness.
Because your lifting the Gantry beam which spindle runs along fixed to solid back plate you remove the Z axis from the equation altogether so then can have very high clearence and still have low deflection because there isn't a Z axis to deflect only the length of tool sticking out the spindle.

This is going off-thread, but you have got me thinking. Particularly as I've committed to buy a Chinese pack of rails and screws, delivered in the next week or so (as much to get me motivated to get off my arse and do something). I'm not a fan of spending a weekend playing with F360, but, and considering this is a fragment of a design (until I get the bits and validate the generic CAD models I've been using are broadly appropriate), is this the sort of thing that you're talking about, and is this a viable starting point for a design.



Basic design centred around the 400/700/1000 lengths sold cheaply. Steel box for the end cheeks (okay, only one in the images, but you can imagine the mirrored cheek), as well as the gantry. 100x50x4 mild steel box section. Spindle mount on 15mm aluminium tool plate. Some poetic license at this time with the mounting of the rail direct to the box section - but I don't think I need to model the epoxy at this time. Clearly I'd need to slave the Z-Axis drives and source a second 400mm screw but this isn't the end of the world.

I don't want to waste more time in F360 on this design if it's a non-starter, so would appreciate your view if this addresses some of the rigidity issues.

[JAZZCNC]

This is going off-thread, but you have got me thinking. Particularly as I've committed to buy a Chinese pack of rails and screws, delivered in the next week or so (as much to get me motivated to get off my arse and do something).

Well I'm surprised at that.? Thought you'd been around long enough to know better than that.! . . . . Design first then buy later is golden rule.

I'm not a fan of spending a weekend playing with F360, but, and considering this is a fragment of a design (until I get the bits and validate the generic CAD models I've been using are broadly appropriate), is this the sort of thing that you're talking about, and is this a viable starting point for a design.

Yes that's the idea thou wouldn't fasten between those two gantry uprights as it will make setting up difficult with little room for error.
Fasten rails on front, ballscrews on side and put brace between gantry uprights at top. Connect two ballscrews with belt and single motor running along top of brace.
Put rails on front of gantry and ballscrew on top.

[Doddy]

Well I'm surprised at that.? Thought you'd been around long enough to know better than that.! . . . . Design first then buy later is golden rule.

I seriously need to commit or I'll end up chasing squirrels. It's not big, not clever, but it works for me. Also, this is entirely hobby work for me which gives me a slightly different outlook than some others.

Steel is cheap (although my time is precious - to me at least)

Yes that's the idea thou wouldn't fasten between those two gantry uprights as it will make setting up difficult with little room for error.
Fasten rails on front, ballscrews on side and put brace between gantry uprights at top. Connect two ballscrews with belt and single motor running along top of brace.
Put rails on front of gantry and ballscrew on top.

Yeah, was already thinking about the top brace. Wasn't worried about two steppers, slaved, but I suppose rotate the screws 180 and belt across the top - yeah, that'd work and save some complication later on.

Ballscrews on (in)side - I lose a lot of Y travel - struck me during this the depth of the pillar blocks is pretty significant. I'll draw something up and have a ponder. Rails on the front sounds a lot easier.

Cheers - I appreciate the input.

[Doddy]

Yes that's the idea thou wouldn't fasten between those two gantry uprights as it will make setting up difficult with little room for error.
Fasten rails on front, ballscrews on side and put brace between gantry uprights at top. Connect two ballscrews with belt and single motor running along top of brace.
Put rails on front of gantry and ballscrew on top.

Dean,

I've been thinking about this a bit, and been through different iterations of a design for the gantry, as always trying to work out the best compromise. Most recently I've been measuring the width of the shed door opening - I don't want a machine that I have to dismantle to get in/out of the shed.

Can I pick your brains on the following - it does, I think, have problems but I simply don't know if they are problems that I need to be worried about.



My concern is the projection of the spindle substantially away from the rails.

For info: Y-Axis beam is 80x120x5 box steel. Uprights are 100x100x5. Braces and bottom skids 100x50x4 . Sizes chosen on the "big is good" rule of thumb but also with a mind towards availability and ease of mounting stuff. Shiny stuff is 15mm alu tool plate. There's imaginary bracketry involved tbd. Of course there'd be a brace across the top, made from similar big-steel.

The obvious solution for the projection is to place the spindle assembly on the other side of the Y-gantry (this was my previous design), but I have a real-life constraint (the shed door width, and sensibly the amount of real-estate that this can take up) that means I'd seriously compromise the available space in the Y-plane (this current design obviously impacts the X-axis a bit - but I can tolerate that more than the Y).



Kind of shows the impact having the ball screws on the inner face gives me. There's also another issue that the (invisible) top-brace then impacts the available Z-height with the spindle impacting the brace (something I can bodge around somewhat inelegantly by offsetting the brace).

I have looked at replacing the Y-Axis with just a big block of tool plate (20mm), but the deflection calculators suggested a pretty terrible performance by comparison to box section.

My real question is one of opinion - of whether the spindle offset will impact the performance enough to make this design impractical - if you have any thoughts I'd appreciate them.