Big-ish router concept review

Hi,
I've been lurking for some time, this is my first post. I'm considering building a cnc router from scrap parts I've collected over time and I'd like to get comments on the concept, frame mostly - it's a bit different than what I usually see around.
This machine would be used for wood and aluminium work, routing and 3D. Travel area should be 1250x2500x200mm. The idea behind is to use welding and machining only where absolutely necessary, and instead use epoxy putty and bolts for connections and self leveling epoxy for getting everything planar. I'd also like to avoid complicated solutions (like diy rotating or diy dual ball screw nuts) if I can.
Attachment 21426
1.8kW/24kRPM square spinde would be mounted on "U" section built from bolted alu plates (25mm back, 10mm sides, sides can have extra supports), 460mm SHS15 rails are bolted on the back (Z bed). This assembly would slide through 3 blocks per rail, mounted on 25mm alu plate (X bed) which could also have uprights making "U" section. Z axis would be driven by 400W(3000RPM/1.27Nm) AC servo through 20:30 10mm AT5 belt reduction to 5mm lead ball screw (200mm travel). I haven't figured mounting yet. Total Z sliding mass would be ~25kg, and no counter balance would be used.
X axis would be assembled from 1460mm LH20 rails, and 4 blocks (long, if I manage to repair/reball set of preloaded BN blocks - few balls fell out due to bad packing/shipping and now the blocks jam), and would be driven by 200W(3000RPM/0.63Nm) AC servo through 20:60 10mm AT5 belt reduction to 2020 ball screw. 1250mm travel leaves 210mm width for the X bed. Ball screw will be mounted on the front near lower rail. Total X sliding mass would be ~35kg.
Gantry would be 250x6mm square steel section leveled with epoxy, and instead of welded rail mounts I'd raise X bed on blocks if required. Section would slide on 20mm HG/LG rails with 3 or 4 blocks per rail (400-500mm support length). I have to figure out best solution for the support (currently 25mm alu base plate with 20mm triangle plate), but it should also be bolts/putty solution.
At first I thought implementing rotating nuts for driving the gantry but now I'm thinking that I could get away with rotating ball screws and avoid extra complexity. Total Y sliding mass would be ~120kg, and it would be driven by 2x 150W(1120RPM/1.27Nm) DC servo through 20:40 10mm AT5 belt reduction to 3220 ball screw. Rails would be mounted to epoxy leveled 250x6mm square steel section 3m in length. This again would be bolted/glued to 2 welded "U" sections made of 250x6mm square steel which would allow rough leveling of the rail supports (epoxy putty + leveling screws). "U" sections would be filled with sand or concrete and capped, long rail support sections probably too.
My main consideration is unsupported overhang of Y rail supports (650mm) which would probably vibrate and resonate.
I'm considering UC400ETH controller so I can handle high encoder rates.

What do you think?

Thanks,
Mihael

ps. If someone has experience with repacking balls into NSK LH preloaded blocks, I would appreciate any advice.

Quote:

---

Originally Posted by mbele

What do you think?

---

I think your servos are too small. I think that you'd want at least 400 watt everywhere. And the price difference is minimal to go from 200w to 400w
I thing a 10mm belt is too small.

Why are you mixing ac and dc servos?

Quote:

---

I'm considering UC400ETH controller so I can handle high encoder rates.

---

What do you plan on doing with the encoders? You didn't mention if you'd be using Mach3 or UCCNC, but neither really does anything with encoders.

Hi Gerry, thanks for looking at it.

Quote:

---

Why are you mixing ac and dc servos?

---

I would like to use parts I already have (I'm missing Z and Y ball screws assemblies only).

Quote:

---

What do you plan on doing with the encoders? You didn't mention if you'd be using Mach3 or UCCNC, but neither really does anything with encoders.

---

I should've put it differently: high pulse rates. Eg. Z motor generates 3000RPM x 10000ppr(4x2500) / 60s = 500kHz pulse rate is needed to spin the motor at highest RPM with highest resolution. UC400ETH should be able to handle 400kHz.

Rotating nuts are the correct choice for that size. 2020 screw is small for the long axis / according to me/ . I used on mine 2510 and rotating nuts. 400w servos are just the right choice in that case.

Bolts? Mehhh, for that size. It will be difficult enough to level everything and bolts will lead to the everlasting suspicion if sth has not moved. You can use bolts and pins, but anyway, my advice is weld it.

1.8 kw is small for that machine. 2.2kw as minimum, 3kw better.

Maybe read my second build log so you will get all the info there on 2600x1300x200 machine.

Again- for such a big machine to be precise, its not a small feat. Dont start the build before informing yourself. You will see that i have spend 3x 200 euro on epoxy before i learned 100% how to make it properly. its one thing to think that you have made it properly, its another thing to compare it against 2m straight edge and dial indicator.

There are only 700kg in my machine frame and it will stand on 2 or 3 legs without any distortion or twist / less than 0.05mm measured/ . Gantry weights 180 kg and flies 20m/min at 3000cm/s2 accelertation. Its not absolutely necessary to make the Z like mine overbuild, but gantry must be similar for perfect aluminum finish.

That frame you show is at the most for foam and wood. Not aluminum. Its incredible how shaky and bendy become things the longer they are. a piece of steel that 2 people can not bend at 1m at 3m you can bend with one finger.

Hi Boyan, thanks for the reply!

I'm still at the concept phase...

For the long axis, I figured that 3232 screw pre-stretched between 2 BK mounts with 20:48 reduction could maybe work, screw would turn 330 rpm to get to 10m/min (i think that's below critical speed), and 2048ppr encoder would then allow 0.001628mm/pulse.
It's just that I'm trying to get by with the low rpm/high torque motors I already have (I would also have to reduce moving weight). Rotating nut would be better but more complicated and probably more expensive solution. And I'll probably end up doing it "your" way after all, but I want to check all the options before investing.

I saw bolts and epoxy putty(eg. jb weld) leveling method in "principles of rapid machine design", so I assumed it worked fine.

1.8kW is what I already have, and I haven't ran into problems not having enough torque (it gets to max at 12kRPM).

I read your build threads, and I was amazed :) Thank you for sharing all the info, it's invaluable resource. Without the epoxy leveling solution you've adapted for your machine I wouldn't even think about making something this big.

Since the space is not a problem, and machine will not move, I'm thinking about pouring concrete foundation(frame) and ditch the steel frame completely, so there will be no overhangs and unsupported parts at the base.

I'm also considering aluminium gantry to reduce weight, same outer dimensions but with thicker walls and ribs, made of 250x10 plate screwed and glued together.

I will draw the complete machine with all the features, and when/if I'm sure and happy about all the details, I'll buy what's needed.

Thanks again for looking at it!

I think you are mistaken about the spindle. Ideal speeds with most of the most popular bits are 16k-24k.

Quote:

---

Originally Posted by Boyan Silyavski

I think you are mistaken about the spindle. Ideal speeds with most of the most popular bits are 16k-24k.

---

Again, I should have been more clear in what I'm writing... sorry. Spindle reaches max torque at 12kRPM, but it can spin up to 24kRPM:
https://www.damencnc.com/en/milling-...rospindle/1226
It works well with both alu and wood. With constant TEA paths and smaller TiAlN coated endmills it has enough torque for light work in steel (min. 6kRPM), but I don't do that often since the machine is not rigid enough.