Large Delta Build help needed.

Good afternoon everybody and since this is my first topic on this forum, let me say happy prints to all of you.
I am planning to build a large and fast delta, where i can print from detailed orange sized prints to large plant pots.

I have small experience with 3d printing (2 months) and own two deltas, a Geetech G2S (royal piece of junk) and an Anycubic Linear plus (excellent kit for the price).
On the contrary, I have great experience in rc building (as i imagine many other forum members) and especially helicopters, from 150 size to 800 size, so i am confident about the quality of this built when it comes to fitting, mechanics, wiring, precision etc.

Now to the built and my questions to all of you knowledgeable people...
My target print volume is roughly base 40-50cm diameter and 80-100cm height.
I really dont understand how is everybody measuring speed on a printer, but what i would like to get is decent 0.3mm layer results at speeds of 100mm/sec (outside perimeter speed).
I think i will go with 2060 extrusions and linear rails, carbon square profile rods instead of tubular (i need opinions on this), and hopefully someone can guide me where i can find frame corners for the 2060s. Are the 2060s overkill or should i go bigger?

I will appreciate all comments and arguments, but i dont need cartesian vs delta arguments, in my mind the delta setup with the small moving masses just makes a lot more sense.

questions:

1. could someone point me to a write-up about the theory and calculations for rod lengths and spacing vs build volume and frame?

2. what size, step and power motors would be right for such a setup?

3. what type of extruder would be up to such a task (and what kind of stepper for it)?

4. Should i go with magnetic effector or ball joints and why?

5. whats the minimum psu i would need?

6. Are normal belts and pulleys up to the job or do i need sth more specialized and where could i find these if needed?

7. Most important of all...which board?? Open source of course, 32bit of course, it has to be up to the job, super quiet drivers, upgradable, fancy trinkets like super high end lcd screens dont fascinate me, but wifi connectivity would be nice. Also easy firmware alterations would be nice, but i think all 32bit boards have that..

Now the problem with the boards, is that i still dont know the difference between all that RAMPS, BAMS, BOOMS and the rest of the jargon, but i want a board that cuts it, and probably with external drivers so i dont have to throw the whole board if a driver fails (not sure if burned drivers is sth that happens with good boards, i need your feedback on this).
Also please keep in mind that the board has to be sourced either form a vendor located in the EU, OR a vendor that is willing to ship it as regular personal mail (otherwise i get hacked by customs, almost 100% markup on the purchase price)

8. temperature control sensors?? no idea what to use...

9. Endstops? Is it important to go all exotic on them?

10. which hotends can handle this load? I plan to use mostly pla but would like to be able to print all materials.

11. What are the largest heatbeds one can find for such size? I suppose it has to be 220v? Who sells them? I will be printing on 5mm mirror

You have probably noticed that i dont mention my budget, and this because i plan to use this printer commercially, so i am not going to cut corners on it. This doesnt mean i am willing to overspend on it, but i do want a machine that will work flawlessly for a long time once its setup. And if it does i will be building more...

Everybody's help, argument and opinion is welcomed, but PLEASE speak out of personal experience.
And you dont have to answer all my questions at once, just the one's you are certain about.

I apologize for the wall of text
and i am in advance thankful for everybody's help in the matter.

My suggestions:

0. Forget square section CF arms, round ones are optimal. For a printer of that size I suggest 8mm diameter minimum.

Have you considered using 3030 extrusions? Robotdigg sells vertices for them.

1. Choose rod length so that the rods are at no less then 20 degrees to the horizontal when the effector is at the opposite side of the bed. This means the rod length typically needs to be 1.1 to 1.2 times the bed diameter. For example, I use 360mm arms with a 330mm bed (of which 300m is printable).

Rod spacing: wider is more stable. I use 55mm.

2. 48mm long NEMA 17 motors will be adequate. For best precision use 0.9deg ones.

3. Up to you. I use Bowden but you can also go with a flying extruder or RDD.

4. Yes, buy the rods with Delrin ends from Haydn Huntley.

5. With a mains voltage bed heater you won't need much power from the PSU. A 24V 120W Meanwell will be sufficient.

6. Normal belts and pulleys will be ok but consider 9mm wide ones if you can find them.

7. I am biased, but nothing matches the Duet for this application. Take your pick between the WiFi and the Ethernet versions.

8. PT100 for the hot end. Thermistor is adequate for the bed.

9. No, microswitches are adequate. But you need a good Z probe. Consider using the Duet3D Smart Effector (again, I'm biased, I designed it).

10. To support a wide range of materials, I suggest a genuine E3Dv6.

11. Both Keenovo and Shenzhen Ali Brother Technology will make a silicone bed heater to your size, power and voltage specifications as well as having a wide range of standard sizes. Choose a heater a little smaller than your aluminium bed plate to leave room for fixings. Aim for a power density of about 0.4W per square cm.

More at [miscsolutions.wordpress.com].

Edited 3 time(s). Last edit at 05/24/2018 05:00PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Wow, that was fast and helpful.
ok some questions
0. I guess u mean 8 mm solid rods? I dont think i can find 3030s around here and i am not sure if they would be stiff enough
8. do i need an extra board for that or it goes straight on the duet?
The rest is clear,
Thank you dc42

I used 2550 extrusions for a meter high delta with 310mm diameter bed. Printer is all metal, no plastic. Top and bottom are torsion boxes which makes the printer extremely rigid.
1. Rod length will depend on whether your bed is inside the horizontals between the towers or extends outside. 60/20 degree rule applies either way.
9. Optical endstops will be more accurate than microswitches.
11. Keenovo also sells genuine Fotek SSRs.

Quote
GreekPilot
Wow, that was fast and helpful.
ok some questions
0. I guess u mean 8 mm solid rods? I dont think i can find 3030s around here and i am not sure if they would be stiff enough
8. do i need an extra board for that or it goes straight on the duet?
The rest is clear,
Thank you dc42

I meant 8mm CF tubes.

To use a PT100 with a Duet you need the PT100 daughter board, see [www.duet3d.com]. There is also support for PT1000 sensors, which don't need the daughter board, but the accuracy and resolution is not quite as good as for PT100.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I use 10mm hollow CF tubes with Haydns Delrin Mag-rods. It's a perfect fit and strong enough. There are only push/pull forces on the arms, almost no sideload. IMHO solid CF rods are overkill.
You could use robotdiggs 2040 vertices with linear rails. When you mount the delta on a solid baseplate, it should be stiff enough.

One problem with big diameter Deltas hasn't been solved yet ( AFAIK ):
The carrier speed could be much higher while printing in the center of the bed, but for wider parts you have to reduce the overall speed. That makes printing wide parts extra slow.
I wish there would be a variable speed parameter or at least a two-zone setting ( inner zone vs. outer ring )
Deltas are good for tall prints, but wide parts: not so much.

About belts: Whatever vertices you'll use, make sure the wider pulleys will fit ( 9mm pulleys ).
To reduce ringing from belt stretch you could replace almost 50% of the belt with something less stretchy, like fishing line. That's still experimental, but sure worth it for an extra tall Delta.

Edited 2 time(s). Last edit at 05/25/2018 03:03AM by o_lampe.

ok magnetic rods it is and duet wifi decided...

i will go with 9mm belt and pulleys, also Zesty Nimble seems like a great extruder.

Genuine e3d Volcano v6 hotend with tip set to print all materials and layer heights.

PT1000 sensor to start with, still undecided about endstops but i dont know if its that crucial one way or another.

Not sure i really need a z-probe though, i have achieved far better leveling results manually on both my deltas, but if you think its a necessity on a large bed i will consider it.

Custom heatbed is the only choice i guess for a 500mm diameter plate.

I am still not sure what kind of extrusions i will source locally but i will probably know by monday.

The fishing line thing seems interesting, and googling for it i run into the cerberus delta design which seems really interesting (especially because it uses printed corners and the theory behind it makes sense) If anybody has any experience with sth similar, please chime in.

Are you guys sure the nema17 long steppers will be enough for a 2 meter delta? Does size of the motor has anything to do with motor speed in steppers? i have no idea.
Also what kind of stepper would you use on the Zesty Nimble, that could work both on 1.75 and 3mm filament?

Thanks again for everybody's contribution in this

You definitely want a Z probe, it makes calibration a doddle. Choose one that responds to the nozzle contacting the bed, for example Smart Effector, Precision Piezo, or FSRs under the bed supports.

The size of the motor needed depends on the mass you are trying to move. The extra mass of the belt for a 2m high delta won't be much more than for a 1m high delta, which is why I think 48mm long nema 17 motors will be adequate - unless you expect that your carriages or effector will have a lot more mass than normal. Larger motors also have higher rotor inertia. My 1m high delta uses 48mm Nema 17 motors, but I only need to run than at 60% of rated current, so there is plenty of torque to spare for a larger printer.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I have something similar under construction. I don't understand the recommendation for 24v power supply over a 12v.

Edited 1 time(s). Last edit at 05/25/2018 04:13PM by cwaa.

more volts is always better because you get smaller currents for same wattage, so things run cooler

Niiiice, is it solid enough like this? Do you have a build thread somewhere?

One more equipment question guys, what kind of filament finishing alert or module should i use?

Quote
cwaa
I have something similar under construction. I don't understand the recommendation for 24v power supply over a 12v.

24V power allows higher speeds, especially when using 0.9deg motors.

Edited 1 time(s). Last edit at 05/25/2018 05:38PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Why does the 3d printing community specify stepper motors by their body length instead of torque and current?

1. Frame
If you want to stay on a delta build try this Page to simulate the printer to get the right rod lenght . even with 2030 extrusions the fram may not be rigid enough. 1m build high means roughly 1,6m total height..so maybe do a Hexagon build?

Board: as DC42 i vote for the duetwifi..used it for my delta too and no board offers better support in case of problems. beside the excellent quality of the board. Also use the smart effector. Sure..a perfect mechanical build would make manual calibration easier but with that build space it can be real pain in the ass.

Endstops: use simple but good quality mechanical endstops. They are not affected by Sun..bright light at workspace ( like optical could)..magnetic fields..temps ( like hall endstops)..they simply work. Guess here is also some threads about problems with high tech endstops that the trigger point can fade over some tenth of a mm because room temp etc changed.

Heatbed: I use a 300mm dia 8mm casted aluminium bed. fortunatly you can buy that size already cutted. In you size you shoud take a casted plate and let lasercut it. after that kenovo should be your heat supplier.

all the rest simply use DC42´s answers.

Thanks barracuda. Well I changed my mind about the size, going with 1200mm 2040s for the verticals and 500mm 2060s for the horizontals. Will try to make the frame as rigid as possible with any add-ons needed. If I manage to produce a super rigid frame, I will then scale it up on my next build. I was thinking of not using a metal bed at all. Mirror-keenovo HeatBed and cork is what I'm thinking..
And yes I'm going with Duet - smart effector--volcano combo.
All part have been ordered, just waiting for them to come in and start building

hmm how thick is the mirror tile? many user uses casted aluminium beds 8mm and thicker..because the heat will spread better. Sure you can heat up a thin mirror tile quite fast but..if heater pad turns off, the tile will get cold also very fast..together with that usual thin tiles ( 3-4mm?) you will get bad prints. Will look like you have Z-wobble. Take a look at other big delta builds..

I havent actually checked how thick a mirror i can source, but isnt glass actually slower to dissipate heat, since its a worse conductor than aluminum? Slower to heat up and slower to cool down?

yes..glas is a good insulation...but..as a glas or mirror bed is MUCH thinner theres less material to store the heat.
At least its your decision..but there is a reason why most if not all big printer uses a thick casted aluminium plate

There's another parameter you might want to consider: thermal capacity. The mirror has a lower thermal conductivity, but with a (much) lower thermal capacity it would cool down faster in the end. I'd be afraid, it'll crack without a solid alu-heatspreader underneath...

i will keep your points in mind and look into them. as for why all use a cast alu bed? It could be that cast aluminum is cheaper than boroscillate glass? dunno!

One thing I'd say with this, looking at your desire to print outside perimeters at 100mm/s keep your effector as light as possible. I'm using smart effector on 360mm rods, same sized delta as David's, but adding watercooling meant I had to slow down the machine a fair bit to prevent ringing/overshoot on the corners. I'm also using Nimble extruder, it's light but the cable has some inertia like the watercooling tubing.

If I tried to make a really fast delta next I'd go for a pivoting bowden extruder with capricorn tubing as short as possible and some type of micro hotend (deltaprintr mini, microswiss etc..). Lightweight fans no massive 50mm blowers. I'd not make the machine too big, theres too much moving mass if the machine is large. I quite like the new approach going around with the hylite folded material that looks really light, but its very new so maybe it does everything claimed or maybe not.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Quote

I quite like the new approach going around with the hylite folded material that looks really light, but its very new so maybe it does everything claimed or maybe not.

Whatever it does, you need good machinery to engrave hylite. It sounds like M_Xeno made the whole spider in one piece. ( but I didn't follow the links he provided )