Mega Rostock fiction? Maybe possible! Lets start the game!

[Elysio]

Hi

In the last few weeks I really tried hard to find a affordable printer with a bigger buildvolume than the Max2, esp. in diameter.
Indeed there are some interesting models, but all of them are in a price range near or over the frontier of 10.000$. I think they are worth every penny, but for many of us not affordable.

So why not doing it ourself, as we do always.

My intention is now at the first stage to gather ideas to create a imagination how this fictional printer could look like. Completely fictional specification.

The only frontier is the amount of money, 3000$

Build Volume :
- min. 400mm to max 500mm in diameter ?
- height can be flexible, but maybe it is an issue of rigidity? For me 1000mm shell be enough.

Structural issues :
- high rigidity for the frame and for the arms
- high flexible joints but with minimal tolerance and high stamina.
- Heat chamber
- Extruder right on the print platform for optimal filament feeding
- Heated printbed with sectoral heat areas (energy saving measure)
- Flexible use cases for the print platform (milling, printing, etc...)
- separated electronics

Electronical issues :
- Controller with a higher CPU performance for better print performance
- USV will save your print in case of energy loss
- Pause and resurrection function
- Plane function (to recreate a failed print)
- Stepper motors with higher torque, because aof a heavier printplattform, and for compensation higher lateral acceleration and inertia.
- SD Card printing
- separated electronics
- Open source electronics
- Heated printbed with sectoral heat areas
- Webcontrol

Constructural issues :
- brouder belts to stand the higher torque
- coated nozzle to prevent clogging (teflon etc....)
- Adjusting ballbearings, to keep the belt on position, similar to the old transmission belts in the history of industry (If it works)
- flexible tube to cool the stepper motor of the extruder and the extruder itself in the heat chamber
- If possible the most moving parts shell not inclosed in the heat chamber, to avoid material stress which will prevent loss of precision.
- preciscly guided channel for filament refill near the stepper motor
- Bouden tube shell be inclosed in cooling tube for the extruder cooling to prevent filament weakening

Print issues :
- same precision as the Max2 in x and y, maybe a bit better brought about tighter arm an skates construction ?!
- Same z height capabilities
- Nozzle size from 0.35mm to 1mm
- Printmaterial : ABS, PLA, Nylon etc... most common materials
- 360° print area cooling, for a constant shrinking.

Software issues :
- firmware ???? I dont have any clue about such things
- Webcontrol
- Webcam
- Time lapse shooting

Safty issues :
- fire fighting device, print erea and electrical erea.
- covered moving parts to prevent injuries
- electrical parts covered to prevent electrical shocks
- 12 Volt to prevent life endangering electrical shocks

Allright my brain is empty now.

Maybe you have some ideas or objections, so feel free ....................

[Xenocrates]

Some nitpicks first:
12V is not inherently safe. there are people who have killed themselves with a 9V in a multimeter. 24V is more useful, and gives better performance on steppers and beds, and can use thinner, cheaper wire, and smaller connectors (for a given wattage, rather than a fixed resistance)

Extruders on the platform increases head mass, which slows printing speed and requires more torque, while also placing them inside a heated chamber if you have one. They also make it much harder to do multi-extrusion (which would be a desirable quality in this large and expensive of a machine)

Delta geometries are not well suited to milling, or other high force applications. The lever arms typically used are small (to save mass for higher speeds and push less on the available torque), and larger ones have associated trade offs. If you want to do milling, you'll need rod ends that will handle that sort of load, which pushes the price up to 230$ for rod ends (Three piece high precision rod ends, QA1 precision HMR-6, and SN6-46 high-misalignment bushings). Note that if you want really high precision long lasting arms, you'd be wanting similar rod ends anyways (unless you went magnetic/tensioned, which is not an option for milling) laser cutting/engraving, or wood burning might be doable though.



Now, my input as to how I'd like it designed:
grounded full metal sub frame: More thermally and mechanically stable, and is not a floating ground or an ungrounded shield.

Cable management room: For something this large, serviceability is important. I want room, either in the towers or in dedicated ducts (preferably both) to run wires.

SSR driven bed: High current in the control side is a Bad Thing. Properly speaking, all power systems should be isolated from control (in most industrial systems at least) As this is impractical, and controller boards were designed to drive hotends and can do so well, high power systems such as the bed, and chamber heaters should be isolated.

External air source for cooling: As you have a heated chamber, a separate profiled cooler with an external air source is desirable, as it allows you to have more control over your cooling profile. The Berd-air may be ideal for this, among available solutions.

Beagle Bone Black/Replicape controller: As the only controller which can currently run a touchscreen(or standard display), Octo/Astroprint, the printer itself, and a time lapse camera on a single integrated solution, with extensible firmware written in Python, and the fastest hardware on the market, I think it wins. Not to mention that it costs less than the Rambo (and is more expandable with the Replicape reach that is planned)

Terminal blocks for electrical connections: These printers would not be small, cheap, or particularly useful to people who don't know what they're doing. As such, terminal blocks, preferably DIN rail (to go with a DIN rail 5V PSU, and circuit breakers) should be designed into the upper and lower decks, so that any given run of wire can be quickly replaced, re-routed, or re-purposed, with little more than a screw driver and a set of wire strippers.

Quick change hotends: Hotend mounts and wiring harnesses should both support rapid swap. I'm a great fan of Travelphotog's mounts (713maker.com), and they work for that pretty well. Another consideration is magnetic mounts, such JJpowelly's, or Raymond Ma's (I lean more towards Raymonds, as it's more precise)

Water cooled hotends: With a heated chamber, water cooling for the Hotends, if not the extruders, should be a priority for best results, especially with PLA. Radiators with fans should be designed in, so that you don't just have a large water container somewhere, and so you can potentially run many more things on the same loop (although the pump would need a high enough pressure to handle it)

Now, I've been working on a design for a printer sorta like this (although i've kept to the Onyx bed, as a small run of PCB heaters of that size is cost prohibitive, and for the most part silicone pad heaters aren't available in the size basket). It's not quite a finished design. It's missing things like the various fasteners, the actual laser cutting files for the shell, and finalizing cable management. Current BOM cost: 3432 (at today's exchange rates). It will be difficult if it's possible to get everything you want at the price point you're after, even assembling it yourself, much less as a commercial product (When it comes to engineering this, if you aren't making at least a 50% profit on the cost, you probably are undervaluing your time/effort/logistics, as you will sell few units) Which kinda explains the major cost difference. SeemeCNC, by keeping the cost under 1K, have hit a major sweet spot, and so can do much more volume than something like this, allowing them narrower profit margins, and allowing them to increase bang for buck relatively speaking. (or at least, that's what it feels like to me. It's a damn sight better of a value than almost any other printer)

[Tincho85]

Nothing to contribute to the post sorry.
Just wanted to say that I really enjoy your contributions Xenocrates.

Thanks for sharing.

[Elysio]

Hi Xenocrates

Thank you for your brillant ideas and objections.

Milling is not a "have to have" function. We can discard this feature.

- But high precision rod ends are a !must have !
- Higher torque also, so I think Nema 24 or 32 is the choice. But there is the question which kind of amps should be taken for such high torque steppers? I guess we will need extra Amps between the controller and the steppers.
- High rigidity, such Alu moulds profiles are very sturdy.

http://aluprofile24.de/product_info.php ... hnitt.html

- For the heat bed I have seen many month ago someone who crated a heat bed by himself. He used a selfmade ceramic bed, it was a molded bed with heatspiral inside. This should be adaptable to a sectoral controlled heatbed.
Not very similar to this, but it can be hint.
http://www.soliforum.com/topic/2544/my- ... eated-bed/

[Elysio]

Hi

Okay I really thought that some of you guys had some experiences with stronger stepper motors ?
Any ideas to the higher NEMA steppers.

Beside that, I guess the steppers NEMA17 from MaxV2 is using them with 200steps? Or is it less. Is there a way to increse the steps to increase the precision of the print?

[Polygonhell]

Your trading off high speed performance for low speed performance with larger steppers, they'll also have bigger issues with fast direction changes.
Gecko has a good discussion of stepper motors here http://www.geckodrive.com/support.html, it's the only really condensed description I've ever seen.
When you've read that, calculate the optimal voltage for the Nema 17 on the Max (or any other 3D printer) and you'll discover the vast majority of them heavily under-drive the existing steppers.

Short version is if you go much bigger you'll need to go to better drivers and MUCH higher voltage.

The High torque Nema 23's on my router are running from a 48V PSU and the Nema 34's on my mill are driven from 70V.

The Max steppers are 0.18 degrees/step or 200 steps per revolution, the driver provides "micro steps" between those steps (16 of them) for 3200 "steps"/revolution. However microsteps are not real steps, they are there to smooth motion, the full holding torque isn't available at the intermediate positions (microsteps cannot be "lost" though), and they are not accurate in the same way positionally.
There are 0.9 degree steppers which are 400 steps/revolution, but they less common and tend to be more expensive.

There are a lot of tradeoff involved, the stepper drivers on my mill will run 256x uStepping, but I run them at 4x, because the ballscrew doesn't move very far per revolution, and I want decent travel speeds. Plus you get into the backlash on the axis being bigger than the step seize anyway.

[Xenocrates]

There are several.

.9 degree motors have 400 steps per revolution as a base. Then there's Microstepping (which can go to 256 microsteps per step, but you lose a lot of torque). You can also use a gearbox or something, which increases torque, but decreases speed. You could also use smaller pulleys, which increases torque and precision, in exchange for having less clearance on your uprights. If you want to use that profile for increased rigidity (I personally prefer the 1030 profile, as it has more internal structure that can be used for cable guidance), then you already need both top and bottom idlers.

As far as other motors go, if you want to go up to NEMA 23, you gain 50% more torque fairly easily over the top end of NEMA 17. But most drivers aren't able to supply the amperage needed, at least of the ones commonly used in printing. It makes more sense to me to go with the highest end of NEMA 17, as you'll be able to drive it closer to it's theoretical capacity. (For example, the Replicape's drivers can output 2.5A peak, and 1.2 continuous. the high torque NEMA 17 motors want 2.0A per phase, where the NEMA 25 want 2.8 (276 oz/in) or 5.6 (286 oz/in). NEMA 34 give 434 oz/IN, at the cost of 6.3A. Meaning you start needing external drivers fairly quickly if you want the big motors, and most of the external drivers are expensive. (230 each for 5A drivers), and want at least 24Vdc.

[Reikal]

Should also point out that laser cutting is not an option in delta format either.
You could go with something like a Blu-Ray laser mounted to the effector, but any hope of full on laser cutting requires a fixed laser to target the machining head using mirrors - there isn't a viable mechanism to do this on a delta.

Sticking to 3d printing, this sounds like a great plan!

[Windshadow]

i saw a youtube video of a small cartesian printer with a laser head that was cutting cardboard (I think it mentioned shirt cardboard not corrugated ) so I guess if your needs were fancy paper folding you could implement something of that sort...

if i was going to add a laser I would want to be ablel to cut foam core for making architect model shapes

[Xenocrates]

If one made something like the part-daddy, you would have sufficient mass budget to have the laser tube (including water cooling), and a lens/air assist mechanism. Build your own CNC has had reasonable success with that sort of method. A diode laser is usually fairly low power, but it could take several passes (And diode lasers are both increasing in power and decreasing in cost as time goes on), and is much lighter, smaller, and usually needs only air cooling, or minimal water cooling.

[626Pilot]

Mostly Printed CNC: http://www.thingiverse.com/thing:724999

With this, you can build a Cartesian robot as large as you want. 1 meter cubed? 1 meter XY and 3 meters in Z? Whatever you want. It will probably cost under a thousand bucks. Regular MPCNC 3D printers can be built for about $300-500, depending on how fancy you go with the electronics. The structural elements are the round rail you get from a home improvement store. It's rigid enough for an actual CNC mill, so 3D printing is no problem. I have been thinking about building a large-format 3D printer using this platform.

[Captain Starfish]

If one made something like the part-daddy, you would have sufficient mass budget to have the laser tube (including water cooling), and a lens/air assist mechanism. Build your own CNC has had reasonable success with that sort of method. A diode laser is usually fairly low power, but it could take several passes (And diode lasers are both increasing in power and decreasing in cost as time goes on), and is much lighter, smaller, and usually needs only air cooling, or minimal water cooling.

Given the continual shaking and so on the effector platform is subjected to, I'd not be keen on throwing a multi-thousand dollar metre long glass laser tube onto the effector. Have seen some yout00b footage, though, of handheld lasers cutting through inch plus steel. Given the thickness of the cable running to the torch part, I'm guessing they've worked out a way of sending the laser beam down a fibre without slagging the fibre. Would make a neat switch for a boden tube

[barry99705]

Mostly Printed CNC: http://www.thingiverse.com/thing:724999

With this, you can build a Cartesian robot as large as you want. 1 meter cubed? 1 meter XY and 3 meters in Z? Whatever you want. It will probably cost under a thousand bucks. Regular MPCNC 3D printers can be built for about $300-500, depending on how fancy you go with the electronics. The structural elements are the round rail you get from a home improvement store. It's rigid enough for an actual CNC mill, so 3D printing is no problem. I have been thinking about building a large-format 3D printer using this platform.

That looks familiar...

If one made something like the part-daddy, you would have sufficient mass budget to have the laser tube (including water cooling), and a lens/air assist mechanism. Build your own CNC has had reasonable success with that sort of method. A diode laser is usually fairly low power, but it could take several passes (And diode lasers are both increasing in power and decreasing in cost as time goes on), and is much lighter, smaller, and usually needs only air cooling, or minimal water cooling.

Given the continual shaking and so on the effector platform is subjected to, I'd not be keen on throwing a multi-thousand dollar metre long glass laser tube onto the effector. Have seen some yout00b footage, though, of handheld lasers cutting through inch plus steel. Given the thickness of the cable running to the torch part, I'm guessing they've worked out a way of sending the laser beam down a fibre without slagging the fibre. Would make a neat switch for a boden tube

The fiber is the laser tube.

[IMBoring25]

The designs I've seen for tube lasers rigidly mount the tube and accomplish the motion by moving a series of mirrors.