Very ambitious for a first foray. I like the idea. So let's get down to the meat and bones of it right away.
mikew67 wrote:Hi, folks,
But now, would love to draw on the accumulated wisdom of this group concerning our starting point and how we might proceed. Here are a bunch of specific questions:
- Starting point: I've downloaded all the available info on the Rostock Max V2 (Thanks, SeeMeCNC, etc.). For example, grabbed the Solidworks CAD file and imported them to Autodesk's Fusion360 system (cloud-based, built for group collaboration, free for non-profit use, check it out). Is this the latest published starting point (especially for the CAD files), or is there something later I should be using?
As far as I am aware, the Github is current to the last public revision. There is a V3 in the works, but that is not available nor final yet, and so the Github hasn't got it.
mikew67 wrote:
[*] Mechanical -- Case: No problem here. Expect to use the high-quality rails (aluminum "T" extrusion, plus hardened steel 1/2" dia. rod) for the uprights.
For reference, is it planned to be all metal (Kossel mini with machined corners, Flux), metal frame with other skin (as before, but with perhaps a lasercut body around it), metal sub-frame for kinematic elements with a different main body( IE the towers are connected together with metal, but all the other parts are attached to something else), or other material with metal towers(Max V2).
All things being equal as far as quality, the all metal designs are best for accuracy mechanically, the kinematic frame is sorta in-between (I don't like it as an idea because there's not much room for structural redundancy, and different materials might pull and warp each other), and the Max is a low cost, mechanically sound idea.
mikew67 wrote:
[*] Mechanical -- Platform: Heated? (Expect to do mainly PLA, but allow for other materials.) Material? (e.g., commercial stuff I saw at the Bay Area Maker Faire last month?)
If you plan to do anything beyond PLA, and want a good experience with PLA, I would definitely heat it. How is dependent on the size/volume of manufacture. Silicone heat pads are fairly readily available in a variety of sizes, shapes, and voltages. PCB heaters are more limited, but don't require as much of a bed structure. If you make a bunch, or often have leftover space on PCB orders, you might be able to make your own. I would design it to at least match the stock max's W/in^2, but at something like 24VDC (It makes currents more sensible, and means steppers can go faster)
mikew67 wrote:
[*] Moving parts -- Skates: Expect to design to fit our rails, but what size roller bearings? Have ordered some smaller ones from Banggood ("China, Inc.") to experiment with. Goal: reduce moving mass. Re noise: do I need to coat bearings or put them in a plastic/rubber roller?
I suggest if you want minimal noise and moving mass, the IGUS polymer bearings in the aluminum casing, which would also let you use their aluminum linear rods, which means thermal expansion is consistent over the entire printer. LMU88 is a standard linear bearing size.
mikew67 wrote:
[*] Moving parts -- Rods and balljoints: Expect to use carbon fiber tubing and metal balljoints (ordered samples from Banggood). Specific advice?
I would consider using Seeme's OEM solution here. I've heard that Traxxas rod-ends can have a lot of slop in some cases, and that it does accumulate. If you do go with the Traxxas solution, I would talk to Brian at Trick-laser. He makes most of the carbon fiber arms/rod ends, and has just revealed adapters to connect his arms to Seeme balls. If you want something that's a little more off the wall, my personal favorite of those is the as yet little tested tensioned ball and cup arms, where a steel ball rides in a machined Acetal (or other material) cup, and retained by a series of tensioned strings connected to the carriage and effector.
mikew67 wrote:
[*] Moving parts -- Belts and pulleys: Same as current Rostock?
I would go for the same GT2 belt, although the smallest available driven pulleys that give you a round number resolution (IE, a 16 tooth pulley on a 1.8 degree stepper gives you .16 MM's per full step, or .01 at the microstepping the max is set to. a .9 degree stepper turns that to .005 per micro step, and a 12 tooth pulley gives you a .0075 mm step on a 1.8 degree stepper. figure out what your layer height multiples you want are, and find a combination that gives you that ((teeth*pitch)/(steps/rev*microstepping))
mikew67 wrote:
[*] Moving parts -- Flying platform: Designed to fit selected hot-end, but would like to make it adaptable to other attachments (e.g., use the printer as a pick-and-place system with vacuum pickup head for PCB assembly).
I suggest that unless you have a concrete reason not to, you want to use the same mounting pattern as the Rostock max. It actually doesn't have issues with other hotends with the right (Already designed and available) parts. There are even tool holders for pens/etc. If you find this limiting in size, I would just hang the tool below your effector.
mikew67 wrote:
[*] Moving parts -- Steppers: Same as current Rostock, or larger to allow faster acceleration and movement?
The size isn't the issue with acceleration. Current and voltage are. I would use either 1.8 (Cheap) or .9 (precise) NEMA 17 steppers. Anything larger needs more expensive drivers. with 24VDC, and some amount of heatsinking from the drivers/steppers, the NEMA 17's are more than fine. Standalone drivers, as you mention you want to use, have a tighter current limit than all in one boards, as they sink to the PCB, and there is less PCB mass there than on say, a Duet. They are also intending to primarily sink through the bottom. So if you're designing the driver mounts, I would go for a nice sized hole for a heatsink to connect to, with the driver on the underside of the stick (if you follow normal driver stick form factor), so that you can have a heatsink catch airflow and be actually connected to the driver.
mikew67 wrote:
[*] Moving parts -- Extruder: Would like to use case-mounted extruder and Bowden tube to minimize moving mass. But this may make fiber control too difficult. A floating extruder hanging above the hot-end with a short Bowden tube may be a good alternative. Comments? Given all that, what's the "best" extruder design to use (given that I can machine extruder shafts with sharp 60º grooves and peaks on my mill)?
Low mass, the Titan by E3D wins. it can use pancake steppers pretty easily, and has a number of other nice features. For the absolute best, the Bondtech QR is king. Dual driven hardened idlers wheels, with a fully constrained filament path. But it's class leading performance comes with a comparable price. If you're making it yourself, I would go with a geared dual drive design, and cut your stepper mass as much as possible.
mikew67 wrote:
[*]Thermal -- Platform: Heated? How? Material?
As I mentioned above, PCB and silicone are the best options. You could also make a nichrome (Toaster wire) heat element and bond that to a surface if you want.
mikew67 wrote:
[*]Thermal -- Hot-end: Probably the biggest unknown. What's the current advice on this? Should I machine my own from specialized materials (high-temp plastics, SS, etc.) or order it complete from "China, Inc."? Or start with the latter and evolve?
I very much like the E3D V6. I would neither buy it from China, Inc, nor make your own. The domain experience and machining techniques in the leading hotends put them in a performance class far above most garden variety home-brews, and far above cheap chinese clones.
Prometheus by Distech Automation for most total flexibility, although the nozzle size is annoyingly slow to change compared to an E3D.
E3D V6 is the latest revision of E3D's rather famous hotend, and it's most flexible for day to day use (And the new cartridge temp sensors are a great improvement, you can swap the nozzles and now temperature sensors very quickly and easily, both to fix defects, or to move to more specialized ones for different purposes.
Genuine Hotends.com (No Acceptable imitations or substitutions) J-heads are the kings of PLA. They are temperature constrained, and oft cloned/namechecked, but never beat on their specific use case.
Those are the big three. There are others, like the Pico-hotends, that have smaller user-bases, and other pro/cons.
mikew67 wrote:
[*]Electronics -- Power supply: No problem. Will build into base.
I suggest 24VDC here. the power supplies aren't that expensive, and it makes your steppers/heaters more manageable. I suggest using a stepdown converter to feed fans 12V's though, for ease of sourcing.
mikew67 wrote:
[*]Electronics -- Controller board: The RAMBo board looks like a nice solution, but not for me. Too expensive and not modular. Looks like the best solution is to break down the board into more cost-efficient and flexible pieces:
If you don't like the Rambo, the RAMPS (Modular version of it), or the RADDS (modular Duet) are both modular boards that are available. But the integrated package comes with some very tangible benefits. I suggest the Duet or RADDS board as an alternative for you, as the MCU is faster and more capable, while the boards are cheaper/
mikew67 wrote:
[*] Stock Arduino Mega2560 or compatible (less than $10 from Banggood), and can be upgraded at any time.
Other than firmware for one processor doesn't transfer very well.
mikew67 wrote:
[*] Either one board or a board with daughter cards for stepper drivers (already have) and heater drivers.
As mentioned earlier, the stepper drivers like big masses to sink heat to. Same for the heater mosfets to a degree. I would either sink each one individually, or buy an integrated board. Otherwise this can bite you later.
mikew67 wrote:
[*] Misc. cables and connectors.
[*]Electronics -- LCD: Compatible with Arduino and firmware (below)?
The Smart LCD controller that SeemeCNC ships works fine, the Paneldue touch LCD driver is even nicer (but isn't well supported by the Rambo/Ramps/non-Due based firmware)
mikew67 wrote:
[*]Electronics -- SD reader: Compatible with Arduino and firmware (below)?
The smart LCD controller mentioned a moment ago includes one. The Paneldue paired with a suggested panel and the new beta Duet board also provide another (The Duet .8.5 and .6 can't talk the the SD on the panel, as they lack the pins/hardware support as yet.
mikew67 wrote:
[*]Firmware: Is Marlin the best firmware to use for this kind of printer? (Already have downloaded the Arduino IDE, Marlin repo, and MarlinDev repo.) If so, I can take my firmware issues to the Marlin group....[/list]
Marlin is far from ideal for a delta printer. Repetier offers better EEPROM integration for calibration without re-compiling, but the Rep-Rap firmware for the Duet family of hardware is best for deltas
mikew67 wrote:
Lots of questions
Sorry. Hope they and the forthcoming replies serve to crystalize thinking in the group.
TIA,
Mike