What needs improvement the most?

[lordbinky]

What I am wondering is what you guys think would be the most beneficial attribute or design feature that you would like improved.

Personally I haven't gotten my kit yet so my imagination is running wild brainstorming all kinds of crazy mods/upgrades and re-inventing of the wheel (mine has blue LEDs so it's better, duh)
EX:

• Inductive heating on a modified nozzle/hot end.
  I've seen it suggested a couple times with little outcome. My main concern is that I would be investing a significant amount of time into a driver board for the induction coils. My intent with that is to provide everyone with a black box to accept the temperature setting from controller boards as a simple voltage input, this way I can design a driver board with a proper feedback system to achieve consistent temps in the element with a quick response.
  Pros
• potentially lighter hot end
• QUICK heating
• finer heat control
Cons-
• ? big investment of my time for the control system on the driver board
• Increased component costs
• The shit I'm forgetting
[size=130]Improved position sensing[/size] (or actual position sensing from what I can tell) The two ways that interest me are optical distance sensors or a magnetic field sensor array. Pros
• Real position feedback - I don't know if the software makes use of this feedback, but making the data available is the first step to auto-correcting
• Less calibrating or more consistent movement
Cons
• The shit I'm forgetting

[626Pilot]

There are a lot of easy things that could be done to make the build easier. The docs are written by people who work with this stuff every day, and can be a bit daunting for people who are either out of practice or just never worked on something like this before.

Hardware

• There are holes suitable for routing the extruder & hot-end wires on two of the three sides of the top plate of the base. Would be nice to have them on all three, so we have more flexibility.
• Include a MODULAR power supply, like this one. These come with pluggable wire harnesses! You get the PSU and a bunch of harnesses in the box, and only plug in the ones you need. Would really do a lot to clean up that rat's nest in the back of the printer.
• Ship the Rostock with pre-made connectors for the RAMBo, rather than making us crimp stepper wires and insert them into housings. It's error prone and it takes WAY too long for the payoff.
• Someone on the forums put together some connectors for the hot-end wiring so that it could be removed without having to unscrew the wires from their connectors. That's a great idea.
• Speaking of the hot-end, give us some nice 8-conductor 16- or 18-gauge wire. While we're at it, the flex tube could stand an upgrade to 3/4" or 1".
• Support mechanism for the hot-end wire harness, so it doesn't flop over and get caught up on a carriage. Perhaps an overhead strain relief.
• Replace the stock delta arm/bearing mounts with something better (there are two alternates I found on the forums, one with magnetic ball bearings and another with carbon fiber arms.) There is too much possibility that you will screw up the hour's worth of sanding you have to do and have to order new parts.
• Ship a glass build plate (rated for continuous high temperature) for the Onyx heated bed. It bows, so you can't build on it without a plate anyway.
• Engrave the diameter of the hot-end nozzle on the nozzle!
• Add an option for all the extras a hobbyist (as opposed to a mechanical engineer who already has all this stuff lying around) would want - silicone RTV, heat shrink tubing, extra screws and nuts, a dial gauge, a digital caliper, pen holder platform, hemostats, Allen wrench set, wire strippers, sand paper or files, a small needle-nosed vise grip (well suited to this build in lieu of a wrench.)
• Any screw that has to be touched during calibration or after assembly, or which can only be got at from a weird angle, should be Allen-headed.
• Make a "maintenance pack" available with common wear items (acetal bearings, etc.)
• Redo the LCD controller standoff design from scratch. It's awkward and it allows the controller to slide around to the point where it falls out.
• The calibration screws don't always raise or lower at the same rate. Perhaps a nut embedded in the carriage would give it a smoother, more predictable threading.

Build Process

• Labeling all the bags of screws and nuts would eliminate a LOT of head scratching.
• Show the finished part at the beginning of the build instructions, so we know what it is we're assembling. Use multiple camera angles if it's complex.
• Refactor the instructions to have zero forward references, e.g. "You might want to do this first..."
• Remove all the redundant steps, such as fastening the heated bed to the Rostock and then removing it later.
• I don't know if they're selling the acrylic kit anymore, but a couple pieces of acrylic designed to be broken would be good to introduce builders to the material. For example, a couple of plates you are supposed to screw together at right angles with a screw and a locknut until one of the pieces breaks. A few exercises like that would teach the builder about the strengths and weaknesses of the material, and how far they can push it before it breaks. (Otherwise, you have to bug SeeMeCNC for replacement parts, which wastes your time and theirs.)
• Some of the instructions seem to imply that you have to bend down the tines on the T-nuts if you have the acrylic kit, but you don't.
• Put up a warning not to rewind hot filament into the Bowden tube, or it will bond with the tube and permanently sieze.
• Add a section on building the extruder drive. There are times when I just don't want to have to scroll through a 30-minute video to try to find something, which brings me to the next point...
• It's 2013. We have broadband, even on our phones! It's okay to upload videos to YouTube in 1080p. :) Some of the instructional videos are so low-res that they make it very difficult to figure out what's going on. I would suggest re-shooting the assembly video in a couple of months, to capture better construction methods and illustrate some of the changes that have gone into the kit since last year.

[/list]

[dbarrans]

The calibration screws don't always raise or lower at the same rate.

How can this be? It makes no sense. Threads don't change pitch.

The only explanation I can think of is the screw head moves a bit because the screw is bent, changing where it hits the end stop switch lever.

- dan

[Polygonhell]

I'd look at software improvements first, but I'm a software engineer. I also think that means a different electronics platform, which is specifically why I am looking at an ARM based solution.

Less issues with movement hickup's
and Auto leveling

Would be first on my wish list.

[626Pilot]

How can this be? It makes no sense. Threads don't change pitch.

They don't, but the screw threads "eat" into the plastic, and the plastic will flex and compress etc. as the screw is moved within it. I have observed this several times during calibration - turn the screw, observe that the head doesn't actually raise or fall.

I'd honestly like to see some options on the carriages other than that stuff. They are so crucial to getting an accurate print.

Also, agree about leveling. I liked the knurled screws I saw on a RepRap, where you could just rotate them a little by hand to adjust the bed level. Granted, the Onyx plate has six screws so that might get a little weird.

[lordbinky]

Ok, so the seemecnc guys are either really slick or my imagination likes to inject numbers into my brain, both are possibilities, but the front page says they're not coming back to the shop until the 27th,which I thought said the 20th last week, either way this means I won't start my 3d printer adventure until june *sigh*. The most dissapointing thing was that with the holiday I ended up getting a freebie four day weekend.

So back to theorycraft for me. I was thinking on how the material warps, and since I don't have personal experience with this stuff yet, so this may be one of my unfeasible ideas that makes things worse.

So time for a game of good idea/bad idea. Would it be useful having a heater ring around the nozzle to prepare the previous layer for the new material from the nozzle? With the idea that this allow for additional cooling to be applied to the print as a whole while countering the delaminating that would occur from cooling the print like that, would this reduce or increase warp?

Thanks for your thoughts.

[MSURunner]

I would be suuuuuuper hesitant to apply more heat to the part if printing in PLA. PLA, generally speaking, doesn't have delamination issues with proper extrusion values. ABS, generally speaking doesn't have too much delamination issues if you allow the natural heat of the print to remain in the area. I find it kind of odd that the Max's hotend seems optimized for ABS when closing off the build environment works best in printing ABS (which would be a PITA with the Max). Now, the radiant heat from the hotend, in conjunction with the heated bed, appears to be enough for ABS prints in the Max, from what everyone else is saying. I'm a PLA guy, so I have no direct experience in the Max with ABS. If you plan on PLA though, I wouldn't do it at all...

[JohnStack]

I like the idea of 8-wire to the hot end. Everyone needs a fan if you're printing PLA. Perhaps a fan bracket or fan kit.

I also like the idea of a standard light kit (although I had fun making mine up.)

Both of these mean squeezing out a few more bucks for @SeeMeCNC - which I think helps them build more great stuff!

And my #1: Let's solve dual extruders!!!! I would love to have the time to solve it but I have to be happy being a PITA instead.

[MSURunner]

I honestly don't think dual extrusion in PLA or ABS is going to be super workable just yet. I think there needs to be another material jump first before we get anything with major precision.

[lordbinky]

What's the problem with dual extruders? The biggest hurdle I could think of is software support.

[Polygonhell]

What's the problem with dual extruders? The biggest hurdle I could think of is software support.

It's mostly a software problem, but all current solutions require the height of the print heads to exactly match, and most drag a second hot nozzle dripping plastic through the layer your printing.
The Dimension machines solve the latter by mounting the heads on a seesaw, to lift the unused head out of the way.
I think because of the Bowden and fast Z travel, the rostock design lends itself to a toolchanger like approach where the unused head can be parked and retrieved.

The existing Slicers only support dual heads for support.
No current slicers AFAIK support printing in multiple colors, and STL files have no support for color, though I believe some professional software uses comments in ASCII STL files to basically add color support. This means you pretty much have to split the models into multiple single color models, slice them separately and combine the results.

[lordbinky]

Thank you, I'm glad I wasn't far off in my mental model. The toolchanger approach you mentioned has the best scaling for the number of heads, but I think something similar to the see-saw idea would be quicker to implement and smooth out the dual-extrusion software wise.

I can't see it being especially difficult to enumerate a color to a head #, which can be handled with trivial changes in software stages. That would let you assign a color for a layer as a whole, which gives a good starting point on handling multiple colors within the same Z slice.

My delivery delay may give me a good chance to dig into the software stuff.

[MSURunner]

The problem with housing multiple heads in a Rostock is the additional weight in conjunction with needing to be level. The seesaw approach could handle one of these but then we have even more weight. The arms are already flexing a bit, so adding more weight and expecting high speed prints doesn't sound like it would work. Plus, keeping a material ready to print like PLA or many of the dissolvable support materials requires it be near liquid, which causes oozing. AMF file format could allow for different colors if it goes that way, but at the moment, you need to have a side script running to interject into the g-code at various points to accomplish dual colors (read major PITA). I know AMF is being researched for Cura though, so maybe it will come. Ultimaker is also working very diligently on dual extrusion, so if they get it manageable in RepRap form, maybe it will diffuse.

[Polygonhell]

I think the primary issue right now is that the value of dual heads doesn't out weigh the inconvenience of actually using them.
2 colors is mostly just a curiosity, and although having separate support has value most prints can be arranged to avoid the need for support altogether.
It's on my list to experiment with, but simpler setup is much more valuable IMO.

[626Pilot]

I've always wanted to see color printing for the Rostock and other DIY 3D printers. I did some research, and there is a guy who's done some experimenting with mixing two colors at different ratios in one hot-end (it has two input ports and he has a small motor mix the plastic inside the head.) The results were okay, but not great.

I think it would be better to use the whitest plastic you can find, and rig up something that would take cyan, magenta, yellow, and black (CMYK) ink and mix it right at the end of the nozzle. I'm not sure, but I think that's how 3D sandstone color printing works. I couldn't find any information on how those printers actually work, but they are full-color and 3D. You can get models printed by websites like Shapeways for a reasonable price, ten to twenty dollars US for a piece a few inches tall, so it can't take THAT much ink. I think if something like this could be done, especially with a delta-arm platform, it would be revolutionary. I'd even be willing to fund research, if anyone has the know-how to make something like that happen. The ink could be fed from small jars that would ride along with the head, which wouldn't weigh much, or they could be fed overhead from very small diameter tubes. I thought about just using CMYK color printing cartridges and having them only print color to the top "pixel" on the cartridge's print head, but I don't know how I'd make that fit next to a hot-end head.

[JohnStack]

This was a six head prototype by someone who showed up at Makerfaire. He was so caught up in BS (sorry) that I took six or eight photos of him waving around this head.
[img]http://forum.seemecnc.com/download/file ... &mode=view[/img]

Also, in terms of interchangeable heads, Hyrel has a cartesian that's head-plug and play.

Photo of it (not a very good one) here:

[img]http://forum.seemecnc.com/download/file ... &mode=view[/img]

Each head just slides over a key and by doing so, plugs into an ipad-like connector. The software auto-recognizes the head. Pretty impressive.

(Met the product designer, a nice guy. Not affordable for the individuals)

[Flateric]

I've run dual extruders for well awhile now. It's not really that big o dealio really. Just takes a little more setup, yet a little more calibration and tada your there. It's not magic, just fairly logical.

I find it most useful for combining black with clear etc. can't mix mediums of things you want to combine obviously.

[cambo3d]

i was wondering what that was in your photos, 6 heads. looks purdy though.