Rostock Max V3 - requests, suggestions, etc.

[geneb]

Build a v2 before you continue to go on about how bad the melamine frame is.

g.

[Eaglezsoar]

Gene, what happened to your Avatar?

[JFettig]

Build a v2 before you continue to go on about how bad the melamine frame is.

g.

V2 is probably much better than a V1 but melamine is not a desired material for a frame of a precision machine. simply moving the machine from one work space to another and it needs re-calibration.

[Glacian22]

Build a v2 before you continue to go on about how bad the melamine frame is.

g.

V2 is probably much better than a V1 but melamine is not a desired material for a frame of a precision machine. simply moving the machine from one work space to another and it needs re-calibration.

I have a V1, which I've moved around my place quite a few times, and it holds its calibration. I haven't thrown it in the back of a truck, but I imagine that's not a typical use case.

[geneb]

I move my machines all the time and I've yet to have to recalibrate unless I've changed the machine on purpose.

Eagle: A power outage last Wednesday killed the power supply in one of my main servers. That machine just happened to run a number of my websites, including the one where the avatar image was held. As soon as the machine is back online, the avatar will appear again.

g.

[626Pilot]

Build a v2 before you continue to go on about how bad the melamine frame is.

Build a printer like the MAX METAL before you continue to go on about how good the melamine frame is.

[nitewatchman]

Build a v2 before you continue to go on about how bad the melamine frame is.

g.

simply moving the machine from one work space to another and it needs re-calibration.

not consistent with my experience

[Xenocrates]

I largely agree with Nite here. Ask mike about the time he took his printer out in a convertible. Especially with the new self adjusting carriages, there is getting to be less and less sources for error that do not automatically adjust.

That being said, a T-slot structural sub-frame would be excellent. It's something I plan for my own printer design, and it makes grounding the towers much easier, allowing them to act as better shielding for the wires (I'm not an electronics guy, so that might be wrong) I would prefer to use 80/20's 30 degree bent extrusions though, to frame out an enclosure, and have a vaguely hexagonal form factor, then tie the towers in to the rest of it with other pieces, as it would allow for their to be room to adjust them slightly, and mean that you aren't depending on the towers for primary structural strength so much, as you want them to be rigid and straight, while loading them down can interfer with that.

[Eaglezsoar]

Hey Nitewatchman, what happened when you took the Max V2 for a ride in your convertible?

[nitewatchman]

Wasn't me, mine just gets moved around the house, occasionally taken to a school classroom or to work. My RMv2 just sees the inside of a Ford Pickup. Biggest problem with this is the dog hair from my Golden Retreater, fans are full of it.

Last convertible I had was a 73' Mustang. Essentially, I ended up burying the Mustang in the front yard. Sold it a few years back and used the money to put in lawn sprinklers. Sold the last Mustang about a year ago, it was a 69' 428CJ Mach 1.

Only thing that I will have to do on the RMv2 is reset the z height when I change temp ranges significantly. Like if I am running PLA at 205C/70C and switch to ABS at 235C/100C, the z zero is raised by .1mm to maintain the first layer thickness I like to run.

[jesse]

Option for e3d v6 w/ volcano hot end. Option for auto bed leveling.

[626Pilot]

To be fair, I don't know what they did with the tower constraints on the v2. If they have the towers constrained from the top, bottom, and all four sides, by the frame itself, then it's not an issue. I would use melamine for the finishing touches, like putting a cover around the electronics and isolating the bottom of the heated bed from fan air in the electronics area.

Edit: Looked at some photos, looks like the towers aren't constrained from the bottom.

[FatedReason]

Hi,

I don't post much on these forums, but I own a Rostock Max v2 I built from a kit about a year and a half ago, and seeing as it is still my favorite printer after all of this time I thought I'd weigh in on my preferences here.

First I think it's worth pointing out why I bought the SeeMeCNC Rostock Max, and in a word, it was "value". I was very happy about the sub $1k price, which included a huge print volume (which has been important to me), and the heated bed. I also thought the delta thing was cool, and I do enjoy not having to wait for Z travels.

With that said, things I have found most irritating were/are:

Delta calibration! It is so hard to get these things dialed in, and if you don't have them dialed in correctly they break stuff, then you have to fix it and then go through that process again. Fortunately SeeMeCNC released new cheapskates which should at least help with getting those set correctly (which having the proper tension is difficult, and can get knocked loose on the old design). However, it is inherently difficult chasing set screws trying to get your bed level (which mattercontrol has helped alot with, though it's not perfect), but then you have to deal with the whole radius issue, and even then you have weird bed/head distances anomalies which occur in different places. I still haven't managed to get my printer to print uniformly from end to end, which detracts from value of it's size when I can't use part of it. Anything that can simplify or altogether eliminate calibration and/or improve accuracy would be a huge win. I have heard that part of the problem with deltas is that arduino controllers are under powered to do the associated math, and that approximations have to be used. Perhaps an ARM based controller would help here, along with some sort of bed detection (for auto calibration).

The heated print bed is awesome! I love how easy it makes print adhesion, and I love the removable pane of glass (so useful). However, I can't really print large ABS parts because of layer delamination issues. Now I completely understand and appreciate the need to keep the costs down, as price was a driving consideration when I was buying (and still is), however I would like a new design to be designed with the idea of a heated chamber in mind. I would also like to see a heated chamber kit be available.

The next issue is really the assembly. It took forever! I would love to see the design be simplified. I would especially like to see part elimination by combining parts into single stamped units (much like PrintrBot has done with their newer stuff) which would simplify assembly as well as increase rigidity and robustness. Perhaps this is cost prohibitive, but in the case of Printrbot it actually proved to be more cost effective than cut out pieces.

To summarize what I'm most looking for are the following three things:
I would like to see print quality go up with the difficulty of getting it go down.
I'd like to see the idea of a heated chamber baked into the design, with a kit to follow.
I'd like to see robustness of the platform improve with a decrease in assembly steps.

On my "it would be nice" list:
More area in the x/y.
24v system for faster bed heating (especially important if x/y increases).
Provision for a second extruder in the design (preferably with wiring included in the original kit, so the second extruder is plug and play ready without need for disassembly), with a kit to follow.

I do want to say, I love my Rostock, it's definitely my workhorse printer, and I'm looking forward to picking up a second one when the v3 comes out.

[626Pilot]

Those are good points.

I read in another thread that SeeMe is looking at moving to a BeagleBone Black-based solution for the controller. I was in contact with the guy behind the Replicape, and I spoke with him about the possibility of bringing my calibration system to the BBB controllers. The conversation trailed off because he didn't have a fully working prototype for his controller yet.

During my research, I discovered that MachineKit's linear delta robot support is fairly primitive. It has no support for having individual tower radius/angle offsets, for example. I think it's also necessary to program for this thing in Python, a language I have zero hours of experience with. If I had a reason to believe it was worth my time, I could add support for those things. However, it seems like Smoothie is a better thing for me to be working on. It's concrete, people use it right now, and it's going to get a lot better with the Smoothieboard 2 coming out in 2016.

We do know that SeeMe is working on an accelerometer-based probe that sits on the effector and converts sudden changes in acceleration into a consumable endstop signal. I would be awfully surprised if they were doing that for the sole purpose of setting the bed height. I think it's more likely that they're interested in using it for some kind of calibration, either dumb Z correction (probe in a grid and use bilinear interpolation to tell the effector to follow the contour of the bed) or a full-on calibration that adjusts the kinematic variables. The latter is better because it makes prints more dimensionally accurate. The former is adequate if you just want to print statues and other things that will never have to interlock with another part.

Bottom line, auto-calibration is an important feature and the BBB doesn't have one for linear delta robots. If they move to that platform without auto-calibration already in place, it will be a waste of their accelerometer probing solution. (Of course, I'll still buy several from their online store because it seems better than all the other solutions I've tried.) I left them a message to ask whether they were interested in going in the BeagleBone direction, because if they had been, it would've been more motivation for me to pick it up. They never called back, so I will keep working with Smoothie.

[U.S. Water Rockets]

Bottom line, auto-calibration is an important feature and the BBB doesn't have one for linear delta robots. If they move to that platform without auto-calibration already in place, it will be a waste of their accelerometer probing solution. (Of course, I'll still buy several from their online store because it seems better than all the other solutions I've tried.) I left them a message to ask whether they were interested in going in the BeagleBone direction, because if they had been, it would've been more motivation for me to pick it up. They never called back, so I will keep working with Smoothie.

It seems to me like going with the Beaglebone black solution would be in everyone's best interest. The AM335x is an extremely capable processor and with this they could really make a printer that not only performed exceptionally well, but also incorporated features that people have come to expect from their "devices". Ethernet, USB Host, Webserver, WiFi, Bluetooth, HD Capacitive Touch Panel, etc. are all things which 3D printers should already be doing, but the roots in Arduino have held the designs back to a great extent. ARM Cortex M based solutions like Smoothieboard are a massive improvement, but they don't come close to the AM335x (ARM Cortex A8) or the level of integration seen in it. The dual Programmable Real-Time Unit (PRU) cores in the AM335x are really the kind of thing that would make a truly great printer. The cost difference between the Smoothie and a BBB based solution is really not that big.

If there were a poll, I would recommend that the V3 were based on something more advanced than a Smoothie (like the Beagelbone/AM335x or the newer Beagleboard X15) and they work with brilliant people such as yourself to develop innovative features like your bed calibration. I hope they didn't write back simply because of the holidays, and you hear back soon. I'd love to buy a V3 with your work built in. That would be a dream come true!