Poll: Which 3D printer controller do you use, or plan to?

[bvandiepenbos]

@626Pilot
I do wish you would reconsider your decision to abandon the smoothie platform. You have been a valuable contributor to the smoothie project. Nothing wrong with looking to the future and testing new options and ideas, that's what makers/open source is all about. But, you have so much invested it is a shame you are jumping ship due to your bad experience with one person.
I can respect your decision whatever it is.
Actually your poll of platform choices prompted my reply.
Very interesting and informative.
It has sparked good discussion.

[JFettig]

I absolutely believe Machinekit can be done as seamless install and operation, all the pieces that are scattered all over just need to get put together. It has brilliant interfaces that if worked easily and properly, could be awesome. Control it from computers, phones, tablets, etc. Configuration could all be simplified and everything, but working with the Machinekit or linuxcnc guys will never get it put together, it'll only get forked 10 million times.

I agree, I'd love to stick with smoothie, its simple and works well. Maybe port another board that is possibly better and more capable, more defined and less scattered.
That is what I like about some of the Chinese knock-offs actually but I cannot bring myself to buy one.
MKS SBASE is the one I'm thinking of.

[jmpreuss]

Also, the Smoothie-compatible boards have pulled ahead of RAMBo, etc. in the poll. That's impressive. When I started the poll, I thought almost everyone would answer that they had an 8-bit controller. SeeMeCNC, if you're reading this, you might wanna talk to UltiMachine about their RAMBo successor. Survey respondents are self-selected, so this is not a scientific poll. However, there are enough samples that it's increasingly hard to ignore. There's also the "do you want to upgrade" question - seven yea, one nay. People seem to want 32-bit controllers!

I talked to Oly at the grand rapids mini maker faire this past weekend. It seems pretty apparent that seemecnc has put some serious effort into teaming up with bbb to be the next generation controller for their printers.

[626Pilot]

@626Pilot
I do wish you would reconsider your decision to abandon the smoothie platform. You have been a valuable contributor to the smoothie project. Nothing wrong with looking to the future and testing new options and ideas, that's what makers/open source is all about. But, you have so much invested it is a shame you are jumping ship due to your bad experience with one person.

They don't have the after-sale support I need, or the hardware I want. I would have left the Smoothieboard 1 behind for its successor anyway. The RAM limitations make the platform harder to develop for than it needs to be, and people have to turn off their LCDs, Ethernet, etc. to run my calibration. It's not a good user experience.

I talked to Oly at the grand rapids mini maker faire this past weekend. It seems pretty apparent that seemecnc has put some serious effort into teaming up with bbb to be the next generation controller for their printers.

Reeeeeeeeeeally! Would be interesting to hear their take on this. I have a feeling that a lot of people will follow their lead. I might be one of them. I haven't made my mind up because neither the BBB or Duet stands out as the clear winner. Each has a weakness that the other does not. They're still neck-and-neck in the polls, too. If only there was a clear sign, such as adoption by a serious player like SeeMeCNC.

[bot]

In your eyes is the weakness of duet its limited CPU speed and limited memory?

[626Pilot]

The CPU speed doesn't bother me. The memory does. It has 64+32K RAM, which is more breathing room than Smoothie's 64K. However, it leaves so little room for future additions. I like the challenge of programming on embedded processors with limited resources, but at some point the challenge stops being interesting and starts being an unnecessary pain in the ass.

In other news, here is Tom's review of the BBB+Replicape solution. His verdict: Way nicer to just use the Angstrom image (rather than trying to install a bunch of crap on top of Debian), but not enough features (like autolevel) and it costs way too much. However, he does point out that you get what you pay for - the electronics are very well-designed.

[youtube]http://www.youtube.com/watch?v=XiHoCD9oeeE[/youtube]
The BBP is roughly half as expensive, so it's possible to put together something nice for less money. If SeeMeCNC is looking at a BBB solution, my guess is they'll hand out an Angstrom image rather than expecting their customers to plumb MachineKit into the existing Debian OS. He said that was WAAAAY too much of a hassle, and didn't work all that well in the end. Angstrom was bone simple - he flashed the image, and it let him configure everything (at least, what there is to configure) by hand.

[bot]

I'm going to be doing a smoothie vs duet shootout soon, and publishing the results. I'm liking Duet a lot right now, but as you say it might not be "future proof" -- then again, look at how far 8 bit has taken us! I'll be doing the tests by swapping the boards into this machine: http://imgur.com/a/mgFyH

[Nylocke]

If there was a Smoothie equivalent that was designed and built as well as the Replicape, I could justify purchasing it one way or another. I have a weakness for well built/engineered things.

[626Pilot]

If there was a Smoothie equivalent that was designed and built as well as the Replicape, I could justify purchasing it one way or another. I have a weakness for well built/engineered things.

I wouldn't hesitate to spend $250 on a controller if it would last forever. The Replicape has extra components that aren't "must-have" necessary, but which protect the circuitry and allow it to run more efficiently. The insulation on two of my part cooling fans wore down over time, and last week I wound up blowing the fan FET. I have to get a new one and solder it in. The Replicape has a standard push-fit 20A automotive fuse, so if that happens, presumably you just have to get a new fuse from any automotive store. (I think I'll protect the new fans' wires with heat shrink and superglue.)

[626Pilot]

I left a message on SeeMeCNC's voice mail asking about the BeagleBone. Hopefully they get back to me.

[ZakRabbit]

Have you looked at CRAMPS instead of the Replicape? Any thoughts between the two?

[626Pilot]

CRAMPS: 6 axes, micropots.
Replicape: 5 axes, digipots.

I like digipots, but I need six axes. The Replicape doesn't seem to be available from the vendor specified on the RepRap wiki. They claim to be working on revision B, and there's going to be a Kickstarter, etc.

I don't know what to think about that. They stop production before the successor is even funded? Doesn't make sense. Maybe someone else is selling them. The Reach expansion board adds another three axes, but it doesn't seem to be available for sale - just "review."

[bubbasnow]

https://groups.google.com/forum/#!topic ... aYZ0jUaGkY

https://github.com/modmaker/DECAMUX

[626Pilot]

SeeMeCNC never returned my call. They have said in the past that they don't like spilling the beans on stuff in development in case they have to change course. I guess we'll have to wait and see.

If I decide to go with the BeagleBone/MK platform, that platform will get full auto calibration (including autolevel), and I'll adapt the segmentless delta calculation idea from RepRapFirmware. This might involve modifying the PRU code, but I've developed in assembly language before. Segmentless positional calculations actually require keeping track of fewer variables. The segmented solution is better for controllers that run at 16MHz. With a fast CPU that can handle floating point math without choking on itself, it's not as big a deal.

I'm leaning closer to the BeagleBone solution at this point, especially if we can distribute a single image containing the OS and MachineKit so that the end user doesn't have to care whether it runs Linux or not. Developing for a processor with 96KB RAM seems like... not saving for retirement. 96K is barely enough for today, and maybe not even that. It seems more based on convenience to developers who only know how to write Arduino code, than a realistic consideration of what people will want from their firmware in the coming years.

[bubbasnow]

there is an snapshot image posted here

http://elinux.org/Beagleboard:BeagleBoneBlack_Debian

[geneb]

So after all the research, what's the consensus on which cape to use with the BeagleBone?

g.

[bubbasnow]

How many drivers do you need?

[JFettig]

I did choose CRAMPS for a reason - I like the Pololu drives, I can choose the drives I want to use vs being stuck with what someone wants you to use. Its also set up for 3 extruders(and 3 axis) where most others are naively set up for 2 and 3 plus its very basic. Its also available from an online store for a reasonable cost!

I'm really excited to see what you come up with, I'd be more than happy to share my opinions but I won't be of much help for actual programming.

[626Pilot]

Got a little question about these BeagleBone capes, namely that they seem awfully hard to find. I couldn't find anyone actually selling the BeBoPr++ or CRAMPS (except for bare PCBs), and the Replicape is listed as out of stock because they're getting ready to kickstart revision B.

Am I incorrect about this? Is someone actually selling the BeBoPr++ or CRAMPS, in an actual online store, and not "send me an email to join a group buy" type stuff? For me to be able to take the BBB solution seriously, I have to be able to buy all the parts from an online store. If that isn't the case right now, then the BeagleBone hardware producers have made my decision much easier.

[JFettig]

http://pico-systems.com/osc2.5/catalog/ ... ucts_id=35

[teoman]

Bbp 1s can do 3+3 axis.

It has 2 thermocouples and 3 thermistors.

I should be getting one soon. Ill report back once i get it.

[626Pilot]

http://pico-systems.com/osc2.5/catalog/ ... ucts_id=35

That's more like it. Are you using a display?

Bbp 1s can do 3+3 axis.

Do you know when they'll start selling to people who weren't in on the kickstarter? Also, are you using a display?

[JFettig]

You can use a display just as you would with windows or linux via HDMI output, I used one when I was setting it up then I'd use a different PC and log into it to run it. There are many ways to do it.

[0110-m-p]

Replicape Rev B2 is now up on Kickstarter with a November delivery...

https://www.kickstarter.com/projects/19 ... ontrol-boa

[626Pilot]

Replicape Rev B2 is now up on Kickstarter with a November delivery...

https://www.kickstarter.com/projects/19 ... ontrol-boa

I have been talking to Elias Bakken about this. He says they have a "Reach" board planned that will add more extruders, but it's still in development and he doesn't know when it might be available.

Whatever I use in production has to have at least six stepper drivers. I told him about this thread, and shared some challenges people have had with the BBB (there's no one-stop shop where you can get all the hardware you need, and no centralized place with the firmware and all the info you need - everything is scattered everywhere.)

Meanwhile, I ordered a Duet with five ports, an expansion board with another four, and a beautiful touch-screen display, all from one website, and it was shipped from Merry Olde England to my place in the States in about a week. I don't have to wonder where to get the firmware, either. I'm going to install that and look at development stuff this week. I offered to port my delta calibration and work on segmentless delta kinematics for MachineKit/BBB, but only if he can get me the Reach board. I have a Diamond hot end, and I didn't buy it to look at it.

If Elias is willing to send me a Reach board for testing, I will start working on MachineKit/BBB when it arrives. (Whenever that might be.) I thought about going the CRAMPS route instead, but the Replicape has better electrical engineering behind it, and I would have the convenience of setting stepper trim through digipots. I need to set up a fleet of delta printers, and that means they're all going to be as close to identical as possible. If I use something with trim pots, I introduce the potential for one printer to run slightly different from another. When I set up a server farm, I set up all the machines identically because it makes administration FAR easier. Any variable I take out of the system could potentially save me days of head-scratching, bad-word-saying, and bug-investigating. To me, this is no different.

If Elias is uninterested in my offer, and if SeeMeCNC continues to be silent about whether they will go with a BBB solution, I will use (and develop for) the Duet until the Smoothieboard 2 comes out next year. At that point, I'll be going back to Smoothie, having never contributed a single line to any MachineKit/BBB repository.

The Duet is my last choice because it has only 96K RAM, but for the moment it seems a lot easier to wrangle.