Open source 32bit boards - Alternatives to Duet

Edit: Formerly titled: "Re-ARM Control Board & Marlin 2?"

Hi all,

A couple of my printers are running RAMPS and I have just seen an advert for a Marlin 2 and RAMPS compatible 'Re ARM' board. Have any of you got any experience in this?

I am happy with my Duet boards and am likely to upgrade them to newer generation duets when cash allows, but I would be more comfortable if I were familiar with two different 32bit control boards and firmware which are in current development.

Thanks.

Edit: Previously said "I am happy with my Duet boards and am likely to upgrade them to newer generation duets when cash allows, but in the interest of security and stability of my business I feel it is critical that I am familiar with more than one family of control boards/firmware." which was missleading.

Edited 3 time(s). Last edit at 05/22/2019 05:09AM by WesBrooks.

In the interest of security and stability of your business, I'd recommend getting away from RAMPS and any controller that uses it or the same type of motor driver modules. The driver modules can't dissipate heat, which drives reliability down the toilet. Oh, and then there's those motor current pots...

There have been a couple attempts to make linux board based controllers, but they kept the driver board the same size as the Beagle Bone board (see Replicape, for example) so it suffers from the same heat limitations as the RAMPS board.

When you're looking for controllers, look for lots of surface area for the driver chips to get rid of heat, and then proper mounting of the drivers to a ground plane in the board. If you don't have that, the rest doesn't matter much. Beware of heatsinks glued to plastic IC packages.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

By stability and security I am referring to supplier redundancy. I just couldn't remember the phrase this morning. The machines are a tool for the business, not a product we trade.

With respect those points have little to do with that, and relate to the performance of the control board. In fairness that wasn't clear in ky original post.

To be a little clearer I already run Duets, and RAMPS (marlin 1.1.9), and am aware of the limitations of RAMPS. I'm interested in having a machine running 32bit and Marlin 2. I did look at linux cnc but the cost involved with reaching the step rates attainable on Duet boards was steep, and the system complex. I welcome other suggestions.

proper cooling should be done for any controller used.
even with ramps it is a good idea for a fan to be used if enclosed, and it is a good practice to have active cooling of parts anyway.

ramps does have one benefit that many other controllers do not,
that is the ability to replace parts, rather than the entire board when a failure occurs.
and this often does at some point in time usually more frequently for beginners.



i would look at your machine more realistically. it needs to go down for maintenance and preventative care anyways.
if the machine has an uptime of 90% and you can meet business needs, then i think a ramps based board would be fine and cheaper.
from what i read the marlin 2 and 're-arm' are in a form to replace the atmega boards used for ramps. this meets a considerable market.
The main issues i see so far with this are the lack of knowledge of users on the product.
there are also some pins that are not 5v tolerant, or output 5v, so some other controllers and display parts might not work without additional parts (voltage shifters, lower voltage chipsets, *)

i see that someone here on the forum has stared a link for 're-arm' here. be gentle with them as they are still working on editing and spelling issues. if anything help them out
[reprap.org]



do you have a vendor that has a complete solution for you in mind?

I say look into marlin 2 sure, but its very much still beta, I would not put it on a business printer... home printers sure.

I don't know if this is up to date, but its still open so... take a look at [github.com]

Re-ARM also runs RepRapFirmware. See [github.com] and [reprap.org].

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I was assuming that the printers were used to produce some product. A hobbyist doesn't really care if they have to wait a month for a part to come from China, but how much does it cost you when a printer goes down?

I wouldn't presume to tell you how to run your business, but if I were going to try to run a business whose revenue was based on selling 3D printed widgets, I'd buy the best and the exact same controller board I could find for every machine I was running (and the machines would be identical). I'd keep some spare boards on hand so they could be dropped in on a moment's notice to keep the machines running. I'd probably keep a spare printer or two handy, too. Dead/damaged boards could then be assessed/repaired/replaced as needed without worrying about missing a deadline. If the margins from the business are too slim to keep spare boards on hand, it probably isn't a sustainable business. But I have no experience with running that type of business, so maybe I'm full of crap.

---

Ultra MegaMax Dominator 3D printer: [drmrehorst.blogspot.com]

Mentioning the business was completely unnecessary on my behalf. All I'm going to discuss with regards the business is I don't want to ever be in the situation where all of my eggs are in one basket. In this case I want to be sure there isn't a huge learning curve to overcome if I need to use the latest version of Marlin.

Trying to rephrase again in order to avoid ambiguilty.

I use Duet board equiped machines and am happy with them. I am looking forward to being able to update these to Duet 2 or the next generation boards as and when they are available.

I have an 8 bit ramps set up which is a reserve for running builds but predominantly so I can gain first hand experience of running Marlin, and be aware of both its strengths and weaknesses.

I am acutely aware that work on the 8 bit compatible Marlin 1.1.9+ is now bug fix only. In order to stay current I would like to have a machine running Marlin 2.0. Thanks for the beta warning I appreciate that and I am not reliant on this machine remaining opperational a high percentage of the time.

I have seen this re arm system is a platform that is capable of running Marlin 2 and I welcome other suggestions. I would work well withing the recommended current limitations of the board, and add cooling as recommended by the relevant manufacturer or community.

Edit: As an aside since it has been mentioned I am also interested in Linux CNC if there is a lowish cost entry into running 32bit and a comparable step rate to current production boards.

Edited 1 time(s). Last edit at 05/21/2019 10:47AM by WesBrooks.

You can put smoothieware on a re-arm... use the CNC build... I've not tried it on a re-arm but works well on genuine smoothie boards. (can't see why it wouldn't work on re-arm)

Yeah, I'm guessing the main driver for the memory card on the board is configuration files from Smoothie ware. I'm not aware if Marlin 2.0 has changed from firmware config to fule based config.

Marlin2 is still config files. Which is a good thing in my view, less likely to lose all your setting when the sd card dies. because people don't backup!

Edited 1 time(s). Last edit at 05/21/2019 11:20AM by Dust.

Ok, I've had a little longer to look at this. Here's the list of boards that are compatible with Marlin 2.0 to a greater or lesser extent:

[github.com]

Shields/capes/whatever that bolt on these boards seem a little hard to come by off the shelf. There are big issues reported with the pre v2.A version RAMPS_FD. RAMPS 1.7 would be an option but again that seems to be hard to come by. I need to do a little more reading but some had got digital current limits working via SPI on RAMPS 1.4 boards and TMC2130 drivers (with the heat sinks on the right side of the carrier).

As others have hinted this is all very beta level stuff.

I did come across the combination of BeagleBoard and the Replicate add on, which has since evolved into a single board solution that is currently under development - Revolve. My previous interest in LinuxCNC was pretty much shelved when I saw the cost of FPGA boards that were needed to get a reasonable step rate. This appears to have promise. Max of eight steppers is more limiting than I'm accustomed to, but better than the RAMPS versions I had previously considered.

So, gut feeling has swung a bit back to LinuxCNC unless I have missed some nice shields to run on top of the SAMD51 Grand Central from adafruit?

Edit: Regards the Revolve option I see you can't buy one yet, but development does seem active!

Edited 1 time(s). Last edit at 05/22/2019 05:07AM by WesBrooks.

I've changed the title to be more representative of my slight change in direction rather than restart the thread.

Edit: My initial post was a quick question to see how much experience there was out there and I hadn't read much into the topic. My intention was to start gathering information from the community in parallel to going away and reading up on whatever I could find. I'm a little uncomfortable about some of the things I've been coming across with regards Smoothie, and hadn't realised the connection with Re-ARM. I don't want to use smoothie, but at the same time appreciate many will be using it and be very happy with it.

Edited 2 time(s). Last edit at 05/22/2019 04:58AM by WesBrooks.

... we've developed a shield for ArduinoDue to adapt to our 24V CNC breakout boards, what's essentially compatible to RADDS, but only for 4 steppers on the board (Pololu-style or pin-out for common stepper drivers).

This is more meant for "industry" use, so the price is way above the "DIY-range".

Here the shield under the ArduinoDue and the BOB:





And here the comparison of the RADDS (bottom left) and our "common driver type":





Have complete "families" and various adapters ("intelligent" with Arduinos or ESP's or "discrete") too:

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Nice. I spy something that resembles a propper attempt at an E-Stop circuilt with force guided contact relays. Did you just break the power to the stepper/servo drivers or do they have interlock loops rated to an appropriate level that you were able to just break the enable? Or was it to interlock something else?!

Step rate? I've seen that CNC controllers tend to use little to mo microstepping in order to maximise torque, so understand your design isn't trying to compete directly with printer controllers.

... yes, there are differen "safety levels" to cut the power or enable to the steppers or other "cut-off" structures -- it's used for industrial CNC-milling amd laser-cutting, so we have some "serious" security measures to implement (especially with the Arduinos)

With the ArduinoDue and Marlin4Due or Marlin2 as firmwares the speed/frequency limits are set by the stepper drivers - AFAIK the Due can run 4 axis synchronous with up to 300kHz ... didn't run this fast yet ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Yeah, I come from a background of leading the development of the control systems of a laser based metal powder processing system. It was sold as capable of processing reactive metals such as commercially pure titanium and various aluminum alloys. 13849 was coming in and getting close to manditory as I was leaving that and it was tripping up many in the field. Since the laser powder process is relatively new, there wasn't much in the way of guidance in ways to demonstrate compliance with the machinery directive. Naturally this led to many conflicts between the project leaders/finance bods and those doing the development work.

Edit: I don't think the required performance levels for stepper interlocks on most printers would be that high, they're not that powerful. With regards to the heaters I would just look to consider ensuring two things need to fail (aside from the controller) before the hotend/heatbed can exceed their specified working temperature range. Edit 2: I'm seriously digressing here.

Edited 2 time(s). Last edit at 05/22/2019 07:42AM by WesBrooks.

... I'm actually too in the proces of developing a "laser-based additive metal 3D-printing head" with metal wires - here it is also a little more appealing than with common FDM printers ... but not this much - have a servo-system running with 0,0025mm "step" resolution and up to 100kHz bandwith at the encoders ... some of the FDM-printers are not this far away form my specs

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

@WesBrooks-

Re-ARM will run RRF. I think I am one of the links that dc42 pointed you to. You should probably go to the main thread though, as my git is my rough notes and files to date.

LPC Port of RepRapFirmware

Why did I choose to try RRF- I know a number of people who use Duets, and I also already had a working printer with a RAMPS /mega already.
I got tired or Marlin's goofiness. I tried looking into smoothie, but I got turned off to their attitude and community ethos.

Some caveats-

Is it the same as running on Duet HW-no. Will it work- Yes. its 100Mhz processor. Its memory challenged, but the main functionality is there.

I have been using the LCD menu system and it really does work well. I was using it with the LCD and Octoprint, but Octoprint is not playing well with it consistently.

I just this weekend got the networking up and going so I am getting into DWC for the first time.

Expect to put in some time to set up your printer. You can use the online configurator to some degree, but expect to hand code some of it. It sounds worse than it is really, and it give you opportunity really learn your printer and how RRF works.

If you are already used to the ways of RRF using it on Duet, then it might be an easy upgrade for you. $45 for the real deal, $~30 for a clone that seems to be just as good. I would but the real deal though and support the folks who developed it.

Edited 1 time(s). Last edit at 06/03/2019 12:43PM by sinned.

Interesting read thanks. Just spotted ooznest are shipping the RADDS 1.5 board. That and the grand central from Ada fruit would be interesting if I could somehow use the SPI from the board at the same time. It would be an interesting learning curve, a little gentler, and cheaper than Beagleboard Black + a suitable cape.

All that said if I got it running with SPI all I have really achieved is Duet 0.8.5 like board with quieter drives? I don't think there is anything the re-arm + RAMPs or Grand Central + RADDs couod do that the Duet 0.8.5 couldn't aside from driver related changes?

This kicks a few spots off the gran central:

[www.mouser.co.uk]

As the SPI and arduino form connectors are unpopulated it maybe easier to put a low profile connector that can bring the SPI out from under the board.

Edited 1 time(s). Last edit at 06/06/2019 04:44PM by WesBrooks.

Quote
WesBrooks
This kicks a few spots off the gran central:

[www.mouser.co.uk]

As the SPI and arduino form connectors are unpopulated it maybe easier to put a low profile connector that can bring the SPI out from under the board.

We have a build of RepRapFirmware for that board, because we used it to prototype ideas for Duet 3. The support for it is still in the RepRapFirmware source code. It can support Duet Web Control on both the Ethernet port and an SPI-connected WiFi module simultaneously.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

That all seems very accessible then. Is it stil possible to achieve high step rates with spi control of the step sticks? Potentially a bit of a birds nest but I can see the spi connectors are made available on the top side of the RADDS board.

Hadn't realised there'd been a serious bit of poo slinging on the forums with regards the RADDS board. Still the only commercially available 3.3V shield aimed at the 3D printer market?

With disappearing Docs it is a little hard to see what the difference is between the v1.5 and v1.6 RADDS boards.

Edit: Thinking further is it possible to get the SPI pin states copied or remapped to save needing to flip pins on the stepsticks?

New TMC5160 driver looks very interesting. Doesn't look like the Tmc2660 ever made it to a step stick form factor.

Edit 2: Thinking further the spi connectors are probably on the pins normally connected to the jumpers used to set microstepping. If so they are not remappable at the chip. Will look at schematics before asking more questions!

I've seen threads lamenting the purging of information from the reprap wiki, but finding it odd that all information seems to have been purged from the web! Found a user manual which will help a little!

Final edit!: Is the disappearing v1.5 and 1.6 docs something to do with the original RADDS board having a non-commercial license?

Edited 5 time(s). Last edit at 06/07/2019 04:02AM by WesBrooks.

Thanks for the comments. Parts on order for yet another project with the side line aim of bringing my coding back upto speed.

I would like to make an add on board for power distribution, earth - DC negative connection, fuses, and power cut relays. I'm hoping this will be a simple introduction to KiCAD. If anyone has favorite tutorials on line please share!

... most complains 'bout RADDS were the insufficient OS documentation -- (they did this mainly to "protect" RADDS from chinese cloners).
Could be, I'm a bit biased, as some friends are developers/vendors of the RADDS, so I have short links to the sources and hardware bits (this was my point too for developing our "comercial" variant).

Another (comercial, more pricey) alternative are the E1701C and -M controllers from HAlaser, based on BeagleBoneBlack (using them for my laser systems) -- [halaser.de]

And another (comercial too) alternative is Editask, actually based on STM32F4 -- [www.editasc.com]

Again, maybe slightly "biased" - was (are) cooperating with the developers and too have some special editions and permissions working in some of my machines ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

The Beaglebone and derived firmwares does look very interesting but after a prolonged think I decided I am probably asking too much of myself to be tackling that right now, particularly as the number of users seem fairly light.

I have been able to cobble together parts (M7 dev board, RADDS, TMC5160) that should be comparatively straight forward to get running and still offer a overall package that is not light years behind the current commercial state of the art.

I would like to make an alternative to the RADDS board that is open source that better suits my requirements. This again will be shelved for a later date, and after sorting out a much simpler power distribution board that is useful to all of my machines, and potentially portable to other systems.

So much I want to tinker with, so little time! :-D

... instead of ArduinoDue, the Teensy3.6 could eventually be a more "powerfull" microcontroller - faster, more flash and memory, running with Marlin2.0 too ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]

Arduino Due and Duet v0.6 and v0.8.5 are Cortex M3 based.

Duet 2 (wifi/ethernet) and Teensy 3.6 are Cortex M4 based.

The dev board I have picked and the proposed spec of Duet 3 is Cortex M7 based.

... depends on the work you want to "invest" ... or using already running community sources ...

---

Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]

Call for the project "garbage-free seas" - [reprap.org]