Migrating an i3 Mk2 from RAMBO to Smoothieboard  

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

...How hard is it to migrate a Prusa from the RAMBO mini to a Smoothieboard? Has anyone done it?...
😕 acutally i doubt that it'll be successful.

➡ it's really a pain to get the prusa i3 mk2 firmware working at an ordinary arduino mega 2560 with ramps module, but at least it's the same processor and development platform (arduino). if you try to change the board type at actual releases (you can't even choose Big Rambo anymore), you'll end up with problems over and over compiling it...*). i guess the prusa guys have implemented some stumbling blocks on purpose.

so now cross compiling it for a completely new type of processor and hardware seems to be very challenging...
and furthermore: what's the point ? how do you think that it'll improve the performance of the printer ? the rambo1.3a board fit's like a glove for the mk2. everything works just fine.

*) only way that works for me is that you've to fool arduino with a modified pins.h (so the firmware / marlin still "thinks" it's a rambo mini 1.3a)

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

Hey !

Converting a Prusa i3 MK2 to Smoothieboard is fairly easy, and has been done by many people. It's actually a fairly common conversion.
Theres about a dozen configuration options you'll need to modify, typically.

Before doing the conversion, you absolutely need to read, in it's entierety, the 3D printer guide :  http://smoothieware.org/3d-printer-guide
This is not something you can skip ! It takes a few hours, but if you don't do it, you will run into trouble, and even risk damaging your board or printer.

As you will be coming from the Marlin firmware ( right ? ) it also makes sense that you read http://smoothieware.org/from-marlin

If you run into trouble ( and after you have read the documentation ), please don't hesitate to contact me via email at [email protected] for any help you might need.

You might also want to ask on the Smoothie forum or mailing list if somebody else has already done the conversion, and can share their config file with you.
This isn't an alternative to reading the documentation though ! If you are going to use a Smoothieboard you definitely need to read the 3D printer guide first.

Cheers.

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

@arthur.w3: with all your ways described at the links you've mentioned, you'll still end up in running a smoothieboard firmware (with a configuration file that fits the mk2 hardware), not the prusa i3 mk2 firmware.

so you'll only use the remaining hardware components (assembly, steppers, heatbed, pinda, psu, endstops etc.) of the i3 mk2 to run as a smoothieboard controlled 3d printer.... but with different behaviour.

➡ but in my mind: that's no i3 mk2 anymore.

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

➡ but in my mind: that's no i3 mk2 anymore.

Sure, it's a i3 mk2+

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

...
Sure, it's a i3 mk2+
i would say that it'll be some kind of i3 mk2 zombie, controlled by a smoothie driven brain

in my opinion the brain defines the personality, not the limbs. so i'ts more a smoothie-mk2-z (with "z" for zombie 😈 ).

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

...
Sure, it's a i3 mk2+
i would say that it'll be some kind of i3 mk2 zombie, controlled by a smoothie driven brain

in my opinion the brain defines the personality, not the limbs. so i'ts more a smoothie-mk2-z (with "z" for zombie 😈 ).

Zombie brains are dumber than normal brains.
Smoothie is way smarter than a marlin driven board.
So it'd be more of a i3 mk2 «Einstein edition»

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

I'd call it way too much hassle for i dunno what gains it'll give me.

Or to keep the naming game going: i3 MK2 I_wish_i_had_that_much-spare_time Edition

No offense meant.

Seriously, while I do not in any way, shape or form want to put in question the smoothieboard, what benefits would I get from this conversion? I never felt the current MK2 was seriously suffering from its maybe not quite leading edge RAMBo & Marlin brain. I'd have about a dotzend other optimization ideas before i'd consider the brain transplant. Starting with frame, motion, extruder.. What am I missing?

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

So it'd be more of a i3 mk2 «Einstein edition»

Are you crazy ? 😆

Einstein : general and special theory of relativity

So, if you have an Einstein-Edition you get much Time dilation and warping :ugeek:

Thomas

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

The Smoothieboard platform would indeed be significantly better than the RAMBo.

However, the question posed about Marlin is interesting. The Prusa firmware is not so far removed from the main Marlin branch it is now debatable as to whether it is any longer "Marlin".

Chunks of the Marlin code has been removed to make way for the significant number of additions (the MM code is now also present in 3.0.10) required by the Mk2.

A firmware port to Smoothieware would be possible although most likely not feasible.

Peter

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

The Smoothieboard platform would indeed be significantly better than the RAMBo.

Peter,

better how? What would we gain by doing this? Trying to understand..

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

Just look at stepper drivers used on Smoothieboard (v2 or v2-pro) and you'll understand. 😉

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

better how? What would we gain by doing this? Trying to understand..

Moor umph from the CPU as well. At 16MHz, the RAMBo is very under-powered and just cannot handle some things (like MAX end-stops - according to Vojtech).

Also very memory-limited which means that it cannot hold many future moves in memory, so (for example) top speeds are very limited.

The stepper drivers are 1/32 step, giving twice the resolution (in all axes), but that also has a downside in that the extruder would not be able to retract quickly enough for the multi-filament upgrade (this may not be true; I haven't look into the capabilities of the stepper drivers yet).

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

better how? What would we gain by doing this? Trying to understand..

There are hundreds of points on which Smoothie is better. We don't have a list anywhere because there is so much and it evolves all the time.

Some are related to the increased computing power, which gives you smoother stepping, better acceleration and planning, less bugs and edge cases, less print surface artefacts.

Some are related to features, like easy configuration, extremely modular system ( see smoothieware.org/switch ) allowing you to do things the way you want very easily. We also have ethernet support, you can access the sd card via usb like you would for a usb thumbdrive which makes it easy to drag/drop config or gcode file.

We also have the best documentation around, by far. And the hardware is designed very carefully.
We also have the most active community of any other system. I talk to more than a hundred volunteers monthly, with at least two new persons spontaneously offering to help every week.

We are also working on the Smoothieboard v2 project, which is the most ambitious controller ecosystem ever devised in the open-hardware cnc/3D printing community. Pretty much everything we are going to do has never been done before, and it's more work than anything that's ever been tried.

Oh also :
https://docs.google.com/document/d/1JEcIxRvH6L158aHFcFJtFToKWCGPPMwlF1e7OzqG57g/edit#heading=h.lbb0ckgykko9

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

Arthur

Thing is, if the firmware could be directly ported to Smoothie, then it would be a workable solution and give much more power to this printer.

But you should also remember that the "originals" with the RAMBo do work. And they will work with all future upgrades.

As things stand, Smoothie is simply not a solution for 99.9% of "original" owners.

Peter

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

Arthur
Thing is, if the firmware could be directly ported to Smoothie, then it would be a workable solution and give much more power to this printer.

I don't think you understand how advanced the Smoothie firmware is.
You marvel at Prusa's version of the Marlin firmware ( which is really great btw ) because the original Marlin sucks and it is much better than that.
But long before Prusa's version of Marlin, there was already something much better than Marlin, and that was Smoothie 🙂

There are actually a lot of users who own the prusa i3 mk2 who move to Smoothie, it's not a "0.01% of users" thing.

It's not that difficult, you just have to follow the documentation, and it gives a lot of tangible advantages.

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

Arthur

Everything is relative. The PE Marlin is exceptionally good, considering the hardware it is running on. No point in debating numbers as they are not known. All I can say is that I am not aware of anyone on these forums having switched. Yes, I have been here for some time and have read a whole load of posts.

Being a software developer, I do marvel at other people's coding, especially PE Marlin and KISS - doing things I would have thought impossible without seeing it. Yes, it's a lot easier to get badly-written code to work fast on a much more powerful processor.

But the fact remains, Smoothie cannot currently do what the MM PE Marlin can; it would require way too much work.

If you think it can, then please provide me with a loaded test board so that I can check it out for you. Please note that the extruder must be capable of 280 (1/32) micro-steps per mm and run at a maximum speed of 100mm/s.

And FYI, I am currently Beta-testing the MM upgrade and Alpha-testing KISS 1.6. I have also developed a post-processor for KISS/MM.

Peter.

P.S. I note that Smoothie is "Event-Driven". How does that compare vs. "Interrupt-driven" firmwares?

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

But the fact remains, Smoothie cannot currently do what the MM PE Marlin can; it would require way too much work.

I'm sure you'd be able to find some differences, for example Smoothie has different menus, is configured differently ...
But I have a strong feeling you don't know enough about Smoothieboard and the Smoothie firmware here ...

If you think it can, then please provide me with a loaded test board so that I can check it out for you.

We do provide free boards, but only to users who actually contribute to the project with code, documentation, etc.

Please note that the extruder must be capable of 280 (1/32) micro-steps per mm and run at a maximum speed of 100mm/s.

This is trivial for Smoothie, it can go 4 times that high. At those speeds, Marlin ( even Prusa's ) is using all sorts of hacks to be able to follow ( like quad-stepping, simplifying planning math ), and those have a negative impact on path planning and step generation quality, which translates into print surface artefacts. Smoothie has none of that.

Is this the sort of thing you think Smoothie isn't up to the task on ? I think you need to read up on Smoothie ...

P.S. I note that Smoothie is "Event-Driven". How does that compare vs. "Interrupt-driven" firmwares?

Smoothie is event-driven in the non-real-time part of it's work, and interrupt driven for the real time part. It essentially makes the code much easier to read, extend and develop on, and much less likely to become a "I'm going to kill myself if I have to code in this one more minute" mess like Marlin has become ( I'm not saying that, the Marlin devs are ... )

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

Arthur, just a curious question. Does Smoothieboard v2 (with Trinamic stepper drivers) detect lost steps? It would be awesome.

Re: Migrating an i3 Mk2 from RAMBO to Smoothieboard

Arthur, just a curious question. Does Smoothieboard v2 (with Trinamic stepper drivers) detect lost steps? It would be awesome.

The drivers have stall detection and we expect to take advantage of that, however you'd need an encoder to actually detect lost steps ( which is something we'll explore on v2-pro ).

I'm going to sound like a grump old engineer, but you really should build your machine ( rigidity, motor power etc ) in a way where you are guaranteed you won't ever loose steps. It's really not difficult, and you shouldn't be using clutches ( like lost step detection ) to work around the fact your machine isn't properly designed/built.