Looking to switch from Spinter to Marlin - any tutorials or help out there??

Hey guys -
Been seeing some buzz here and there of great prints coming out with marlin. Ive been using sprinter since I started, and was thinking of trying to switch over to Marlin. However, it has been a while since I configured my firmware (forgot how), and after looking at Marlins configure file it looks like a different animal to set up.
Im running A prusa with a wades extruder, ramps 1.4 on a SAE machine. Anyone have a similar machine with a Marlin config I can work from to start with?? Or anyone know of a tutorial?
thanks!
Jeff

The setup is almost identical, I copied all my settings (manually, one by one) from sprinters configuration.h to marlin, and was up and running within an evening. I get superb quality with marlin and "slic3r.exe --gcode-arcs", and I highly encourage you to switch!

The config file is heavy commented, so you should be able to read through the file and understand what you have to change, and as you read through it I'm sure you'll start to remember things from last time.

I was pointed in the direction of this config file from jmil: [github.com] And others reported good results with that as a base.

Just remember to change the acceleration settings (to what you use in sprinter now). The defaults are meant for an ultimaker, and it's a tiny bit brutal on a prusa. I use 900 on x and y now, which is a bit lower than with sprinter, but because of lookahead there's no total loss in print time, but a big increase in quality.

---

--
-Nudel
Blog with RepRap Comic

Ok, cool...
So the only file i need to change then is the config?? Just copy all my sprinter settings over and flash the board? Now on that acelleration - where do I change that? is that max acell?? I would like a fairly painless changeover.
I have tried slicer and it seems better than sfact - what is it that you listed above? "slic3r.exe --gcode-arcs" is that something diffrent?
Thanks!
Jeff

Yes, you only have to change configuration.h. The acceleration setting is the same as in sprinter, but the variable name is different. My settings (prusa, sang1.3, gregstruder) are #define DEFAULT_MAX_ACCELERATION {900,900,80,10000}

The changeover was pretty painless for me, though I had to do a few tweaks and upload the firmware a couple of times.

I found TEMP_RESIDENCY_TIME to mess up the first layer, and not needed at all, so I turned it off. (By default it waits 30 seconds for the temperature to stabilize before prints). I only have to find a way for it to start printing as soon as the temp is +/- 1C within target temp, since it can sit and wait for a while now until it reaches the precise temp. Other than that, marlin's pid-settings (I used ultimaker defaults) keeps the temperature incredible stable for me.

--gcode-arcs is a command line option for slic3r, which makes it generate arcs in the gcode where applicable, thus making super-smooth curves.

---

--
-Nudel
Blog with RepRap Comic

I just did the change your asking about on an identicle (by your description) machine and yes my prints are much better and faster than with sprinter.

I opened my sprinter config.h and marlin config.h side by side and went through making the necassary changes.
I had trouble compiling though; errors in the SD files. I ended up deleted all the SD files (since I don't need SD support) and edited the Marlin.pde (I think it was) to not include the SD files. It then compiled with Arduino 1.0

If you're still having trouble I could send you my firmware and it should work for you. You'd probably need to edit in your steps per mm, though.

Micah

Micha,
I appreciate that!! That would be a great stepping off point since I see you had issues compiling. Im not using SD support either... So just copying over my steps from sprinter to the marlin config file should work?
Ill PM you with my email if you could send it to me..
Thanks!!
Jeff

I recently gave a first try at Marlin. Prusa Mendel with RAMPS 1.25 and Arduino Mega 2560, heated bed, GregFrost's extruder, ParCan - supplied thermistors on both the bed and the extruder, current Sprinter configured and printing well: quite standard. Also copied settings one by one from Sprinter to Marlin configuration.h
Marlin 1.0 just starts and quickly refuses to run and halts everything because it senses some temperature is wrong.
I suspect it is not reading from the right pins, am under the impression there was a change between RAMPS 1.2 and 1.3, but haven't investigated this further. Is anyone trying Marlin having similar problems?

With the firmware I sent follow these directions and you should be set up.

First set you steps per mm. In Marlin they look like this (half way through configuration.h in a section labeled mechanical settings):
#define DEFAULT_AXIS_STEPS_PER_UNIT
In Sprinter (near the top of configuration.h):
float axis_steps_per_unit
Just copy the bracketed numbers from sprinter into marlin.

Now compile and upload your firmware using Arduino 1.0 (older versions didn't work for me)
When uploading is complete shut down your arduino software and open your host software. Make sure your baudrate and port is correct before connecting. Marlin uses 115200 baudrate by default. Connect to your machine.

If that works, turn on your power supply and try moving your axes 1mm at a time to see if they move in the correct direction. Heat up your hot end and make sure it moves the right way. Note which axes are moving the wrong way. Shut down your machine and quit your host software.

Open your configuration.h and find the following line (about half way down):
#define INVERT_X_DIR true
For each axis thats moving the wrong way change from true to false or false to true. E0 is the extruder (E1 and E2 are not used).
Compile, upload, shutdown arduino, open host, connect to machine. Your axes should now be moving in the correct directions.

My firmware is setup for mechanical endstops. If you're using optos there's some settings in configuraton.h (under mechanical settings) that may need to be changed. Changes are described in the firmware.
Next, with one hand on your power supply switch, try to home each axis individually. If they don't move towards your endstops then shut down your power supply to prevent crashing. Note which axes are home in the wrong direction. Shut down, open configuration.h

Find (right bellow the invert settings):
#define X_HOME_DIR -1
For each axis that home in the wrong direction change -1 to 1
Note: right below these settings is the max length for each axis. you may want to adjust these for your build envelope.
Compile, upload, shutdown arduino, open host, connect to machine.

You should be configured now, but let me know if your still having problems and I'll do my best to help.

Edited 1 time(s). Last edit at 01/01/2012 12:07PM by miro87043.

Lanthan Wrote:
-------------------------------------------------------
> I recently gave a first try at Marlin. Prusa
> Mendel with RAMPS 1.25 and Arduino Mega 2560,
> heated bed, GregFrost's extruder, ParCan -
> supplied thermistors on both the bed and the
> extruder, current Sprinter configured and printing
> well: quite standard. Also copied settings one by
> one from Sprinter to Marlin configuration.h
> Marlin 1.0 just starts and quickly refuses to run
> and halts everything because it senses some
> temperature is wrong.
> I suspect it is not reading from the right pins,
> am under the impression there was a change between
> RAMPS 1.2 and 1.3, but haven't investigated this
> further. Is anyone trying Marlin having similar
> problems?

I had the same problem. Your pins should be set correct if you set your motherboard at the top of configuration.h
here's what fixed it for me.
uncomment //#define THERMISTORHEATER_0 3 and change the 3 to the coresponding thermistor listed right above that setting
comment out #define HEATER_0_USES_AD595
uncomment these to lines
//#define THERMISTORBED 1
//#define BED_USES_THERMISTOR

re-upload and see if that helps

Edited 1 time(s). Last edit at 01/01/2012 12:05PM by miro87043.

Thanks for sending that!! Ill give it a go here in a bit - after i finish recovering from last night! LOL!!
Thanks as well for the detailed setup instructions. I also have mechanical endstops and they are at the min positions. Im using a heated buildplate as well, so ill have to make sure thats a go too.
Ill let you know my success or failure!
Jeff

Well, I got it configured and working!!
Thanks!!
However im having a slight problem... as it was listed above, it pauses for about 10-20 seconds after executing my start gcode. I really dont like this as it looks like it oozes some material out .
I checked the solution listed above and it was already commented out - any ideas how to fix that?
thanks!
Jeff

@miro87043, jmir and all: A big THANK YOU, I have it running and printing with your instructions!

First looks: at equal speeds, Marlin is much softer on the mechanics, the machine runs quieter, vibrations reduced. Maybe also less risk of bolts unscrewing, parts loosening etc.

Printing quality is substantially improved, cleaner.
Adopted!

A note of cauton: the Z homing speed MUST be reduced to something around 80 mm/s or less. I had left the configured speed, and of course while homing one of the Z motors stalled while the other kept running and crashed the X axis obliquely on the bed. Am I glad I had the extruder block just fastened to the X carriage with magnetic clamps, it popped up without damages.

Parabolic Wrote:
-------------------------------------------------------
> Well, I got it configured and working!!
> Thanks!!
> However im having a slight problem... as it was
> listed above, it pauses for about 10-20 seconds
> after executing my start gcode. I really dont like
> this as it looks like it oozes some material out
> .
> I checked the solution listed above and it was
> already commented out - any ideas how to fix
> that?
> thanks!
> Jeff

I think the pause is the software waiting for your extruder to hit temperature. Mine does this too. Slic3r or sfact can print a skirt which allows the filament to catch up before starting my print. Thats the only solution I've found. I would think the only way to get rid of this pause would be a setting in sfact/slic3r/skeinforge, but I don't know for sure. You could add a G code to extrude 5 mm after hitting temp.

Anybody else know a solution for this problem?

Edited 1 time(s). Last edit at 01/01/2012 05:58PM by miro87043.

Lanthan Wrote:
-------------------------------------------------------

> A note of cauton: the Z homing speed MUST be
> reduced to something around 80 mm/s or less. I had
> left the configured speed, and of course while
> homing one of the Z motors stalled while the other
> kept running and crashed the X axis obliquely on
> the bed. Am I glad I had the extruder block just
> fastened to the X carriage with magnetic clamps,
> it popped up without damages.

I had a similar problem with my Z axis not working when changing to marlin. Mine were inconsistently working together. I recall reducing max and default Z speed and changed minimum Z feed from 0 to 1. Not sure why the last change was necessary, but it was.

hmm - thats odd though - as I dont start a print until I see the head temp at operating temp.....so what its waiting for I have no idea...
is there a way to take out that wait command??
Jeff

If it's doing a 30 sec countdown (to stabilize the temperature) before it starts printing, set TEMP_RESIDENCY_TIME to 0, as I wrote above...

I'm not sure, but HEATING_EARLY_FINISH_DEG_OFFSET (default 1) could be the option that controls how close it has to be to target temperature before it starts. For me it waits to do the switch from 159 to 160 before it starts, while it could just as well start at 158 or 159. The pid control is incredible at keeping the temperature stable, but it does mean it takes longer to reach the exact target temperature when ramping heat, and 10-20 second of ooze can damage the first layer.

---

--
-Nudel
Blog with RepRap Comic

What is the last number that Marlin returns in response to M105? It gives the usual hot end and bed temperature followed by another number preceded by "@:".

Never mind. It looks like it's the hot end heater output value. It goes to zero when I turn off the hot end heater and the value goes from 0 to 127 depending on the temperature when the heat is on.

Edited 1 time(s). Last edit at 01/03/2012 06:54PM by brnrd.

So now I have the pirated cupcake (has ramps electronics) and the prusa-ish mendel running smoother than ever and producing beautiful, smooth parts with Marlin. I set the max acceleration for x and y at 900.
Of course, I am tempted to increase the speeds for printing, that is the
next question. Currently I am using in skeinforge 44 mm/s for the mendel and 40 mm/s for the cupcake, with perimeters and first layers printing at 0.5x those values.
What speeds and settings have you found optimal for Marlin and a given hardware?

Current setup: slic3r/pronterface/marlin/ramps 1.4/prusa with linear bearings, 1.75 mm hot end, .35 nozzle
I'm continuously trying different speeds, but I've been getting consistently good prints with 40 mm/sec perimeters and 100 mm/sec fill at .28 layer height. I even went .1 layer height an printed this: [www.thingiverse.com] at the same speeds.
From my understanding and limited experience it's slic3r that's allowed me great prints and small layer heights at some what quick speeds. I spent about 2 minutes setting up slic3r and got results that would take me a couple days configuring with skeinforge or sfact.

Ok,
So the conversion has gone pretty painless for me to move to Marlin - thanks to Miro's help. However, I rna into my first stumbling block with Slicer last night. I had to reprint a ring I made before, and figured Id try it with slicer - It didnt print well half way up, and the sections started falling over due to there was no fill (5mm thick sections only had the 3 perimeters - no fill in the center. So I had to restart it with the Sfact settings I had and it printed perfectly. So im still not sold on Slicer. Plus I tried doing the Mo-Ring at some stupid layer height <.1 and it looked good to start, but again - alot of holes due to not enough fill. Maybe its just my settings?
Im running a .4mm tip with a .28 layer height - would like to get around.1 or less - what do i need to also change to acomplish that?
Thanks!
Jeff

Parabolic Wrote:
-------------------------------------------------------
> Ok,
> So the conversion has gone pretty painless for me
> to move to Marlin - thanks to Miro's help.
> However, I rna into my first stumbling block with
> Slicer last night. I had to reprint a ring I made
> before, and figured Id try it with slicer - It
> didnt print well half way up, and the sections
> started falling over due to there was no fill (5mm
> thick sections only had the 3 perimeters - no fill
> in the center. So I had to restart it with the
> Sfact settings I had and it printed perfectly. So
> im still not sold on Slicer. Plus I tried doing
> the Mo-Ring at some stupid layer height <.1 and it
> looked good to start, but again - alot of holes
> due to not enough fill. Maybe its just my
> settings?
> Im running a .4mm tip with a .28 layer height -
> would like to get around.1 or less - what do i
> need to also change to acomplish that?
> Thanks!
> Jeff

I should start by letting you know that I'm using pla on heated pcb (60 c) with glass on top. Also my experience is fairly limited, but I'm more than happy to share what I've found.

As for slic3r, it is important to use the most recent version (0.5.7) and it is still early in development and has some limitations. However, I'm not sure what's causing your issue. Compare your settings between sfact and slic3r. Are they set to the same fill density? Is sfact only doing 2 perimeters, allowing more room for infill? If your using pla, do you have a fan on your hot end? Printing onto layers that haven't cooled enough can cause messy and unstable prints. Slic3r may be less tolerant of non-manifold stl's. You may try running your file through cloud.netfabb.com
I've printed 0 infill with slic3r, just 3 perimeters, without issue.

Printing thin layer heights is something I have very little experience with, but, as I mentioned above, making sure each layer cools sufficiently prior to the subsequent layer seems to more important with thin layers. This can be done by slowing your feed rate and/or putting a fan on your x carriage.

Hopefully this helps,
Micah

Well,
Im also using a heated bed and glass with Kapton tape on it - was using blue tape for a while - spotty results - sometimes sticks sometimes lifts..
I also made the switch to ABS as i find it prints alot better than abs. No Im not running a fan either. Maybe the perimeter difference is why the infill wasn't present. I do like how slicer builds the perimeters from the inside out - makes more sense to me.
Jeff

I couldn't get Marlin 1.0 to even compile. Running the lighter nongen6 version was easy though. I have a sanguiolouloulou board.

jkomp316 Wrote:
-------------------------------------------------------
> I couldn't get Marlin 1.0 to even compile.
> Running the lighter nongen6 version was easy
> though. I have a sanguiolouloulou board.

I had trouble compiling too. I kept getting errors in the sd* files. To remedy this I got rid of them as I don't need SD card support.
I believe the following is what I did:
In Marlin.pde I deleted the line: #include "cardreader.h"
In the Marlin folder I deleted all files starting with sd and the 2 cardreader files. This may not be necessary as you delete reference to them in the first step but I did it anyhow as I have very limited knowledge of programming in arduino.

After these steps Marlin compiled in Arduino 1.0

Edited 1 time(s). Last edit at 01/05/2012 09:39PM by miro87043.

The SD library needs to get updated for Arduino 1.0(arduino.h issue), The creator of the SD library already posted an update but it still has to be incorporated into the reprap firmwares.

I compiled and uploaded Marlin from Arduino 0018 without any problems.

---

--
-Nudel
Blog with RepRap Comic

I am using Slic3r -> Pronterface -> Marlin.

If you set the temperature in Slic3r to 0 it does not wait for the extruder to get to temperature.

I then use Pronterface to set the extrusion temp. and manually wait for it to get to the correct setting before I print.

Geoff

---

---

How often I found where I should be going only by setting out for somewhere else. - R. Buckminster Fuller

[apapageek.com]

It's all starting to make sense. This is good to know as I was wanting to add a SD slot to my board.

Thanks Nudel and Royco

Anyone have a configuration.h I can take a look at? I am having issues with my Bed Thermistors showing up in Replicatorg 26

I am running Sprinter with no issues except it shows Unknown: x:0.00y:0.00z:0.00e:0.00 when I am in control panel, so I know the thermistors are ok.

I am running:
RAMPS 1.4
Thermistors
ReplicatorG 26


Thanks

here it goes: Marlin Configuration.h and my current skeinforge (46) settings
give a try to pronsole and/or pronterface!