My first Smoothieboard prints, wow!

[Flateric]

I had a thought today while considering a side project I will start on today.

I'm currently in the process of switching in my 0.9 degree steppers again after having removed them over a year ago due to problems with heat and being able to feed data throught the arduino fast enough to keep up. These 0.9 steppers have the dampers on the shaft and I will also use my mounting dampers as well. Essentially a dual damping for each motor.

I suspect the patterns we can see sometimes have improved (less visiible) due to the way the smoothie drives the steppers and because of the ability to process data and pass this data as steps to the motors much more evenly and consistently.

Then I remembered a thought I had before the 0.9 steppers were experimented with.

I suspect if we were able to drive the motion of our machines with servo motor solutions it would be optimal. I am not talking the servos from the RC hobby industry of course. But the servo motors from the cnc machining world of course. This is the reason that any type of decent plotter or vinyl cutter is always driven by servo motors and not steppers. Speed, fluidity of motion and accuracy.

I am however stopped at that point because I can't think of an easy way to drive proper servo motors with any current hardware or firmware for 3D printers on the market. Everything is step and pulse based.

Servo's also offer the benefit of giving accurate positional feedback, no more lost steps.

Cost is higher however on both the motor and controller and drive side typically.

An additional plus to servo driven movement is the machine can be scaled up in size without the concerns of step resolution becoming very problematic fast as stepper motors.

Just my thought of the day, take it for what you will.

I lucked into some very nice THK linear bearing guides a couple of days ago that are perfectly sized for me to replace my cheapskates with. Which got me thinking along these lines and scaling the machine and problems with scaling.

I thought I had seen or used a delta sizing utility once somewhere that let you calculate optimal size for the delta based of the step size/pulley/layer height desired ratios. Or am I crazy, anyone know of one hiding somewhere on the net?

[Broose]

I am however stopped at that point because I can't think of an easy way to drive proper servo motors with any current hardware or firmware for 3D printers on the market. Everything is step and pulse based.

i believe some servo amplifiers can take step and direction as commands- check out Copley Controls as a possibility:
http://www.copleycontrols.com/index.html

[snoman002]

I can understand the desire for servo's vs steppers, but I think its wasted money. Steppers have proven to be a viable option, even in the metal working world, and the 3d printer world has SO much less force to deal with lost steps are far less likely. Heck, it is only this fact that delta robots are viable in the first place, the lack of stiffness in certain conditions makes them less than perfect for metalworking. I think that until the print heads get so fast that nema 23 steppers cannot handle changes in direction (so, like, an order of magnitude faster), servos will offer little in added value.

On the other hand, stepper servos are the cats meow, and in reality servos are only a motor with an encoder. They aren't black magic, the stock steppers could be made into stepper servos relatively quickly. You could use linuxcnc to run a stepper/servo based system if you wanted, even could put glass scales on the towers. But I still think its overkill for these systems. Pushing a 3/4 corncob endmill through Stainless Steel in a production environment where every minute of machine time hurts, yep big old servos are the ticket, but here, I think even a nema 23 would have FAR more power than is needed in nearly any FFF printer.

[Flateric]

Not the power that we are after by going to servos.

[snoman002]

Not the power that we are after by going to servos.

Then what is it your going after? That's the main advantage. Or, it boils down to power....

OK, let me clarify, a servo has a feedback loop through the encoder so if you miss steps the system has the feedback to take care of it. A stepper does not. That only matters if your running the risk of missing steps, which you really only run if your asking too much of your steppers. Since 3d printers don't have super stiff ways, or heavy spindles, and are not running a big bit through thick metal they have a FAR lower risk of missing steps and loosing position.

Steppers do have the nice position of having more resolution. A 1.8° degree stepper has 200 steps, that's a fairly precise encoder on a servo. A .9° stepper has twice that which is a very high precision encoder. And that's all before micro stepping.


Edit:
It seems the resolution of a servo can be much higher than I was familiar with, still this number is variable depending on how much you want to spend. Still, its really only important if your loosing steps. Heck, glass scales on each tower would give you more resolution than servos as that's a complete loop (actual position of the carriage, not motor position).

[Flateric]

On a finely tuned and dialed in the is a sort of moire pattern phenomenon that can occur. The tighter the tolerances on the machine, the less slop on the moving parts there is the more this pattern makes itself visible.

The encoder on the servo is really just a nice bonus that comes along with them. They are also generally faster and more accurate when going faster than the stepper counter parts. With smoothie we have pretty much reached the peek speeds and sustain them that the extruded plastic can achieve before sheer and other undesirable problems happen.

So the reasons for higher res steppers and the next logical improvement servos would be for even further improved print quality.

[JohnStack]

Geesh, I can't wait to try the smoothie. All I'm waiting on now is rails for my mini-kossel. They are so incredibly brainless to put together. I think I spent more time finding parts than I did actually building it. With power supply, it is roughly $600 - depending on which hot end you use.

I have to thank someone in the forum for this killer nuts and bolts site:

http://www.boltdepot.com/default.aspx

And if you want to build your own, this site has most of everything else:

http://www.tridprinting.com/BOM/Kossel-Mini/

And the rails are coming from:

http://www.robotdigg.com/

[mhackney]

John, my mini kossel is almost done. Using the smoothieboard too - came in yesterday. Just figuring out where to mount the board now. Any ideas?

I think I was the source to bolt depot. Been using them for years.

Cheers
Michael

[mhackney]

Forgot to add, if you are using an ezStruder on a mini kossel, this mount rocks: http://www.thingiverse.com/thing:261017

Unfortunately, the guy posted the STL upside down. I fixed it, see attached.

Kossel_EzStruderMnt-FLIPPED.stl

(316.88 KiB) Downloaded 560 times

[Eaglezsoar]

Thanks for flipping that over. I have had all the parts for the mini kossel except for the extrusions. They will be here Monday.
Please let us know how the smoothieboard works for you and yet to be answered is if the smoothie firmware will allow dual
extruders.

[gestalt73]

You know, I don't know when they added it, but it appears that newest firmware supports dual extruders:

UPDATE: that's what I get for not reading thoroughly. It's in the multitool branch as testing, not part of edge yet.
https://github.com/logxen/Smoothie/tree/multitool

http://smoothieware.org/3d-printer-guide#toc21

You need to configure the corresponding temperature modules, and I'm not clear how toolchanges work, but it looks like they merged the fork from multitool.

Code: Select all

# Extruder module configuration
extruder.first.module_enable                       true             # Whether to activate the extruder module at all. All configuration is ignored if false
extruder.first.steps_per_mm                        400              # Steps per mm for extruder stepper
extruder.first.default_feed_rate                   600              # Default rate ( mm/minute ) for moves where only the extruder moves
extruder.first.acceleration                        500               # Acceleration for the stepper motor, as of 0.6, arbitrary ratio
extruder.first.max_speed                           3000             # mm^3/s
extruder.first.step_pin                            2.3              # Pin for extruder step signal
extruder.first.dir_pin                             0.22             # Pin for extruder dir signal
extruder.first.en_pin                              0.21             # Pin for extruder enable signal
delta_current                                    1.5              # Extruder stepper motor current

#settings for second extruder
extruder.second.module_enable                       true             # Whether to activate the extruder module at all. All configuration is ignored if false
extruder.second.steps_per_mm                        400              # Steps per mm for extruder stepper
extruder.second.default_feed_rate                   600              # Default rate ( mm/minute ) for moves where only the extruder moves
extruder.second.acceleration                        500               # Acceleration for the stepper motor, as of 0.6, arbitrary ratio
extruder.second.max_speed                           3000             # mm^3/s
extruder.second.step_pin                            2.4              # Pin for extruder step signal
extruder.second.dir_pin                             1.23             # Pin for extruder dir signal
extruder.second.en_pin                              1.22            # Pin for extruder enable signal
epsilon_current                                  1.5              # 5th axis

[lordbinky]

I made progress this weekend, but I'm fighting with my thermistors atm ( I just installed the screw-in thermistors) and I'm getting min-temp errors, oddly even the bed thermistor got this error a few times. I question the thermistor's connector so I'm just going to snip it off and solder it to remove it as a possibility. Other than that, the 70 odd layers before the heater cuts off have been looking good, and I have two calibrated profiles to play with, one with an improved center point using X/Y center offsets and one without (I do prefer repetier's tower offsets to changing the center point offset...it's maddening...). If I get that thermistor issue cleared up I'll be able to post information this weekend. If it goes REALLY smoothly, I'll swap in a RAMPS board I have and show comparison (killed one of the Rambo's drivers so I can only do an equivalent comparison). Still, I should be able to post something this weekend

I've been using software endstop offsets, and with .9° steppers it seems changes of .01mm are on the edge of ~= 1 step in height. I figure this because it usually, but not always, changes the number of steps for the offset by 1 step when I change it by .01mm. I find it entertaining on how 1 step in offset changes the feel in a paper test.

[gestalt73]

Hey lordbinky,

Are you using the software endstop offsets to make bed level adjustments instead of moving the setscrews?

[mhackney]

I'm finding the lack of documentation on the Smoothieboard production layout to be frustrating. I purchased the optional 5V switching regulator that is supposed to install on production boards and can't for the life of me figure out where it goes. I also read on the smoothieboard page that there are jumpers to power the MOSFETS from VBB without requiring a separate supply at the edge connectors or jumping (with wires) to them. Again, can't for the life of me find these. The "documentation" refers to a "non-border connectors section below" which does not have a diagram. With a little google detective work I could find the equivalent diagram for the beta boards but it doesn't really match the production boards. I'm thinning I should have gone with the X5, it has some very nice docs and connection diagrams. the smoothieboard came with 3 strips of pins that I have no idea what they are for.

Anyone find any good diagrams/photos/docs?

[lordbinky]

I swapped out the endstop switches for just the hall effect sensors (just the sensor itself, ~$1 ea.) and glued the magnets to the top of the cheapskates, so I only have software adjustments for endstops. I did this when testing repeatability of movements and calibrations with a mounted dial indicator. I did see improvement in the consistancy which is fantastic, which itself may not be directly because of the sensors themself but because I am not making physical contact with switches when homing. Using software adjustments wasn't an issue with the repetier firmware or the smoothieware firmware, besides just hoping it was implimented well in the firmwares since I didn't bother to check the code.

Also, new precompiled edge firmware was released the other day with settings changed to minimize thermistor noise and cleared up my thermistor issues.

Now I just have to figure out why the print stopped at ~700 of 900 lines and the board locked up...

[lordbinky]

I found these pages helpful, they also have more info on github but I can't get to that right now.

http://smoothieware.org/smoothieboard-schematic

http://smoothieware.org/lpc1769-pin-usage

I'll post pictures tonight of the jumpers on my board since I know which ones your talking baout

[mhackney]

The developer answered some of my questions so I started putting this together for him to share.

[lordbinky]

I read somewhere that they say to not use the jumpers if you will pull more than 1-2 amps over them. Were you just populating the board or were you planning to use them?

[mhackney]

Yes but there are 2 MOSFETS - small and big. With the smalls the pins for the jumper are already installed and since they are used for fans, using the jumper is no issue. That's what I plan to do.

[JohnStack]

Reading this with the unique smell of dread, fear, excitement....lol

[Eaglezsoar]

Reading this with the unique smell of dread, fear, excitement....lol

Just follow along with what has already been posted, what Michael posts and you will be fine.
You guys are blazing some new trails again and helping everyone in the process because these boards are
going to be used more and more in the future. It will be curious if the smoothieboard will be the board of the
future or the beagle form of the card. Time will tell.

[lordbinky]

Ah ok, I haven't used the small mosfets and I just use a terminal board to power fans/lights/etc. I don't have a layer fan setup so controlling the fans isn't a need yet.

[mhackney]

One interesting "feature" of the small MOSFETS - if you do add the jumper, the small MOSFET power connector actually outputs 12 volts when the board is powered on. You can use this to turn on a cold end fan automatically.

[prussiap]

I just started following this. I'm total Newbie with my R2. haven't really done anything fancy or tweaked too much. I just got my diamond hot end in from the kickstarter so by necessity i've been wanting to do something now .

Looks like azteeg x5, smoothieboard, or what i saw in this post beaglebone black might make for a good 3 extruder upgrade. I have no idea how you guys got to the config settings you needed or the ports for the machine?

How did you all learn what settings you needed. Years of experience and tweaking? Any blogs or tutorials?

I feel way behind the curve and this smoothieboard is both terrifying and exciting to see be implemented but i'm afraid i might screw up my working build and never get it fixed. Any tips, literature or thoughts I wouldn't mind some help. I'd like to do some of this for our local Hacklab too.

p.s. i didn't mean to hijack the thread, I'm curious want to follow this thread and could use some pointers on my own conversion.

[Generic Default]

I have three smoothieboards that I operate on a weekly basis, one on a 4 axis Taig mill, one on a sherline mini lathe and one on my Rostock Max V1. Soon I'll have another on my scratch built delta printer.


When you get it, there are only a few things to worry about.

First, get some thick gauge wire for the power input to the Smoothieboard. Clamp down on the ends with the screw terminals, and plug them in to the terminals on the board.
Make sure the polarity is right; a red light should come on when you power it on. It it doesn't, power it down immediately.

Second, connect all of your stepper motors. The white plugs that come with the smoothieboard suck, but they may be using better ones now. Make sure the connections are VERY good.

Thid, connect all of your thermistors and endstops. This is easy.

Fourth, plug in the smoothie with the USB to your computer and format the config file with the right numbers. You might need to download the newest config file from their github to run a delta machine since the default for smoothieware is cartesian.