Marlin: Mostly awesome ...

I started experimenting with the Marlin firmware last night. I ran my printer (makergear / ramps 1.2) with default speeds settings from SFACT and the operation was very smooth.

For comparison I ran the same gcode from SFACT with Sprinter and the operation of the bot was rougher. When doing the infill Sprinter seems to reverse directions at edges very rapidly creating more vibration. Marlin on the other hand is doing something to make the direction changes much smoother. I am hoping Marlin will allow me to keep my machine in tune longer since it won't shake things loose as rapidly ...

Of course I do have one problem. The movement in the Z direction is not reliable. Sometimes the Z axis will rise with layer changes and sometimes not. I have changed the default max acceleration in Configuration.h to match my Sprinter settings but this has not solved the issue.

I am able to reflash the arduino back to Sprinter and I have no such problems. Layer changes are reliable so I'm nearly ready to rule out a basic hardware issue.

I have also noticed a Z_LATE_ENABLE setting that sounds promising that I will try this weekend.

Anyone had similar issues?

I think I had similar issues, but it was a year ago and I've forgotten the specifics. It may be that you're tripping the thermal shutoff on the stepper drivers. Try a long manual Z move and see if it pauses intermittently on the way, stopping short.

Some firmwares shut off the Z steppers between moves by default to keep them cooler. Others don't. I think most of them have a setting, and "late enable" sounds something like it. I saw a discussion once on the merits of disabling Z between moves, but I can't remember what the outcome was. One side argued against it, saying that you would lose any partial steps between moves with this setting. I can tell you from experience with my MG Prusa that you can find stepper driver current and acceleration settings with the steppers wired in parallel that will work reliably without tripping the thermal safety. Unfortunately, I can't give you any pointers, because it's been working for a year. I don't remember what's set to what, and I can't check on it now.

It's good to hear that Marlin is working with RAMPS 1.2 again. It tried one of the release candidates for 1.0, but one of the tables I needed was marked "fix me". As soon as pronterface connected, it would shut down on a thermal sanity check. I'm running ancient software. It's a few months old!

try sprinter experimental

I had an issue with vibration with sprinter and on narrow areas of cross hatching it would try to reverse so rapidly that it would miss some steps and the layer would get shifted. I reduced the acceleration settings in configuration.h and it has been fine ever since.

George

Marlin RC2 has Z-axis motors enabled during idle, by default. I believe Sprinter disables the idling Z motors by default. There's a line in Configuration.h to change this.

Also try reducing Z motor speed and acceleration by about 30-40%.

Did you end up getting this solved? I was having very similar issues with marlin and I ended up just rolling back to sprinter. On mine the Z axis would sometimes refuse to move. The typical case was when I was preparing for a print and would home all axis and then raise Z a few mm to preheat. When I started the print the first layer would be at the height I raised it to for preheating. When it was in this error state I seemed to need to reset the board to get the z axis to move again. A few of the ‘successful’ prints I had also looked like they had dropped z steps.

No, I never have gotten past my problem with Marlin. I still would like to make it work but I have moved onto sprinter experimental as se5a suggested.

Sprinter experimental actually incorporates some of the Marlin code (like the planner stuff) which makes it run very nicely. My printer is running much smoother with Sprinter at this point.

The latest things I tried to prevent Marlin from skipping z level changes was
* M201 Z50 (set maximum z acceleration to 50 mm/s^2) - matches what was in my original Sprinter
* M203 Z2 (set maximum feed rate) - matches what was in my original Sprinter

I had great hopes for these two gcodes. I could tell the z axis was moving slower (like my original sprinter settings) so I thought I might have solved it. Printing quickly showed me though that Marlin was skipping at least half the z level changes.

I actually believe I am facing a z motor driver overheating issue. On extended moves up/down of the z axis the motors will pulse ... on for a little ... then off ... then on ... etc

Yes sounds like the current is set too high and it warms up and shuts down for a while. Whether the Z moves is just luck if it is in thermal shutdown mode or not at the time of the move.

The difference between firmwares is probably whether Z is disabled between moves. In which case it would not have time to overheat.

---

[www.hydraraptor.blogspot.com]

nophead Wrote:
-------------------------------------------------------
> Yes sounds like the current is set too high and it
> warms up and shuts down for a while. Whether the Z
> moves is just luck if it is in thermal shutdown
> mode or not at the time of the move.
>
> The difference between firmwares is probably
> whether Z is disabled between moves. In which case
> it would not have time to overheat.


Thanks. This is totally my problem. When I inspected my Z axis I see that the trimpot on the Pololu is FUBAR (free spins all the way around). I assume I'm over currenting the axis. Since the stepper doesn’t get hot and I have a fan on the electronics I'll just roll with it in Sprinter till it fails completely.

Some pots do spin all the way round. The best way to set it is to measure the voltage on the wiper and set it to 0.4V on a Pololu for 1A, up to about 0.6V if you have a fan on it and your Z motors are in parallel.

---

[www.hydraraptor.blogspot.com]

I was thinking of trying Marlin but seeing the mention of Sprinter experimental I thought I would give that a go, seems much smoother.

Printing at 60mm/s

Turn up your non-printing movement speed to save some time, a Prusa can do 200-300 mm/s just fine with acceleration.

Also, the acceleration rate affects the smoothness a lot. 500 mm/s*s is silky smooth, 1000 mm/s*s still nice but snappier and at 2000 mm/s*s you really start to feel that the printer is being stressed but at the same time the print head just seems to disappear from one place and appear at another instead of moving...

Quote
Turn up your non-printing movement speed to save some time

Yes I forgot to increase that when I speeded up, Sprinter experimental seems have a simplified acceleration setting with the default at 1000 mm/s^2
What is the Prusa speed limit? Also depends how fast you can melt the filament I guess.
60 200 1000

The top non-printing speed is limited by the maximum step rate of the electronics, which depends on the firmware implementation, and how many steps per mm are needed. 320 mm/s is one figure I have read.

The maximum printing speed is usually limited by the extruder. I personally don't think it's worth it to try to break 100 mm/s, printing thicker tracks and layers instead brings better quality in less time.

ttsalo Wrote:
-------------------------------------------------------
> printing thicker
> tracks and layers instead brings better quality in
> less time.

I find almost the opposite. Our speed is limited by the volume of plastic we can melt per second and when you increase your layer size the volume goes up dramatically and limits your printing speed because you can't melt the plastic fast enough. And with the larger layer sizes you do not get the layers squished down as much so the adhesion is not as good. Plus you have more hot plastic on the part which takes longer to cool because of how thick it is. But if you double your print speed and half your layer height you will be printing with the same volume of plastic per second but it will be squished more giving better adhesion and cool faster because the majority of plastic extruded on each pass is exposed to the air. Further more it lowers the amount of ooze when you print a lower layer heights for some reason (tested and proven on three different bowden cable machines). So in my experience I can print an object at 0.15 in the same time or less than the same object at 0.3 . The only issue I ever have with printing at lower layers (and why I print at 0.25) is bridging does not work well when your w/t is greater than 2.

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver