Optimizing an A1's Print Speed.

[Frank]

There are a bewildering number of parameters in any 3D printer that affect the print speed.

I'm a novice at 3D printing, even after over a year of owning an A1 (#45), I can identify the following:

Machine & Component Design Choices. Including:

(a) the types/ratings of the stepper motors, the mass of the moving components,

(b) the choices of machine type eg. Fixed bed, moving bed, etc.

(c) the drive electronics (for the A1 – RAMPS).

Unfortunately, we're stuck with most of those choices unless we make significant mechanical or electrical changes.

While not impossible (for example: drilling holes in the Y mounting plate to lower it's mass), let's not consider those for the moment.

The Electrical setup of RAMPS:

(a) The current settings for the stepper motor drivers. I'm using the following settings:

X and Y .33 volts, Z(1) and E1=Z(2) .42 volts, E0 .47 volts

consult the Polulu documentation for info on how to set the current.

Motor current affects torque developed by the stepper. Given fixed mechanical/physical constraints this limits

the maximum speed you can achieve without loosing steps.

(b) RAMPS may also have a computational limit on the feed rate based on it's architecture and available computing speed.

That's a rumor, I don't have a reference, my closest comparison is Linux EMC which runs an interrupt/speed test on setup

and limits step rate based on the hardware.

However, I think recall seeing it mentioned as one of the reasons for going to the Smoothie Board.

The Firmware Settings in Marlin:

These are the numbers displayed by Marlin when you enter M503 in Pronterface. I'm using:

SENDING:M503

echo:Steps per unit:

echo: M92 X128.00 Y128.00 Z1600.00 E441.00

echo:Maximum feedrates (mm/s):

echo: M203 X400.00 Y400.00 Z5.00 E45.00

echo:Maximum Acceleration (mm/s2):

echo: M201 X5000 Y5000 Z100 E10000

echo:Acceleration: S=acceleration, T=retract acceleration

echo: M204 S2500.00 T2500.00

echo:Advanced variables: S=Min feedrate (mm/s), T=Min travel feedrate (mm/s), B=minimum segment time (ms), X=maximum xY jerk (mm/s), Z=maximum Z jerk (mm/s)

echo: M205 S0.00 T0.00 B20000 X8.00 Z0.40 E2.50

echo:Home offset (mm):

echo: M206 X0.00 Y0.00 Z0.00

echo:PID settings:

echo: M301 P11.69 I1.10 D31.02

presumably the limits on acceleration and step rates were derived by exhaustive testing during development.

But that may have been putting too much faith in Ezra, as the RAMPS was a substitute for the promised Smoothie drivers.

Software Settings in Slic3r: There are two places for setting print speed in the current release of Slic3r.

[I'm assuming that other popular slicing software has similar settings]

From: Print Settings-> Speed page

Speed for Print Moves

Perimeters 70 mm/s

Small Perimeters 50 mm/s

External Perimeters 75%

Infill 100 mm/s

Solid Infill 90 mm/s

Top Solid Infill 70 mm/s

Support Material 70 mm/s

Bridges 30 mm/s

Gap Fill 20 mm/s

Speed for Non Print Moves

Travel 110 mm/s

Modifiers

First layer Speed 30%

Acceleration Control (advanced)

Perimeters 0

Infill 0

Bridge 0

First Layer 0

Default 0

From: Printer Settings -> Extruder 1

Retraction

Speed 45 mm/s

Logic says that the Marlin settings above would override any attempts by the Gcode to go above/beyond their settings,

so the Slic3r settings have to track, to a certain extent, the Marlin settings.

One thing I never understood [or got the opportunity to question]: Slic3r has 'first layer' settings for print speed.

Why is it then that Ezra would recommend manually (via the knob on the RAMPS front panel) cranking up the

speed to 120% after the first layer was printed? Why not just set the first layer speed to 80% of “normal” ???

So, Given what I've posted, would any of you who have been operation consistently 'above and beyond' the

factory provided numbers care to share the numbers you are using, or the methods you used to get there?

Frank Pirz

Media Conversions

[John D]

Hey Frank - having lived with #15, some comments -

1] Marlin / Repetier stall out on an AVR at about 30/40 Khz - so you are hardware limited to the maximum speed you get out of your RAMPS board.

2] From PBC, the carriage type, screw, nut, and motor combination we purchased from Ezra had a design tolerance of about 330mm/s.  This did not include a calculation for the mass of the carriage or Y plate/bed.

3] As nearly as I can tell, the 6+ months of "extensively tested and tweeked" firmware from TL looked a lot like the Marlin defaults at the time.  Read into that what you will.

In the "hobby" 3D printer world, I stick with what the folks who wrote the firmware say - the firmware config is supposed to protect your printer hardware from a bad command from the slicer / control software by representing the maximum safe limits  - so in other words, it's supposed to be configured in such a way that if you send it something insane - like a "G0 X300 F40000000000" it does not make a nasty noise ans skip steps. In the CNC world, there are more dire consequences than in the 3D printing world, but the same theory exists.


What I've personally found - and please do take this with a grain of salt, as I've just moved an A1 over to Smoothie and am finding serious differences - for Marlin is :

1] My current settings for my motors are a bit different - I don't have my notes in front of me - but I used the PCB specs.  I'll update this thread later when I'm off the beach! ;-)  Current setting 100% affect the speed and acceleration settings.  I also moved to 8825 drivers.

2] I can safely move at 220mm/s on X with the stock setup, and 180mm/s on Y with acceleration set at 2200 for Y and 2600 for X.  "Safe" means I don't skip steps at all.  As mentioned above, I'm away from printers at the moment, and pulling a ccecil by posting from the beach - I don't remember my jerk setting - I want to say 12, but...

3] I don't use the Ez method of increasing speed with the jog dial.  I configure my slicer to print a the speed I want to achieve the quality I want.  I use silc3r's or KISS's ability to limit first layer speed to achieve good bed adhesion.  I did / do use the jog dial to test quality setting / maximums after a firmware / slic3r upgrade.

4] I can print reliably with "OK" quality results with PLA up to about 180mm/s as configured in slic3r/KISS.  If I print at this speed I see artifacts (ringing) in the exterior walls.  If I go over 180, I start starving my extruder.  I've played around a bit with increasing heat to allow for a higher flow rate, but the quality really starts to drop off.  I'm still running a .4mm j-head with the Phidgits 1:5.18 gearhead.  I did replace the TL hobb with one of blddk's straight cut ones. 

5] I generally print between 100-120mm/s exterior perimeter speeds for a good compromise of speed vs quality in PLA, marginally faster in ABS.  100mm/s really gives me smooth exteriors with very limited ringing.  ** Since I've moved to Smoothie, I'm having different results.  I think that this is more a firmware difference, but have not spent enough time yet to be sure.

I would also say that speeds and acceleration in Smoothie vs Marlin / Repetier are 100% "different".  I've not had the time to play with Smoothie enough to really say what's going on - speeds "feel / sound" different, and what worked consistently on Marlin does not translate to Smoothie.


Finally - I moved from one continent to the other late last yeah, and that saw me crating the A1 up and shifting it 6k miles.  We set it up when it arrived, but ended up having to pretty much rebuild / re-align it.  From that experience, I can tell you that very slight mechanical alignment issues can really cause you heartburn - my Y was consistently skipping, and it took a through cleaning, lube, and alignment to get it back into shape.  I complicated my life by switching over to Smoothie, which has - I think - given me better print quality at a higher speed, but also limited the "top end" speeds I can get out of the printer without skipping steps. 

This is something the better half wanted to print - one of our first really long prints with Smoothieware.