Re: Does someone want to know how fast they can print/extrude? look here
I assume you're referring to the Print Settings->Speed->Acceleration control (advanced) settings. Those settings control how quickly the printer can accelerate between speeds. Prusa has provided a tuned set of values for their printers. These are obviously maximums based on the motor speeds, but it's also a good idea to not accelerate or decelerate so quickly when printing fine details and corners. The various settings for infill, first layer and perimeters give you more granular control for these features that may matter more or less to you in terms of quality.
Can anyone tell me what Accelaration Control does? The description is very poor...
It seens the infill print alot faster (More motion) at 40mm/s than a perimeter at 40mm/s, is the Accelaration Control responsible for that?
Raw speed isn't going to matter as much on a small model as acceleration. The speeds are maximums, but it's how much time required to hit those speeds that matters. If your print contains small surfaces, you may never hit those maximum speeds. Acceleration is also behind a lot of the bad time estimates you get from different slicer software. Unless the software is using realistic values that match the printer, estimates will be way off. In the worst case, they estimate print times based purely on speed, not how long it takes to move between speeds. The result is overly optimistic.
To answer your question: Yes, acceleration likely is why you're observing infill printing much faster than perimeters. Given a long enough line, both might come up to the same speed (e.g. 40mm/s) but with a higher acceleration factor, infill will do so faster.
Factor in extrusion rates and it all gets a lot more interesting. How much filament does the printer lay down while accelerating up to those speeds? If you can lay down twice as much filament, even at lower speeds, your print may finish much more quickly.
and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan
Re: Does someone want to know how fast they can print/extrude? look here
I assume you're referring to the Print Settings->Speed->Acceleration control (advanced) settings. Those settings control how quickly the printer can accelerate between speeds. Prusa has provided a tuned set of values for their printers. These are obviously maximums based on the motor speeds, but it's also a good idea to not accelerate or decelerate so quickly when printing fine details and corners. The various settings for infill, first layer and perimeters give you more granular control for these features that may matter more or less to you in terms of quality.
Raw speed isn't going to matter as much on a small model as acceleration. The speeds are maximums, but it's how much time required to hit those speeds that matters. If your print contains small surfaces, you may never hit those maximum speeds. Acceleration is also behind a lot of the bad time estimates you get from different slicer software. Unless the software is using realistic values that match the printer, estimates will be way off. In the worst case, they estimate print times based purely on speed, not how long it takes to move between speeds. The result is overly optimistic.
To answer your question: Yes, acceleration likely is why you're observing infill printing much faster than perimeters. Given a long enough line, both might come up to the same speed (e.g. 40mm/s) but with a higher acceleration factor, infill will do so faster.
Factor in extrusion rates and it all gets a lot more interesting. How much filament does the printer lay down while accelerating up to those speeds? If you can lay down twice as much filament, even at lower speeds, your print may finish much more quickly.
So let me check if i understand this well, allow me that compare:
Two cars, both can reach 200km/h
Car 1 have a high accelaration power and will reach the 200km/h in 10 seconds (Infill)
Car 2 have a lower accelaration power and will reach the 200km/h in 20 seconds (Perimeter)
That only if the the road is long enough to reach that speed otherwise they will only reach the speed they can on that run length.
Is that correct?
If so does the printer deaccelerate when they have corners or it will continue to accelerate up the perimeter ends? ex: 60x60mm square
How can i calculate how much time will take based on accelaration and speed to reach the maximum speed?
For example: 1500mm/s^2 accel and 40mm/s print speed
sqrt(1500) = 38,729
40/38,729 = 1,032 s
is my math correct? 1,032 s to hit the 40mm/s?
Re: Does someone want to know how fast they can print/extrude? look here
How can i calculate how much time will take based on accelaration and speed to reach the maximum speed?
For example: 1500mm/s^2 accel and 40mm/s print speed
sqrt(1500) = 38,729
40/38,729 = 1,032 s
is my math correct? 1,032 s to hit the 40mm/s?
No sorry your math is wrong.
Acceleration (A) at a given time (t) when acceleration is constant is given by
(I) A = a where a is the acceleration at time t = 0
Velocity (V) at a given time (t) when acceleration is constant is given by
(ii) V = at + b where b is the velocity at time t = 0
Distance (S) at a given time (t) when acceleration is constant is given by
(iii) S = ½at² + bt + c where c is the distance from the origin at time t = 0
As the initial velocity is zero and we are measuring from the start point b and c become zero giving
(iv) V = at
(v) S = ½at²
Rearranging (iv) to give t becomes
(vi) t = V ÷ a
Substituting (vi) into (v) to get S in terms of V and a gives
(vii) S = ½at² = ½a (V ÷ a)² = V² ÷ 2a
So the time taken to reach the target velocity of 40mm/s with an acceleration of 1500mm/s² starting from a standstill is given by
t = 40mm/s ÷ 1500mm/s² = 0.2667s to 4sf
Distance covered during this acceleration is given by
S = (40mm/s)² ÷ (2 x 1500mm/s²) = 0.5333mm to 4sf
Regards,
Martin
Martin Wolfe
Re: Does someone want to know how fast they can print/extrude? look here
Thank you Martin for your deep explanation, it makes more sense now.
I only have a few questions left:
1) Accelaration values at Slic3r are only applied to X and Y axis right?
2) Are Max Accelaration values found by testing or by motor specs like torque?
3) Can two different accelarations be used? Like: X1000 Y800? Or will have issues? If so i guess X and Y motors should also be same model right?
4) There is an accelaration value to every axis, including the extruder, if E set to E10 for example which is very low compared with the default, does the firmware manage that and will start the line much slower or will only respect the XY acelaration and perform a underextrusion?
Re: Does someone want to know how fast they can print/extrude? look here
The gcode interpreter will scale back on XY acceleration in order to not exceed the maximum extrusion speed that you set (either in gcode or in firmware). You can prove it to yourself by setting X and Y accelerations to the highest possible values but setting a very low extrusion speed. The extrusion speed constraint won't be violated, and accelerations will be muted as necessary.
So, one way to do it, actually, is to set your maximum volumetric extrusion speed limit, and then set literally everything else to absurdly high limits. It will then manage things (maybe not optimally) so that your print doesn't violate the volumetric extrusion limit that you set. I did this once just to see if it would work, and it does.
I'm not claiming that it's the best way,but it is a way.
One could imagine a better firmware/slicer that could plan ahead and really optimize to minimize total print time. Not sure if those exist yet, but that would be the ultimate I think.
Re: Does someone want to know how fast they can print/extrude? look here
The restriction moderates speeds as well. You can really see this at work in the preview mode in Slic3r after you save the gcode file. The real power of this approach is that it will work even if you change nozzle sizes. You can also set a maximum for your extruder in your print settings in the unlikely event you find a filament that can be shoved faster than the E3D V6 can handle.
The gcode interpreter will scale back on XY acceleration in order to not exceed the maximum extrusion speed that you set (either in gcode or in firmware). You can prove it to yourself by setting X and Y accelerations to the highest possible values but setting a very low extrusion speed. The extrusion speed constraint won't be violated, and accelerations will be muted as necessary.
So, one way to do it, actually, is to set your maximum volumetric extrusion speed limit, and then set literally everything else to absurdly high limits. It will then manage things (maybe not optimally) so that your print doesn't violate the volumetric extrusion limit that you set. I did this once just to see if it would work, and it does.
I haven't set anything to higher limits, but I have been able to create a single printer profile and printer settings profile using the already aggressive Prusa defaults in Slic3r to use with all filaments. I set the max volumetric speed in the filament profile, and it throttles all of the other settings. Since discovering this a few weeks ago, I've gone back and been able to successfully print with all of the materials that caused me problems in the past, including: TPE, NGEN Flex, Fillamentum Extrafill Vertigo Gray, eSun woodfill, 3D Fuel Wound Up and others. With a new filament, I just copy an existing close profile, set max volumetric speed low, and work upwards. I have nearly eliminated extruder clicks, and haven't had the filament extruder spaghetti mess since.
I only have to create a printer profile for each nozzle size. It's made managing profiles much easier with Slic3r.
I'm not claiming that it's the best way,but is is a way.
It is certainly a quick way to get some usable results. From there, it's easy to identify issues and tweak other settings if desired.
One could imagine a better firmware/slicer that could plan ahead and really optimize to minimize total print time. Not sure if those exist yet, but that would be the ultimate I think.
Unfortunately, Slic3r seems to be the only one supporting this approach right now. I'd really love to go back to using Cura. At least now, I can get things close using Slic3r, then reverse-engineer those settings (speed, acceleration, etc.) for Cura. Unfortunately, Simplify 3D insists on using percentages for speeds, so I'd need to pull out my Excel spreadsheet to do anything with it.
and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan
Re: Does someone want to know how fast they can print/extrude? look here
Before you get over excited, the results are far from optimal if you rely only on the extrusion limit to govern your print. The reason is simple: you have to set it artificially low so as to not derail when going around tight turns, and that leaves you extruding much less (and slower) than you could at other times.
Re: Does someone want to know how fast they can print/extrude? look here
No need to set the extrusion limit artificially low. As I discovered in the speed tests when I introduced the increasing acceleration reducing the X-Y jerk values from 10 to 2 fixes the corners by forcing them to use gradual acceleration. The straight runs still runs still accelerate quickly. It is only when going above 300mm/s would the jerk values need to be raised bit. However with high accelerations on straight runs you will still get ringing.
Regards,
Martin
Martin Wolfe
Re: Does someone want to know how fast they can print/extrude? look here
What else remains to be done?
Re: Does someone want to know how fast they can print/extrude? look here
I am working on a new print fan nozzle at the moment. The default Prusa one puts far to much air on the volcano heater block and nozzle. At the moment my PETG Benchies are crap. Once I have a print fan nozzle that gives good results I'll rerun the PETG speed test. Hopefully will show an improvement on glossiness.
Regards,
Martin
Martin Wolfe
Re: Does someone want to know how fast they can print/extrude? look here
These firmware settings are potentially slowing things down. They are defined in the file configuration_prusa.h. I upgraded them, but maybe they can be set even higher. Here I show both the original settings (commented out) and the upgraded values:
//#define DEFAULT_MAX_FEEDRATE {200, 200, 12, 120} // (mm/sec) max feedrate (M203)
#define DEFAULT_MAX_FEEDRATE {500, 500, 500, 240} // (mm/sec) max feedrate (M203)
// #define DEFAULT_MAX_ACCELERATION {1000, 1000, 200, 5000} // (mm/sec^2) max acceleration (M201)
#define DEFAULT_MAX_ACCELERATION {9000, 9000, 9000, 9000} // (mm/sec^2) max acceleration (M201)
//#define DEFAULT_ACCELERATION 1500 // X, Y, Z and E max acceleration in mm/s^2 for printing moves
#define DEFAULT_ACCELERATION 9000 // X, Y, Z and E max acceleration in mm/s^2 for printing moves
//#define MANUAL_FEEDRATE {2700, 2700, 1000, 100} // set the speeds for manual moves (mm/min)
#define MANUAL_FEEDRATE {9000, 9000, 9000, 9000} // set the speeds for manual moves (mm/min)
//#define NORMAL_MAX_ACCEL 2500 // Y-axis max axxeleration in normal mode in mm/s^2
#define NORMAL_MAX_ACCEL 9000 // Y-axis max axxeleration in normal mode in mm/s^2
//#define NORMAL_MAX_FEEDRATE 200 //max feedrate in mm/s, because mode switched to normal for homming , this value limits also homing, it should be greater (172mm/s=9600mm/min>2700mm/min)
#define NORMAL_MAX_FEEDRATE 900 //max feedrate in mm/s
Note: I'm not claiming that these are the best values to set them to. This was just a quick pass on my part from when I previously looked into this. Also, the upgraded values need checking to make sure they are not out of range.
Re: Does someone want to know how fast they can print/extrude? look here
It is not the fan that is the problem. It moves plenty of air. Just with the existing fan nozzle design far too much of it is wasted on the heater block and not enough hitting the printed part where needed.
I am hitting a problem with my Volcano only print fan nozzle. It's air path is so efficient that at 12%min speed 14%max speed I am getting under and over cooling in the same PETG Benchy.. Prusa default for PETG is 30% - 50% fan speed. I am trying at 13% - 13% at the moment but it looks like I will have to deliberately include some sort of damping in my design so that t can actually have a range. Definatly shows that the default Prusa print fan nozzle design wasted far to mach of the air flow on the heater block and just missing the part.
The core of the design is a plenum with a linear nozzle wrapped around the Volcano nozzle. Unfortunately as I said it has very little air resistance and all off the air it moves is aimed at the part being printed where the new filament is being added and only a tiny fraction of the air hits the nozzle tip and none on the heater block.
The first photo shows an early draught being printed illustrating the basic concept. The final two photos show it fitted to my printer. Note that it has a passthrough for the PINDA probe.
[EDIT]
The 13%-13% fan speed also shows over and under cooling. Definitely not enough control at 13% with that low air resistance. Oh well inbuilt VCDs (volume control dampers) here I come. 😛
Regards,
Martin
Martin Wolfe
Re: Does someone want to know how fast they can print/extrude? look here
Cool! (literally)
Say, not to overload you, but when you get the dampers working, and the cooling to your liking, how about 1-4 LED's to light up the extrusion as its coming out? I suppose it could be a separate "thing", but somehow integrating it seems like it would be ideal.
Anyhow, grist for the mill. I would agree that getting the cooling right is higher priority, not to mention a lot harder to figure out.
Re: Does someone want to know how fast they can print/extrude? look here
Defiantly got to get the cooling correct. At the moment only PLA prints are any good. The default Prusa fan nozzle with the plated copper volcano and nozzle was giving a to cold bow wave on a Benchy. My universal nozzle was better but needs work once I have the volcano one correct I will feed back what I learned from it into my universal one.
Regards,
Martin
Martin Wolfe
Re: Does someone want to know how fast they can print/extrude? look here
I continue to have worse speed from PLA than from PETG. For instance, with PETG I can set speed=12000mm/minute on this test cube:
https://www.thingiverse.com/thing:2892579
and have it print acceptably. However, if PLA, I can't do much better than speed=5000mm/minute on the same test object, and that's with setting XY jerk to 1.
Has anyone else noticed anything similar? Maybe I'm overgeneralizing from the one spool of PETG and the one spool of PLA that I've been using for testing.
Re: Does someone want to know how fast they can print/extrude? look here
I'd expect it to vary by brand. Some of the better filament brands provide recommended maximum print speeds. You may be getting into "you can do that, but you're out of warranty" territory.
[...] Has anyone else noticed anything similar? Maybe I'm overgeneralizing from the one spool of PETG and the one spool of PLA that I've been using for testing.
and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan
Re: Does someone want to know how fast they can print/extrude? look here
My volcano arrived today and already installed on MK3.
My current limt is 30mm/s^3 for volumetric speed.
Running 0.6mm nozzle up to 0.45 Layer Heights
For 0.45mm layer i use 92mm/s without problem so i will settle at 29mm/s^3 to give a small margin. (PLA tested)
0.10mm layer 200mm/s
0.15mm layer 200mm/s
0.20mm layer 200mm/s
0.25mm layer 166mm/s
0.30mm layer 138mm/s
0.35mm layer 118mm/s
0.40mm layer 104mm/s
0.45mm layer 92mm/s
Re: Does someone want to know how fast they can print/extrude? look here
I took a look at some of the high-speed PLA filament manufacturers:
and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan
Re: Does someone want to know how fast they can print/extrude? look here
I took a look at some of the high-speed PLA filament manufacturers:
Interesting! I never knew there was such a thing as high speed filament, let alone high speed PLA. I wonder what makes it so? Also, according to the first one, my PLA speeds are about normal for "regular" PLA.
Hmm.... Here's one that claims it can print at up to 400mm/sec: https://www.imakr.com/us/en/all-filaments/779-fiber-force-high-speed-pla-filament.html
Re: Does someone want to know how fast they can print/extrude? look here
Interesting. Not even that expensive.
[...] Hmm.... Here's one that claims it can print at up to 400mm/sec: https://www.imakr.com/us/en/all-filaments/779-fiber-force-high-speed-pla-filament.html
and miscellaneous other tech projects
He is intelligent, but not experienced. His pattern indicates two dimensional thinking. -- Spock in Star Trek: The Wrath of Khan