Modified Taz Speed trials - 40mm Calibration cube

Anyone want to know about how fast my modified taz can print things at? Me too. It turns out, if my math is right, it can go at least 400mm/s, faster if I figure out how to make it go faster than that.

Anyways, here’s a long boring video showing it printing really fast

http://youtu.be/DLWHBJgYJD4

Cube pictures:






bottom:

FWIW, I have tested TazStiff at 800mm/s travel speed with 12000mm/s^2 (yes, ~1.2G) acceleration.

EDIT: Not to sound unimpressed of course. Achieving 400mm/s print speed is really impressive.

I’m pretty sure I can push it faster, I just need to adjust thinges to allow it to go faster. You should post a speed comparison video.

Will do. Can’t print fast with this machine yet though, still using an ancient extruder.

EDIT: Also try printing larger parts, and try the extruder advance feature in the Marlin firmware. You may find acceleration is slowing things down overall at the speeds you are using.

I’m totally in on this. I’ll do a speed trial this weekend on mine also!

I doubled all the settings:
http://youtu.be/rajsR6qSiWg

acceleration was at 6,000 m/s^2 in that video. I think I could still go faster on the mechanicals, but there is no way in hell the extruder is going to keep up any faster than that at the moment without some modification.

I gotta do your mods!

How accurate were the dimensions at that astonishing speed?

I actually ended up 0.1mm under on all sides after it cooled down except top to bottom, which was dead on at 40.00 on the second higher spoeed one. The first one was 0.2 over on all sides except the top which was .05 over. Normal speed prints I usually end up at 40.00 on all sides. definitly some extruder oddness at the higher speeds though.

What was your acceleration for the first 400mm/s video? On a 40mm cube, you need at last 4000mm/s² to reach 400mm/s at least on the middle of the perimeters.

Anyway, you prints are looking amazing and the speed is incredible :sunglasses:

first video was 3000 mm/s^2. Second one was 6000 mm/s^2

piercet, I have to bump your thread. As I’m not satisfied with my RJUM bearing mod (I think the rods itself may introduce the ringing…) I remebered your speed test and your perfect 200mm/s benchy at thingyverse…

So I’m thinking about to go the Openrail route. As it will be very hard to source some of the parts here in germany and additional taxes if I order from US, I have some questions to you to be shure I get the quality I want in a single part order step.
To clarify what I want, my first priority is surface quality. At the moment I’m limited to the stock jerk and accel. settings (8mm/s jerk, 500mm/s² accel.) to get at last only a small amount of ringing, and therefore to low speeds (45mm/s for perimeters, 50% for external perimeter, 15-20mm/s for solid infill) to keep the edges sharp and the squished filament buildup at top infill direction changes at a minimum.
The only way to solve this, as long as Marlin has no working nozzle pressure control features, is to allow higher jerk and acceleration.

As far as I can see on your pictures, your cube shows no signs of ringing even at 3000mm/s² in your 400mm/s test? Do you have some other examples of higher speed prints where we can see top surface and sharp edges quality?
Which of your mods (X,Y,Z) bumped the quality the most, would it be OK to stay with the rods on the Z axis?
My biggest concerns are about the heat inserts. While I found some shops that have a small selection of Openrail profiles, I found not a single one where I can get heat inserts :confused: Hopefuly I can modify most of your parts to be used with normal nuts in some way…

I haven’t printed many things at that extremely high print speed because I didn’t want to put very much wear and tear on the machine. If you would like to see a specific part printed at high speed, let me know and I can print it. At the higher speed pass on the cube, I was under extruding a bit because the hot end couldn’t keep up, but there was no ringing or wobble present at all even at that speed.

The modifications I have made to my printer that have had the most effect on improving surface quality at speed are the following, in this order:

  1. The Openbuilds X axis
  2. The Openbuilds Y axis
  3. The Taller Extruder mount (eliminates forward nozzleflex)
  4. The reinforced extruder carriage (or the Taz 6 carriage)
  5. Reinforcing the corners with this bracket: (http://i-t-w.com/parts/brackets-misumi-hblfsnb5-series-5-aluminum) - stiffens the frame
  6. The Anti wobble modification
  7. The aluminum bed plate to main frame connector pieces sourced from Techsavvy34 in the forums here
  8. The Z axis openbuilds modification - does help, but it is a minimal amount (less than I was expecting actually) and the Anti wobble has a similar effect

The openbuilds X and Y make the most immediate and dramatic difference in printed part quality. The carriage modifications are incremental. The anti wobble can be pretty dramatic difference as well if your machine has wobble. if it does not have noticeable wobble, its less important than the other reinforcements. The corner frame reinforcement pieces are required for very high speed prints, as the frame will attempt to rack otherwise, but are less important for normal speed printing.

The Z axis doesn’t hurt anything, but it doesn’t help as much as I was expecting that it would. After giving it a lot of thought, I believe one of the biggest sources of offset in the x and y direction, which the Z axis should constrain, is actually flex in the Y bed rods. As the belt moves the bed in and out, it tends to exert a small but noticeable offset force on one side or the other. The X axis, being on stacked rods has a similar but less noticeable kick up of the nozzle. Since Z moves much slower and less frequently, the cheapest effective modification there is likely going to be replacing the bushings with a more precise set, and maybe hardened rods.

There are two other modifications that I also recommend strongly, that don’t necessarily improve surface quality. The Aluminum heated build plate to replace the glass plate allows for faster printing and much better ABS part adhesion to the bed. the glass bed is heavy and doesn’t spread heat as well as aluminum. The tricky part is finding a 12" x 12" aluminum plate that is 3/16" thick and not warped in any way by cutting.

The other modification I recommend is the Z axis endstop micrometer target modification. if you aren’t going to install auto leveling, it’s the best way to maintain a highly precise Z axis height

For the heat set inserts, it is possible to order them from one of the manufacturers in Germany. http://www.spirol.com appears to ship there. You want the series 14 tapered inserts, M3, and m5 (a couple M2 are also used but can be worked around.) Their minimum order quantity is 150 inserts, and I have no idea what the cost is, but that might be an option. They are also called ultrasonic weld inserts, so they may just be under another name auf deutsch.

One option, if you cant find the actual inserts, would be to modify an “acorn nut” or one of those nuts with the dome on one end. If you carve some grooves in the surface of the nut to rough it up a bit, then insert it like a heat set insert, it may hold.

Hope that helps!
Tim

Thank you very much for that detailed answer!

I found this picture of nopick in the X axis thread as a comparison between stock and X-Y mods installed: https://forum.lulzbot.com/download/file.php?id=4847&mode=view
No further questions about print quality :sunglasses: :laughing:

I’m not planing to print at 400mm/s, but 60-100mm/s with good surface quality would be nice. I know the hotend will reach its limits somwhere between 10-15mm³/s.
What’s your print speed, accel, jerk where you feel comfortable not to damage the printer with no ringing?
One last question, you wrote you designed the metal back plate because it’s not possible to print a backplate with the necessary strength that’s lightweight enough. I assume lightweight means heavier than you want it to be in this case? Or is it a real disadvantage to use the printed one.

I will start with my part list and where-to-source list today :smiley: Thanks for sharing your cool projects with us!

You’re welcome!

That picture Nopick uploaded is about the best advertisement for the X axis modification in the history of printer modifications!

I generally print around 150mm/s and 1250 acceleration. I could go much higher than that and not risk excess wear, but above that speed, it starts getting a bit noisy so unless I need something in a hurry, I generally leave it there.

The metal backplate is much lighter than the plastic one. It’s about 1/10th the weight. That being said, the plastic versions are still pretty light. I’m just a firm believer in making the extruder carriage as light as possible to preserve quality and speed. The plastic ones are Just as strong as the metal one, just heavier.

This is awesome! I was curious if you could share your slic3r settings files. I am interested in tuning my TazStiff as well and it would be useful to have some target values to aspire to.

Sure, here’s my general go to configuration, and then the speed trial config for comparison.
mainconfig_0.50mm_nosupport.txt (3.59 KB)
speedtestconfig.txt (3.59 KB)

Thanks for sharing your setting Pierce! I’ve been meaning to ask. I’ve been attempting to replicate your speed tests on my own machine. I’ve had great success with everything except my corners are a little “off”.

I can’t speak highly enough of Piercet and his open rail mods. The print quality improvement is fantastic… Plus it’s nice not feeling the metal rods flex when I change filament… The only flex I’m seeing now is actually in the X carriage bracket/connecting parts and not on the open rail. I’ve used Piercet’s diety of a printer as a template for my own. If lulzbot isn’t paying him, then they should be.

Z wobble nut, X-axis open rail, Y-axis open rail, reinforced extruder carriage, reinforced extruder mount. I have skipped the Z axis upgrade for now.

I see now I’m still missing the reinforced corner brackets and the aluminum bed to main frame connector? (not sure what that is, I’ll have to do some research when I get home).

Honestly the printed back plate for the X-axis is legit and not overly heavy. Myself and many other have gone with the printed version and there seems to be no problem. I will just say that the current posted BOM is for the thinner aluminum plate So get longer screw :wink: The biggest factor for weight was when I upgraded to Pierces extruder carriage. Seriously when I pull my tool head off, it’s a monster…