High Speed TAZ4.5

I have recently been speeding up the parameters on my TAZ because it think that it has the structural ability to handle much greater forces than the current settings generate. I have only done 2 upgrades to my TAZ4; I have the Hexagon nozzle installed and this upgrade that erv posted.
I found that even at slower speeds the bearing upgrade was a necessity to remove a lot of chatter and slop in the tool head. The hexagon nozzle has made a world of difference in my print quality and I think is why I can lay down plastic accurately at higher speeds. All these prints have been done with ABS. Could it be done with PLA or other filaments? I’m assuming that specialty filaments need to run slower based on their unique qualities.
I’ve printed 2 different sets of pieces at what I think could be close to the upper limits of how fast the TAZ can move around with somewhat mixed results. The first prints are of two types of E3D nozzle mounts this one that ogrethetoymaker put up in this threadhttps://forum.lulzbot.com/t/e3d-hot-end-first-pics-and-how-to-guide/704/1
and the type that mushoo designed on page 2 of the same thread.
here is the brim and first layer going on @100mm/s

here is the 3rd layer to get a feel for top and bottom print speed–160mm/s

here is everything in between @240mm/s–it takes about 0.15 seconds to reach max speed!

These parts came out impressively strong. They were really hard to break with a 12oz hammer and didn’t shear very easily. There were very few cracks along a seam and nothing snapped apart. I would say they were as strong as similar parts I’ve printed with LulzBot settings.

The second print was half a chess set. Sorry I don’t have better pictures I accidentally gave them to a friend before I got individual closeups. The pieces came out pretty clean, I think I have to slow down the perimeter speeds some for more intricate parts.

I did notice–you can see it in the videos–that when making some of the weird perimeter curvatures the machine would seem to stall a bit like the processor couldn’t load gcode blocks fast enough. What I think is really happening is that we are seeing the acceleration slowing the nozzle down when it is moving out to the next perimeter layer or going around a tight i.e. high directional acceleration area.

My settings:
I’m printing with this ABS right now. http://www.amazon.com/Filament-Printers-Makergear-Makerbot-RepRap/dp/B00BSW4HDA/ref=sr_1_1?ie=UTF8&qid=1425243143&sr=8-1&keywords=3mm+abs+filament
It’s only $17 a kg so I figured it would be good to use while messing around.
I realize that I have the travel speed set at less then the print speed, I have since upped the travel speed to 250mm/s.
Cura_Work_Speed.ini (10.6 KB)
Additional info on trigonometry and some physics on acceleration–in case you don’t know, if you do skim across:
I have the acceleration set to 2000mm/s^2 which is the most important setting. What happens is the ramp up time is shortened with increasing acceleration and the head can reach its intended speed sooner. For example if you are printing something with lots of curve and hairpin corners and the acceleration is set very low the print head may never be moving at the print speed that you assigned it. This is because any time direction is changed–which is constantly in an arc or circle–there is an acceleration associated with this. This is a good example of what the X and Y axes are doing graphically when they make a circle.
When either axes crosses the 0 plane they are at max velocity and minimal acceleration–even though the line is on a slope it is straight therefore no acceleration is happening. At the troughs and peaks the axes are experiencing max acceleration and minimum velocity–most curved parts of graph, most acceleration. Speeding up the image creates higher velocities due to the higher accelerations through the peaks and troughs–the curves get sharper and the straight lines get longer. Since forces are what cause strain on the physical machine the only way to change these forces are to either change the mass that is moving or the acceleration that it is moving at–F=ma.

My questions:
I’m making the assumption that the acceleration setting units are in mm/s^2 can anyone confirm this?
Are these settings, particularly the acceleration sustainable? Can the TAZ components handle these kinds of forces–I’m thinking the gt2 belts will be the first things to blow? Did I miss anything?

I was just wondering why you use such a large brim on your prints? I have seen a few people that do this and I was just wondering the reason behind this.

That is… surprisingly fast! I’m stuck at about 50mm/s (with something like 700 for acceleration) right now, for a variety of reasons: Dual extruder chassis is heavy as all heck, I’m using an E3D Volcano (at 3mm it just can’t heat the plastic up fast enough to go very fast - someday i’ll be switching to 1.75mm to address that), and apparently my X-axis is sloppy too. Just ordered some hardened rods and ball bearing bushing… things to address that.

I will say this: I frigging love how modifiable these printers are. Yeah, I’ve probably dumped a lot more than the $2200 I dropped on the stock printer, but… it’s just so much fun!

Oh, one other thing. I picked up some of these: http://www.matterhackers.com/store/printer-accessories/3d-printer-motor-vibration-dampers-4-pack the other day. I only put them on the X and Y motors (Z moves are ‘rare’ enough that I don’t feel the need to silence them at all). I had to get some longer screws to make it fit, but it is NOTICEABLY quieter. Not sure how that might affect speed, though.

Just use some common sense, think about how you’re are motors are liking those speeds…measure the temps, they’re probably ready to catch fire…I wouldn’t run those speeds for an extended period of time without added cooling for the steppers.
Hard on the belts and bearings also…all for what, saving an hour on a four hour print?

I was just wondering why you use such a large brim on your prints?

I don’t use LulzJuice or any type of bed prep so the big brim helps keep everything held down. Plus it heats the glass up a bit more and makes the temp a bit more even. And I like to just heat up the printer and hit start.

I’m using an E3D Volcano (at 3mm it just can’t heat the plastic up fast enough to go very fast - someday i’ll be switching to 1.75mm to address that)

If you have access to a machine shop you could have them make the heater block out of copper. It has about double the thermal conductivity of aluminum–2X as much energy can be transferred across the same volume.

and apparently my X-axis is sloppy too. Just ordered some hardened rods and ball bearing bushing… things to address that.

I’ve actually been thinking about moving my X axis to 12mm hardened tubing to increase the axis rigidity, I’m not going to bother till I make the move to a dually but it may be something you want to think about. Maybe rails would be a better upgrade than shafts altogether.

I picked up some of these: > http://www.matterhackers.com/store/prin > … ers-4-pack the other day.

I’ve had a look at those before and would like to pick up a pair for the X and Y axes. I’ve just been focusing on some of these other upgrades lately.

think about how you’re are motors are liking those speeds…measure the temps, they’re probably ready to catch fire…I wouldn’t run those speeds for an extended period of time without added cooling for the steppers.

I totally agree with you here. I’ve been thinking about making blower cages and putting active cooling on the X&Y steppers even before I started doing this project…I wonder how hot the steppers get inside all the enclosed TAZes out there. I don’t think that even at this acceleration they are getting pounded that hard. I’ve heard that the Ultimaker2 has an acceleration setting of 3000mm/s^2, granted the tool head is much lighter but the weight of entire x y carriage, rods, and blocks probably comes close to equalizing the forces between that printer and mine. The stepper controllers on the rambo limit the max current so they can’t ever get too much juice–which is my shotty safety net right now. I like your idea of actually measuring the temperatures of the stepper motors–why didn’t I think go that. I can measure all five and use the Z axis as the “normal” since they are pretty much just holding for most of the print. I’ll get on that right away.

Hard on the belts and bearings also

Hard on belts, very true. The belts are probably the components that are taking the biggest beating but there is plenty of room to put in 8mm wide belts and pulleys. I use lmu10 bearings and hardened rods so the bearings don’t have much of an issue with these speeds. I think they are a common part in laser cutters.

…all for what

True true…reality is that I’m doing this mostly for sport, to see what the TAZ can do. For the hobbyist this means little but for the people out there with a cluster of printers making production run parts speed may be everything–say you can squeeze in one more print a day on each machine. I’ve noticed that there is a fair amount of talk about maximizing print speed–bigger nozzles, higher temps, bigger heater blocks, etc. So I’ve gone the route of faster travel speeds and higher acceleration.

saving an hour on a four hour print?

Yeah, I understand you here but more importantly is that what Cura or slic3r give me for a print time is pretty close to the actual time it takes–I finish sooner on prints quoted to take more than 6 hours in Cura and I always finish faster than what slic3r says. This is important for me because I do a fair bit of freelance engineering–I have a lab in a building that houses a lot of small companies–and I charge customers my time to design the parts they need, machine usage by the hour, and filament. I also try to get customers parts or models the same day or next day–my biggest R&D selling point–so every hour counts. If I can make a tweak on a part by 10am and have it printed out and ready for the customer to retest after they get back from lunch they are usually really impressed. Plus when I say that printing cost is $25 an hour and their part is going to take 16 hours to print it becomes a much harder sell.

I forgot to post some closeups of the finished parts! And also the frame doesn’t flex all that much either although I may put in some larger gussets. Thanks all for the fantastic comments so far.

Ok, I see all your points…interesting…yes, if you’re doing production work, it possibly help a little, if you have the ball bearing style bushings, then thats’ a plus…so biggest concern is belts and steppers…just keep playing around and see what you find… if anything, faster for abs should be better, the layers have less time to cool. Keep it up and share the results1

Just to keep everyone updated with what I’ve been working on. Plus I was having fun with a high speed camera.
Here is a fun video.
Are the steppers and belts really working that hard?
Here is my first test with trimmer line. See more info on that here https://forum.lulzbot.com/t/for-those-who-want-to-print-with-105-trimmer-line/1489/1

Did you ever take a temperature reading on the stepper motors to see how hot they were actually getting?

No, not yet but I can wrap my hand around them even after 20+ hours into a print. So they can’t be too hot.