So recently, Albert from 247 printing posted a video on his sub 9 minutes 3DBenchy that he printed on his basically stock but input shaper tuned VORON 0. I thought that this was the opportunity to also just play around with the slicer and make a video for you on what I had to do to get below my 10 minutes mark. Turn's out, it's not that easy and at this moment I still didn't succeed because there is obviously more to reaching that goal besides some slicer trickery.

By the way, if you remember my first video on the VORON you'll know that I had huge problems with the Carbon Fiber Nylon that I used for my parts because they slowly deformed over time. People were asking if I already fixed that. Well, I could have just printed all parts from ABS and be done with it, but that's not how I and also you will learn a lot. Instead, I'm currently working on a material creep test that I came up with. For that, I printed test pieces of a bunch of different materials that I load in compression by tightening a bolt with a torque wrench. Torque and the pretension in the bolt are proportional to each other. So, when the material creeps away under the load, the pretension will reduce, and I will be able to retighten the bolt again by some amount that I measure with these dials. This is what I'll be doing in regular intervals at ambient and at slightly raised temperatures to compare how different materials behave. If you have any thoughts on that procedure, please let me know and also make sure to be subscribed and have selected the notification bell to not miss the results. Despite these tests that are ongoing, I actually got the Carbon Fiber Nylon to work as well. How, you might ask! Well, I annealed some of the parts at 160°C for two hours and had them now installed on my VORON for several hours of high temp printing without any sign of deformation so far, pretty impressive! The cool thing about annealing Carbon Fiber Nylon is, that it basically doesn’t deform during that procedure in contrast to PLA! I'm still not sure what I'll be using for the re-build and will decide that after I finished the creep test.

But let's get back to the 10 minutes 3DBenchy problem. If you do the maths, you can already see why getting to that time might be complicated if you don't have the proper setup. On average, a 3DBenchy that was printed per the #SpeedBoatRace rules in ABS weighs around 9g. This equals to 14mm³/s extrusion rate if you'd extrude at a constant rate. Since you constantly accelerate and decelerate and have travel-sections in-between, you'll probably need at least double of that amount peak in reality because the high-speed sections need to compensate for the slow extrusions during speed changes. You can check your Slicer at which extrusion rate you’re running and if you are still in the capable range of your extruder. A normal E3D V6 hotend can handle around 10 to maybe 15mm³/s if you increase temperatures. For anything more that you want to extrude, the feeder will start skipping or stripping. So, I have a SliceEngineering Mosquito on my Voron 0, and I intentionally started with the standard version. There is also the Magnum heatbreak available that increases the meltzone but I wanted to find the printer's bottlenecks one after the other as a learning experience for all of us.

I already noticed with fast test prints that my standard setup could not extrude as fast as I needed, which showed up in a lot of underextrusion. To find out where the limit of the regular Mosquito was and how the High Flow variant compared, I did an extrusion test, which you might have already seen on the channel. For that, I use a simple GCode to first prime the nozzle and then, feed a specific amount of filament through the hotend and adjust the speed. By weighing the extruded plastic, I can determine if there is slip in the system due to slipping or stripping of the filament.

I did the extrusion test not only for the recommended temperature range of the DuraPro ABS from Extrudr that I used, but also at 280°C nozzle temperature to find out how raising temperatures helps with melting the plastic.

During these extrusions tests, you can not only see how the shape of the filament spiral changes, but at some point, you can hear the extruder skipping. With the normal-flow Mosquito, we can see that there is a significant difference between the upper and lower recommended extrusion temperature. Using 210°C, we can see the first big decrease in extrusion rate at around 15mm³/s, with 240°C extrusion temperature, this shifts to 20mm³/s. Using the high flow heatbreat the picture is pretty similar though the longer meltzone helps to extrude around 30% more material. I was honestly a bit disappointed about those results because I thought the longer heating zone had a significantly higher impact. If we take a look at the melt-zone it's plausible, though, because it also only increased by maybe 30-40%.

Interestingly, at 280°C, which is quite a bit above the recommended temperatures for ABS and should only be used when printing very fast, the impact looks different. With the normal flow hotend, we again dip drastically at 25mm³/s, whereas with the high flow hotend, even at 35mm³/s we still only under-extrude by around 10 to 15%. An interesting behavior I currently still can't wrap my head around is, that with the Magnum heatbreak, the extrusion amount constantly reduces and doesn't stay flat in the permissible extrusion range and only dips down more when it can't transfer enough heat into the materials anymore.

If you're by the way wondering how the print quality differs between the normal flow and high flow hotend at "normal" speeds, check out this Benchy. The general quality is the same, and I'm only able to see a bit more stringing, but that's honestly because I used the same GCode with the same retraction values. Increasing them for the longer melt zone variant might already do the trick. I'd also like to compare other normal flow and high flow hotends against each other, so if you have any suggestions for contestants, let me know in the comments!

With those results, I thought, that I did have the setup with the high flow hotend now to attempt the 10 minutes Benchy. Since printing ABS is quite smelly, I moved the printer from my office down to the basement. Unfortunately, all of my fast printing attempts failed after this relocation. Even those that still printed with the old setup. The printer was able to heat up to the set temperatures, started printing, but then stopped the job after a bunch of layers with a thermal runaway warning. First, I thought I screwed up the wiring while exchanging the heatbreak, but then it dawned on me. The room in which I printed before has around 22°C, my basement rather something like 15°C. This causes more convection heat loss that the heater cartridge needs to supply. This is usually not a problem, but the longer meltzone of the Mosquito Magnum is now capable of really melting a quite significant amount of material. And this energy for heating and melting the plastic needs to be supplied as well. If you do the math, for prints on a standard printer with standard parameters, the amount of energy you need to melt the plastic is negligible and usually below 5W. So the 40W of your heater cartridges are only necessary to heat up the hotend in a sensible amount of time and, of course, compensate for the convectional losses. When we're speed printing, the energy required to melt only the plastic can be as high as 20W, which is almost half the cartridge capacity. Combine that with the convectional losses, especially also due to the cooling fans and the heater is not able anymore to handle that task. Honestly, I even think that the PID loop might be a problem as well because we tune only during the task of heating the heater block. On a normal setup that might be sufficient, because melting requirements are negligible but on this high speed setup probably more than 50% of the total energy is used for the melting process and that's not considered during tuning.

Anyway, enough talking. I tackled that issue with a silicone sock that many love and some also hate. Initially, the OG E3D silicone sock was intended for keeping the hotend clean from sticky PETG but it turns out that silicone socks are also great for insulating the hotend, so reducing the convection losses. SliceEngineering doesn't sell socks for the Mosquito, and I wanted to quickly continue with my tests, so I just grabbed a mold design off Thingiverse that's actually intended for use with high temp sealing compound. Since I still had some two-part high temp silicone around that I also used in the past for my other silicone socks, I injected this material into the mold and had a nice silicon sock the next day that fit okay around my heater block.

And look at that! Using the sock, I didn't have any problems anymore with dipping temperatures and thermal runaway as before, so it clearly works. Because I was curious, I did another extrusion test run to evaluate the influence of the sock on the extrusion rate, which was really interesting. I made another reference curve at 240°C without a sock and then ran exactly the same code after installing it. As we've seen previously, the plot without the sock, slowly reduces in extrusion rate until it finally dips down. The one with the sock installed is basically level until 20mm³/s and only then starts dipping down, which is the behavior we're looking for because we want to have constant flow regardless of the print speed! I suspect this improved behavior might be due to the temperature gradient in the heater block. If not isolated, the temperature will slowly drop starting from the heater towards the nozzle. Isolating it will reduce convection and therefore keep the temperature more constant within the whole block, which results in more equal melting. So again, a nice prove that shows how effective these small silicone socks can be if they don't fall off and give a bad day!

Finally, there was the question how different materials extrude at high speeds. Not every ABS or PLA is the same. There are tons of different compositions for each polymer class on the market that differ in properties. One property that might be interesting for us is their viscosity while molten because we’re not only getting extrusion problems when we’re not able to melt a plastic properly but also if the viscosity of the filament stays too high once we push it through the small nozzle orifice. A thicker melt means more back-pressure which means more work for the feeder which results in stripping and skipping. For this test I compared ESun ABS+, Extrudr ABS, Redline PLA, DasFilament PETG and Tiamet3Ds Nanodiamond PLA that they say is actually really well suited for fast printing due to it’s composition and thermal conductivity. I again performed the same extrusion tests at 280°C which is a temperature you might use for fast printing and weight the samples afterwards. What we can see right away is that the materials melt quite differently. Whereas some are already quite viscous even at the highest extrusion speed, some show a significant amount of die-swell which is an indication that there are still significant internal stresses in the molten material probably due to high viscosity. The high flow hotend still performed quiet well and we only have around 10% of underextrusion at 35mm³/s for most materials. ABS+ seems to perform a little better than the standard ABS also judging from the extrusion shape. PETG and PLA are also very similar. The only material that perform significantly different is the NanoDiamond PLA because that’s almost extruding perfectly up to 30mm³/s and only then dips a little. I really need to take this material more into consideration for further speed printing tests. Which material to use in the end though doesn’t only depend on the melting rate because if the material doesn’t solidify fast enough it doesn’t really help us with our goal and the cooling topic is something we’ll take a look at in a following video of this series. I wanted to test even more, for example, how the pretension on the hobbed gears impact extrusion rate but, unfortunately, the printhead of the VORON didn't like the abuse, and at some point, the collet constantly slipped out when the backpressure from the nozzle was too high. Since replacing this part requires quite a bit of re-wiring, I'll continue testing once I re-built the machine!

In summary, a high flow hotend and probably also raising the printing temperatures will be necessary if you try to print really fast. A silicone sock usually improves temperature stability and might result in even better hotend performance and depending on your application also think of your material choice, because some materials are better suited for this application than others.

And why did I not yet succeede in printing a 10 minutes benchy, like Albert? I still have some problems with my motion system and will need to tune the input shaper among other things because I currently still face quite a lot of layer shifts on my print.  But that’s something for another video. If you have any questions, comments, or suggestions on this video, please drop them down below!