Would it be possible to add something like a peristaltic pump in between the two heater blocks so that you could retract fully liquid plastic much more easily? I'm not sure what kind of tubing you could use that would withstand the heat of the melted plastic, since PTFE will probably deform fast if it's being squished at over 200C--but if you could figure out a way to pump liquid plastic to and from the first heater block into the second without any backlash (hence the peristaltic pump), that might be the key to unlocking the quality.
Your lack of insulation is likely a huge problem. With heater set to 240-260 but no insulation, the nozzle and end of heat zone is probably more like 170-190. It's impossible to get consistent quality extrusion if the temperature control isn't actually working because of gigantic internal temperature gradients due to lack of insulation. If you don't want to mold silicone socks, just wrap everything in rock wool and kapton tape.
The 9 minute benchy is significantly worse than a 9 minute benchy on an Ender 3 with cooling & extruder upgraded to be able to do that. At that speed you're only averaging 17 mm³/s. It looked better once you turned the cooling up, so I think your cooling is way below the level you need for a fast boat.
So pleased to see someone utilising the benefits of cable robots for 3D printers. Well done! What kind of stepper motors are you using? Any chance of upcycling some old CD drive motors for those of us on a budget? Also it looks like the right blower is catching the wet plastic in places. Please check the blower tips for tell tale residue.
Sorry, my internet is not great so I couldn't get through the whole video. Are you pumping air under your moving bed like an air hockey table? I hear what sounds like air but I can not spot it, maybe it's just because it's sped up. If not then maybe it would help with friction if the bed does ride on air.
16:00 That's speed and 9 min?😑 All videos in the youtube, something like "3 min benchy" not even close to such speed. Now I'm thinking that all they are fakes
Hi, I'm Baran. Let me give you a great idea: First, you should lower all four motors by 5 mm along the Z-axis. This way, the bed will be pulled downwards, but there won't be any movement restrictions. Second, replace your bed material with an oily plastic. You're welcome! ;)
We've gone from top-to-bottom straight path hot end geometry to some stuff with really interesting CHT internal geometry, but we are still pushing straight down, leading to bad blobbing/leakage even with high retraction. I wonder if you could improve the hot end design with a u-bend like in a toilet to help reduce pressure from gravity, or if this would be difficult for an extruder to push, or just disconnect from retraction settings entirely. If you had a sideways S shape, you could put heater cores between the loops, which buys a lot of surface area. Congrats on first benchy!
few things to consider - i think rattling is a big contributor for motion system not being able to cope with current 5min goal. it is a parasitic dynamic load. - i think two zones of heater should have different temperatures. entry zone should just preheat plastic close to melting temperature. while heater is in two peaces, upper part can be tested separately to determain neaded temperature and power to achieve this. reason being that molten plastic has much more drag on inner walls, thus requesting more push power from extruder motor and reducing overall flow of extruder at the same time. - more speed does not equal more flow. as You said, speed from compressed is making lots of noise, but that is not equal to lots cooling. air ducts could be optimized better for high speed flow and / or last half a meter of hose before entering ducts could be much bigger diameter to decrease air speed and improve volumetric fill of ducts. i hope this makes sense. :)
Ive got a 4:09 benchy that I decided not to post (blobby side) and a 4:56-ish one. Without pressure advance, it consistently blobbed along the z seam bad enough to end the prints. Also, cooling: try offsetting the ducts to create a vortex. Instead of interfering with each other, it spreads the cooling effect
With springs, if you apply pre-tension on the springs you can end up in the better end of the spring curve that is not as steep in Force / distance as an untensioned spring. also pressure based spring might help here.
I don't think you mentioned explicitly that you did it, but just in case: PID calibration! Cranking the voltage of the heater elements up that much will surely need it.
Cooling: easy to improve, just split the shop air line into 3 and use 3 outlets+expanders with an slight offset to create some vortex for a more stable airflow. you dont need an duct like with an fan. Comparing it the other setups: you get the pressure, you got the outlet area, Bernoulli is your friend. measuring other setups is harder tho, but why bother. hotend: next prototype should be as simple as possible. i would go for a single split with a longer heat zone. maybe the next splitter is needed for mixing but a longer heat zone should do the same. 4 heaters should be fine, maybe go for 6 and heat it up gradually. extruder: a modern extruder will to the trick. motion system: thats the hard one. its hard to isolate all the things going on. you currently run an "undefined" system with the top part of the air bearing missing and the outcome is still acceptable.
If you want higher motor performance: 150W Yaskawa SGM7A-C2A 40mm Servo Motor with a Nidec Shimpo 42mm 3:1 gear reducer VRB-042C-3-K3-S8ZG8 and a Yaskawa MP2600IEC Dual Axis Motion Controller (Dual axis servo drive with basic plc functionality to send commands to external z axis drive).
Input shaping, linear advance, a silicone sock, automatic part-cooling control (proportional valve if not a fan), and increasing the spring-tension should all help to get tighter printing. That dynamic hot end temperature thing might be great too.
Really awesome progress. I think next you need to get a version of the stabilized air bearing platform finished. That's the one part missing true proof of concept it seems. Try to go for something as final as possible. Then tighten up the motion system with added stepper cooling and permanent positioning and all. Then move on to the extruder and hot end again. I mean the order doesn't really matter, but going system by system as you have been makes sense. Just my thinking.
Perhaps a pellet extruder could work well on this motion system. They can move far more plastic than a filament extruder. Not sure how well an oversized one would do with a small nozzle. Maybe it's worth considering?
i'm intrested to see how you test the cooling effeciency of different setups. one way could be as simple as a normal hotend (heater and thermistor) set to 240 degrees, and then just reading off the duty cycle when cooling air is being blown directly onto it. i'm also qurrious to see if a cheap leafblower can outpreform your shop air compressor. Btw, is there any thing to gain from pre cooling the compressed air with a peltier or in a ice bath before you pass it through the regulator, just to remove a little more energy from the air...
no. just no to the leafblower idea, there is no gain from this but the cooling part could help but the gains from it are not that big compared to the delta of the molten plastic to the shop air
since you have high pressure air available have you considered just injecting a high pressure low volume jet down into a velocity stack to let Mr Venturi do the work? similar to how dyson has its 'bladeless' fans.
That 260C on the heater block with the PTFE plug ... Getting fairly close to the limits of PTFE. For the cooling, you could have a look at the air knife from Streamtek. That is a way to increase air flow and cooling without using too much shop air - but I would safe this for a much later stage. First important things to increase the performance are the heater blocks and the movement system, especially the table. That additional plate @17:20 and 2 air bearings, one from the bottom of the table and one through that plate could create a very stable setup
found it ! The capstan drive, which is a zero backlash gear reducer using rope : ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-MwIBTbumd1Q.html . This will solve your cables and pulleys issues. The rope is specific, cheap, and called DM20
Have you considered trying something like this variable temperature script to lower the temperature when printing the chimney and increase its quality? ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-P6Y8uUPd3yg.htmlsi=V8f_AAl3tZETAEF6
thats a very interesting bed slinger ! I like the cable friction setup. The last time I saw one it was for a robot leg actuator, and the guy said he used iron then stainless wire but it did put too much wear on the pulley and had trouble with tension. He switched to something different, cant remember exactly, i think it was carbon fiber reinforced nylon. If you get trouble , try finding that one. Also stupid question maybe but ... why dont you insulate the hotend ?
This is very entertaining to watch. Love the raw engineering. NeedItMakeIt is actually running a part cooling duct contest right now, see who's got the best design for high speed cooling
you can make a heating block that feeds 2 filaments into 1, and if that works you might use 3 copies to build a tree, 4 filaments into 2 blocks into 1 block to nozzle
Unsure of what your long term goals are for this but defs consider patenting you work, even as an open patent, just at least so its confirmed as your work. Awesome stuff dude.
29:37 remember what I mentioned a few videos ago with the orifice size thing? you've run into that limit. you can only shove so much at a time through a 0.5mm hole. molten plastic isn't liquid, it has a shear modulus.
