Single large gear with dual bearings to wrap the filament arround the big gear and maximize contact. That would be my design. But i must admit i am super happy with my hextdudort Plus with larger gear. It offers TONS of grip and power to achieve crazy flow and it doesnt have that much artifacts. Very consistent ❤
@Vez3D Prusa could sell their Nextruder as a standalone, it would be interesting to play with it. If you ensure higher tolerances and it is a full metal extruder then I believe it. :) But I am kinda done with testing extruders, there are so many possible variables that it hurts my head to think about how to eliminate them and present the most apples-to-apples comparison… ;D
Was just rewatching this video and I noticed something. Pause at 4:46 and there definitely is a difference for both the Orbiter and Galileo after the Z-wobble fix. Put the video in 4K and go full screen to see it clearest. The easiest difference to spot is the light reflection at the left side of each test print. Before the fix, the line is uneven and much choppier. After the fix it is a nearly perfectly straight reflection. Looking at the overall surface of the print there is a also a slight, but noticeable reduction in any VFA patterns, as well as a more even appearance overall. Pause at 5:09 and the same improvements are even more obvious. The light reflections are much cleaner, especially on the Galileo print, and a visible shifting pattern every several lines is almost completely gone. Overall, the fix was definitely helpful and I'd say there may even be 5% wobble left to completely dial out with a more advanced Z-wobble fix, as you can still see a slight repeating pattern every 26 lines or so, but I think the simple fix was good enough.
@@PrintingPerspective it's only fair, your videos helped me make some decisions when building my printer so thank you for that. Also like someone else in the comments said, it would be interesting to see the results of a belt extruder compared to these gear extruders.
Glad they did. My only fear is that belted extruders will wear out the belt very quickly and that fiber filaments might not be compatible as they most likely destroy the belts even faster. :)
Couple of things to consider for the variations you’re seeing between the Bambu A series and the custom printer would be any resistance in the Bowden tube setup from the extruder back to the spool. I’ve noticed personally a much more consistent extrusion performance using the ERCF plus a belay sensor and synchronised extrusion between the ERCF and the extruder to ensure the filament is fed properly into the extruder with minimal drag. Also another thing to note is that the Bambu printers internally apply some minor smoothing to sharp angles so just be conscious of it when comparing like for like models between them and a klipper printer (that doesn’t apply any). Finally do consider any potential speed variations due to minimum layer time - I’ve found significant improvements by excluding the external perimeters from the layer time slow downs (and have implemented a feature to do so in orca 2.1). Lastly I believe the Bambu firmware applies variable Pa compensation depending on the print speed - I believe I had seen something to that effect in the X1 plus code but can’t 100% confirm it. I’m currently working on a variable PA feature to emulate something similar in orca, here: github.com/SoftFever/OrcaSlicer/pull/5609
I just installed a HGX lite with the large extruder gears and it's a world of difference. I did notice you can have serious problems with binding, you need to be very careful when tightening the assembly as a very small shift will make it bind up. I also used Super Lube synthetic 30 weight with PTFE. What sold me on the LGX lite was the free floating extruder gears that ride on the needle bearings that allow them to naturally settle centered. I can honestly say my printer now looks as good and maybe better than the A1 with helical gears. And i'm still jelly of those helical gears. Secret sauce, i cleaned the bearings out and ran it dry for a bit with no load to break in the bearings then put on the Super Lube. WD-40 is great to wash parts, then compressed air to dry it out. Now, another issue is print speed (max flow).. before the single gear would allow slippage and now there's basically no slippage BUT when it reaches the motors limits you see strange artifacts of basically the motor. Essentially there's a very fine line between flowing "100%" and skipping, and i like running my motor at low amps for when im printing enclosed high temp filaments. FYI that limit was forcing PLA @230C at 100mm3s though a volcano nozzle for about 5 seconds until the heater couldn't keep up, i thought it was going to grind it right away and beardly let any filament though.. i was shocked it pushed so much. Either way, i print mechanical parts and never really need more than 30mm3s. But it was cool to see the filament sponge back to size coming out of the nozzle lol Best $17 spent on Aliexpress lol Other extruders i've played with are the Sprit and the E3D REVO H2 V2S, but not much better than a MK8 extruder when printing at the optimal melting point. I just wanted something that was smoother, and the HGX lite with large gears did the trick
I had been running the OMG (a dual single-gear)for a while and it printed flawlessly, and tried the HGX and the HGX oblique both with slightly worst extrusion consistency. Im now designing an OMG style extruder with two HGX gears in a dual single-gear configuration.
In my Voron 0, I couldn't feel much difference between the Sherpa-MINI, Orbiter 2, and LGX-Lite, but this Galileo 2 made an overwhelming difference. It has excellent extrusion consistency.
I love how it sounds like your saying that all other extruders aren't real extruders but "proper" is just the name 😂 I would love to try a belt extruder one day though. Seems interesting.
You need to take a look at the extruder motor drivers(and see if they are the same across all the printers) and see what settings they are working with because microstepping, amperage, frequency hysteresis (this setting is not seen on a typical printer config file but can be put in for better control at least in klipper), etc. all affect the torque curve of the motors. Also the spring preload will affect the torque output as well. All these factors affect the torque consistency and therefore output.
I have found that using CNC machined gears that use ball bearings rather than needle bearings greatly reduces the layer extrusion errors. I use fysetc, but there are others out there (rigda, mellow etc). The best result I have had so far is using the Sherpa Micro "fixed idler" extruder. There is no tension screw and the tolerance of the print determines the meshing of the gears.
DIY hevo corexy dual z leadscrews with sherpa mini. I resolved z banding enabling z hop on prusa slicer. For me it was the backlash on the leadscrews. I would try using oldham couplers sometime in the future.
There are definitely some words I pronounce that are harder to understand, but if a viewer follows the context and the things shown on the screen, I feel like most people can intuitively get wtf I am saying, you know. ;D Plus there is a full 1:1 subtitles from a script. People just complain to complain. But thanks for the positive comment, appreciate it. :)
@@PrintingPerspective Actually my comment was not complaining so go and reread it and the discussion below it, I was attempting to give you feedback in case you had sped up the video a bit as alot of youtubers are doing. Don't take my comment wrong, I like your videos and will comment both to stroke the algorithm and to give you feedback to help you improve.
I don't understand dual gear drives. As the filament is loaded, the gears separate to allow it to pass between them. Surely the drive teeth also separate at the same time, so it's probably filament friction doing a fair amount of work to turn the second drive wheel. The drive teeth will not mesh properly when filament is there.
They are designed in a way when filament is loaded both gears are in the optimal position and both gears are driving the filament. But because we don't have perfect tolerances one gear might be pushing more or the other less and we get into a stage where the filament is not consistently driven = introducing inconsistent extrusion. That is my theory about it.
@@PrintingPerspective Thanks for the reply. I agree with everything except the gear mesh. Optimal can't be with filament present or it'd not go closer without filament.
@@saddle1940 Exactly. If the gear mesh were right, it'd stop the drive teeth from getting any closer and you'd have effectively a fixed-distance gearing system rather than fixed-tension.
@@daliasprints9798 ' If the gear mesh were right' You realize this is easy to fix, right? These are all old ideas, and hashed and rehashed years ago.. You'll find my and other comments on old gear and dual gear extruder videos. You MAKE THE MESH NOT CHANGE. Press the bottom side shafts together with a spring, and keep the gear side at full engagement. Then press the top side of the shafts together with a spring as well. Let the top side float further away so the groove can separate to adjust for filament thickness etc and the grip you want on the filament. The gear side is held fully engaged, it just angles very slightly instead of moving apart. THE PROBLEM is merely holding the two shafts too fully parallel. You can even 'fix' the fixed parallel shaft dual gear extruders merely by loosening up the alignment a bit and let the whole moving side float more to an angle. MORE looseness lets that moving side adjust to an angle, instead of rigidly fixed making the gear mesh stay exactly parallel and open up. More imprecision makes it work better, it's why literally the problem was more noticed on better quality dual gear extruders, while cheap knock offs had enough play that they floated closer together and had less problems. This is essentially like using an independent rear suspension so you can adjust for unevenness in the road, instead of a solid axle that breaks traction on one side when the other side lifts it.
Would be nice to see the dual single drive extruders like the OMG I wonder if the quality would improve or would it be the same as a dual drive extruder?
I have two KP3S 3.0 printers with ballscrew ;D This one reached his "final form" where I just don't want to touch it anymore, but I have another one for doing that if I find time to test. But I want more to salvage the SK1 and make it my main test printer with Hermit Crab V2 print head tool-changer plates to make the designing/testing process way quicker.
@@PrintingPerspective I would just say from a marking and business owner, if you asked prusa for one to do tests and see if it lives up to their hype, then maybe a poll on your social network they could realize they have a market for them to sell it. Or better some Chinese manufacturer will see it and see a demand and sell it. Just a thought
There is an unspoken hero of single drive extruders - the Titan, one of the first popular patterns - A1 mini looks alike. It works like a charm for me with TPU. How would it score in this comparison?
how about the TPU? I went from a single gear to dual and I was extremally disappointed. now I got a galileo2 and still not happy I even think about the "proper extruder", just to get the strongest thing so I can't no longer blame that part of the process
I went from CW2 to Galileo 2 and was quite happy, but I was thinking about going Xol and Orbiter 2. Now I'm wondering if that's a mistake, and I should go Xol/G2SA instead. I'll probably build it as I planned first, and test G2SA later, since I won't have a use for the old CW2-style Galileo 2 extruder after that anyhow.
I don’t remember if you did or if somebody else did a test on the concentricity of the extruder gears but that made a big difference in the print quality.
I switched from clockwork2 to galileo and at first got slightly worse results. But twas most because I had rebuild and tweaked and tuned my cw2. After rebuilding my g2e and throughly wearing in the planet gears for mesh consistency (DO NOT skip that step in the manual, grab the drill and wear the gears in) my results were the best I'd ever seen
That is probably the reason why it didn't perform to its full potential because I skipped that step. I didn't even read about it because of the title LOL "YOU KNOW WHAT REALLY GRINDS MY GEARS?". Thanks for mentioning this! I will have to make a pinned comment to address this. :)
I am following in your footsteps and also have slightly worse results and im running out of ideas of how to tune it out. I didn't see anything in the manual about wearing in the gears though, I need to check that out Edit: I read the latest G2E manual again, and still don't see a break in process. Could you share the details?
@@PrintingPerspectivethis step is less relevant for the injection molded gears that come with the kits since Christmas or so. As they don’t have the same tolerance issues as the mjf printed ones. What is super important though is to make sure the hobbed gear and the filament path are perfectly aligned. If they are not you’ll get filament walking on the gear and rubbing on the filament path. This is pretty obvious it’s happening if your rotation distance is different from the recommended one on GitHub.
@@CorySimpsonthe break in process was for the original batch of Galileo 2 kits made from SLS parts. The new injection moulded gears and housing don’t require running in.
I am seriously considering trying to design a single-gear/dual bearing extruder that can fit my k1 Max. The stock updated extruder is really good, no issues there, but a single gear one would eliminate one potential source of artifacts, and it would be one less thing to worry about or poke at.
@@h.y-chen Yup, the A1 has an eddy sensor in the head that does "active flow rate compensation". I don't think that's pressure advance. I think that compensates for filament that varies in size over the roll. That said, the picture they show on their sales page *looks* like it's an alternative for pressure advance, so maybe I'm wrong about exactly how it helps.
@@h.y-chen Wow, okay. I just read up more on it. It doesn't work *during* the print, it only works before the print. It just optimizes the K-value for the filament at the start of the print. So it's just an automated k-value setting, and shouldn't affect anything during PrintingPerspective's tests.
I love powder-bed printing, I've led the purchase and install, and initial operations of one for an employer (using the same PA12GB mentioned in the vid), but I would like to emphasize that PA12GB is about 1/13th as stiff as 6061-T6 by mass. You can get parts as stiff by designing box sections, etc, but just saying parts printed in it are as stiff as aluminum inaccurate.
Would you be able to test out both cnc aluminum vzhextrudort versions out?. There are also slanted and non slanted teeth versions. Comparing the VZHexturdort and VZHexturdort Plus and effects of the slanted and non slanted gears interrests me, could also try bearing as idler gear but that is alot of variables. The Plus version should have lower pushing force but if this is still enough for high flow hotends (enough meaning if you push more you get bad layer adhesion anyway) i can see the advantage of the bigger gear with your earlier video about big filament gears. Love the videos btw. Keep it up!.
I wonder about nozzle influence. Is it not perhaps better to set extrusion width ever so slightly wider than the nozzle? Like say when you have a 0.4mm nozzle to always use 0.42mm outer wall?
Worm gear extruders are interesting too. Very simple and high reduction makes them ideal for single large drive gear extruder designs. Only issue is when you have too much reduction and start to get beyond the optimal point on the stepper torque curve, and their need to be super rigid.
Also a shit ton of friction on the worm gear. In cnc kitchens extruder test the nf cannon (lgx with worm gear, over 1000 esteps) didn't even cross 5kg of pushing force, meanwhile a orbiter 2 (around 700 esteps) was doing nearly 7kg and the bmg extruder was traction limited at a bit over 5kg
I have a couple different machines with a StealthBurner toolhead. The one that uses the Galileo 2 produces far better prints that the one with Clockwork 2 or one with the LGX Lite.
Do you mean the gray part? It was CNC'ed from metal, but you can also 3D print it and secure it with screws - www.printables.com/model/457053-bambu-labs-hotend-adapter-to-v6-size
I'm in the design stage of my own DIY printer and at first this video made me waffle over my choice of the HGX Lite for the extruder, but that printer is meant to exclusively use a 0.2mm nozzle to print minis so I breathed a sigh of relief when you pointed out the extrusion force requirement
For small nozzles you want extruder to have the best filament grip and high extrusion force. Especially if we are talking about high speed 3D printing. :)
If you have the money, you can go with vz-hextrudort as it was designed to compensate for gears eccentricity inconsistencies. It should be the best dual gear extruder and will out live your printer. Otherwise I would go for quality single drive gear extruder. This is just my opinion, so don't trust me :)
VZ hex has some issues though. The angled gears mean you have forces pushing the gears back and forth on the axle they run on. Honestly id go with printed sherpa or orbiter
@@skilledcamman4752Totally agree with you, but for me price diff to orbiter was like +10eur /usd. When my printer dies or when I need something bigger, extruder will be the only part that stays.
@skilledcamman4752 depends if you have inconsistent extrusion problems with the current setup. For slow printing, even the stock KP3S PRO extruder should be fine.
Could you do this again but with TPU? I was ordering a G2E last night but canceled it in the end because I realized it's only a single gear extruder and I've always seen dual as better for TPU.
Well with TPU layer consistency is kind of out of the scope, so Orbiter V2 might be the better choice because it has very little space between extruder gears and the filament path - www.orbiterprojects.com/wp-content/uploads/2021/11/filamentguideinsertion-2.png I think this is the main aspect of what you want from an extruder to be good for TPU, as it is not easy for the flimsy filament to escape elsewhere.
@@PrintingPerspective That's kinda what I had guessed so why I decided to keep the orbiter 2 I have. I mean it works great with tpu so far. Almost no stringing, no filament getting wrapped around the gears like I had with a BMG, and i really love the sensor addon. I was HOPING to find a better extruder which the G2E may be for most people but with me printing TPU fairly often sticking with the Orbiter 2 may be a better idea for now.
@@YourBuddyDinec I also own the orbiter v2 and when printing tpe 83a from sunlu I always get the filament wrapped around the left gear when trying to go quicker. At what extrusion rates are you printing your tpu and what's the shore hardnes of it? maybe I could adjust my prints then :`)
Seems like the general consensus here is that dual drive is best suited for small nozzles, but single drive probably gives the best quality for standard and larger ones (depending on the setup)
The other big difference is flexible filaments. For things like TPUs, the level of tension you need to put on a single drive to get enough force into the hotend can cause jams or can result in the filament sticking into the drive gear itself and tangling itself inside the extruder. The reduced bite and more equalized force imparted on the filament in a dual drive means this is less likely to happen for these kinds of filaments. And the softer the filament is, the worse this gets. This part is anecdotal, but I've also seen far fewer full extruder jams in dual gear systems, but when they do happen, they're much harder to fix. Used to see the semi-circle gouge in a jammed filament all the time on single gear drives, but the stepper motor seems to stall before doing that on dual systems.
Interesting video and a great comparison. 👍👍👍👍 This is, however, a comparison of designs from different manufacturers and so the use of better design solutions and the accuracy of the design can have a significant impact on the outcome (a premium manufacturer may have a quality advantage and not necessarily a design advantage). It would be interesting to compare the same drive type in a configuration as a dual-gear and in a configuration as a single-gear then you could compare which version has the advantage. I saw on printables a new version of the HGXlite extruder in a lever configuration modelled on the Prusa Nextruder. It would be interesting to compare this design with the dual-gear version of the HGXlite. Ideally they should be in similar housings.🙂😁
I don't see under which circumstances one would need the torque provided by a second drive gear. If an extruder needs this much torque, it is because there is an issue somewhere that needs to be resolved. Under normal circumstances, one drive gear is plenty.
@@PrintingPerspective Yet most (all?) cheap titan clones are horribly bad in concentricity. Good extruder requires precision manufacturing, Bambu Lab probably having that :)
Please do try to adapt the bambulab extruder to the kp3s or other cheaper printers. Id love to see how good they are on cheaper printers, maybe they will be adapted like the tz hotend's.
Z wobble isn’t wobble. It’s variation in z travel due to slight binding of the screw nut as it rotates thru its revolution. The variation in z causes under or over extrusion and the bands you see in prints
To compare across different machines makes the test rather useless (P1S vs. A1). The only falid test is between Gal2 vs. Orb2 as it is on the same machine with hopefuly the same settings. And there is no visible difference.
Yeah whatever his results were in this video you can chuck it out the door. Also, 10 years ago we had perfect prints with hobbed bolts on 3D printed Wade's extruders.
@@megatesla Out of curiosity went to google "Wade's extruder" - I was scarred for life x) it's incredible you managed to print anything with those kind of tools back then, we are so spoiled now days!
@@Inveterat Not sure, never had an issue with his other vids so wondered if he sped this up. apparently alot of utubers are doing that for the adhd kiddies crowd.
@@YourBuddyDinec Might want to stop being a DB. There are multiple comments about this, so apparently some people are having trouble UNDERSTANDING him. Didn't say anything about not being able to hear him.
@@kf4hqf2 If you have bad hearing then understanding people speak often accompanies that 🤷♂ I'm just worried people might not realize they are developing a hearing problem and giving a recommendation.
I can understand what he is saying, but the constant tone changes in every sentence is very distracting. High, low, high, low. Nothing like ending sentences high or low that most native English speakers do for different types of sentences.
Single large gear with dual bearings to wrap the filament arround the big gear and maximize contact. That would be my design. But i must admit i am super happy with my hextdudort Plus with larger gear. It offers TONS of grip and power to achieve crazy flow and it doesnt have that much artifacts. Very consistent ❤
How would that work with very brittle CF/GF filaments?
That's what I'm doing. 😎
What is stopping you from making a new extruder with this concept? :P
@Vez3D Prusa could sell their Nextruder as a standalone, it would be interesting to play with it.
If you ensure higher tolerances and it is a full metal extruder then I believe it. :) But I am kinda done with testing extruders, there are so many possible variables that it hurts my head to think about how to eliminate them and present the most apples-to-apples comparison… ;D
Just a matter of time until some chinese will copy the prusa nextruder concept.
Was just rewatching this video and I noticed something. Pause at 4:46 and there definitely is a difference for both the Orbiter and Galileo after the Z-wobble fix. Put the video in 4K and go full screen to see it clearest. The easiest difference to spot is the light reflection at the left side of each test print. Before the fix, the line is uneven and much choppier. After the fix it is a nearly perfectly straight reflection. Looking at the overall surface of the print there is a also a slight, but noticeable reduction in any VFA patterns, as well as a more even appearance overall. Pause at 5:09 and the same improvements are even more obvious. The light reflections are much cleaner, especially on the Galileo print, and a visible shifting pattern every several lines is almost completely gone. Overall, the fix was definitely helpful and I'd say there may even be 5% wobble left to completely dial out with a more advanced Z-wobble fix, as you can still see a slight repeating pattern every 26 lines or so, but I think the simple fix was good enough.
Not sure if you ever saw this comment, but I keep coming back to this video, hahaha.
Thanks for all the testing you do!
Oh, wow, appreciate your support! Have a great day! :)
@@PrintingPerspective it's only fair, your videos helped me make some decisions when building my printer so thank you for that. Also like someone else in the comments said, it would be interesting to see the results of a belt extruder compared to these gear extruders.
Glad they did. My only fear is that belted extruders will wear out the belt very quickly and that fiber filaments might not be compatible as they most likely destroy the belts even faster. :)
Couple of things to consider for the variations you’re seeing between the Bambu A series and the custom printer would be any resistance in the Bowden tube setup from the extruder back to the spool.
I’ve noticed personally a much more consistent extrusion performance using the ERCF plus a belay sensor and synchronised extrusion between the ERCF and the extruder to ensure the filament is fed properly into the extruder with minimal drag. Also another thing to note is that the Bambu printers internally apply some minor smoothing to sharp angles so just be conscious of it when comparing like for like models between them and a klipper printer (that doesn’t apply any).
Finally do consider any potential speed variations due to minimum layer time - I’ve found significant improvements by excluding the external perimeters from the layer time slow downs (and have implemented a feature to do so in orca 2.1).
Lastly I believe the Bambu firmware applies variable Pa compensation depending on the print speed - I believe I had seen something to that effect in the X1 plus code but can’t 100% confirm it. I’m currently working on a variable PA feature to emulate something similar in orca, here: github.com/SoftFever/OrcaSlicer/pull/5609
I just installed a HGX lite with the large extruder gears and it's a world of difference. I did notice you can have serious problems with binding, you need to be very careful when tightening the assembly as a very small shift will make it bind up. I also used Super Lube synthetic 30 weight with PTFE. What sold me on the LGX lite was the free floating extruder gears that ride on the needle bearings that allow them to naturally settle centered. I can honestly say my printer now looks as good and maybe better than the A1 with helical gears. And i'm still jelly of those helical gears.
Secret sauce, i cleaned the bearings out and ran it dry for a bit with no load to break in the bearings then put on the Super Lube. WD-40 is great to wash parts, then compressed air to dry it out.
Now, another issue is print speed (max flow).. before the single gear would allow slippage and now there's basically no slippage BUT when it reaches the motors limits you see strange artifacts of basically the motor. Essentially there's a very fine line between flowing "100%" and skipping, and i like running my motor at low amps for when im printing enclosed high temp filaments. FYI that limit was forcing PLA @230C at 100mm3s though a volcano nozzle for about 5 seconds until the heater couldn't keep up, i thought it was going to grind it right away and beardly let any filament though.. i was shocked it pushed so much. Either way, i print mechanical parts and never really need more than 30mm3s. But it was cool to see the filament sponge back to size coming out of the nozzle lol
Best $17 spent on Aliexpress lol
Other extruders i've played with are the Sprit and the E3D REVO H2 V2S, but not much better than a MK8 extruder when printing at the optimal melting point. I just wanted something that was smoother, and the HGX lite with large gears did the trick
I had been running the OMG (a dual single-gear)for a while and it printed flawlessly, and tried the HGX and the HGX oblique both with slightly worst extrusion consistency. Im now designing an OMG style extruder with two HGX gears in a dual single-gear configuration.
Great writeup, explained a lot of things. What would you choose between LGX lite and HGX lite (oblique or standard straight gears)?
In my Voron 0, I couldn't feel much difference between the Sherpa-MINI, Orbiter 2, and LGX-Lite, but this Galileo 2 made an overwhelming difference. It has excellent extrusion consistency.
i´d love to see the same tests on a "proper extruder" the belt extruder designed by Jón from proper printing
Yup, that dude rocks.
the proper extruder also doesn't bite into the filament which will help to rule out variables.
and the papilio lite extruder by kevinakasam as well.
I love how it sounds like your saying that all other extruders aren't real extruders but "proper" is just the name 😂 I would love to try a belt extruder one day though. Seems interesting.
@@YourBuddyDinec lol, I got the same feeling when I was writing, so I added the quote marks to sound less like that... Seems like it didn't work.. Lol
You need to take a look at the extruder motor drivers(and see if they are the same across all the printers) and see what settings they are working with because microstepping, amperage, frequency hysteresis (this setting is not seen on a typical printer config file but can be put in for better control at least in klipper), etc. all affect the torque curve of the motors. Also the spring preload will affect the torque output as well. All these factors affect the torque consistency and therefore output.
Yeah his tests aren't very good
I have found that using CNC machined gears that use ball bearings rather than needle bearings greatly reduces the layer extrusion errors. I use fysetc, but there are others out there (rigda, mellow etc). The best result I have had so far is using the Sherpa Micro "fixed idler" extruder. There is no tension screw and the tolerance of the print determines the meshing of the gears.
DIY hevo corexy dual z leadscrews with sherpa mini. I resolved z banding enabling z hop on prusa slicer. For me it was the backlash on the leadscrews. I would try using oldham couplers sometime in the future.
Thank you so much for the testing and information, learnd alot from the subject
Not sure what those other comments are complaining about, you're perfectly understandable.
There are definitely some words I pronounce that are harder to understand, but if a viewer follows the context and the things shown on the screen, I feel like most people can intuitively get wtf I am saying, you know. ;D Plus there is a full 1:1 subtitles from a script. People just complain to complain. But thanks for the positive comment, appreciate it. :)
@PrintingPerspective Subtitles are always appreciated, thank you!
@@3dexperiments Great all around since I'm sure many have heaving problems in one way or another.
@@PrintingPerspective Actually my comment was not complaining so go and reread it and the discussion below it, I was attempting to give you feedback in case you had sped up the video a bit as alot of youtubers are doing. Don't take my comment wrong, I like your videos and will comment both to stroke the algorithm and to give you feedback to help you improve.
I don't understand dual gear drives. As the filament is loaded, the gears separate to allow it to pass between them. Surely the drive teeth also separate at the same time, so it's probably filament friction doing a fair amount of work to turn the second drive wheel. The drive teeth will not mesh properly when filament is there.
They are designed in a way when filament is loaded both gears are in the optimal position and both gears are driving the filament. But because we don't have perfect tolerances one gear might be pushing more or the other less and we get into a stage where the filament is not consistently driven = introducing inconsistent extrusion. That is my theory about it.
@@PrintingPerspective Thanks for the reply. I agree with everything except the gear mesh. Optimal can't be with filament present or it'd not go closer without filament.
@@saddle1940 Exactly. If the gear mesh were right, it'd stop the drive teeth from getting any closer and you'd have effectively a fixed-distance gearing system rather than fixed-tension.
@@daliasprints9798 ' If the gear mesh were right'
You realize this is easy to fix, right? These are all old ideas, and hashed and rehashed years ago.. You'll find my and other comments on old gear and dual gear extruder videos.
You MAKE THE MESH NOT CHANGE. Press the bottom side shafts together with a spring, and keep the gear side at full engagement. Then press the top side of the shafts together with a spring as well. Let the top side float further away so the groove can separate to adjust for filament thickness etc and the grip you want on the filament. The gear side is held fully engaged, it just angles very slightly instead of moving apart.
THE PROBLEM is merely holding the two shafts too fully parallel. You can even 'fix' the fixed parallel shaft dual gear extruders merely by loosening up the alignment a bit and let the whole moving side float more to an angle. MORE looseness lets that moving side adjust to an angle, instead of rigidly fixed making the gear mesh stay exactly parallel and open up.
More imprecision makes it work better, it's why literally the problem was more noticed on better quality dual gear extruders, while cheap knock offs had enough play that they floated closer together and had less problems.
This is essentially like using an independent rear suspension so you can adjust for unevenness in the road, instead of a solid axle that breaks traction on one side when the other side lifts it.
Would be nice to see the dual single drive extruders like the OMG I wonder if the quality would improve or would it be the same as a dual drive extruder?
Love your testing! I would love a video where you compare the A1 extruder to a run-in Galileo 2 on that same ball screw printer you have.
I have two KP3S 3.0 printers with ballscrew ;D This one reached his "final form" where I just don't want to touch it anymore, but I have another one for doing that if I find time to test. But I want more to salvage the SK1 and make it my main test printer with Hermit Crab V2 print head tool-changer plates to make the designing/testing process way quicker.
Yes, please make that a1 inspired extruder, that would be awesome 😊
It would probably be easier just to use the parts from the A1 series extruder, but yeah, I would love to make my own, if I will find time to do it.
@@PrintingPerspective great, good luck!
Would love this!
Bambu lab parts can be easily purchased in most part of the world, so this would be great.
You look at the new prusa extruder it uses one big gear with two bearing pressure points
I just wish that they would release the Planetary Gear Carrier, ie the drive gear to the public. Id love to play with one on a test setup
@dalton11231 Yeah, I know, but they doesn't sell it as a standalone. :/
@@PrintingPerspective I would just say from a marking and business owner, if you asked prusa for one to do tests and see if it lives up to their hype, then maybe a poll on your social network they could realize they have a market for them to sell it. Or better some Chinese manufacturer will see it and see a demand and sell it. Just a thought
I wonder how the papilio extruder would fair
There is an unspoken hero of single drive extruders - the Titan, one of the first popular patterns - A1 mini looks alike. It works like a charm for me with TPU. How would it score in this comparison?
how about the TPU? I went from a single gear to dual and I was extremally disappointed. now I got a galileo2 and still not happy
I even think about the "proper extruder", just to get the strongest thing so I can't no longer blame that part of the process
I went from CW2 to Galileo 2 and was quite happy, but I was thinking about going Xol and Orbiter 2. Now I'm wondering if that's a mistake, and I should go Xol/G2SA instead. I'll probably build it as I planned first, and test G2SA later, since I won't have a use for the old CW2-style Galileo 2 extruder after that anyhow.
Very interesting results!
Thank you! Cheers!
I would make this comparison (orbiter & galileo) vs Bondtech LGX & HGX lite/oblique
I don’t remember if you did or if somebody else did a test on the concentricity of the extruder gears but that made a big difference in the print quality.
Yes, it does, that is why dual-drive gear extruders can suffer a lot if high manufacturing tolerances are not ensured. :)
I switched from clockwork2 to galileo and at first got slightly worse results. But twas most because I had rebuild and tweaked and tuned my cw2.
After rebuilding my g2e and throughly wearing in the planet gears for mesh consistency (DO NOT skip that step in the manual, grab the drill and wear the gears in) my results were the best I'd ever seen
That is probably the reason why it didn't perform to its full potential because I skipped that step. I didn't even read about it because of the title LOL "YOU KNOW WHAT REALLY GRINDS MY GEARS?". Thanks for mentioning this! I will have to make a pinned comment to address this. :)
I am following in your footsteps and also have slightly worse results and im running out of ideas of how to tune it out. I didn't see anything in the manual about wearing in the gears though, I need to check that out
Edit: I read the latest G2E manual again, and still don't see a break in process. Could you share the details?
@@PrintingPerspectivethis step is less relevant for the injection molded gears that come with the kits since Christmas or so. As they don’t have the same tolerance issues as the mjf printed ones. What is super important though is to make sure the hobbed gear and the filament path are perfectly aligned. If they are not you’ll get filament walking on the gear and rubbing on the filament path. This is pretty obvious it’s happening if your rotation distance is different from the recommended one on GitHub.
@@CorySimpsonthe break in process was for the original batch of Galileo 2 kits made from SLS parts. The new injection moulded gears and housing don’t require running in.
I am using an orbiter v2 on my 5 plus corexy build with rapido hf hotend. I'm putting a cyclops(orbiter v2 clone) drop in for my k1 max.
I am seriously considering trying to design a single-gear/dual bearing extruder that can fit my k1 Max. The stock updated extruder is really good, no issues there, but a single gear one would eliminate one potential source of artifacts, and it would be one less thing to worry about or poke at.
Hey, can you test the Papilio Lite extruder? Maybe even with some parts printed in sls or slm by pcbway?
did you disable A1 active flow rate compensation? does that make difference?
That feature is for Pressure Advance which eliminates bulging on the corners, I have manual values and I don't run it before the prints.
@@PrintingPerspective doesn't A1 detect filament width and auto adjust flow rate when printing? Not pressure advance
@@h.y-chen Yup, the A1 has an eddy sensor in the head that does "active flow rate compensation". I don't think that's pressure advance. I think that compensates for filament that varies in size over the roll. That said, the picture they show on their sales page *looks* like it's an alternative for pressure advance, so maybe I'm wrong about exactly how it helps.
@@h.y-chen Wow, okay. I just read up more on it. It doesn't work *during* the print, it only works before the print. It just optimizes the K-value for the filament at the start of the print. So it's just an automated k-value setting, and shouldn't affect anything during PrintingPerspective's tests.
@@Raytenecks ok, so that just K-value thing, thanks
I love powder-bed printing, I've led the purchase and install, and initial operations of one for an employer (using the same PA12GB mentioned in the vid), but I would like to emphasize that PA12GB is about 1/13th as stiff as 6061-T6 by mass. You can get parts as stiff by designing box sections, etc, but just saying parts printed in it are as stiff as aluminum inaccurate.
Would you be able to test out both cnc aluminum vzhextrudort versions out?. There are also slanted and non slanted teeth versions. Comparing the VZHexturdort and VZHexturdort Plus and effects of the slanted and non slanted gears interrests me, could also try bearing as idler gear but that is alot of variables. The Plus version should have lower pushing force but if this is still enough for high flow hotends (enough meaning if you push more you get bad layer adhesion anyway) i can see the advantage of the bigger gear with your earlier video about big filament gears.
Love the videos btw. Keep it up!.
I wonder about nozzle influence. Is it not perhaps better to set extrusion width ever so slightly wider than the nozzle? Like say when you have a 0.4mm nozzle to always use 0.42mm outer wall?
I use 0.42mm for regular prints.
Worm gear extruders are interesting too. Very simple and high reduction makes them ideal for single large drive gear extruder designs. Only issue is when you have too much reduction and start to get beyond the optimal point on the stepper torque curve, and their need to be super rigid.
Also a shit ton of friction on the worm gear. In cnc kitchens extruder test the nf cannon (lgx with worm gear, over 1000 esteps) didn't even cross 5kg of pushing force, meanwhile a orbiter 2 (around 700 esteps) was doing nearly 7kg and the bmg extruder was traction limited at a bit over 5kg
@@kilianlindlbauer8277Just need a big motor driving the input. 😎 My cannon gears have LDO 2804 driving them.
It's crazy how all your videos are so good !
Thanks! :)
I have a couple different machines with a StealthBurner toolhead. The one that uses the Galileo 2 produces far better prints that the one with Clockwork 2 or one with the LGX Lite.
That sounds about right. :)
1:19 How did you mount the hotend to the groove mount bowden adapter?
he showed it in this video
th-cam.com/video/dQ3kQ2w9D2s/w-d-xo.htmlsi=PhJ9p7mACAe3oGkY&t=113
Do you mean the gray part? It was CNC'ed from metal, but you can also 3D print it and secure it with screws - www.printables.com/model/457053-bambu-labs-hotend-adapter-to-v6-size
I'm in the design stage of my own DIY printer and at first this video made me waffle over my choice of the HGX Lite for the extruder, but that printer is meant to exclusively use a 0.2mm nozzle to print minis so I breathed a sigh of relief when you pointed out the extrusion force requirement
For small nozzles you want extruder to have the best filament grip and high extrusion force. Especially if we are talking about high speed 3D printing. :)
I like relatively slow printing on kp3s pro, should I go single-drive or dual-drive?
If you have the money, you can go with vz-hextrudort as it was designed to compensate for gears eccentricity inconsistencies. It should be the best dual gear extruder and will out live your printer. Otherwise I would go for quality single drive gear extruder. This is just my opinion, so don't trust me :)
VZ hex has some issues though. The angled gears mean you have forces pushing the gears back and forth on the axle they run on.
Honestly id go with printed sherpa or orbiter
@@vanher282 imo opinion such an expensive extruder isnt worthit using with a cheap printer.
@@skilledcamman4752Totally agree with you, but for me price diff to orbiter was like +10eur /usd. When my printer dies or when I need something bigger, extruder will be the only part that stays.
@skilledcamman4752 depends if you have inconsistent extrusion problems with the current setup. For slow printing, even the stock KP3S PRO extruder should be fine.
per curiosità hai provato con ingranaggi stile bmg monoblocco o con ingranaggi fissati con grano???
can you test a WWBMG vs a WWG2 that would be a very interesting test
Nice comparision, i'd love to see how hextrudort low plus with helical dual gears performs... in theory that is also wobble free
Could you do this again but with TPU? I was ordering a G2E last night but canceled it in the end because I realized it's only a single gear extruder and I've always seen dual as better for TPU.
Well with TPU layer consistency is kind of out of the scope, so Orbiter V2 might be the better choice because it has very little space between extruder gears and the filament path - www.orbiterprojects.com/wp-content/uploads/2021/11/filamentguideinsertion-2.png
I think this is the main aspect of what you want from an extruder to be good for TPU, as it is not easy for the flimsy filament to escape elsewhere.
@@PrintingPerspective That's kinda what I had guessed so why I decided to keep the orbiter 2 I have. I mean it works great with tpu so far. Almost no stringing, no filament getting wrapped around the gears like I had with a BMG, and i really love the sensor addon. I was HOPING to find a better extruder which the G2E may be for most people but with me printing TPU fairly often sticking with the Orbiter 2 may be a better idea for now.
For the record I've printed TPU with a bearing on a Bear extruder without issue.
@@YourBuddyDinec I also own the orbiter v2 and when printing tpe 83a from sunlu I always get the filament wrapped around the left gear when trying to go quicker. At what extrusion rates are you printing your tpu and what's the shore hardnes of it? maybe I could adjust my prints then :`)
@@crypton1782 Not super fast. 60-80mm/s. Make sure the extruder isn't tightened TOO much as too much grip can do that.
Seems like the general consensus here is that dual drive is best suited for small nozzles, but single drive probably gives the best quality for standard and larger ones (depending on the setup)
The other big difference is flexible filaments. For things like TPUs, the level of tension you need to put on a single drive to get enough force into the hotend can cause jams or can result in the filament sticking into the drive gear itself and tangling itself inside the extruder. The reduced bite and more equalized force imparted on the filament in a dual drive means this is less likely to happen for these kinds of filaments. And the softer the filament is, the worse this gets.
This part is anecdotal, but I've also seen far fewer full extruder jams in dual gear systems, but when they do happen, they're much harder to fix. Used to see the semi-circle gouge in a jammed filament all the time on single gear drives, but the stepper motor seems to stall before doing that on dual systems.
Interesting video and a great comparison. 👍👍👍👍 This is, however, a comparison of designs from different manufacturers and so the use of better design solutions and the accuracy of the design can have a significant impact on the outcome (a premium manufacturer may have a quality advantage and not necessarily a design advantage).
It would be interesting to compare the same drive type in a configuration as a dual-gear and in a configuration as a single-gear then you could compare which version has the advantage.
I saw on printables a new version of the HGXlite extruder in a lever configuration modelled on the Prusa Nextruder.
It would be interesting to compare this design with the dual-gear version of the HGXlite.
Ideally they should be in similar housings.🙂😁
I been using gallileo 2 on my 2.4 since it launched, it's awesome
I don't see under which circumstances one would need the torque provided by a second drive gear. If an extruder needs this much torque, it is because there is an issue somewhere that needs to be resolved. Under normal circumstances, one drive gear is plenty.
The A1 extruder looks like a clone of the old E3D titan extruder
That's what I was thinking as well. Straight up the same mechanism for it.
Because probably it can't get simpler than that. ;D
@@PrintingPerspective Yet most (all?) cheap titan clones are horribly bad in concentricity. Good extruder requires precision manufacturing, Bambu Lab probably having that :)
Kinda but with helical (oblique tooth) on the pom (i assume) gear
Try the single peace bmg gear (RIDGA) by mirageC
Wont the PAGF35 creep ?!
Please do try to adapt the bambulab extruder to the kp3s or other cheaper printers. Id love to see how good they are on cheaper printers, maybe they will be adapted like the tz hotend's.
Z wobble isn’t wobble. It’s variation in z travel due to slight binding of the screw nut as it rotates thru its revolution. The variation in z causes under or over extrusion and the bands you see in prints
Good stuff.
To compare across different machines makes the test rather useless (P1S vs. A1). The only falid test is between Gal2 vs. Orb2 as it is on the same machine with hopefuly the same settings. And there is no visible difference.
that's what I thought the whole time.. really couldn't agree more..
Yeah whatever his results were in this video you can chuck it out the door. Also, 10 years ago we had perfect prints with hobbed bolts on 3D printed Wade's extruders.
@@megatesla Out of curiosity went to google "Wade's extruder" - I was scarred for life x) it's incredible you managed to print anything with those kind of tools back then, we are so spoiled now days!
great video
Had no Idea bambulab only used single-drive gear extruder
Only the a1, x1 and p1 use dual gear ones
@skilledcamman4752 A1 and A1 Mini has single drive gear and P/X series have dual-drive extruders.
@@PrintingPerspective ye I knew that I meant only for the a1 seres.
So it appears that the A1 is the printer with better extrusion *and* less VFA that's available now.
He.
Who prints at 25 mm/s? Really push some plastic to make this real world. This test means nothing.
My prints can peak at 30mm/s2
@@brianfox8439 30mm/s² would be acceleration.
And where K3D-Minifeeder? = )
Did you speed up the video? I had to slow it down to .9X to understand you.
As a non-native english speaker I undestood him perfectly. Perhaps it's the trained ear from working together with people from many countries.
@@YourBuddyDinec It is getting worse from all the years of concerts and working on a flightline!
@@Inveterat Not sure, never had an issue with his other vids so wondered if he sped this up. apparently alot of utubers are doing that for the adhd kiddies crowd.
@noanyobiseniss7462 makes sense, one must appease the algorithm.
@@Inveterat And those types don't know how to speed something up without their hand being held.
So hard to understand you sometimes....
Can you slow down please. It’s hard to understand you. Great video
Might want to have your hearing checked.
@@YourBuddyDinec Might want to stop being a DB. There are multiple comments about this, so apparently some people are having trouble UNDERSTANDING him. Didn't say anything about not being able to hear him.
You can always play the video in 0.75X or 0.5X speed. I listen at 2X and have no issues, but hope this helps.
@@kf4hqf2 If you have bad hearing then understanding people speak often accompanies that 🤷♂ I'm just worried people might not realize they are developing a hearing problem and giving a recommendation.
I can understand what he is saying, but the constant tone changes in every sentence is very distracting. High, low, high, low. Nothing like ending sentences high or low that most native English speakers do for different types of sentences.
just speak english whatisthisnonsensephrasesyousayicantunderstand
Just use your ears, hes talking english but you're too incompetent to understand it.
@@NimVim LOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOL almost laughed
@@fadybenmoussa2114 In this case consider checking your brain.
I understand him perfectly at 2X speed.