Is the Nextruder hot end design choice a mistake?
So forgive the seemingly click bait title but it's the question I've been asking myself and I think the answer is yes. I've had my XL a few weeks now and I've come to a similar conclusion to these two videos. The 0.4 nozzle is the size of choice for the XL, most of the time....
There are many reasons covered in these videos and we can discuss variables like print speed, temps, layer height etc. What I want to focus on is flow rate. My thread title comes from the limitations of the Nextruder hot end to melt filament fast enough. The videos make reference to this as well and by modern standards it seems to me the stock setup is already limited in terms of flow. When the printer ships with 0.6mm nozzles and the flow is the limiting factor in your print speed (the profiles are slower for the 0.6 v 0.4) then what good are the core xy kinematics? Even if you can move the tool around quickly you're quite severely limited by that flow rate.
We know E3D produce the nextruder nozzles and they've very similar to the Revo ones. So why didn't Prusa just adopt Revo?? There was plenty of time to do so and they could have taken advantage of the better heaters etc as well as Revo developments (Revo high flow). One has to assume going proprietary was a business decision. I know it's meant to be open source but I've not seen specs published etc although I'm sure it's easy enough to reverse engineer.
I think for the XL to take advantage of IS and leverage the benefits of Core XY we will need to see some sort of CHT style nozzles, certainly for 0.6 and up. I know there is the adaptor but what I'm talking about here are the design decisions and how they will impact the Nextruder for the foreseeable future.
Interested to know what others think...
RE: Is the Nextruder hot end design choice a mistake?
Personally i changed 3 out of 5 toolheads with 0.4. I did see some improvement but to be honest i didnt run thorough test as the guys in the videos. I print mostly with petg and what i noticed was that the sometimes stray small blobs in multimaterial prints where eliminated and also an improvement in stringing. This small blobs were the result of nozzle drip during travel. It usually end on the tower but still they were there. Maybe as the guy on the 2nd video said its a matter of tuning the profiles. For the time being i dont see a huge benefit of staying with 0.6. Even in simpler models if there are supports i mix pla and petg so the printing layer size is .2 to .25 (to be able to use soluble options). What im hoping to see soon and the reason i kept the 0.6 nozzle on 2 of the heads is multi nozzle size printing. And maybe if i can use the soluble options with .4 layer size for prints that dont mind a stray blob here and there or a bit less on the quality size.
Maybe a future option will be a high flow version of the 0.6 nozzle.
I think they didnt use the revo probably to make it simpler for people to use the old e3d nozzles with the adapter. Or maybe it was a matter of cost versus benefit.
RE: Is the Nextruder hot end design choice a mistake?
I honestly do not really see the reason why the nozzle should not be good? With the 0.6mm Nozzle the XL is SOOO fast at printing. Yes for everything high detailed a 0.4 is kind of a must I admit.
I personally do not really see why the revo would be a better choice. I actually think there is a good reason the nozzle of the nextruder is that long and that would not have been possible with revo right?
IMHO it would have been a clever move to just give people the 0.4mm nozzle since it is much easier for beginners. Apart from that I am super happy with the Nextruder.
RE: Is the Nextruder hot end design choice a mistake?
The 0.6mm isn't really that fast though is it? The speeds are lower than the 0.4mm nozzle, likely due to the limitations of the flow rate. The 0.6 is leaving performance on the table. If it could melt plastic faster then the hardware could certainly allow faster printing IMO.
As they start to use the built in accelerometer I would expect the input shaping could allow for faster print moves but the hotend side will limit print speed from what we're seeing.
RE: Is the Nextruder hot end design choice a mistake?
The 0.6mm isn't really that fast though is it? The speeds are lower than the 0.4mm nozzle, likely due to the limitations of the flow rate. The 0.6 is leaving performance on the table.
I think you are making the mistake of looking at the Prusa Slicer profile right? These profiles are not always optimally tuned. I was able to lay down a lot of filament very very fast with a 0.6mm nozzle, but to be honest I can not say how fast since it was some weeks ago I did that. IMHO you will not get to limitations of the flow with a 0.6mm nozzle unless you will accept very rough prints with not a lot of details.
I mean this is with 0.8mm, but I think it shows the nextruder can handle quite a lot. I think the problem for most prints will be the printfan not cooling down the filament fast enough.
RE: Is the Nextruder hot end design choice a mistake?
The Revo has started to see issues with nozzles slowly backing out and ultimately being snapped off when they come into contact with a print at high speed.
Yep. The default profiles for the XL, even the IS ones; are very slow. Prusa do this deliberately to have stock profiles that with generally result in success and quality and they made it clear that speed was not a high priority for them.
The Nextruder will happily run WAY faster with a lot higher volumetric flow rates and still produce amazing prints. I am running way faster than the stock profiles without issue.
Speed comes at a price in other areas, depending on how fast you go.
I've only had my XL for a week and am producing awesome multicolour prints with the stock 0.6 nozzles teamed up with one extruder running a V6 adapter and a 0.6 NozzleX for PETG CF.
Drying your filament is critically important. I live in a very hot and humid area. Before printing, I dry my filaments till they are stable at 20% or less for 12hrs.
99%+ of my prints are in PETG and after some tuning, my XL punches out print with little to no stringing at all.
Also, printing and installing these brush holders has been a massive improvement.
https://www.printables.com/model/462473-prusa-xl-series-nozzle-scrubber-brush-update-inclu
RE: Is the Nextruder hot end design choice a mistake?
As already written, a 0.4 nozzle is not always slower as a 0.6 nozzle. Especially if you print multicolour parts you will also save some filament.
As long as there is no official support from Prusa for multi-nozzle prints (details with 0.4 and large areas with 0.6). I know, that there is a work around for this setup.
This week I will receive my remaining 0.4 nozzles for a full XL set. Black Friday sale with 35% off 😀
RE: Is the Nextruder hot end design choice a mistake?
[...] One has to assume going proprietary was a business decision. I know it's meant to be open source but I've not seen specs published etc although I'm sure it's easy enough to reverse engineer.
On the other end if you're wrong about this and their plan was to open source everything eventually as they said they would, then it make sense to develop your own solution as the Revo hot side is proprietary.
RE: Is the Nextruder hot end design choice a mistake?
It’s a fair comment. Time will tell. I wonder if we’ll see a CHT licensed version of the nozzles. That would be interesting.
RE: Is the Nextruder hot end design choice a mistake?
You could try one with the V6 nozzle adapter I guess. Would be interesting to see if it can bring the 0.6 speed to 0.4 level.
RE: Is the Nextruder hot end design choice a mistake?
What do you mean by "bring the 0.6 speed to 0.4 level"?
Given a 0.6 is laying down 31%+ more plastic per pass, you need to effectively multiply any speeds you are using by this amount to get a more accurate comparision to the speeds you're running a 0.4 at.
I can assure you any job I'm running will be drastically faster with a 0.6 than it is with a 0.4.
RE: Is the Nextruder hot end design choice a mistake?
The 0.4 profile has print moves faster than the 0.6 profile. Of course the 0.6 has to push more plastic at the same layer height and even more at an increased layer height. At the same height you can print a part faster with 0.4 nozzle although it doesn’t mean the parts are equal as there will be more plastic in the 0.6 one. The videos I link to focus on that specific situation. What I’m saying is we have a motion system that can move at that speed but the 0.6 nozzle must be at its melt limit else why limit the print speed? So in theory, if we remove that limit you’d get even faster 0.6 nozzle prints. There is definitely a tend towards faster flow and improved melting of the plastic to allow increased print speeds. With the build volume of the XL the faster Prusa can get it printing the better in the current market, IMO.
RE:
What he said. Just look at the for 0.20mm layer height default speed profile for both 0.4 and 0.6mm.
The perimeter print speed is 125 for 0.6 and 175 for 0.4. That's a pretty huge difference.
Surprisingly the infil speed is the same in both profiles.
I use 0.4mm most of the time because of the reduced stringing and better print quality. But now and then I want to print something large and I always compare the 0.4 vs 0.6 to see how much time I will save to determine if I should bother switching the nozzles back.
The difference has always been negligible to the point I don't use 0.6 at all right now.
Go try it in prusa slicer. Create a cube and slice it with the same 0.20mm layer height with the speed profile and compare the print time between 0.4 and 0.6mm profile.
Here what I got.
180x180x36mm, 15% infil, 0.4mm = 5h 50m
180x180x36mm, 15% infil, 0.6mm = 5h 36m
180x180x180mm, 15% infil, 0.4mm = 22h 11m
180x180x180mm, 15% infil, 0.6mm = 21h 37m
180x180x180mm, 50% infil, 0.4mm = 2d 13h
180x180x180mm, 50% infil 0.6mm = 2d 10h
355x355x355mm, 15% infil, 0.4mm = 6d 0h 42m
355x355x355mm, 15% infil 0.6mm = 5d 23h 13m
Notice how it barely changes anything. For the small one, it would take me more time to switch nozzle than the time it would take. Is it worth the lose in print quality?
Even for the large cube you would save less than 30 min. I need to go to 50% infil to save 3h. But even then I don't think I would switch over to 0.6 unless I didn't care about quality and had to print multiple times the same thing.
And then I tried filling the whole bed and surprise.. barely over an hour saved over a 6 day print.
Those aren't the most realistic scenarios. Your time saved would vary from model to model depending on the amount of perimeter and everything.
But I have yet to find a print that makes it worth it to switch back to 0.6mm even when I don't care whether the print looks good or not.
Why did they ship this printer with a 0.6mm nozzle!? Make no sense.
RE:
I reiterate:
The default profiles for the XL, even the IS ones; are very slow.
The Nextruder will happily run WAY faster with a lot higher volumetric flow rates and still produce amazing prints. I am running way faster than the stock profiles without issue and I have almost zero stringing issues with 0.6 nozzles and I pretty much only print in PETG.
I do use 0.4 nozzle for models with really fine details - mainly small text that is face down in the first 2 layers or very small surface details.
People need to stop pointing to the stock profiles as them being a speed barrier or some sort of limitation and tune their own. There isn't a volumetric flow issue with the 0.6 nozzle.
It will run the 0.6 without any issues at the stock profile speeds for the 0.4. It will actually go much faster without issue.
In an terms of comparing 0.4 to 0.6, do your example, but change the perimiters for the 0.4 to 3 and the 0.6 to 2 so the wall thicknesses are equal.
You must have old profile configs or are not using the 2.7.0 IS profiles?
Even so, I've just gone in and done some of your examples:
36x36x36 box:
Default XL IS 0.4, Default 0.20 SPEED XLIS 0.4, 2 perimeters, 10% cubic infill, 4 top and bottom layers: 51m
Default XL IS 0.6, Default 0.20 SPEED XLIS 0.6 2 perimeters, 10% cubic infill, 4 top and bottom layers: 31m 40% faster
Default XL IS 0.4, Default 0.20 SPEED XLIS 0.4 3 perimeters, 10% cubic infill, 4 top and bottom layers: 52m
Default XL IS 0.6, Default 0.20 SPEED XLIS 0.6 3 perimeters, 10% cubic infill, 4 top and bottom layers: 32m 39% faster
180x180x36 box:
Default XL IS 0.4, Default 0.20 SPEED XLIS 0.4, 2 perimeters, 15% cubic infill, 4 top and bottom layers: 6h54m
Default XL IS 0.6, Default 0.20 SPEED XLIS 0.6, 2 perimeters, 15% cubic infill, 4 top and bottom layers: 5h30m 21% faster
Default XL IS 0.4, Default 0.20 SPEED XLIS 0.4, 3 perimeters, 15% cubic infill, 4 top and bottom layers: 7h7m
Default XL IS 0.6, Default 0.20 SPEED XLIS 0.6, 3 perimeters, 15% cubic infill, 4 top and bottom layers: 5h44m 20% faster
360x360x360 box - which is no issue to get on the bed BTW:
Default XL IS 0.4, Default 0.20 SPEED XLIS 0.4, 2 perimeters, 15% cubic infill, 4 top and bottom layers: 7d23h20m
Default XL IS 0.6, Default 0.20 SPEED XLIS 0.6, 2 perimeters, 15% cubic infill, 4 top and bottom layers: 5d21h6m 17% faster
Default XL IS 0.4, Default 0.20 SPEED XLIS 0.4, 3 perimeters, 15% cubic infill, 4 top and bottom layers: 8d3h24m
Default XL IS 0.6, Default 0.20 SPEED XLIS 0.6, 3 perimeters, 15% cubic infill, 4 top and bottom layers: 6d2h5m 16% faster
Despite the 'slower' print speeds, simple maths wins the day due to the cubic increase resulting from the higher extrusion width x the layer height.
And just for fun, lets run the same settings, but change the 0.6 speeds to match the 0.4 default speeds; which the XL will do without breaking a sweat.
Default XL IS 0.6, Default 0.20 SPEED XLIS 0.6 @ 0.4 speeds, 3 perimeters, 15% cubic infill, 4 top and bottom layers: 5d22h31m 27% faster
No surprises. The 0.6 beats the 0.4 by a mile in every example. Even the 36x36x36 box is 20mins faster.
Nozzle changes take me about 2mins because I have full hotends ready to swap out.
RE: Is the Nextruder hot end design choice a mistake?
Interesting. In the videos I linked to and based on the profiles at the time the max flow for the 0.6 was about 12-15mm3/s which by todays standards is low. Would be good to have someone do some more testing to confirm. Sadly, I currently don't have the time.
RE: Is the Nextruder hot end design choice a mistake?
Apologies to thse that think I'm starting to sound like a broken record...
The default profiles are conservative and should be taken with a grain of salt. Prusa deliberately do this to ensure most users will get good prints most of the time without having to tweak anything. They are nowhere near the max potential of the platform.
If you own the printer, are using it for business purposes and are serious about getting the most out of it (which makes you more money in less time); you spend the time on tuning your profiles. The time spent will repay itself in short order.
I have and have reduced job times by ~50% versus the high speeds I run my MK3S at. Another printer that will got WAY faster than what the default profiles will have you believe is possible.
This is not a home user plug-and-play toaster - it's a prosumer machine with a lot of potential waiting to be unlocked.
RE:
You must have old profile configs or are not using the 2.7.0 IS profiles?
Double-checked this morning and I was definitely using the IS profile and there's no new configuration available. I'm on 2.7.0.
Yet somehow I'm getting vastly different number from yours. Is this on the 5T profiles? I'm using the 2T profile and now wondering if there a big difference between the two.
Nozzle changes take me about 2mins because I have full hotends ready to swap out.
Yeah.. it's a lot more involved without that investment that most people without a business won't do. Most of the time I already have a filament loaded so 2-4 min lost just unloading that. Then a good 15 min waiting for everything to cool down. Then another 2-4 min reloading the filament. Whole thing take 20-30 min easy.
RE: Is the Nextruder hot end design choice a mistake?
I really appreciate your contribution to this thread. What do you make of the test shown in this video at about 3m50s? the 0.6mm nozzle seems to top out such that the flow limits the speeds yet you seem to be able to push the speeds faster? Materials, temp etc may be the reason.
RE: Is the Nextruder hot end design choice a mistake?
@ipind-3d
Installed all 1T, 2T and 5T profiles on 2.7.0 over my lunch break and they gave me identical time.
I launch PS, create a cube with the default 36mm dimension. Everything else is the default, 15% infill, 0.20mm speed profile, 5 top layers, and 4 bottom layers. 2 perimeters.
All I do is change the profile between 0.4 to 0.6.
0.6 give me a 48m print time.
0.4 give me a 33m print time..
On all 3 profiles.
Reducing 0.6 to 1 perimeter only brought the time down to 32m.
Why are you getting vastly different number from this!?
Why is 0.6 that much slower on my side!?
So confused..
RE: Is the Nextruder hot end design choice a mistake?
Now that is interesting.
Yep. Running the 5T IS profiles, but the speeds should be the same regardless of the number of tool heads.
The settings I was using are 0.20 SPEED XLIS 0.6, 2 or 3 perimeters, 15% cubic infill, 4 top and 4 bottom layers. But really, as long as you are using the same settings on both nozzle sizes; you should still see big differences.