Now that I can fully get behind. With an XL type competitor, I wouldn’t expect anything less.
After doing a little more tuning and testing, I’m of the opinion of the 0.8mm nozzle is too much for the hot end. I ran a flow test for both ASA and PLA. For ASA, the test was much closer to the actual ability of the printer, meaning it was tuned fairly close to the number I measured from the test. I tuned the max flow rate of PLA and it was way off. It tuned to 40mm/sec3 but could only print at 28mm/sec3. When running at the max tuned setting it was under extruding really badly on perimeter lines. Which is odd, because the max flow test was beautiful under similar straight flowing lines.
But the take way I have, 0.8mm nozzles have no value outside of the wider line width on this printer. I’ll qualify that with, I’ve only spent a small amount of time tuning it, but I’m not even seeing speed improvements with 0.56mm layer height opposed to 0.24mm. Actually, Orca and Bambu Studio are estimating 0.56mm is actually slower. This is because the kinetic speed with thin layers is more valuable than slow speed with less thick layers. Note, this is with the nozzle tuned max flow rate, not the generic 0.4mm nozzle max flow rate from the filament profiles.
For those that want to try, I did find something helped to optimize the speed, I felt higher nozzle temps were required. I only tried +5C more but it was clearly an improvement when pushing the filament to the limit (at least on ASA). At the base max flow rate, it wasn’t needed, but at an elevate rate it was. This could be an area that may yield a little more performance, but I didn’t explore it very far.
Largely, my feelings from the 0.6mm nozzle was very similar. Not worth the hassle to switch from the 0.4mm nozzle unless you are using vase mode or prefer wide walls and thick layers for layer adhesion. Personally, I believe the 0.6mm nozzle actually had more use, because there were some tangible speed improvements. The 0.8mm nozzle didn’t seem to have any, or at least anything greater than a couple of minutes per hour of printing on specific models.
Use this as a starting point, not an end all. Again this was based on a very small sample size of about 10 prints, but I’m pretty comfortable in saying the printer doesn’t really respond to big nozzles. Its possible the upgraded hot ends may unlock a few more benefits, but the stock hot end seems to be the limit (but I reserve the right to be wrong there, it could be another piece of the puzzle).
What I’d also like to mention, I haven’t really found the nozzle profiles to be the problem. They actually seem well made and competent. Like many others, I did assume they were the problem initially, but no… they are not the problem.
I still feel like to make the larger nozzles work, they will need to employ a better way of limiting the volumetric flow or a way to make the max flow nozzle centric. Flip flopping between filament profiles is annoying, as well as having to manage them separately (basically having two identical profiles with 1 or 2 numbers being different). You can just use the base filament profile, but that leaves a little performance on the table.
1 Like
Maybe even the Prusa XL has had problems. According to Teaching Tech, originally it started with 0.6mm nozzles as the default, but that has since changed to 0.4mm.
On the other hand, you have someone like Tom Sanlanerer who thinks 0.4mm is obsolete and nearly everyone should be switching to 0.6mm. Go figure.
Yeah, the XL had some real issues that they needed lots of time to sort out. I believe there was at least one hardware change, possibly more. But they got it sorted. On the Bambu side, I wouldn’t say the X/P series is broken though, big nozzles are just not an ideal configuration. The printers are simply optimized for a 0.4mm nozzle. Which I think was a good idea. So I’ll just accept it as what it is, better with a 0.4mm nozzle.
Honestly, if it were to have more volumetric flow, it would go faster on the 0.4mm nozzle too, so it would likely end up in the same place. So unless, you go out of your way to optimize for a larger nozzle, I don’t think there will be a need to go that route anymore. Maybe if we get to the point where the available speeds and accelerations are too slow to keep up with the hot end, that may be a place where the bigger nozzles make sense, but I don’t see that occurring. In the past, when 100-200mm/sec was a fast movement speed, I can see why bigger nozzles had a value. But as speeds became higher and higher, the speed benefit from taller layer heights seem gone.
I have to go back and listen to his claims. Because honestly, it didn’t age well based on what I’m thinking.
The larger nozzle enables printing thicker layers and lines, a feature which can be explored to reduce printing time. However, there are restrictions which can limit the printing speed.
Extruding more material translates into a higher flow rate, which means that a larger quantity of material is melted and extruded, requiring a higher amount of energy per time of the hotend.
Similarly, a higher filament MFR is necessary.
In the case the above requirements are satisfied, you will be able to deposit thicker layers at high speeds, but unless the print surface is quite large (increasing layer time), you will face a demand for higher cooling power, as you need to cool down a larger quantity of filament in a similar time.
If the high speed associated with the 0.4mm nozzle comes from using the hotend heat and cooling power near their maximum capacity, the larger nozzle will not be able to print faster.
2 Likes
This is where the problem comes from. The larger nozzle and taller layers are limited by flow. So its a case of pick your poison, higher speeds with more layers, or lower speeds with less layers, both volumetrically limited to about the same time.
That said, I did see some legitimate gains on the 0.6mm (possibly a better sweet spot), but I’m not a fan of losing the detail.
I am ok with the lower speed. In the case of trying to reduce the printing time with larger nozzles, it is essential to optimise the printing settings, such as reducing the number of walls, as the lines are thicker, or the infill pattern and %. However, the potential is likely limited without changing hardware and optimising it for the larger nozzle.
For me, using a 0.6 mm nozzle is limited to the filaments with abrasive fibres such as PA-GF and PLA with significant wood content to avoid clogs and, in the case of wood, also for aesthetics.
Edit: as with everything, there are exceptions. In this case, printing in vase mode benefits from the thicker wall.
2 Likes
I am old fart and started printing with 3mm filaments…
So I come the world of 0.8 and even larger nozzle sizes…
As so many already mentioned the default is 0.4mm.
Bambu had a nice 0.2mm contest to get more data for providing better profiles but anything above 0.4mm these is considered a waste.
Not waste in terms of material but in terms of supporting it.
Take something as basic as the required flow rates into consideration.
Double the nozzle size and you get a much large surface area.
0.126 square mm for the 0.4mm nozzle and 0.503 for the 0.8mm nozzle.
So your volumetric flow rate going from 0.4 to 0.8 has to go up by almost 5 times…
The stepping and gearing of the extruder system are not designed for those speeds.
A stepper motor looses torque at high speeds and we need high forces to print fast.
Plus the lack of thermal mass in the hotend means the tiny ceramic heater has go into overdrive - which is not supported by the firmware in terms of having a dedicated PWM table for the other nozzle sizes.
While printing with tiny nozzles and 3mm filament was a pain in terms of getting an accurate extrusion using 1.75mm filament means being a bit stuck.
Can’t go much finer but going much bigger means the hardware poops out…
I would simply suggest to use a more powerful heater cartridge and to change the gears on the extruder…
But that does not really work without being able to make the required firmware mods.
Our printers are all about speed and shedding weight.
Actually supporting 0.8mm nozzles would mean adding weight…
So being just able o somehow print that wide is seen as sufficient 
I sometimes would love to be able to print strong parts much faster with just 3 single walls instead of many plus infill but my Bambu printer won’t let me…
4 Likes
I had forgot about this aspect. I thought about it while it was printing, but only saw the AMS buffered spin rate, which appeared much more tame. But thinking about it, the extruder is likely the biggest factor of the problems I saw.
Going from 0.4 to 0.6, saw the volumetric rate effectively double, but going to 0.8 from 0.6, was a direct wash. The extruder failing to keep up makes a lot of sense for the test failing as early as it did. Effectively, setting a cap on its performance.
Of course its not all of the problem, but I do think its a major part of it. Thanks for the contribution.
Here it is:
P.S. I did in fact find that turning on Arachne for a 0,6mm nozzle made a huge improvement.
3 Likes
Ok. I remember what the problem was with his logic. It relies on Arachne. I have issues with Arachne because it’s a setting you need to keep an eye on to avoid artifacts, so I choose to not use it, except when its needed. Take a look at the Bambu Wiki on the pros and cons, and see if you mind using it, Introduction to wall generator | Bambu Lab Wiki.
Also, he did this with a MKIII which has a 200mm/sec max speed (and about a 150mm/sec generally accepted max speed for quality). This coupled with the 15mm3 volumetric flow (very solid), and you could potentially be bottlenecked with a 0.4 where you can easily flow enough for 250mm/sec (which the MKIII can’t really do well). So, you have a situation where the 0.4 nozzle is limited by the kinematics of the printer at 150-175mm/sec when the extruder can supply filament for 250mm/sec (on the 0.4 nozzle). In this case, the 0.6 could flow more filament at the taller layer heights and still have headroom in kinematic speed, leading to faster printing. The 0.6 nozzle doesn’t need to go over 90mm/sec to beat the 0.4 nozzle with the taller layers.
But that’s not the case for the Bambu printers (or any new printer). Print speeds of 250mm/sec are the norm. Yes, that is still very fast, but we now see peak speeds that remove the bottleneck that the MKIII had. These new peak speeds largely remove the need to go to a larger nozzle. There are still some situations where you can reach peak profile speeds (300mm/sec) while still having the ability to flow more filament. But if you were to bump the max speeds up 50mm/sec or so, that benefit would disappear too. I think the BBL printers still have great quality at 350mm/sec on most lines other than the outside perimeters. So, in reality I think you could negate the need to bump up to a 0.6 nozzle, but depending on geometry, the surface quality could be a little better with the 0.6, but you would still need to factor in thicker layer lines and lower overall details (or manage the Arachne issues).
All up, I think he was right… for the time this video was made. But the 3D printing world has REALLY moved on since then, and the 0.4 nozzle is back in style for performance reasons. At least for moderately good flowing hot ends and extruders.
2 Likes
Ohh… BTW, thanks for bringing that video into the conversation. It really helped me close the loop on my thoughts. It all makes sense to me, even if I may be missing a piece of the puzzle or over emphasizing one aspect over another. I feel like the picture is very clear now.
1 Like
He did a follow-up video:
where he bizarrely advocates for keeping the 0.4mm nozzle while printing using a 0.6mm profile. It seemed to completely contradict his earlier video, and it left me wondering which approach is really “the best”?
Yes, at least for PLA, this is what I found as well. On the other hand, I’m noticing that some filaments, like ASA and perhaps PETG, seem to have trouble printing at higher X1C speeds even though well within their maximum volumetric flow rate. I don’t understand why. The discrepancy didn’t manifest on an MK3, maybe because an MK3 never got to such high speeds. If it really is a speed issue, then in the case of such filaments, maybe going to a bigger diameter nozzle will let you push more plastic, at the expense of resolution? It’s hard to know for sure because the 0.6mm profiles don’t seem as well developed, so it’s not easy to devise a simple comparison. This is where I got frustrated, threw up my hands, and retreated back to a 0.4mm nozzle, where overall it just seems easier to get good prints without much effort, even if the price is you may have to dial back the speed on some filament types.
Speed in the 3D printing world is a relative thing…
We can’t overcome physics.
All we ever do is to monitor the temp of the hotened, using a temp sensor somewhere in the housing.
But we NEVER check the actual temp of the filament coming, let alone the temp inside the melting chamber.
In the ideal world the melting chamber would be rather high in volume, really thick walled and our extruder very strong and fast.
We make compromises though.
Check the time the filament has inside the short melting chamber and how this time translates to higher print speeds.
You will find that in order to actually get the filament hot enough you need to ramp up the heating power for higher speeds.
Imagine the filament being a thick sausage…
It goes through the hotened and heats up.
But the ONLY thing able to heat it are the walls.
Means the inside takes much longer to heat up than the outside of the filament.
The mix of hot and cold filament causes inconsistencies and can even result in the nozzle blocking up for a moment with a blob of too cold filament.
If you ask me than a hotend does require design changes as much as the extruder in order to deal with much larger or smaller nozzles…
@user_3026326371 Those limits should be captured by the maximum volumetric flow rate number when you run that calibration on any particular filament. Right? You enter that into the filament profile, and the slicer makes sure the generated g-code doesn’t exceed that limit.
You could if you wanted run the calibration for each nozzle diameter, and if it turns out the maximum volumetric flow rate differs, then create a different filament profile for each nozzle diameter. Wouldn’t that adequately cover all your concerns?
I just now thought of a simple experiment that might prove revealing:
- Run the maximum flow rate calibration. It will pick a middle of the road layer height for the test, such as 0.2mm for a 0.4mm nozzle, or 0.3mm for a 0.6mm nozzle, or 0.4mm for a 0.8mm nozzle.
- Plug the measured maximum flow rate number into the filament profile, as per usual.
- Now, using the same model that the maximum flow rate calibration used, try printing at the minimum layer height, and see if the print is still acceptable. The minimum layer height will drive up the print speed, and so, if higher print speed somehow causes a failure, it should happen then and there, even though the previously measured maximum flow rate was not violated. In theory, it shouldn’t fail, so if it does, there’s something wrong with the theory. If it doesn’t fail then, but does fail when printing some other model, then I guess some other factor(s) must be in play and not just speed alone.
If I can find the model used for the maximum flow rate calibration, I’d certainly be willing to give this a try.
I have never done such a test, so this is merely my opinion: it will likely work; the only drawback would be if the test print has a small area but is easily manageable by the minimum later time. Nevertheless, within the test, there is no change in the area and, thus, in the flow rate, so the problem associated with nozzle diameter variation will not be addressed. This is how I imagined what would happen, but I’m not sure it will happen.
1 Like
@JayZay Not understanding what you’re saying. Maybe an example would clarify?
The failure that I found was less about speed and more about flow. Sure… since flow is determined by speed, you can say its speed related, but in all honesty, that isn’t the differentiator.
If the machine had the ability to move perfectly at infinite speed, my issue would still occur at the normal flow rates I tested/calculated. This is because the extruder or hot end (or combination of both) simply couldn’t keep up. Note, my tests were at short and tall layer heights with the large nozzles. This worked because the larger nozzles brought the speeds down so low that they were really a non-factor. For example, at the lowest layer height of the 0.8 nozzle was 0.24, to do 150mm/sec you need to be able to flow 36mm3. I didn’t have any filaments capable of that, so all of my 0.8 tests were done well under 150mm/sec.
Adversely, if the extruder or hot end were infinitely capable, I wouldn’t have had an issue with the speed because the machine is known to be able to work perfectly at those speeds (150-250mm/sec).
I feel like this topic needs graphical explanation because it’s so interrelated. If I were smarter, I could set up a theoretical top speed map/graph based on nozzle size, flow rate, speed and time. Then, I think the whole thing would be super clear. But I don’t have the patience to think my way through that. But as envisioned, it would show you the crossover points where the nozzles make sense and don’t. Obviously, this would be for speed only, not other aspects like layer adhesion and vase mode.
1 Like