I also want to mention that oddly enough whenever I add Brims the layer issue doesn’t exist. The Brim ALWAYS prints fine and doesn’t have those gaps between lines like the actual print object has.
There are flow rate settings for first and top layers under advanced or experimental.
I wonder if that flow rate is absolute or relative to the standard flow rate. It really feels like a slicer issue to me.
In Dev mode there are such settings:
But most probably it will need to be tuned for every object and every combination of materials, nozzles and etc. Which will be madness of testing, and absolutely fail the idea of the Bambu printers.
I absolutely support the idea that if A1 could print something with given quality X2/H2 should be able to do the same or better. With all the better tech inside.
I fully agree. If it works on A1 printers, it should be possible to have it work on all others too.
It still feels like a bug in the slicer somehow.
If it were a Bambu Studio issue, it shouldn’t show up in Orca Slicer. I was told the nightly build supports H2C.
This has been my entire motivation to push on this issue and thrilled to see the discussion over the last few weeks.
I think all of us who have used FDM printers accept there are some precision limitations, but it’s not an acceptable response for Bambu to then simply write this off as “limitations of FDM” printing with the very obvious and apparent better quality differences within the lower-tier (ie. non-flagship and much cheaper) Bambu Lab FDM printers.
I’m still suspecting this can likely be solved with slicer and or firmware updates, which does require ongoing troubleshooting not just by the larger community here, but internally by the Bambu Lab engineers.
For anyone from Bambu who might read this: There does seem to be a push over the last year to release more devices and capabilities, this has likely lead to the trade off in quality over quantity, but anyone in management should know degradation in any one aspect of the balanced scorecard (process perspective) usually result in downstream negative affects in the other parts (customer perspective). As a customer, knowing my very expensive BL flagship printer is producing a lower quality output (even if more capable) than a low-tier device makes me question future purchases (which will affect company financial performance if competitors are are more reliable).
It does appear the H2C has been added back into the OcraSlicer nightly builds from when I last checked a few weeks ago. If I have some time I will try some of my models I have had issues with and the test model above.
Bambu Studio and Orca Slicer have more in common than you may think.
If I hear that first and top layer have issues, and other layers are fine, a hardware issue seems unlikely.
If only the first layer has an issue, it might be a Z-offset issue and therefore be hardware related. But the hardware has nothing specific to say about the top layer imho.
Just a wild guess though.
Yeah, and why the ironing on these devices should be so hard to dial in compared to all lower tiers.
Either they tried to introduce some very smart algorithm, or the old “magics” from the old gen printers are just not applicable on this extruders, and they need to create new algorithms now, but too busy with too many products. Or the only one dev that knows actually how the tune them left. Stuff like this, things from the real life, that we already saw more or less in other cases and products.
Yes, I understand. But when I print the same object on my P1S, it turns out perfectly. You said yourself that your H2D doesn’t give you any trouble, and you’re using the same slicer we are or did I misunderstand that? I still think there are different issues at play occurring at different levels, that result in similar symptoms. If it were easy to fix, surely Bambu would have resolved it by now, after all these months.
As you can see first they said this is a bug then they said this is limitation of the technology…
I just wanted to add my thoughts on this here as well. I also posted over on reddit about my current and limited findings, and agree with others that this is a problem that will likely need multiple things to be fixed in order to resolve it. Probably both software and firmware. The hardware seems solid enough.
First, i’d like to thank those that have kept up on this issue with Bambu. I’ve only had my H2C for about 2.5 months, and have been printing other multicolor prints, but just recently printed some models i designed with mutli-color on the first layer. I had previously disregarded some of the printing issues i saw in the past, but now after reading through this and the bug report, I believe this is definitely a larger issue which will encapsulate a lot of different causes.
First, I truly believe that the auto-bed leveling process is flawed. On an early print, i noticed one corner of the print would lower the bed a huge amount, and the resulting object had very noticeable layer lines and gaps and the part was dimensionally inaccurate on that one side. Thankfully that part at that point didn’t need to be super accurate, so I didn’t mind. I don’t watch all my prints, so I haven’t noticed it to that drastic amount since, but i’m doubtful that it is always correct. If it can be THAT wrong once, without a declared bug being fixed, you can’t be confident that it is accurate and consistent.
Second, after playing with getting a more accurate first layer using a 0.2mm nozzle for mini printing, I noticed that the start gcode for the 0.2mm profile actually raised the nozzle for flat plates causing a 0.01mm higher first layer. So, when i saw this issue and noticed the issue in my prints on the 0.4mm nozzle, i checked the start gcode again. The start gcode for the 0.4mm nozzle is exactly the same as the 0.2mm nozzle. There is actually an if statement checking for a 0.2mm nozzle in the 0.4mm profile. Also, comparing this gcode to my p1s, i noticed that even if the z-offset adjustment was being used, it was only -0.02 instead of the -0.04 on the p1s. This would cause more first layer squish reducing the layer lines. So, it definitely looks like there is an issue with the H series machine profile and start gcode that needs to be addressed. I have changed my gcode to by -0.04 on the h series and will be doing more testing using my models to see how much difference it makes.
Third, after changing the gcode, I did a manual filament flow calibration and found that 1.0 would be better right now for me than the default 0.98. I also checked the bottom of the calibration tiles with a magnifying glass and could see a noticeable difference in the bottom lines and gaps. Granted, even at 1.03, it didn’t completely close all pin holes on the bottom layer, so that’s still a concern on simple objects like that.
I definitely don’t have a fully acceptable result yet, and I think that we will still need help from Bambu to resolve other machine based issues, possibly like bed leveling. I’m looking forward to hopefully getting to be able to print consistently with this machine. The P1S was so great because it was just plug and play. So, hopefully we can get back to that again soon.
Another thing I may try in order to mitigate this issue is to reduce the first few layer heights down to 0.1 or something like that, but that shouldn’t be necessary and would be more of a crude workaround for this issue.
Looks like you have only one top layer. But I might be mistaken.
Might be that the bridging is causing this, but your flow rate seems quite low.
Or your nozzle is clogged.
Not that part had 2 top layers. Was printed with Bambu ABS using their print profile with a 0.6 nozzle. I only have this issue when using the 0.6 nozzles, never with 0.4.
It feels like you need to do some flow calibration here. Your last image show a lot of under-extrusion. Are you running a standard profile?
Standard Bambu profile with their ABS.
Then your nozzle might be partially clogged.
Try a cold pull or another nozzle to see if the issue still appears.
After running a bunch of tests and inspecting and editing the machine start GCode, I have a much better first layer result on my H2C. This change could affect the top layers as well depending on how thin your print is. I’ll try to sum up my testing and findings as best I can. Here is a picture of one of the final results, I labeled it with R2 for the #2 right nozzle being used for the color.
Above shows some small gaps in the yellow D20 outline, but that is mainly because of how small it is and the limitations of a 0.4mm nozzle. In future prints I’m tweaking settings still to get that a little better, however, the biggest thing is that you can’t see all the lines in the black area and the filament is fused between the yellow and the black. You also cannot see any extrusion lines in the yellow letters.
Short description of Fixes/causes:
- Machine Start GCode for the 0.4mm H2C default profile is incorrect. It does not apply any z-offset for the textured build plate.
- Some nozzles cause under-extrusion. I believe this is due to a manufacturing issue causing more resistance inside the nozzle.
- Slight bed leveling issue (Front of bed has the nozzle closer to bed when printing than the middle or back)
Before doing anything, I made sure to cold pull all the nozzles and wash my build plate with dawn dish soap and ensure it was rinsed thoroughly. I’m also using all Bambu Lab filament (Basic PLA and Matte PLA for some colors).
First, I started with default settings on the textured plate and printed a 0.2mm height “first layer test” pattern. I just used primitive objects to create 9 squares connected by lines about 2mm wide. This was to test both the first layer quality as well as the bed leveling consistency. The lines were not connected together and looked like it was “under-extruded” even though it wasn’t. Pictures available of tests if needed.
I then took a look at the Machine Start GCode I was talking about before. I looked at the P1S start GCode and compared that with the H2C start GCode. This is what the Textured plate section of the default H2C machine start GCode looks like in the current version of Bambu Studio:
;===== for Textured PEI Plate , lower the nozzle as the nozzle was touching topmost of the texture when homing ==
{if curr_bed_type==“Textured PEI Plate”}
{if nozzle_diameter[initial_no_support_extruder] == 0.2}
G29.1 Z{-0.01} ; for Textured PEI Plate
{else}
G29.1 Z{-0.02} ; for Textured PEI Plate
{endif}
{else}
{if nozzle_diameter[initial_no_support_extruder] == 0.2}
G29.1 Z{0.01} ; for Textured PEI Plate
{endif}
{endif}
The P1S section looks very similar, just simplified a little bit because there is no 2nd nozzle. The P1S z-offset for the textured plate is -0.04mm, where this would only be -0.02mm. Also, I have no idea why it is checking for a 0.2mm nozzle in the 0.4mm nozzle printer profile.
Via many test prints and measurements with calipers, I adjusted the machine start GCode to have a different z-offset. I tested this with both a flat plat (Biqu Glacier) and the textured plate. After dialing it in with the left nozzle, i also tested and tweaked with the right nozzle. My final GCode is as follows: ***** USE CAUTION when changing GCode, Use at your own risk.
;===== for Textured PEI Plate , lower the nozzle as the nozzle was touching topmost of the texture when homing ==
{if curr_bed_type==“Textured PEI Plate”}
{if nozzle_diameter[initial_no_support_extruder] == 0.4}
G29.1 Z{-0.05} ; for Textured PEI Plate
{else}
G29.1 Z{-0.05} ; for Textured PEI Plate
{endif}
{else}
{if nozzle_diameter[initial_no_support_extruder] == 0.4}
G29.1 Z{-0.02} ; for Textured PEI Plate
{endif}
{endif}
I probably didn’t need to change the checks of nozzle diameter, but I don’t like seeing 0.2mm nozzle stuff in 0.4mm printer profile. Basically I adjusted the z-offset another -0.3mm for the textured plate, and set it to -0.02mm for all other plates. These values resulted in much better first layer tests. I also found that the front of the build plate when using the right nozzle was slightly too close to the bed, but i’d rather have better first layer squish on the entire plate than adjusting for the front of the bed. On top of that, I couldn’t figure out if there was a way to adjust z-offset for the right nozzle separately from the left.
After I got that dialed in, i started printing some objects with multiple colors on the first layer. They started coming out better and I was adjusting the first layer extrusion ratio to fill gaps in the small text and design I was using. I was also printing in some different colors as I went because they quality was coming out a lot better. I then printed in a Matte Baby Blue and noticed the extrusion lines were noticeable again. This is when I realized that by changing colors, I was also using different right nozzles. I realized the first time i printed that Baby Blue it used a nozzle I had flagged and tried to get warranty replacement for a few months ago. This was because I noticed 3 of 6 of my induction nozzles (right side) would not allow me to insert a 0.4mm nozzle cleaning needle through it due to severe metal on metal resistance inside the nozzle. So, it was I moved all 3 of those nozzles to 4-6 slots. The first prints with baby blue was using nozzle #5.
You can see the lines in GAMES and in the D20 especially. Immediately after printing the above, I then printed it again and manually selected R2.
As you can see the blue printed extremely well with a simple nozzle change. I bring this up because different nozzles may deliver different results and this needs to be accounted for or fixed. These nozzles apparently are “expendable” parts according to Bambu and can’t be replaced under warranty. Therefore, I’m still looking for a way to get rid of whatever metal burr or something is causing the issue with the cleaning needles and this under extrusion.
With larger designs, you will have much better results after changing the machine gcode alone. My design is small and has some inherent gaps i’m still fixing with settings like first layer extrusion ratio on just those svg modifiers, wall order, number of walls, line width, etc.
As mentioned above, I think there is still an issue with the bed leveling function due to my results with testing the z-offset and the difference between the front, middle and back of the print bed. I may try to manually tram the bed.
I’ll also probably open a support ticket regarding all my findings as well to get more eyes on this.
I suggest you try adjusting your machine start GCode as well and see the results for yourself.
***** USE CAUTION when changing gcode, ensure you know what you’re doing before trying.
Great work here. Thank you.
On left vs right z-hight: maybe a single value applies to both and the difference (obvious bed movement when switching extruders) is hard coded? I believe the “nozzle offset” calibration also incorporates z axis. Although those numbers aren’t visible.
Your results beg the question: if the fix is a small adjustment to start gcode, why hasn’t Bambu thought of this? Especially for the plate that comes with the printer! (Note, I am not questioning your conclusion. Just interested why BL has this configured incorrectly by default
)







