[Solved] Help with fixing artefacts on prints.
I'm tuning my B1QU-B1.I tried printing a part (please see image below) after tuning my first layer, flow rate, temperature tower etc. I'm however, getting the following artefacts on the print and i don't know what could be the cause of that.
(The red is the one I printed first and the green one after tuning in a few of my settings.The green one had significant improvements after the tuning compared to the red but i still can't get rid of the horizontal lines as seen in the image.)
From what I found it could be the Z wobble issue (i'm not sure what that is though).I found a model online to test it and here are my results;
The left one (Skin colour) is before the tuning and the right one (Silver) is after the tuning.Significant changes again, but , still couldn't get rid of the horizontal lines.
Any suggestions with fixing the same?
PS : I put a rubber mat below my printer to reduce vibrations because the glass table top on which my printer is kept would vibrate a lot while printing.
Unfortunately when the z-axis leadscrew develops imperfections there isn't much you can do to fix it other than replacing the leadscrew. For what it's worth, your print on the right (silver) looks quite decent aside from the two lines about a finger-width up from the bottom. You're not going to get significantly better results unless you buy a precision leadscrew that costs more than the entire average Chinese printer 😆.
So, to make your existing setup more resilient against wobbles, you should make sure the screws holding your brass nut to the gantry are a little loose so the brass nut is able to move a little radially to compensate for the wobbling of the leadscrew. Also, make sure that where the leadscrew connects to the coupler, you didn't have the leadscrew touching the motor shaft before you tightened the coupler - there needs to be a bit of a gap between the leadscrew and the motor shaft so the coupler can do its job (which is to allow a little angular misalignment between the motor shaft and the leadscrew).
It also helps sometimes to wipe off all the old grease from the leadscrew and re-apply fresh grease. The areas with lines might coincide with areas where there is extra grease or dirt in the leadscrew.
One thing you could try (as a workaround/last resort) is taking the leadscrew out completely and turning it upside-down and re-inserting it (this is possible because the leadscrew is symmetrical). That way, the defects that are currently on the bottom of the leadscrew (where you currently use for most prints) move to the top of the leadscrew (which you only use for very tall prints). Based on your column it seems the top of your leadscrew has less defects than the bottom, though you'll have to try it and see whether this results in a real improvement.
One last thing to do is to use layer heights that are multiples of your printer's "magic number" (i.e. where each layer height change is an integer number of steps so you don't get rounding inaccuracies when the firmware converts layer heights to microsteps and so the motor shaft position does not suffer angular inaccuracies between one layer and the next because of the inherent imprecision of microstepping. If you want to learn more here's a very good article on the subject of how one microstep does not necessarily move as far as another microstep to demonstrate the benefits of moving in full steps when possible - https://hackaday.com/2016/08/29/how-accurate-is-microstepping-really/ ).
For many machines the magic number is 0.04mm so the best layer heights to use are 0.12, 0.16, 0.20 and 0.24mm.
If you want to double-check yours, the formula is (1/steps-per-mm) x microstepping.
So for example, for my creality CR-10 mini with 400 steps/mm on the Z-axis and 16x microstepping, my magic number is:
(1/400) x 16
= (0.0025) x 16
(you can get your steps/mm by sending the printer an M503 and looking for the line starting with M92 Z--.-- )
Hope this helps! 🙂
I experimented some more and was able to reduce the artefacts a little.I loosened the screws like you said and used a rubber band to apply a bit of tension on the lead screw so that it would shake a little less while rotating. Here is my result;
I haven't tried other things you've said yet though.
I also found out I'm having trouble printing round parts. 😆
Could this be because of the same issue? or is it due to something else?
Also, could you please suggest me good suppliers/brands for such parts? (lead screws, linear rails, aluminium extrusions, heated beds, stepper motors etc) In future I'd rather buy good quality parts and actually print than tune my printer for days or months by buying cheap ones LOL.
PS : Please let me know whether the rubber band thingy I tried was a good approach toward fixing the issue or a bad one.I felt like doing it and simply gave it a shot 🤣.I hope it won't affect other things while printing.
Thanks and a very happy new year to you and the team!
For the z-axis top, it's an ingenious idea 😆 but I think it might contribute some of its own artefacts when the rubber band moves up and down from the leadscrew's action. It's odd that your printer leaves the top of the leadscrew unsupported like that. On my CR-10 there's a plastic block with a ball bearing inside. The bearing slides over the top of the leadscrew and has about 1mm of play on its outside (so the leadscrew can wobble a bit) and the block is attached to the frame with T-nuts.
You can buy something similar, like so - https://www.amazon.ca/Printer-Upgrade-Aluminum-Leadscrew-Creality/dp/B0822796N1 - this might even fit your printer, assuming yours uses 2020 extrusions and the leadscrew is 8mm in diameter. Alternatively you could design and print one (if you can source your own bearing, bolts and T-nuts), or try and find ones on thingiverse that match the dimensions - for example:
- this one doesn't even use a bearing: https://www.thingiverse.com/thing:3534568
- this one mounts under the frame: https://www.thingiverse.com/thing:4004121
- this one doesn't use bolts: https://www.thingiverse.com/thing:3985879
- this one is conventional: https://www.thingiverse.com/thing:3477209
- this one uses no bearing and no screws (easiest one for you to print and test): https://www.thingiverse.com/thing:3358978
With regards to the issue - I see a drastic improvement in quality already in the first picture (side-by-side comparison). The bolt picture shows overheating or excessive overhang angle on the lower surfaces - you can try and tackle that by lowering the temperature, increasing the cooling, lowering the print speed, lowering the layer height or switching to a higher-temperature filament like PETG.
The cylinder is showing artefacts at the start of each layer. Those are quite tricky to diagnose and are usually caused by ooze control settings - coast, wipe, extra restart distance on unretraction etc. The best thing for that is to upload your profile here so I can analyse it and see if anything stands out, but generally it's safest to turn off all those features or set "0" for the coast, wipe and extra restart distances and see if it improves.
Good suppliers/brands - I don't know of any. The brandnames basically spring up and disappear the next month. I suspect they're all just made in the same handful of factories, then get rebranded and sold online by a great many distributors. I've bought most of my parts from banggood and https://www.3dprintersonlinestore.com/ . I've also used Amazon and eBay, and these days AliExpress too. Unfortunately there's really no way to keep track - many vendors purchase fake reviews to make their products seem more desirable, and companies have been known to switch suppliers mid-production, so the good genuine reviews you might be reading refer to an earlier edition of the product before they switched and quality went downhill. My personal strategy is to buy cheap but temper my expectations and be ready to have to throw things away and buy a replacement. So far it hasn't come to that, all the parts I bought have been functional. Ironically my printer itself arrived broken (the extruder lever was practically broken off and the screw was bent into a banana shape right out of the box).
If you're willing to pay more to get guaranteed good quality, probably the easiest way to do that is to jump onto the Prusa ecosystem and only buy parts from the Prusa website. Prices have actually come down in recent years and a Prusa Mini isn't that much more expensive than an Ender 3 these days.
E3D ( https://e3d-online.com/collections/printer-parts ) also have genuine, good quality aftermarket parts with very solid engineering behind the things they invent themselves. I also hear good things about TH3D but have no personal experience with either of these.
If you want really, really high quality leadscrews you'd have to go to an OEM like Steinmeyer, but I wasn't able to find pricing and availability - I doubt they sell directly to individuals :/ (and at any rate a high-precision ground leadscrew might cost over $200 - you could buy an entire printer for that!).
Hope this helps, looking forward to your feedback 🙂
HAHA thanks and yes the rubber band does have artefacts of its own. I've noticed a slight bump on the front side of the same part not very significant but noticeable.Could possibly show more if I print something else.
Primary issue - I'll try printing one of the parts from the links you've provided.Hope that works well 😀.I'll try other things you've suggested in your first reply too.
Switching to Prusa Ecosystem : I've had a really great experience with my Prusa so far but, importing it in my country pretty much doubles it's cost with import duty and taxes and shipping </3 😫.That's why I bought a B1 and Prusa Mini clone kit from China (which also is yet to be tuned 😭).They simply were cheap and I also thought buying base level printers would help me understand various other aspects/difficulties of 3D printing better. Although I've made decent progress, I still haven't got both of them running properly.I've learnt a lot to be fair though 🤗.However, from here on I'd rather pay more and buy 1 good printer than buy 2 or even 3 cheaper ones for the same price. Saves a lot of time! Pay once cry once kinda thing 😀
Cylinder Issue : Here is my config file for you to analyse. I use Prusa slicer as my primary slicer and I've tweaked the default ender 3 profile in it for my BIQU-B1.
Thanks and Cheers,
I know that feeling! - I live on an island and some things are impossible to get here 😣 . I was trying to get some large LiFePO4 batteries to use for emergency power backup but every logistics company declined because they're considered dangerous goods (as far as shipping goes - they're actually safer than traditional lithium-cobalt batteries but they lumped all lithium batteries under one category). After a few months of trying I gave up and bought a regular lead-acid battery locally (incidentally, a UPS always beats the "recovery" feature of the printer firmware 😉. If you want to protect your prints from power cuts, buy the cheapest UPS in the computer shop and connect it in line with your printer!).
So, to your profile, I have the following suggestions:
- Layers and perimeters - Top solid layers - reduce to 4; bottom reduce to 3 (this will save some time and material and shouldn't affect quality too much)
- Infill density - reduce to 15% (saves time and material, won't affect strength too much)
- Speeds - make them like this:
This significantly speeds up the printing of top layers, saving tons of time. The quality impact isn't so great. For "small perimeters" feel free to keep your current setting of 25mm/s, it'll save some time and might actually improve quality in your case as the speed changes less when making walls.
- Extrusion width - make them all "0" - this way it will automatically compensate for different nozzle sizes (should you decide to change it) - you easily define installed nozzle size under Printer Settings -> Extruder 1 -> Nozzle diameter. No need to come back to the extrusion width page and update ;). Alternatively, for optimum quality, set all widths to 0.4mm for an 0.4mm nozzle. That's the finest you can go with a 0.4mm nozzle and will print finer details better than the default slightly thicker extrusion width (1.125x nozzle diemater = 0.45mm default)
- Filament - I see your extrusion multiplier is exactly 1.0. Have you calibrated your e-steps and done an extrusion multiplier test? If not then start by calibrating the e-steps, then print a 20mm calibration cube and look at the top layer to see if your extrusion multiplier is spot-on. Here's how to tell:
With underextrusion you'll have gaps between the diagonal lines. With overextrusion you'll have a bumpy, raised surfaces around the edges where the top diagonal lines meet the perimeters.
- Cooling - untick "keep fan always on" and set disable fan for the first 3 layers. You want to have a few layers with fan off for safety (less risk of prints lifting from the airflow with a bigger buffer of fan-free layers), and you want the printer to be able to
- Max volumetric speed - set to 0 to disable this feature. Not sure if this could be it but it might be introducing some artefacts when you hit the flowrate limit.
- Printer tab - are you sure your printer requires relative E distances? Unticking this might result in better prints (worth trying, just watch the extruder for the first few lines to see that it behaves as expected)
- Extruder 1 tab - lift z - set to 0.2mm. Firstly this is a "magic number" (integer number of full steps for most 3D printer motors) which means higher positional accuracy, and secondly it's less than the 0.5mm you currently have so it will speed up your printing and put less wear on your leadscrew 😉
- Retract on layer change - untick - you don't need it
- Wipe while retracting - this is probably the cause of your artefacts. You're seeing the result of the nozzle smearing filament in the wall with every wiping move. Untick this too 😀
- Retraction speed - try 90mm/s if your axis speeds allow it. You get faster prints and possibly less oozing. Change the number if your extruder makes any strange noises like skipping steps or resonating (my CR-10 mini resonates at 70mm/s but gets quiet at higher speeds).
Hope that helps - of all that the wiping is the one most likely to resolve the problem, but the rest should help improve quality a bit (or at least make it faster and minimize material usage a little). Let me know how it goes!
Unticking "Wipe while retracting" solved the problem. THANK YOU!!!
I reduced the flow rate and the results are as follows,
I noticed another issue though.I'm not sure but, I think its because I'm printing at a high temperature. Please let me know your thoughts on it.
Here's a benchy I printed after making a few changes you asked me to before
I also printed this ( https://www.thingiverse.com/thing:3358978 ) like you mentioned before and it improved the quality without creating other artefacts 😀
I still haven't understood one thing yet.How does relative E distance affect print quality? Aren't Absolute and Relative E distance supposed to do the same thing, just differently? 🤔 I've always been confused about purpose of the two.Also, from I've read so far, relative positioning has been favoured over absolute.Or am I missing something?
Really happy to hear that disabling wiping solved the problem! 😀
As such, relative and absolute should work identically. However with purely absolute extruder coordinates you start to end up with crazy large numbers. A 1kg spool has about 330 meters of filament (assuming standard 1.75mm diameter), so a 1kg print would end up with commands of the form G1 E330000 by the end of it.
Also, many printers implement a firmware feature that protects from huge extrusion commands (entered by mistake manually or because of some mishap with the slicer settings). In Marlin, you set it in configuration.h via the lines
#define PREVENT_LENGTHY_EXTRUDE #define EXTRUDE_MAXLENGTH 200
It limits the maximum extrusion distance, so for example if you set the to 200mm as above, the printer will ignore extrusion commands that tell it to extrude more than 200mm. For example Sending G1 E205 will have no effect.
For these reasons, a slicer like Simplify3D will not operate an extruder in truly "absolute mode". It will work in absolute mode up to a point, then send a G92 E0 command, which rests the current extruder position back to 0mm. Then it starts sending absolute extruder positions until the next reset. Here's an example:
The logic varies (Simplify3D seems to reset the extruder either after a line or when starting a new type of feature, like infill, inner/outer perimeter or support). As you can see, you only rack up a few dozen milimeters of absolute extrusion before resetting (in this case 47.3mm). So even absolute extrusion is a hybrid of absolute and relative extrusion.
In relative extrusion mode there are no reset commands and every E distance gives you a clear reading of the quantity of filament extruded in that move (since you only need to read that value directly - whereas with absolute extrusion you have to calculate at the difference between the current line and the last to see the extrusion volume). This means that relative extrusion should give slightly smaller .gcode files, almost imperceptibly faster printing (less commands being processed) and easier human-readability.
The real advantage of relative extrusion is that rounding errors do not accumulate - a good explanation of the topic is here: https://www.sublimelayers.com/2017/10/to-extruder-relative-or-not-to-extrude.html
However most firmwares will assume absolute extrusion is the case unless specifically told to use relative extrusion with M83. So it's possible that in some cases unexpected behaviour happens (say resuming from a power cut or restarting printing after a filament or colour change, and you forget to send M83), and some firmwares don't even support relative extrusion - so absolute is the default for most slicers and you have to manually change it if you want to use relative.
The top of your cube looks best with an EM of 0.96. In fact, I think there's still a tiny amount of overextrusion and it might possibly look better with 0.95 or 0.94 - try it and see 🙂
Lastly, the artefacts you see at the top of the X are due to overheating. Either lower the print temperature, or bump up the fan speed, or increase the minimum layer print time by a few seconds so the printer slows down more when printing fast layers to give the filament time to cool before printing another layer on top of it. You'll find this last setting under Filament Settings -> Cooling -> Slow down if layer print time is below...
Great job with improving print quality so far! Try changing these and report back if the quality improves further! 🙂
Regarding the artefacts on the top of X - I had already reduced the printing temperature from 210ºC to 200º for that.This time I reduced the temperature to 190ºC and printed with same slicer settings as before.I also increased the minimum layer print time by 5 seconds like you advised.Here are the results
One thing I've noticed is that the Y part turned out good even at 200ºC with old settings.Could it be some other problem? With the belt or something else maybe? Shouldn't the Y part also have the same problem if it's a temperature issue? I'm a bit confused.
I printed the 20mm cubes with reduced flow rates too.Here are the results
On printing some more I've noticed other artefacts 😆 </3
I changed the flow rate for the above prints to 0.95, all other settings were unchanged. In the fitment test I'm able to move 4 pins.3 without any issue the fourth one I had to use tweezers to break a little bit of plastic stuck inside and free the pin.The fifth one is stuck, couldn't break it loose even with tweezers or pliers 😀.
PS : The first layer didn't stick well due to the bed not being clean on one of the test cubes.
Regarding the X and Y, that makes sense because the part cooling fan is probably blowing in only one direction, so part of the print naturally gets more cooling than the other. Speaking of which, what percentage is your part cooling fan running? For PLA it should be 0% for the first 3 layers (to help bed adhesion) then 100% from there onwards (or you can ramp it up gradually after layer 3, reaching 100% by layer 5).
Another thing, have you calibrated your e-steps? Can you verify again by physically extruding 100mm of filament and checking the distance it actually extrudes? Your cube with 0.93 is still showing quite a bit of over-extrusion and it's not common to need such a low extrusion multiplier to print (normally EM hovers around 0.96-0.97 which is why I suspect an e-steps issue might be involved).
The over-extrusion might explain many of the artefacts you're seeing.
This one is normal:
It's the layer seam (z-seam) and is inevitable when the extruder starts a new layer. You can move it around by using particular settings in the slicer (options include "random" so you get those little pock-marks all over the print, "nearest" or "optimized" which is what you have now, which puts the seam at the nearest position to the last layer end position for highest print speed and least motion, and "aligned" which puts them all in a line so they can be hidden in a sharp corner, or sanded off when the print is done).
As for the gap in the "3" of "io3dp", try toggling "detect thin walls" and see if that gets filled in.
Looking forward to your feedback 🙂
Before all this, I made sure to calculate the e-steps with the help of IO3DP e-steps calculator. The default value my printer came with was 96.00 and I increased the E-steps to 98.03 later.Here is a photo after extruding 100mm of filament. (The red mark being marked at 120mm initially)
Please note that I've have taken the small unmarked gap after the 30cm marking into consideration while calculating the e-steps.
Regarding the fan speed, at the time of printing the cubes, I had the fan speed set to 100% after 1st layer.I'll make sure to change that.
Please let me know if I'm missing something and whether or not I should reduce the flow rate even further.
Thanks as always.
Your e-steps are spot-on 🙂
Your extrusion multiplier needed lowering below 0.93 according to the cubes, so could you print a single calibration cube at an extrusion multiplier of 0.90 and post a photo of the top surface (where the letter Z is)?
If you want to save some time and material, here's a truncated cube with just the top (where the Z is) that I made for such testing 🙂
Here are a few photos of the same.
Do let me know if i need to make any more changes.
It's close, but I'd go with 0.92 to be safe. Below that you can see a defect start to open up at one of the interfaces between the top layer and the perimeter -
If print quality starts to show signs of over-extrusion (especially if you change filament down the line) you could bump it down to 0.91 or 0.90 depending on what the prints start looking like, but for now I'm going to suggest the conservative option of keeping the lowest extrusion multiplier before the defects started getting more visible (0.92) because that way the printer should still comfortably give acceptable print quality even if it encounters a length of filament whose diameter is on the upper or lower limit of the specification 🙂
I'll keep it at 0.92 then. If I come across any issues in future I'll let you know. Thanks a lot!! you've been of great help like always! 😊
We’re a bunch of 3D printing and design nuts. We just want to make great 3D design available to everyone.
Good design takes time, great design takes a process. You can learn this with us, to build your skill set in this rapidly expanding market.