ffhein

Well, it’s just a guess. But also Gemma4 increased to 31B, so I think there’s a possibility that Qwen does the same.

27B dense was a good size for 3.8, but I’m guessing they’re going to add a few B parameters in order to make it noticeably better.

Do that, and I’ll check it if I can find anything strange in it.

Are you saying what we’re seeing in the photo is supposed to be the same layer? Can you upload the exact gcode file you were using when printing, if you haven’t solved it since posting this?

Cheap nozzles from Amazon etc. can also have the wrong diameter, or just be generally poorly machined. Perhaps not relevant for OP’s case, but could be useful to know when helping people debug printing issues.

Thinking that I could do my own research using Google and find a good printer (which I have always been able to do when purchasing electronics in the past), but not realizing how widespread fake reviews had become, and that most youtubers produce heavily biased reviews to stay on good terms with the manufacturers who are sending them free stuff.

If you look at the edge above the lifted corner, isn’t that supposed to be straight, or are the walls not meant to be flat?

My guess is that you need to print slower and/or increase nozzle temp. If the filament isn’t properly melted that also leaves tension in the material, which can cause corners to be pulled off the bed. I.e. cooling issues is not the only thing that can cause warping. What are your current slicer settings? I assume your high-flow nozzle isn’t PTFE-lined, so printing hotter should be safe.

Those kind of uneven layers can also be caused by a mechanical issue on the vertical axis, but I don’t think that would increase the risk of warping.

This one supposedly beats Jev at its own game:

https://huggingface.co/Cloudflare/clef (27B)

https://huggingface.co/Cloudflare/clef-flash (9B)

Requires a modded llama-server to run, but I find this project especially interesting. They have extracted the fine-tuning and decision making head from Clef (which is a modded Qwen3.8-27B), to make it possible to run both Clef and normal Qwen at the same time, while sharing the majority of the parameters.

https://huggingface.co/trevest/Qwen3.8-27B-Clef-LoRA-GGUF

If your printer is fully stock, replacing the bed springs either with silicone spacers or stiffer springs can be a good value upgrade. Keeps the bed from unleveling itself while printing, and IMO it also makes the initial adjustment a tiny bit easier. A bed probe is an immense quality of life improvement though, if you can afford to spend a little more money, since it both compensates for less-than-perfect bed leveling and also for any unevenness in the bed surface.

On the subject of upgrades, stay away from the “all-metal extruder” which used to be popular on Amazon etc. The stock Ender 3/v2 extruder is poorly designed and breaks easily. The metal version of it doesn’t break, but it’s still has the same bad design. If you’re going to invest money in a new extruder there are better options.

You mean with a closed enclosure? Maybe try printing with it open then, PLA generally want as much cooling as you can muster. I don’t have an enclosure myself, but I’ve heard that some people who do have to print with their doors open when using PLA.

Also does it print straight, and then later start bulging while its printing other layers on top? Or did it print the bulge that way?

update: I zoomed in closer on your other photo, and that choppy edge + horizontal lines are also indications that something is wrong. My first guess(es) would be either an extrusion issues (mis-calibrated extruder/slider, trying to print too fast, poorly machined nozzle…) something mechanical (frame isn’t properly assembled, loose roller wheel, hotend isn’t fully screwed to the carriage…).

Btw. I saw you wrote that you have a 0.2mm nozzle. Just to be sure, you did put that into your slicer, right? The default settings are probably 0.4mm nozzle. Also what line width and layer height are you using in the slicer settings?

Right now, around 50% RH indoors. I leave my spools of PLA on the printer for months or even years, and I’ve never had any humidity issues. If I were to speculate, my guess would be differences in additives being the main reason that only some people need to dry their PLA filament.

First of all, flash the firmware and reset eeprom on your found printers. Previous owners might have disabled safety features.

Common e3v2 problems: frame not assembled with square angles. Invisible crack in the extruder arm. Flat spot on wheel. Gantry sag. Bed unleveling itself over time. Bed not flat (not an issue if you’re getting a probe)

as much as I like this idea, I assume it would also magnify the punishment if you get caught.

It has indeed been really good, so much better than similarly sized models a few years ago. Would be interesting to know how far it is from what is possible to fit into a 27b model.

The main reason anyone would “need to” switch to llama.cpp is if they want to do partial offloading, i.e. split the model between GPU and CPU. This works quite well for MoE models, but you didn’t say anything about this, so I’m just wondering what your goals are.

Absolutely nothing wrong with switching to llama.cpp, I also use it, but that’s because I occasionally want to run models larger than my VRAM. It has official docker images and a server with both API access and a decent web-UI.

If you’re only going to run models which fully fit in VRAM, then tabbyAPI is also a good option. However, it uses Exl3 format instead of gguf, so llama.cpp probably makes more sense if you already have a lot of models in gguf format. tabby also comes with docker files and API support, so either should be quite easy to integrate with your setup.

Did you hot tighten the nozzle while putting it back? Don’t think it caused your current failure, but cold tightening can lead to more clogs in the future.

I think layer shifts can be caused by many different things, but the first thing I’d check is that the bed carriage can move freely and there isn’t something mechanically blocking it, like a wire or a misplaced screwdriver. The second thing I’d check would be the belt and gears, so that the bed can’t move without the stepper motor also spinning.

Price is comparable to a used RTX3090 with 24GB vram, which is probably more attractive to someone who is also interested in Linux/Windows gaming (and already owns a pc I mean). I would also guess that the RTX would be faster than the MacBook. IMO unified ram is more interesting when you can get a lot of it

2010 is ancient technology, according to wikipedia Nvidia released the 600 series in 2012… Even if there was some inference engine supporting it then lack of computational speed and memory bandwidth would probably make it not worth the effort.

I bought a used 3090 two years ago, and back then they were usually listed for €800-1000 in my country. I thought I was lucky to find one for €700 after searching for a few months, and I don’t think they’ve ever been cheaper than this here. There are definitely fewer of them available now, but you can still buy one for €950 (and possibly even lower if you’re patient). So prices have gone up, but IMO not by ridiculous amounts like RAM.

When shopping for a cheap 3d printer forget about brand loyalty, most of the common ones have released both a some decent and some garbage printers. That doesn’t mean they’re all the same though, some brands are worse than the rest. Since I started 3d printing I’ve been told to avoid Flashforge because they (allegedly) make printers that are designed to fail, and that they have altered common open designs to force you to buy their extremely overpriced spare parts instead of cheaper/better third party components.

Unfortunately I’ve been out of the 3d-printer-recommending loop for a few years but when I was more active Sovol had an above average track record. However, there is nothing saying that they won’t do the same thing as Creality, and drop any semblance of quality while bribing youtubers for good reviews once they are popular enough.