3000+ LoC changes to implement a 2nd extruder into Marlin, that natively supports 6+ extruders.
I also am disappointed at how difficult it is to implement new features in this firmware. It’s so custom and so intimately tied to the closed-source touchscreen that it’s effectively impossible - a mere mortal like myself cannot possibly dream of busting out 3000 changes to add a second extruder, let alone any other feature of similar, relatively minimal complexity.
It’s my fault for dragging @Atom back into this.
Good to hear from you again Adam!
My point was to point out the quality people that have been lost and alienated by SM’s continued attitude and business practices.
Adam was around for a long list of non-existent or at best delayed support, quality control and build problems, lack of communication, shipping delays etc. etc. etc. Every firmware update introduced just as many bugs as it fixed. At that time the only help and info available was from other users on this forum (and FB). There was little to no info on their website and what there was was either out of date or wrong. So I don’t blame him for choosing to move on.
At least now there’s a lot more info on the website and not all of it is out of date.
Makes me wonder: Would it be possible to disentangle the Controller firmware form the touchscreen, create a new fork from Marlin adapted for the SM2 hardware and then use OctoPrint as interactive interface? Basically just unplug the touchscreen and ignore it?
Actually it’s the very third beta dual extruder commit (The other PRs got closed.). But the fact that they are working for almost a year now on adding this feature (and the code quality is still questionable) pretty much supports your point.
For me,
it is a great pity for anyone who turns their back on the forum. So much is lost to both sides
And yet I can understand every reason for it.
I am a novice in the use of the three possibilities of SM2.
Ok, already a little more than back then in debian Slink and OpenBSD 3.1.
I am a newbie here and know UNIX there.
Today I control my Mac via bash, among other things. I got the A350 the day before yesterday and haven’t even set it up yet because of the lack of a base. And I really don’t have a clue about anything else either.
But I’m more excited than I’ve been in a long time for Christmas.
What I want to say.
I am soaking up a lot of your knowledge at the moment, even if sometimes (no actually - still often) I only understand the rudiments. Part of that non-understanding at this point, of course, is your frustration.
Simply because I am not yet ready. So far away from you guys.
Perhaps one or the other is not only lacking progress but also a little more recognition and respect.
Both from the company and sometimes here in the forum. Unfortunately, since Edwin’s departure, the forum no longer seems to be visited by the SM team. No feedback from staff anywhere for good or bad example. Or a thank you and an attempt to adopt some great approaches from here.
A great pity indeed.
I wonder here?
Newbie = Enthusiast
Senior = Disillusioned
If that’s the case: then I’ll buy you a round of Guinnes and give you back some of your initial enthusiasm
Would not say so. I’m one of the Kickstarter backers and work since quite a while now with the machine. It is certainly not perfect and has some things where the design is questionnable. Also, the connection to the Snapmaker Team feels not very tight and their pledge to be open to the community, also on the software side, looks a bit halfhearted. But nevertheless the machine delivers, and I have a lot of joy with it, and lots of ideas with not enough time to make them happen!
I think a good part of the frustration you read is that Snapmaker Team gives an air to deliver a perfect Pro machine, and the price point reflects this to some extent, but this claim is not justified. If they’d be a bit more humble and actively working on the specific concerns of people, the community might be more forgiving. Classic mismatch of expctations.
For me it would be:
Newbie: You got to be realistic
Senior: You got to be realistic
No not yet, I did some wood carvings, then I got the laser and have been playing with that. It’s amazing how much smoother and more capable the new laser is. Things like Christmas ornaments that used to take 20+minutes and 6 passes now get done in 3 minutes with 1 pass. I will get to milling aluminum and steel when my need for that arises. But right now me and my wife get more use out of our new machines then we did the SM2. but like most machines it took some work to get it there.
It’s always good to here from passionate people like you. I may not be around much but if you need me you can always tag me and I will take a look.
This is so true, witch is why I tried not to turn my back… I just have my headphones on so you’ll have to get my attention first If you wanna talk haha.
Also I think it is more like
Newbie: doesn’t know what they want
Senior: knows what they want
So as the newbie transitions to senior and they start to realize what they really want out of their equipment. And they find that the SM2 just isn’t it. Though some do find it is what they want (like @Hauke). To me it’s about finding the right machine for you.
Possible, certainly. With software it’s always the case that Everything is Possible.
Would it be the best way to go? No. The most important technical fact behind all this is that Marlin itself isn’t a particularly good code base for SM. It doesn’t really have a scheduler proper underneath it all and its lack bleeds through in lots of systematic ways. Marlin is really designed for direct control of a few axes. It’s not particular well suited to CAN bus, which has to be grafted on and then made to work with direct control. As the number of simultaneous things that need tracking increase, Marlin becomes increasingly unsuitable.
Taking code from Marlin might be desirable, but its code isn’t designed with an eye to reuse in mind, so it might not be worth it. Taking inspiration from Marlin, at least in places, would certainly be useful.
I will never turn down a pint of Guinness, as long as it’s a proper pour!
I’m still enthusiastic but also realistic. I see a lot of potential in the SM and it’s extremely capable as long as you know it’s limits. It’s mostly hampered by the Luban software and bad firmware updates. With Cura, Lightburn and Fusion 360 it’s a lot more useable, but the learning curve increases exponentially.
So I stay on this forum to hopefully help others figure out how to get the most out of their machines and also to learn from others. I still hold out hope (perhaps naively) that SM will at some point show up on this forum (and FB) and actually respond and contribute and listen.
Any negative comments I make are there in order to keep pressure on SM. Hopefully they’re fair and honest.
I’m not picky I’ll take yours if it’s not a proper pour. Free beer → no complaints
My progression has closely followed Ken Blanchard’s development levels:
1: Enthusiastic Beginner: Excited at the infinite project possibilities
2: Disillusioned learner: After a few failures the reality this is not a magic machine is setting in. Can’t just throw any model at the machine and expect greatness.
3: Capable but Cautious: After a few months of picking the brains of more experienced makers I’m realizing the machine is capable, if limited, and with reasonable expectations and patience good results are possible.
4: Self-Reliant Achiever: At this point I enjoy helping others work through issues, and it’s increasingly frustrating when the answer turns out to be “that’s a bug in the software that’s existed for months and months”. Being unable to provide feedback to Snapmaker on the forums, as when Edwin participated, feels like a step backwards.
100% agree. It’s possible to both enjoy using the machine while also being critical of its flaws and expressing a desire for improvement.
It’s easy to get to 3k LOC when you do fun stuff like this:
+ float xprobe_offset_from_extruder = X_PROBE_OFFSET_FROM_EXTRUDER;
+ float yprobe_offset_from_extruder = Y_PROBE_OFFSET_FROM_EXTRUDER;
- if (!position_is_reachable_by_probe(rx, ry)) { return NAN;} // The given position is in terms of the probe
- nx -= (X_PROBE_OFFSET_FROM_EXTRUDER); // Get the nozzle position
- ny -= (Y_PROBE_OFFSET_FROM_EXTRUDER);
+ // if (!position_is_reachable_by_probe(rx, ry)) { return NAN;} // The given position is in terms of the probe
+ nx -= (xprobe_offset_from_extruder); // Get the nozzle position
+ ny -= (yprobe_offset_from_extruder);
My main issue with delving into maybe fixing the firmware is that I’m sure it’s just chock full of “wtf” goodies like this.
I bought my A350 nearly one year ago. It was with the old platform and after complainig about my leveling and printing problems I got the new one that sligtly improved my leveling issues but printing big things with good results on the whole platform was impossible. I tried everything from all types of autoleveling to manual hot leveling and even manualy adjusting the leveling matrix. I always had areas where the first layer was up in the air and some cm away the nozle was scratching the bed.
In my opinion its simply the wrong typ of bed for such a big build space. I finaly got my level isues fixed by buying a 4mm silicat glass. Printing big PLA and PETG objects is no problem now. (I never tried ABS at all.) Autoleveling is history but manual leveling is no problem for me. And I never had to bad stiking issues even when printing on the non frosted side (it has a frosted and an non frosted side).
Snapmaker should realy consider designing a good, maybe metalized (for auto leveling) glass bed for the A350.
I have been expecting them to do this for some time, due to all the support time eaten up by the stock bed. It’s a straightforward solution with readily-available parts, and all the SM engineering dept has to do is devise a mounting method.
But hey, let’s have an air purifier!
Gave up, bought a Creality glass bed, have not yet taken it out of the package.
To be fair the air purifier has allowed me to use the laser function of my A150 without the burning smell of paper, wood, leather (horrid smell), or plastic in my apartment. I live in a small 800 square foot apartment so space is at a premium (which is why the 3 in 1 functionality is important to me). I also have no garage or workshop to work out of so I have to make sure not to stink up my own apartment or make myself sick from the fumes. In that regard it’s been extremely helpful and worth every penny (bought it for the $399 sale price)
Looking back at the Snapmaker Original and seeing how the 2.0 models turned out it seems that the A150 (and by extension the A250) simply perform better and have much less hardware/mechanical problems than the A350s. I think the engineers at Snapmaker overestimated the scalability of their current design. I think once you go past the A250 in size there needs to be some design or internal part changes to accommodate for bigger tolerances, added weight, and additional forces. It’s probably why the XL CNC and H-laser design hasn’t yet materialized beyond a prototype. They probably need to rethink some things when scailing bigger.
Based on my own personal experience the A150 has performed almost flawlessly with me. I haven’t had bed leveling issues or a single failed print. The print quality easily rivals my friend’s Prusa Mini, and the laser works well for my purposes. The CNC module is pretty much a toy but it works really well for engraving my PCBs and the tops of lids from little wooden storage boxes. If I need access to a real CNC my friend has a big 4x4 foot monster with a 4.0kw spindle, but for little engraving projects the Snapmaker CNC module is fine.
My only real complaint thus far would be the software, I have years of experience with Aspire/Vcarve Pro, Fusion 360, Mach 3 and other programs so Luban definitely feels very bare bones (and a little finicky) but for 95% of my work it’s adequate and gets the job done.
The A150 is the most similar of the three models to the Snapmaker Original, so I think those of us who bought the little guy got the most proven design—not just in terms of the bed, but also things like filament not getting caught on the linear modules.
The quality of the software (Luban) and the firmware is also definitly not reflecting the high price of the machine.
Just a few problems I run into in the last weeks:
I did some laser jobs:
I run into the background capturing problem (support statement: already known to snapmaker team) that costed me some houres to figure out and find a workaround
I tried different laser powers and run into a problem that the gcode is sometimes faulty by not setting the laser power at all. Then it simply runs with the last set value and I was wondering why I got the same results.
Then I switched to 3D-Printing and as experienced already last year:
When you trie tweeking the Z-Offset from the touchpad before starting the job in nearly 99% the value is not saved correclty. It works well if you change it after pressing start during the heating phase or anytime later.
These are some realy annoing problems I would not expect from such an expensive device.
But now some good words about Luban.
The settings and the generated gcode is realy good tuned to the device. Its not to flexible with the settings (I realy miss manual placement/adjustment of support structures). But generell printing results are realy good.
I also tried simplify 3D. First it took me much efford finding settings that bring nearly as good result as Luban and second is that it has to few settings or possibilities for setting different speeds, wides and extrusion amounts for different structure types (first layer, walls, infill). You have to tweak options that you would never expect to have effect… thats a different story.
So I think Luban is doing a good job in generating the 3d gcode!
That’s because it’s uses the nearly default Cura slicer under the hood. If you like Luban but find it limiting then just use Cura for finer grained control.
Those are great! The micropoors they advertise really do their thing. I hope you can find a good way of mounting it.
Also @digitalfreak I belive there is already a mod to change the SM2’s prox sensor into a physical probe (like a BLTouch). That would work with the glass beds.
