Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Why Prusa is floundering, and how you can avoid their fate

drewdevault.com

Post reply on HN
Print view View original

151–159 of 159 posts

Re: Why Prusa is floundering, and how you can avoid their fate

#151

Post by slau »

sebazzz wrote: post #141
slau wrote: post #35

I think Prusa has had a couple of missteps of late, but they’re not in the gutter yet. On the XL: the 0.6mm makes sense with the build volume, to significantly improve print times, but seems to really not work well with the 5 toolhead models, and introduces massive amounts of zits, stringing, cross-contamination, etc. On the Mk4: at the price point, it’s maybe missing a few features to really hold up against the X1C,…

There is nothing wrong with Marlin, especially not for the purposes Prusa uses it for. Marlins biggest weakness is needing to recompile for some things because they are defined by compile-time. All other things like active input shaping calibration can be added relatively easy.

There is _a lot_ wrong with Marlin. It is a complex codebase, full with ancient artefacts, litterred with preprocessor ifdefs every 2-3 lines of code, dynamic includes in the middle of CPP files, etc[1]. It's about as unreadable as C++ code gets--well, I guess it's not template metaprogramming.

Klipper by contrast is a breeze to read through[2].

I am very grateful for Marlin, for all of reprap, and everyone who has contributed to it. But saying there is nothing wrong with it is straight up misguided.

[1]: https://github.com/MarlinFirmware/Marlin/blob/bugfix-2.1.x/M...

[2]: https://github.com/Klipper3d/klipper/blob/master/src/lcd_hd4...

Re: Why Prusa is floundering, and how you can avoid their fate

#152

Post by agent462 »

commandar wrote: post #79
slau wrote: post #35

I think Prusa has had a couple of missteps of late, but they’re not in the gutter yet. On the XL: the 0.6mm makes sense with the build volume, to significantly improve print times, but seems to really not work well with the 5 toolhead models, and introduces massive amounts of zits, stringing, cross-contamination, etc. On the Mk4: at the price point, it’s maybe missing a few features to really hold up against the X1C,…

>On the XL: the 0.6mm makes sense with the build volume, to significantly improve print times, but seems to really not work well with the 5 toolhead models, and introduces massive amounts of zits, stringing, cross-contamination, etc. More fundamentally, the XL uses an open frame design specifically because Jo wanted to prioritize being able to easily see and remove prints. The entire gantry sits on a cantilevered ope…

Please explain to me why you think there is a fundamental engineering failure in the frame design - I'd like to see the data.

Just so you can get an idea of where I am coming from, I have a Prusa XL and a Voron 2.4 right next to each other.

The Prusa XL frame is very solid and it's also not a quad gantry. The lead screws take the weight of the bed. The Voron uses 20x20mm extrusions and the Prusa XL uses 30x30mm extrusions and it has stamped steel for additional rigidity and support. On the Voron, I've added titanium backers to reduce flex from thermal linear expansion. While the XL would technicall have thermal expansion, when enclosed, it will be much less from the much larger extrusions and additional support.

The toolheads on the XL on the other hand are heavy, they will likely be the larger issue with speed+quality. People also complain about the Stealthburner weight with TAP.

There are faster projects than the Voron just like there are faster printers than the XL. You have to weigh what you want. The toolchanger design, for me, works flawlessly and is so much better than the AMS, ERCF, MMS. More companies will clone it.

Re: Why Prusa is floundering, and how you can avoid their fate

#153

Post by commandar »

agent462 wrote: post #152
commandar wrote: post #79

Earlier quoted context omitted.

>On the XL: the 0.6mm makes sense with the build volume, to significantly improve print times, but seems to really not work well with the 5 toolhead models, and introduces massive amounts of zits, stringing, cross-contamination, etc. More fundamentally, the XL uses an open frame design specifically because Jo wanted to prioritize being able to easily see and remove prints. The entire gantry sits on a cantilevered ope…

Please explain to me why you think there is a fundamental engineering failure in the frame design - I'd like to see the data. Just so you can get an idea of where I am coming from, I have a Prusa XL and a Voron 2.4 right next to each other. The Prusa XL frame is very solid and it's also not a quad gantry. The lead screws take the weight of the bed. The Voron uses 20x20mm extrusions and the Prusa XL uses 30x30mm extru…

This is getting into first principles territory. The short version is that the gantry being on a cantilever and an open frame means it's going to act like a tuning fork under toolhead acceleration. Higher resonance limits toolhead acceleration before print quality nose dives.

I haven't seen any accelerometer graphs posted for an XL, but it doesn't take much more than opening up PrusaSlicer and finding maximum accels set to 3k to see the direct result of that. That's where the rest of the community was 4-5 years ago. There's a reason why Prusa has an entire blogpost making excuses for why their printers' performance numbers aren't up to snuff compared to modern designs (while also cherry picking comparative numbers)[1].

It's not my only complaint about the XL but it's the most obvious one. I haven't been impressed in general with the prints I've seen come off those machines.

[1] - https://blog.prusa3d.com/original-prusa-printers-now-printin...

Re: Why Prusa is floundering, and how you can avoid their fate

#154

Post by commandar »

slau wrote: post #136
commandar wrote: post #79

Earlier quoted context omitted.

>On the XL: the 0.6mm makes sense with the build volume, to significantly improve print times, but seems to really not work well with the 5 toolhead models, and introduces massive amounts of zits, stringing, cross-contamination, etc. More fundamentally, the XL uses an open frame design specifically because Jo wanted to prioritize being able to easily see and remove prints. The entire gantry sits on a cantilevered ope…

I didn’t know about this point. Which printer would be closed frame, so I can understand the difference? From what I’ve been able to observe, the quality of the XL is quite good for single tool head prints, however the quality really nosedives when multiple tool heads are introduced, and a bunch of people have been finding tricks to improve things (more wiping, less lifting, insane retractions, etc).

Pretty much anything that looks like a box. Most anything from Voron, Annex, Rat Rig, VzBot, etc.

Or if you're looking at an XL, add a frame member across the front and a vertical frame member on each side.

Re: Why Prusa is floundering, and how you can avoid their fate

#155

Post by agent462 »

commandar wrote: post #153
agent462 wrote: post #152

Earlier quoted context omitted.

Please explain to me why you think there is a fundamental engineering failure in the frame design - I'd like to see the data. Just so you can get an idea of where I am coming from, I have a Prusa XL and a Voron 2.4 right next to each other. The Prusa XL frame is very solid and it's also not a quad gantry. The lead screws take the weight of the bed. The Voron uses 20x20mm extrusions and the Prusa XL uses 30x30mm extru…

This is getting into first principles territory. The short version is that the gantry being on a cantilever and an open frame means it's going to act like a tuning fork under toolhead acceleration. Higher resonance limits toolhead acceleration before print quality nose dives. I haven't seen any accelerometer graphs posted for an XL, but it doesn't take much more than opening up PrusaSlicer and finding maximum accels…

So your theory is anecdotal.

You should also check out multi-color print times on the XL vs the ASM or any other solution.

I think what Bambu has done is incredible but these are different machines for different reasons. The XL has a massive print area compared to the Bambu printers and the toolhead design I prefer. I just gifted an x1-carbon to my mom for Christmas. I'm not a hater or Prusa shill but people are just spouting opinions without data or truly understanding the differences between the machines.

Re: Why Prusa is floundering, and how you can avoid their fate

#156

Post by commandar »

agent462 wrote: post #155
commandar wrote: post #153

Earlier quoted context omitted.

This is getting into first principles territory. The short version is that the gantry being on a cantilever and an open frame means it's going to act like a tuning fork under toolhead acceleration. Higher resonance limits toolhead acceleration before print quality nose dives. I haven't seen any accelerometer graphs posted for an XL, but it doesn't take much more than opening up PrusaSlicer and finding maximum accels…

So your theory is anecdotal. You should also check out multi-color print times on the XL vs the ASM or any other solution. I think what Bambu has done is incredible but these are different machines for different reasons. The XL has a massive print area compared to the Bambu printers and the toolhead design I prefer. I just gifted an x1-carbon to my mom for Christmas. I'm not a hater or Prusa shill but people are just…

>So your theory is anecdotal.

It's based on basic engineering principles. Prusa can't magic away physics and these are lessons the rest of the community learned years ago. E3D had to course correct on some of the same mistakes with their toolchanger. There are good reasons why nearly every other CoreXY design you're going to encounter uses a closed frame (and the best will use structural panels).

>I think what Bambu has done is incredible but these are different machines for different reasons.

Sure, it's got its niche. But the drawbacks will narrow that niche. Prusa's ever narrowing niche is kind of the focus of the OP.

> The XL has a massive print area

Which makes running slow compared to other designs more painful since print times are cubic with volume.

>just spouting opinions without data or truly understanding the differences between the machines.

If you showed me a square wheel, I can tell you it's not going to roll particularly smoothly without needing to perform extensive testing on that specific implementation of square wheel. I can do that because it's something that's been tried elsewhere and we already know the results.

This style of open frame has been tried before. We know Prusa didn't choose to use it because there's any sort of engineering advantage because Jo has explicitly told us why they're using it.

Re: Why Prusa is floundering, and how you can avoid their fate

#157

Post by holoduke »

riddley wrote: post #44
drewdevault wrote: post #37

Earlier quoted context omitted.

Businesses are more complex than just one facet. Not everything the business offers has to be a direct revenue driver; it builds trust with the brand and improves the experience by putting everything you need with respect to your printer in the same spot. Which of the following experiences delivers better value to the customer: 1. They buy a printer and can immediately slice and print stuff on the vendor's web app 2.…

Again I have to ask if you've ever used a Prusa, no really this time? Prusa users use Prusaslicer, not Cura. Printables.com often hosts gcode files that require no slicing at all by end users. These points completely aside -- do you think 3d printer novices are going to last long without learning a slicer? Is that really your assertion? Also, is installing Cura a "nightmare" ? So much explosive language!

I use both slic3r and cura. Both are installed in 2 minutes. Slicer is good for having better out of the box profiles. But the biggest drawback for me is the auto orientation tool to lay an object flat. For thin complex objects slicer always has issues. Cura does not.

Re: Why Prusa is floundering, and how you can avoid their fate

#158

Post by TheCleric »

yetanother12345 wrote: post #146
TheCleric wrote: post #83

Earlier quoted context omitted.

Agreed. I thought they had some good points up until this and immediately thought “I would RUN from a company that did this.” Also do people REALLY pay for models, let alone $10? I think I can count on one hand the number of times I’ve paid for a model to 3d print. Much easier to find an existing one, hack one, or just make my own.

I can confirm that there is indeed a market for (unprinted) models, just like there is a market for print-to-order parts. The market works such that you pay for the right to print the model even if the model as such has one license or another. Check markets like Fantasy, role-playing, games, if in doubt

That’s where I do most of my printing. I guess I’m usually okay with using the free prints of “generic goblin” rather than trying to get the absolutely best one I can. :)

Re: Why Prusa is floundering, and how you can avoid their fate

#159

Post by neither_color »

rjbwork wrote: post #12
TimMeade wrote: post #10

I think this is missing a little context. I believe Prusa and to some extent the older market leader Creality were sideswiped by a better product. A group of engineers got together and formed Bambu Labs. Their printer are all less than 2 years on market and they have become the new standard for the simple fact they just work. 3d Printing historically has been "early adopter". I'm long term tech. It took me a month of…

Yeah I got into 3d printing a few years ago to start a side business for a niche hobby, got a nicer Creality CR-10S, I think it was. Endless hours tinkering with the thing, and still could only get nice solid prints like half the time. I just don't have the patience to deal w/ stuff like that anymore as I've gotten older. I just want something that works OOTB. I think in the next few years if I get back into it I'll…

I print casually on my cr-10 and I feel the same way about not wanting to invest time in it. I made a little checklist for myself for when I re-visit it every few months I can get right back into it: open filament, heat the nozzle & bed, level the bed, tram it, etc. I'm such a casual I didn't know Bambu existed until this thread. I'd love to just have something that works. Keeping up with software stacks is enough for my tired brain.
Post reply on HN