Hacker News Reader

Comment by carterschonwald | Hacker News Reader

The newest printers out of prusa, the recently launched mk4 and the much delayed prusa xl are absolutely half cocked. They have interesting design ideas, but are shipping without features that we now know are absolutely table stakes for quality machines.

Concretely: the mk4 talks about having input shaping, but doesn’t come with an accelerometer, which is needed for that to work well. The sovol 06 is a much more affordable machine that has similar motion system to the ml3/mk4

The xl really should be a fully enclosed printer like a voron trident or bambu x1 carbon. (The latter two have similar ).

If you ignore all the alternatives to prusa machines that are out there, they’re reliable machines that do the thing. But there’s so much amazing stuff in both community driven open source printer designs and also much more affordable manufacturing of the open source designs than you can get from prusa.

View on HN

Well we don’t really know that an accelerometer is needed for input shaping for a nozzle that is only accelerating across a single plane. The mk4 nozzle only moves on one horizontal axis when printing not two like the XL and the Bambu. Not that I don’t trust your take, but Prusa usually has their ducks in a row and I expect this to be no different.

I don’t care about any features of the Bambu line and would continue with Prusa line. They are nice sure but they aren’t that much cheaper with that much better features. You didnt really mention these jaw dropping world changing features so feel free to fill me. I hope it’s more spaghetti detection.

You’re forgetting that z hops are a thing. Even in a single axis the load from Eg the filament reel depends on the actual weight of the reel, and there’s almost certainly some variation in the cast aluminum frame parts (internal voids are I’m told an inevitable happening in aluminum/metal casting at scale). I assume the input shaping also would be needed on the bed slinger element as well.

As for the bambu, I’m getting voron level performance out of it, though admittedly with a much less open machine. I was using a prusa mini much of the past year but I quickly hit a lot of rigidity issues with its frame that were impossible to address without a total rebuild.

I guess for me the issue is simple: most of the plastics which I want to build stuff with (whether for engineering reasons or haptics of how they flex and feel) really require / benefit from enclosed printing. Pla is an incredibly strong material, but it can’t really handle any flex without brittle failure. And I just don’t like that.

I’m also in the process of putting together a voron v0, which I expect I’ll use even more due to sheer speed and ease of heating.

Nylon and asa and tpu are just so much more gratifying in my use cases.

What do you think about CPE/Chlorinated Polyethylene? Works fine for me on a Prusa i3mk3s in a room heated to 75 degrees F, with limited drafts (make a tent/enclosure if you have to)

I’m not familiar with it! If it’s a fun material great! While the fumes from any filament require serous caution for long term health risks, I’d assume anything with chlorine or fluorine pose some acute health risks? (I make sure I run these machines in a fully ventilated area that is rarely occupied with a person, in my case a spare bathroom )

I built an enclosure for my prusa out of a 20 dollar tub, so I don’t consider an enclosure like a big thing because they are easy to make. If I didn’t bespoke my own I would have made the ikea enclosure.

I can print tou all day, you can’t really compare the mini because it’s not direct drive and it’s a Bowden. I bet if you had a proper MK3 you wouldn’t have any issues with your filaments. The mini is the mini for a reason. I want to do asa and nylon but the fumes… eww

Very true!

What I was trying to articulate is that a bed slinger motion system requires a much larger enclosure than a core xy motion system or other fixed/z axis bed designs. So that if you have any real estate constraints and want an enclosure, bed slinger becomes non viable. :)

I agree that the fumes are bad, I use a spare bathroom with always on ventilation for mine.

The unfortunate thing is that the volatilized chemicals/ and particles in pla and petg are also not super great for people, they’re merely less terrible smelling

>Well we don’t really know that an accelerometer is needed for input shaping for a nozzle that is only accelerating across a single plane.

Input shaping requires an accelerometer for calibration. On a bedslinger like the Mk4, you need two sets of measurements - one for the toolhead and one for the bed, because both are moving and both can experience different resonances. The Prusa solution is to pre-calibrate input shaping from the factory; this is less than ideal, because printer resonances can be substantially affected by the surface that the printer is sitting on. The user could re-calibrate input shaping, but that would require them to connect their own accelerometer module to the motherboard and find some means of jury-rigging it in place.

https://blog.prusa3d.com/announcing-original-prusa-mk4_76585...

I’m not sure what you are trying to say. It sounds like you are saying that because they are factory calibrating it, it’s not needed… so then it’s not needed? What does it matter if it’s less than ideal if it produces satisfactory results? Results are the only thing that matter and Prusa is consistent with producing them so I’m not sure why this is different?

>What does it matter if it’s less than ideal if it produces satisfactory results?

Input shaper calibration is affected by the surface that the printer is standing on. If your workbench happens to be more or less rigid than the surface the printer was calibrated on at the factory, the calibration won't be accurate. The input shaper could fail to correct for some resonances, or correct for resonances that don't exist. This will cause visible artefacts in the print.

Other printers at a similar price point offer one-touch calibration of input shaping, or automatically calibrate before each print as part of the homing and warmup routine. On the Mk4, you need to buy an accelerometer module, figure out some way of attaching it to the printhead and the bed, then run the calibration routine twice.

Yes we do, because the accelerometer is used to measure resonance on all axis when the printer vibrates at speed. Check the Klipper documentation to see how it works.

I'd say the proof is in the product. I've read a lot of complaints over the last couple of days that the Mk4 isn't CoreXY, doesn't have an accelerometer, should be enclosed, should be _____. If the Mk4 can produce the same quality of print, does it matter how it gets there?

To use a metaphor, Prusa made the very best, reliable car you could buy. Then a bunch of similar cars, that were pretty good were released. And a few really amazing Sports cars were released. And Prusa showed up again with a really reliable, solid car. And people are mad because they thought they were going to be bringing out a flashy sports car.

"Table stakes" seems hyperbolic here given that we've lived without input shaping at all for quite a while.

Speed boosts are welcome but that doesn't mean everyone needs to "overclock" to get the absolute best performance. And if you do want to do that, couldn't you add an accelerometer later?

You do not need an accelerometer for input shaping. You only need one to determine the optimal constants (which represent the resonant frequencies of the machine). Once determined, the accelerometer is not needed. All machines made the same way will be able to use the same constants.

Until maintenance is done or the machine is customized, like changing the nozzle or air duct.

Exactly this. It’s almost anti open source/user mod friendly to have a fixed calibration at factory time.

Also the level of qc needed to make a one time calibration at factory be accurate for the lifetime of the printer would require an order of magnitude more expensive machine and a service contract to maintain it.

If you’re looking for a printer to mod, why start with an expensive and well-sorted Prusa as a starting point?

I own 6 printers, 0 Prusas, because I like modding them and don’t see any reason to pay Prusa to give me a better starting point that I’ll just then undo.

Open source == ease of changing it. At least with software, and I don’t see any reason why we shouldn’t have the same expectations for open hardware like 3D printers.

3D printers are already one of the easiest things to modify. I wouldn’t be surprised if over 1/5 of what is printed is parts/mods for printers.

Prusa printers are no different in this regard.

3D printers being modifiable does not make it easy, at least by open source software standards.

A Prusa works great stock. But there's always something you can upgrade. I used my stock mk3s for years, but recently decided to upgrade the hotend for a quick-changing one.

+1 on the sovol sv06, did a bit of digging and this looks like the best featured printer at the moment for the price point ($350 CAD right now on Amazon prime)

Yeah, I would have saved a fair bit of money if I had gotten the sovol 6 as my first machine (though the sovol 5 is a bit more rigid and space efficient)

Reply on news.ycombinator.com