Why camera requires a separate connection?
I noticed that the included camera requires wifi to work. If the printer is otherwise connected with a ethernet-cable, it means the system now requires two separate connections to the same network. And I guess the same applies if the printer itself is connected with wifi.
Correct me if I'm wrong but this seems like a half-assed solution. I guess this post is more feedback and less a question.
RE: Why camera requires a separate connection?
I noticed that the included camera requires wifi to work. If the printer is otherwise connected with a ethernet-cable, it means the system now requires two separate connections to the same network. And I guess the same applies if the printer itself is connected with wifi.
Correct me if I'm wrong but this seems like a half-assed solution. I guess this post is more feedback and less a question.
With Prusa's printers up to the original Core One, the main microcontroller has always been considered the bottleneck. It just can't handle passing all the image data through to its WiFi or Ethernet connection. Hence the camera needed its own WiFi connection as a workaround. The compressed, binary Gcode format is another consequence of these limitations, I think.
In the Core One L, I understand that Prusa has overhauled the xBuddy board -- to accommodate a CAN bus for intelligent peripherals, and maybe further updates? It is disappointing that they were still unable to beef up the computing power and let the microcontroller handle the complete network connection. Probably too big a development step, given that they needed the Core One L quickly as a response to Bambu's H2 series.
RE:
I dabble a bit in "pragmatic" embedded design. The longer I think about it, the more sense it makes.
No doubt a "monolithic" integration can be done, may even approach routine if you hire the right people. But also, no doubt it invites new ways it can crash and burn, like new security vulnerabilities or short-lived consumer components going EOL.
Separating the camera at the network connection looks absurdly low-tech but avoids a whole bunch of problems from interaction between the systems. Such as stepper motor deadlines on one side, camera interrupts on the other. Interrupts triggered by a black-box chip that you didn't fully understand in the first place and now have to replace in a rush because it can't be sourced anymore. And maybe I can get independent µC-cores that are "almost" perfectly independent... hello cache, DRAM, ...
Don't fight a battle if you can avoid it altogether.
Long-lived industrial equipment is usually modularized for a reason, and low-tech in places where it can afford to be.
RE: Why camera requires a separate connection?
Having a printer control and cameras on different networks is a really good idea in certain industries, you know? 😉
See my GitHub and printables.com for some 3d stuff that you may like.
RE: Why camera requires a separate connection?
It seems like a "low tech" solution at first, but thinking of the pros and cons I can't really knock it. Doesn't take processing resources from the printer, and if the printer's own firmware should crash, it's nice to know you can (probably) still look in to see what's happening with the camera. Also means you could update to improved (higher res, etc) cameras without needing increasing resources from the printer itself. Coincidentally, I have either at home or work used 3 other printer brands with built in cameras. They're all inferior in image quality to the Buddy Cam. Prusa really just needs to expand the streaming a bit to work via prusaconnect (harder to secure) and maybe a web widget so it works on a LAN through prusalink.
RE: Why camera requires a separate connection?
Having a printer control and cameras on different networks is a really good idea in certain industries, you know? 😉
That must be different industries than those which would benefit from being able to use the Ethernet connection instead of WiFi? 😉
I get the various arguments that separating the camera data stream avoids technical risks and cost (for hardware and R&D time) for Prusa. The current Buddy camera solution was the pragmatic thing to do, given the hardware already in the field in other printers and given the development timelines for the Core One L. But I struggle with the claim that it provides a customer benefit over a clean, integrated solution where the complete printer-related data stream is available via a single connection, wireless or wired.
RE: Why camera requires a separate connection?
I think It was quicker by getting ready made made hardware and being able to just focus in writing firmware for it. The fact how it ended probably delayed the cameras so long that it was a real problem.
On the other hand if it was integrated into the printer it could have delayed the new printer line - and while you really do not need a camera to print , it was somewhat making things easier, for the price of the common feature among other vendors.
Looking at how much new stuff is in the Core One L maybe that could have been integrated into it, so I guess that may be a good question for Joe on the Formnext 2025 🙂
See my GitHub and printables.com for some 3d stuff that you may like.
RE: Why camera requires a separate connection?
Wow really good discussion guys.
I also lean on the opinion that the current implementation was done because it was faster to do it this way (but I just don't like it). Who knows, perhaps we'll see it in the next iteration of the design (whatever it will be called).
Btw I would love to ask from Josef in Formnext but unfortunately I can't skip work for a week. Perhaps one of these years...