Failing XL, What do you think is the problem?
I had it with support.
I am tired of sending pictures, videos, and waiting in chat queues.
I received my XL before summer, and got it upgrade with 2nd Tool.
In September started crashing the heads constantly, to the point that could not be calibrated without a crash. (I think it passed only once in 6 or 7 attempts). like this (from 1m 30s):
After doing so many tests, videos, photos and belt tensions... they decided to take the printer in. So I sent it back to Prusa.
To my surprised they only did a routine review in which they claim they could not reproduce the issue. Which is very hard if you have done some testing and understand that the point of testing is to reproduce failures so that can be fixed, not to retry until you can tick green... but that is another story.
I got my printer back, and after printer one benchy, the filament sensor starting failing, seems dead (I mount on right side, mirrored).... so after cleaning and testing multiple times a decide that mount on the other port (filament 1-3), and there, the filament sensor works perfectly.
So, clearly the filament sensor and cable are fine... but there is some issue with the PCB or Firmware.
Considering the the printer and its history. I am almost sure that this is a failure originating in silicon. I have noticed other users with hardware failures.
But what is your opinion? where do you think the problem origin is?
Meanwhile I will disable the sensor and brace for further impacts.
RE: Failing XL, What do you think is the problem?
Sadly I can't get the video to run on my end. What FW are you running? Not to sound like support but I would also like to see pictures around both your toolhead and nozzle seals. A good view of the cable path.
Let's just try to focus on getting your xl working. The more info you give the more help we provide with this.
RE: Failing XL, What do you think is the problem?
The firmware is the latest: 6.1.3
Here we can see the video of the hear crashing.
https://drive.google.com/file/d/1Z6r3jz_DZXR3UTp6vMxyPrz1Cub_NP_j/view?usp=drive_link
The side sensor is a no brainer. The connectors are solid a clearly labelled: filament 1-3 and filament 4-6.... on one connector the filament sensor works perfectly, on the other fails 100% of times, and was working last week.
So, sensor and cable are in perfect condition.
And I cannot see how this situation could emerge from a firmware bug.
RE: Failing XL, What do you think is the problem?
My guess is belts or motors. Linear rod screws on your x.
Worse case, you have a bad board. The z grinding can be heard at the end as well. But please check every bolt in those rods. My voron had one loose one and it caused all softs of grinding issues.
RE: Failing XL, What do you think is the problem?
No, I sent the printer to Prusa, they diagnosed it, and they did find all the parts of the printer were "without noticable damage or deformation"
The diagnosed also stated: "The belts are tensioned properly.All pulleys are well positioned and tightened."
I copy the full diagnose on reception:
Initial diagnosticsA) Plastic parts + Frame/extrusion:
The frame parts is straight and has no visual damage or deformation.Plastic parts and all extrusions appear to be free of visible damage or deformation.B) Linear rails + linear carriages + trapezoidal nuts + Z motors:
Linear rails and carriages appear to be without noticable damage or deformation.The trapezoidal nuts and Z motors are without signs of damage (minimal play).C) Steel sheet + heatbed:
Steel sheet Satin. The Steel Sheet has signs of use. Steel sheet can be used for printing.The Heatbed segments are properly connected and the cables appear to be undamaged. Heatbed appears to be free of visible damage or deformation after cleaning with Isopropanol. The insulation is not damaged.D) Tool changer + Extruder + hotend: Number of extruders: 2
Tool changer appears to be free of visible damage or deformation (Will be tested).1)
Plastic parts are without visible damage.The docking station - OkFilament senzor - Ok. Loadcell - Ok (will be tested)The print fan and heatsink fan are without any significant visible debris.The nozzle is brass type Prusa Nozzle (0.4). The nozzle and heaterblock aren´t clean. The heater and thermistor cables are not damaged.2)
Plastic parts are without visible damage.The docking station - OkFilament senzor - Ok. Loadcell - Ok (will be tested)The print fan and heatsink fan are without any significant visible debris.The nozzle is brass type Prusa Nozzle (0.4). The nozzle and heaterblock aren´t clean. The heater and thermistor cables are not damaged.
Plastic parts are without visible damage.The docking station - OkFilament senzor - Ok. Loadcell - Ok (will be tested)The print fan and heatsink fan are without any significant visible debris.The nozzle is brass type Prusa Nozzle (0.4). The nozzle and heaterblock aren´t clean. The heater and thermistor cables are not damaged.E) Extruder cable (PTFE, Cables, Cable guide):
The cable and PTFE tube, including the leads, are not damaged.F) Belt Tension + Pulleys:
The belts are tensioned properly.All pulleys are well positioned and tightened.G) XLBuddy (Board, Cover, Sandwich board) + XL modular bed board:
XLBuddy verze: 09/2023 is without visible damage and is functional.XL modular bed board: is without visible damage and functional.XL sandwich board: is without visible damage and functional, the cables are correctly connected.XL splitter board: is without visible damage and functional.H) XL side Fil. sensor: Number of sensors: 3
The filament sensor is functional and without visible damage.I) PSU:
Number of PSU: 2The PSU has no visible external damage. The printer turned on without any issue.J) LCD (USB, knob, reset button, Speaker):
Knob, speaker, display, reset button, USB port - everything Ok.K) LED strips (board, LED strip)
The LED light is functional.L) Accessories: flash disk. - 1 Spoolh., No PWRC, Anthena, wifi adapter, No tools, calibration pin
1) FW version: 6.0.32) Bootloader version: 2.3.53) Calibrations + Tests: OK
So, after a chat with the Engineer, the persistent behaviour that you see on the video could not be explained or reproduced.
They did 2 prints, which attached with the printer and returned it.
Then, after reception, I printed a benchy and the filament sensor became "dead".
RE: Failing XL, What do you think is the problem?
I am inclined to think that both the head crash (which Prusa could not reproduce or explain) and the dead port, share a root cause.
@cotano A bad board would actually explain both things. Say a failure on the main micro-controller. It could also come from poor PCB design, i.e a BUS not properly pulled up, etc... but Prusa should have found out this by now. Note that the engineer is recording in the protocol the version of the Buddy board (2023/9) but it is not recording the version of Dawrf, Sandwich... and many other boards involved).
Both things could also be explained by poor firmware causing memory corruption. This is not uncommon... accessing memory already freed, freeing a pointer twice... etc. However, it would surprise me that Prusa ships firmware in this state, this would be acceptable during alpha testing of the product, not years after launch.
The internet is full of videos of PrusaXL where suddenly the head just crashes, shift layers significantly, etc. Most people tend to think that it is a mechanical issue, but I am starting to think that it is one of the previous two.
Curiously enough the support engineer suggested a "factory reset" to as a potential fix to the dead filament sensor.
RE: Failing XL, What do you think is the problem?
That grinding noise. Having similar problems with my XL 5T too, only a few months old. When I kept getting asked about firmware and even coerced into a factory reset, I knew that the troubleshooting process was designed to be of no help.
How can I spend $3500USD on a printer from a brand that is supposed to "just work" and I have problems on problems and get the runaround from support.
And not to mention, tool 2 has always struggled. But it's intermittent and there's no way to prove to support that there really is a problem. I just get the runaround. Ugh. I wish I could just send the whole thing back for a refund.
RE:
I have been reviewing the microcontroller family (STM32F4) and the program is in flash... only under strange electrical conditions the flash would be altered (I have never seen such a thing, and I have manufactured 1000s of pcbs). And in any case, the integrity of the FLASH memory can be verified any time... it is possible to test if it differs from the current version that has been flashed.
Then there is the possibility of a firmware bug (i.e a wild pointer corrupting memory), the problem is that the filament sensor is already dead right after turning the printer ON. Which code could have corrupted the memory? navigating the menu until the tests? printer initialization? perhaps the most probable would be the Network/WIFI stack, as the printer will connect to WIFI on startup. However there is an ESP chip that deals with WIFI, so, it wont be a lot of code. I guess I could disable networking and re-test, if that is possible.
There are recent/opened bugs describing head crashes, filament detectors failure, etc:
https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/4235
https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/4329
https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/4304
https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/4259
However I find VERY concerning that since October the 23th the Budy Board Firmware repo is basically stale, with two minor commits.
And then there is the possibility of the PCB having an issue. Here I cannot find the schematics, the test engineer noted the version of the Buddy Board, which implies there is active work on it, or recent work on it. But again, I don't know how many versions there are and what changes were introduced since mine, in order to asses if any impacts the functionality at play.
Considering the Prusa is letting the QA of Prusa XL to the users the least they could do is open the schematics completely. Any body knows what is the latest version of the buddy board?
So it seems that after all this wait time both Firmware and PCB are pretty beta-ish
RE: Failing XL, What do you think is the problem?
I have filed this BUG report: https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/4353 sumarizing the issue
RE: Failing XL, What do you think is the problem?
My mistake, I downloaded the buddy repo, and diffed 6.1.3 and the latest alpha and there is a lot going on... that seems like a relief.
I will disable the filament sensor for now and keep checking as releases are launched.
tool 2 has always struggled. But it's intermittent
It's harder to check on 2h, with 5h you just look along the line for a head a millimetre out of line ... Make certain the dock for the offending head is pulled fully into place by the mounting screw, see: https://help.prusa3d.com/guide/5-extruder-accessories-assembly_523539
- look from step 14/15 onward and the same for the other dock(s). The dock screws MUST be tight, it's easy to miss the last quarter turn and leave it not-quite-home when tool changes are affected and the head may not be picked up correctly.
I am unable to reproduce any of these issues, like most users my XL really does just work so I cannot help with your underlying problem.
Cheerio,
RE: Failing XL, What do you think is the problem?
tool 2 has always struggled. But it's intermittent
It's harder to check on 2h, with 5h you just look along the line for a head a millimetre out of line ... Make certain the dock for the offending head is pulled fully into place by the mounting screw, see: https://help.prusa3d.com/guide/5-extruder-accessories-assembly_523539
- look from step 14/15 onward and the same for the other dock(s). The dock screws MUST be tight, it's easy to miss the last quarter turn and leave it not-quite-home when tool changes are affected and the head may not be picked up correctly.
I am unable to reproduce any of these issues, like most users my XL really does just work so I cannot help with your underlying problem.
Cheerio,
For that step I took off the honeycone panel to give me more space. Those are not easy to tighten.