Use "test suite" rather than "test case" in error about mixing TEST and TEST_F

[RedBeard0531]

I'm pretty sure that this was missed when changing terminology since moving to another test case doesn't seem to make sense in context.

[Voxel3DPrinting]

Similar issues happen on other nextruder machines; CORE one;

• https://forum.prusa3d.com/forum/prusa-core-one-assembly-and-first-prints-troubleshooting/z-axis-trouble-but-only-when-printing/

MK4((S)

• https://forum.prusa3d.com/forum/english-forum-original-prusa-i3-mk4-hardware-firmware-and-software-help/mk4-probing-for-plate-in-mid-air-issue-homing-z-axis-when-starting-a-print/
• https://forum.prusa3d.com/forum/english-forum-original-prusa-i3-mk4-hardware-firmware-and-software-help/loadcell-issue-mk4/
• https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/3392

[Voxel3DPrinting]

Following my posts above, I have proceeded to try and fix my issue (since I really want to use the machine for continued development of my own projects)

Left-over possible root causes of the issue, based on concluded findings above;

• 1. Thermal expansion of nozzle section could result in changed strain
• 2. Position of the printhead during probing could result on different strain
• 3. Electrical Noise in the loadcell signal

I proceeded testing all of the above; 1)

• Tried removing the heatsink fan to scale expansion, no effect.
• Tried to let the heatsink soak and stabilise for 40minutes, no effect
• Excecuted custom code that set the heatsink temperature target to 80C, to prevent fan from interfering, no effect

• Probed via handwritten gcode in 9 different spots, no effect ((makes sense since the HX717 detects peaks instead of threshold

3. (was more plausible for me, since I have been involved in 3D printer PCB design before, where electrical noise exceeded 70V peaks at maximum, when unfiltered.)

• Tested by referring to the J13 connector pinout of the LOVEBOARD found here. I learned that the data (and E-motor wires are twisted white/black, heater wires are black only, I separated the twisted from the untwisted wires and found that it did not do much, which makes sense since digital reading is way less sensitive to noise than analog.

Now my assumption is that some noise interferes inside of the Loveboard or Xbuddy

[altaic]

Might be #4571

[Voxel3DPrinting]

Might be #4571

oh my god, look at my reply I just posted, was writing the last sentence when I saw your comment!

So then yes, it indeed looks like it is related to noise caused by the heater PWM. Unfortunately, tho, I cannot compile code (hardware guy, not a huge firmware fanatic).

[altaic]

oh my god, look at my reply I just posted, was writing the last sentence when I saw your comment!

So then yes, it indeed looks like it is related to noise caused by the heater PWM. Unfortunately, tho, I cannot compile code (hardware guy, not a huge firmware fanatic).

You could try taking out pin 7 (analog ground) from the nextruder wire connector (connected to the xBuddy board) and wire it to a better ground somewhere on the xBuddy board (schematic). You could fry something if you don't know what you're doing, so be warned.

[Voxel3DPrinting]

oh my god, look at my reply I just posted, was writing the last sentence when I saw your comment! So then yes, it indeed looks like it is related to noise caused by the heater PWM. Unfortunately, tho, I cannot compile code (hardware guy, not a huge firmware fanatic).

You could try taking out pin 7 (analog ground) from the nextruder wire connector (connected to the xBuddy board) and wire it to a better ground somewhere on the xBuddy board (schematic). You could fry something if you don't know what you're doing, so be warned.

ohkee good idea, Pin7 at Xbuddy side I then assume? But where inside of the machine could I find a better ground? also, wouldn't it be easier/ less risky to upload a less sensitive firmware?

[altaic]

ohkee good idea, Pin7 at Xbuddy side I then assume? But where inside of the machine could I find a better ground? also, wouldn't it be easier/ less risky to upload a less sensitive firmware?

Yeah pin 7 of the wiring harness connector on the xBuddy side. I'd probably go for the ground on one of the thermistor connectors (J15 or P2). According to the schematic, the connector mounting pads are connected to ground.

That said, in https://github.com/prusa3d/Prusa-Firmware-Buddy/issues/4571#issuecomment-2774248450 apparently v6.3.1 already made the load-cell less sensitive, so in theory this issue should already be mitigated. Not sure what else to say other than wait for the Prusa guys to weigh in.

[Kachidoki2807]

Keeping an eye on the loadcell output from the sensor info menu gaves me a lot of interesting data. Especially the value variation on a finger press is around 250. But when moving the head around the bed it can goes way above this threshold. When heating the nozzle too.

My idea is that the cable bundle plus the pneufit and PTFE are hardly constraining the loadcell because they are attached to the top of the aluminium heatsink. You can easily see yourself that slightly shaking the cables leads to very big variations on the loadcell output.

See my related ticket #4555.

[CZDanol]

Internal ticket: BFW-7075

[Shikadi297]

I have the same issue, a workaround is to go to the tune menu and set the nozzle temperature to 0 when it's trying to home. It will succeed quickly and set the temperature back to another value on the next step. Turning it off for bed leveling also sped up the process as it managed to detect every point on the first try

[CZDanol]

Could you guys please test the following?

• Have the "log to txt" in Settings → System enabled during the failing bed leveling, send the log file here
• Check that all connectors to the toolhead are properly clicked in, and that none of the wires are pinched
• Check that the heatsink is attached to the 3D printed gantry piece with a plastic spacers, not metal ones

[swaggner]

Could you guys please test the following?

• Have the "log to txt" in Settings → System enabled during the failing bed leveling, send the log file here
• Check that all connectors to the toolhead are properly clicked in, and that none of the wires are pinched
• Check that the heatsink is attached to the 3D printed gantry piece with a plastic spacers, not metal ones

All connections and wires look good. I’m not sure about the spacers you’re referencing, but I got an assembled version of the printer so I’m assuming PRUSA used the correct ones. I’ve attached the log. Thanks for your help. I’m anxious to get an official solution.

log.txt

[XSIS-Energy]

I have exactly the same problem, any solutions for this bug?

[altaic]

Check that the heatsink is attached to the 3D printed gantry piece with a plastic spacers, not metal ones

Please post a pic so we can figure out which parts you're talking about.

[klibardi]

I can confirm exactly the same behavior on my machine. It must be interference from the pwm as solely setting nozzle temp to zero solves Z-homing and nozzle cleaning issues (very frequent failures due to detected collision)

[a-pushkin]

Exactly same issue with firmware 6.3.2: severe Z-axis problems (problems whenhoming, failure to clean nozzle, retries when doing bed leveling) when the nozzle heater is on.

Turning off the nozzle heater solves all these problems even though nozzle is still hot.

Here's my forum thread about the issue - https://forum.prusa3d.com/forum/prusa-core-one-assembly-and-first-prints-troubleshooting/z-axis-trouble-but-only-when-printing/#post-749398 It seems that I'm not the only one having it.

[Prusa-Support]

Could you guys please test the following?

• Have the "log to txt" in Settings → System enabled during the failing bed leveling, send the log file here
• Check that all connectors to the toolhead are properly clicked in, and that none of the wires are pinched
• Check that the heatsink is attached to the 3D printed gantry piece with a plastic spacers, not metal ones

Said spacers (heat-sink — x-gantry) are supposedly the ones depicted in the assembled step 6.23. https://help.prusa3d.com/guide/6-nextruder-assembly_849034#849995

Thanks for your cooperation.

Michele Moramarco Prusa Research

[XSIS-Energy]

log.txt

i checkt the complete nextruder but anything is fine, I post the txt file i have plastic spacers and all connectors are clicked in

[XSIS-Energy]

Any updates ?

[beatwinthewave]

On the forum, two users report having replaced the main cable and having resolved the problem https://forum.prusa3d.com/forum/prusa-core-one-assembly-and-first-prints-troubleshooting/z-axis-trouble-but-only-when-printing/paged/2/

[drg407]

On the forum, two users report having replaced the main cable and having resolved the problem https://forum.prusa3d.com/forum/prusa-core-one-assembly-and-first-prints-troubleshooting/z-axis-trouble-but-only-when-printing/paged/2/

That was me, support had me replace the heatsink/loadcell with no luck. Then they had me try m104 s0 before mbl but that didn't seem to work. Then I tried it after every m109 command and that allowed me to complete homing/nozzle cleaning without errors. Support then determined that the main harness could be faulty/have interference and sent me a new one. I replaced the harness, and tightened the plastic posts for the heatsink with a wrench just in case hand tightening was not sufficient. Besides that I didn't deviate from the assembly instructions. It's been about a week since I've had the new harness in with zero collision errors on z homing and nozzle cleaning.

[RPospisil]

Hi, I've just rebuild to Core One and exactly the same issue. Tried all hints mentioned here and in referred links, but only what works is to turn off heating during homing.

[aclii]

I haven't assembled my Core One yet, but it was interesting to read about this bug so I decided to check Loveboard grounds and the main cable. It looks like there is no electrical connection between the heater ground and analog ground on the Loveboard. Looking at the main cable you can see some twisted black/white wires and straight black wires:

• extruder wires (4) are twisted in pairs for each phase to reduce EMI on other wires in the cable.
• RS-485 serial communication (4) are twisted in pairs to improve noise rejection on differential signal lines
• For some reason heater wires are just straight even though we have quite powerful low frequency PWM running on them along 1+ meter cable.

I would try to twist heater wires in 2 pairs (+24 and GND) and then twist pairs to minimize possible interference on other wires in the cable.

[RPospisil]

Hi, I've just rebuild to Core One and exactly the same issue. Tried all hints mentioned here and in referred links, but only what works is to turn off heating during homing.

Update: I've received a replacement cable from Prusa and the issue disappeared.

[Prusa-Support]

As far as I can tell, we have actually seen multiple assembly or hardware solutions to similar problems, and replacing the main cable is actually one of them, where it makes sense. The cable replacement is not necessarily the solution. Maybe yes, maybe not; let us help!

In fact, minor tensions due to assembly and wiring concerns are probably some of the trickiest aspects to debug so far, but our Customer Support seems to be able to solve virtually all of these problems, even those unfortunate cases that, due to a lack of more or less obvious clues, may require more than one replacement and longer than expected troubleshooting sessions. 🙇‍♂️

So far, there doesn't seem to be a direct correlation between the problem and the firmware, but we keep investigating. For the time being, please thoroughly discuss your case individually with our Customer Support before providing further feedback on GitHub, and keep us posted about the final result of the troubleshooting if you don't mind. 🙏

Michele Moramarco Prusa Research

[mach0gr]

I did what @aclii suggested above and I no longer have the issue with not able to home the Z axis reliably or having many false-positive triggers. I removed the cable harness, unpinned these 4 wires from one of the connectors and twisted them as described in aclii's post. Put back the harness and is working like a charm !

Unpinning connections from the clickmate connector requires a small needle where you have to bend a plastic catch (from the connection side) by less than one mm to release the pin. Do it are own risk if familiar with handling these stuff or else you may break it and it won't pin back again.

Thanks !

[bkerler]

I just tested on my machine and running

M109 S200
G28

works fine as expected on my Core One + MMU. Thus this seems to be no SW-related bug as it seems but HW issue that should be resolved using Prusa Support (most probably cable harness issue as already stated). So I assume that this fw issue can be closed as it is hw related ?

[GWdd]

Is there a repository of HW bugs requiring action, as in a 'recall' where users can know by serial number if their cable harness is bad? I have a new Core One and I'm experiencing nozzle cleaning errors. It'd be nice to know if I have one of the earlier 'bad' cables.

[bkerler]

@GWdd I don't think there is such thing. If you don't have any issues, you are good. If you have issues, contact support or fix the cable on your own.

My cable is stock and unchanged and my printer hasn't got any of the mentioned issues, so there is no reason to fix the "bad" cable. In most cases, the issues are related to bad belt tension. If you are sure the belt tension is fine, follow the instructions by Voxel3DPrinting to make sure it is indeed the same issue.

[GWdd]

@bkerler Factory assembled printer. No reason to doubt the assembly or blame a "bad user" of things. And since Prusa has acknowledged changing the cable to a better and more reliable design, I am simply asking do I have the ealier questionable design or the new design And I am having problem that match the symptoms listed above. And since I bought an assembled printer, I am not going to do work that should be covered under warranty.