Z-axis referencing table detection problems
Hello,
I am reaching out to you here in the hope that you may be able to assist me in identifying and resolving the issue more quickly.
Some time ago, I noticed a problem with the Z axis, namely that the rear screw had about 1 mm of play and the prints on the rear part came out terribly. Support did their job, I got a new motor and nut. In addition, I updated the software during the replacement. And here the problem begins.
1. Z-axis calibration - done
2. Head referencing - problem
The printer moved the table up to a height of about 7-8 cm from the head and stopped, slowly moving up bit by bit, constantly stopping. The nozzle touched the table - hurrah! But not so fast - a collision was detected. Printer reset, table lowered, homing and collision again... after 4 minutes it finally homed. First print - homing, collision detected, homing, table search, collision, homing - works. Second test print, same story.
After contacting support, they suggested that the hotend, which I hadn't touched, was too tight. I went through the entire load cell troubleshooting procedure, and it got worse. I tightened everything back up my own way with a torque screwdriver, and it's a little better, but every print has at least one collision during homing.I'm slowly starting to regret my purchase because for the price of the printer + MMU + camera, I could have bought the same set from a competitor. My friend has a Chinese printer for a fraction of the price and doesn't have as many problems as I do with my Prusa.Has anyone else encountered this problem and managed to solve it? Or is selling the printer and replacing it with something else the only solution?
RE:
The erratic Z homing is due to false readings from the load cell -- it reports touches where there are none. Two reasons for this have been discussed in various forum threads:
- Electrical interference from the nozzle heater. Since the heater current is modulated (PWM), its pulse edges can interfere with the load cell signal. Only a few printers seem to be affected to such a large extent that Z sensing becomes unreliable. We can discuss countermeasures if needed -- but I understand that your printer shows problems also when you select "Auto Homing" from the menu, without the nozzle heater being on, right? In that case this can't be the issue.
- Mechanical strain on the print head. Prusa's theory about the hotend being tightened too much points in that direction. More common causes I have seen discussed here are: (a) nozzle being pushed in too far, such that its rear end pushed against some internal stop in the print head, or (b) cable and PTFE tube tugging on the print head because they don't have enough free length. See here in the assembly instruction, and also the subsequent steps.
Could you confirm that Auto Homing also has problems while the nozzle heater is off (to rule out cause 1), and have a look at potential mechanical issues (cause 2)?
You may have found this already: You can see the current load cell readings under Control → Calibration & Tests → Loadcell Test. That may be helpful to see whether e.g. the load cell responds strongly to touching the cable harness, or whether it becomes more quiet when you pull the nozzle out a tiny bit.
RE: Z-axis referencing table detection problems
Thank you for your quick reply.
1. It's better with the heater turned off. I don't actually use Autohome every day, only every few prints.
2. I replaced the nozzle with a new one and moved it slightly closer, without applying any special pressure or resistance. The PTFE tube is different from the original because I have an MMU3, but the printer was working normally before and I did not notice any particular problems with homing.
3. I did a cell test where you have to touch it with your hand, and it passed.
When I get home, I will try to do homing with Z-axis calibration without heating the nozzle, and then I will try to run the programme so that the printer is based with the nozzle turned off.
RE: Z-axis referencing table detection problems
Yesterday, as part of a test, I did the Z-axis positioning and then homing, which took 2-3 touches of the table.
Come to think of it, this printer has always taken quite a long time to detect the table, but there have never been any collisions, etc. It usually took 30 seconds, and then another 30 seconds to clean the nozzle. Even when I installed a new nozzle, it cleaned itself in two rows.
Support gave me an answer today that it looks like wiring. I received some test files and they are asking for feedback on which one will work better.
RE:
2-3 taps sounds like the normal sequence for Z homing. If I recall correctly, the bed first moves up relatively quickly to get it up to the nozzle, resulting in a first (less accurate) determination of the touch position. The firmware then moves it down a little bit and back up at a slow speed for an accurate measurement, and repeats that once more to double-check.
So without the heater on your Z homing works normally. That indeed points to electrical interference and issues with the wiring -- routing or bad connections. Some users have reported that twisting the pair of heater wires which runs through the "umbilical cord" to the print head has helped for their printer. But that is a pretty intrusive modification to make, and it is not clear why it would work: The load cell signal that is transmitted back to the main board is already digitized there and should hence be quite robust.
The wiring to the load cell or heater right in the Nextruder unit seems a more likely source of such crosstalk. There is also the fallback solution of disabling the nozzle heater via Gcode before the homing and bed probing steps.
Please keep us posted what Prusa support recommends and whether it helps!
RE: Z-axis referencing table detection problems
I found this thread and it fixed my issue. It does seem to be electrical interference from the heater and causes multiple random issues.
RE: Z-axis referencing table detection problems
Hello colleagues!
Yesterday, I lost my mind because I was getting annoyed with printing nonsense... I had to waste a lot of time before anything was printed...
I disassembled the nextruder piece by piece, cleaned everything, and then put it back together. I also replaced the PTFE tube because I had replaced it before the problems occurred. When I was installing the nozzle in the nextruder, I pressed it down and then pulled it back by the thickness of a sheet of paper. Later, I received an email from Prusa with a print test, where the heating for levelling is turned off - it passed successfully.
I turned on my first standard file, the print went well, but the wires were catching on the exhaust housing. After fixing the wire, the problem returned on the next file, and now it works correctly sometimes and has problems other times.
I don't have the patience for this printer anymore, sometimes I feel like throwing it in the bin.
RE: Z-axis referencing table detection problems
I find it a bit difficult to keep track of the status of your printer. Is it that Z homing (and bed probing) work reliably when the nozzle heater is off, and erratic when the heater is on?
RE:
Either electrical interference with the load cell signal or a bad load cell.
Anyway, as for cheap Chinese printers, they have their problems as well. For example: QidiTech3d Permanently banned me for warning people after my family lost everything from a fire
This won't help you but my Core One is a solid printer with no issues. It didn't start off that way but then I did build it from a kit.
RE: Z-axis referencing table detection problems
We need more clarity from @kamiloss about the state of affairs. The original post described that pure homing (during which the heaters would be switched off) was also showing the problem. Hence I mentioned interference as one of the potential causes in my first reply, but largely ruled that out due to the initial description. But later posts suggest that homing with the heaters off works, which would point towards interference.
I think there is a clear "tree" of diagnostic tests and conclusions, but at the moment we are not able to walk the OP through that.
RE: Z-axis referencing table detection problems
All right, let's start from the beginning.
1. I had a problem with homing when starting printing. First, 30 attempts to detect the table, a few knocks, a collision message. The second time, homing was successful, sometimes on the third attempt. Support suggested that it was too tight, so I loosened it, but it got worse, so I went back to tightening it more, which was better than after the ‘repair’.
2. I tested homing without heating – 3 touches of the table and that was it.
3. I did homing with heating turned on – a small problem, but it worked without a collision message.
4. Load cell tests are successful.
5. Irritated, I did a full maintenance, washed everything and put the old PTFE tube back on - I replaced it because I got a transparent one from work - I like it better.
6. In the meantime, Prusa sent me test files and, depending on whether it will be OK without heating, they will send me a new cable for the nextruder/if the problem persists, we will replace the load cell.
7. After maintenance, I tested the files from Prusa, everything was fine, then I ran a larger print, everything started successfully, but the PTFE tube was catching on the fans, so after printing I corrected the position of the tube.
8. The printer is having homing problems again. I worked on the cables a bit and it's better now. It usually needs 5-6 touches of the table, but it takes longer to clean.
9. I am currently waiting for a response from Prusa on what to do next. Maybe they have another idea, because for me it's not clear that it's the cable/load cell.
RE: Z-axis referencing table detection problems
Thank you! So it seems fair to summarize that (a) when the heater is off, Z homing works robustly, but (b) when the heater is on, Z homing behaves erratically -- works sometimes, but if e.g. you move the cable harness and PTFE tube a bit, it has problems again.
That would indeed suggest electrical interference from the pulsed nozzle heater current. The main cable harness to the Nextruder seems to be a key contributor here -- some users have opened it up and twisted the heater supply wires, fixing the problem. So please take Prusa up on the offer of providing a replacement cable. Fingers crossed!
RE: Z-axis referencing table detection problems
In my case, it was specifically during the nozzle probing to the corner of the bed to "clean the nozzle" where I would get failures. The work-around was to turn off the heater during the process. Later I removed the nozzle cleaning step entirely in favor of a nozzle wiper mod.
The 6.4.0 firmware made improvements to bed leveling and nozzle cleaning sensitivity which may have worked around this issue (I don't know in my case since I still have it turned off). I never bothered to twist the heater wires even though that was on my to do list. I did read in other threads that Prusa sent out new cables to anyone who asked. This makes me think they are aware of the issue at least.
RE: Z-axis referencing table detection problems
For now, I won't be experimenting because I have to print a few promised items for a friend before Christmas. During the holidays, I can play around with modifications or twisting wires. @hyiger could you send me the hardware modification you have to disable cleaning? Maybe I could use it. Usually, when I'm nearby when the printer finishes, I clean it myself with a brush.
Additionally, I'm waiting to hear what support has to say.
RE: Z-axis referencing table detection problems
@hyiger could you send me the hardware modification you have to disable cleaning? Maybe I could use it. Usually, when I'm nearby when the printer finishes, I clean it myself with a brush.
Can use whichever of these you like: Core One Nozzle Wiper
You then add code to do the wipe and disable the stock nozzle cleaning which starts with G29.
... ;G29 P9 X208 Y-2.5 W32 H4 M117 Wiping Nozzle G0 X249 Y-17 Z5 F5000; G0 X211 Y-19 Z4 F5000; ...
RE: Z-axis referencing table detection problems
Hi,
In connection with the holidays, the New Year and the arrival of winter, everything has been slightly delayed. I received a parcel with a new cable and instructions to attach the load cell to the carriage, first until I felt resistance, then 1/8 - 1/4 of a turn.
I did everything as instructed and the printer is working much better, checking the table much faster. Now I will test it and let you know in a while if it is still working properly or if there are any problems.