How many Quad Gantry Level passes ?

[lukics]

I have a freshly built 2.4 350x350 from FormBot.
I have questions regarding the gantry and quad gantry leveling:

• Is it normal that the gantry drops a few milimeters when the z-motors are powered off? (on the side of the gantry motors)
• how many passes of quad gantry level are regular? I have more or less 3+ passes, sometimes (seldom) also giving up with an error of too many reties.
• Is it ok when during the quad gantry leveling the probe retires are done ("Probe samples exceed tolerance. Retrying...")? I tried the inductive probe, but also the installed Kklicky behaves the same.
• Should I heat up the bed and then do quad gantry leveling, or is it even ok to do leveling during heating up?

Thanks in advance for your experienced input.

 

[NoGuru]

If the motors are not powered it is common to have the gantry fall a little.
For passes, just whatever it takes to pass. I normally have 2 to 4 passes.

I always heat the bed first. If I even think I am going to print I have a macro that is called PREHEAT and I let that run for about 30 minutes before starting.

 

[lukics]

For passes, just whatever it takes to pass. I normally have 2 to 4 passes.

What about: „Probe samples exceed tolerance. Retrying..." during probing?

 

[NoGuru]

Then I would say there is something wrong with the probe. I bet the mount is loose. Make sure it does not move when it touches the bed. I have seen the mount become loose after some prints so that is where to look.

 

[lukics]

Then I would say there is something wrong with the probe. I bet the mount is loose. Make sure it does not move when it touches the bed. I have seen the mount become loose after some prints so that is where to look.

The strange thing is, that from the start I had the originally supplied inductive attached. Once I saw these issues I installed the klicky PCB probe, experiencing the same results. I then switched back to the inductive one. Then I changed the raspberry pi3b to a raspberry pi4B4G, thinking that the processing power may change something, same as before.
I also followed the instruction for V2 Gantry Squaring. One thing that can be a problem is the wiring. I have the Hartk SB Toolhead Board, which reuses common wires like ground and 5/24V, not having separate wire pairs for fans and sensors. this should not have an influence as far as my knowledge tells me.
I will also heat up the printer to max for an hour and retighten all possible gantry screws, crossing my fingers to find the reason.

 

[steve/]

I would say the 2 to 4 time is about right. I am set for 0.016 max different. Now for the "Probe samples exceed tolerance. Retrying...")? I have had this, but I have normally been the cause, like changing the filament or working on some part and moving the frame.

 

[lukics]

These are my settings:

## Select the probe port by type:
## For the PROBE port. Will not work with Diode. May need pull-up resistor from signal to 24V.
#pin: ~!PB7
## For the DIAG_7 port. NEEDS BAT85 DIODE! Change to !PG15 if probe is NO.
pin: PG15
## For Octopus Pro PROBE port; NPN and PNP proximity switch types can be set by jumper
#pin: ~!PC5


#--------------------------------------------------------------------


x_offset: 0
y_offset: 25.0
z_offset: 0
speed: 10.0
samples: 3
samples_result: median
sample_retract_dist: 3.0
samples_tolerance: 0.009
samples_tolerance_retries: 3

Wondering if doubling the samples_tolerance is ok to do?

 

[NoGuru]

Try it and see what happens, worse case scenario is you put it back to how it is now.

 

[Mike Eitel]

Are you aware what samples_tolerance: 0.009 means in reality. Human hair is 0.05 - 0.08

m5c
Mike

 

[Neil_I]

It used to be 2-3 passes for me, with a visible droop on rear right side when turning motors off. Since fitting the GE5C mod to the Z joints QGL is only 1-2 passes and there's no visible droop. Definitely the best mod I've done so far. There was a nice improvement for input shaper graphs and max accelerations too

 

[lukics]

I am now getting much better results.
I have performed QGL, loosened the Z joints, and retightened. Now the gantry does not visibly drop when disabling the motors.
Later I am going for the solution with the GE5C bearings.

 

[lukics]

Since fitting the GE5C mod to the Z joints QGL is only 1-2 passes and there's no visible droop

I can confirm this. I have added this mod and now things do look totally different. I do not have the "exceed tolerances" error and the passes are reduced, too. Additionally, I have added TAP.

 

[Zdenek]

Hello gentlemen. Can I ask you for some advice? I just completed Voron2.4 and have been struggling with QUAD_GANTRY_LEVEL for a week. I have a BTT Manta M8P + CB1 board, I used a BLTOUCH switch, I changed it to an induction one and it still gives me the same error. Home XYZ is fine, then I give a level, the head goes around all the corners and finally pops up ERROR and a lot of long nonsense numbers. I've probably tried everything and can't figure out what's wrong. I am quite new to KLIPPER. I am attaching a picture of the error. Thank you for your opinion.

 

[Durahl]

I have a freshly built 2.4 350x350 from FormBot.
I have questions regarding the gantry and quad gantry leveling:

• Is it normal that the gantry drops a few milimeters when the z-motors are powered off? (on the side of the gantry motors)
• how many passes of quad gantry level are regular? I have more or less 3+ passes, sometimes (seldom) also giving up with an error of too many reties.
• Is it ok when during the quad gantry leveling the probe retires are done ("Probe samples exceed tolerance. Retrying...")? I tried the inductive probe, but also the installed Kklicky behaves the same.
• Should I heat up the bed and then do quad gantry leveling, or is it even ok to do leveling during heating up?

Thanks in advance for your experienced input.

Here's my CHF 0.50

• I don't think this ever happened to my V2.4 so I'd argue no but maybe I'm just the odd one out
• Depends on how sensitive your Values are... I reduced mine to usually be done with like two rounds if things got really out of whack after like carrying it around for maintenance.
  I set mine to retry_tolerance: 0.03 instead of the default retry_tolerance: 0.0075
• I'd wager a No?
• Not the worst Idea since you'll want to measure the state at which you're printing.
  Personal Advice is to initiate the Heating, followed by XYZ the Homing, followed by Quad Gantry Leveling, wait for things to reach desired temperature, then do the Bed Mesh, followed by a single Z Homing to get the best results.

 

[lukics]

Here's my CHF 0.50

• I don't think this ever happened to my V2.4 so I'd argue no but maybe I'm just the odd one out
• Depends on how sensitive your Values are... I reduced mine to usually be done with like two rounds if things got really out of whack after like carrying it around for maintenance.
  I set mine to retry_tolerance: 0.03 instead of the default retry_tolerance: 0.0075
• I'd wager a No?
• Not the worst Idea since you'll want to measure the state at which you're printing.
  Personal Advice is to initiate the Heating, followed by XYZ the Homing, followed by Quad Gantry Leveling, wait for things to reach desired temperature, then do the Bed Mesh, followed by a single Z Homing to get the best results.

Things changed dramatically on my end after I made two mods, GE5C and TAP. Now, I typically have two passes with about three taps per probe. The only issue remaining is that occasionally, a tap pauses in the down position for an inexplicable second.

 

[Laserbea4k43]

It used to be 2-3 passes for me, with a visible droop on rear right side when turning motors off. Since fitting the GE5C mod to the Z joints QGL is only 1-2 passes and there's no visible droop. Definitely the best mod I've done so far. There was a nice improvement for input shaper graphs and max accelerations too

GE5C mod made things a little better for me, but still get the same errors when starting a new print in a machine that's been running for a few hours.

 

[Fabreeko]

My stock 2.4 takes about 3-4 tries and if its been sitting for a while it will fail due to gantry sag in the back. My other one with ge5c is definitely better at 2 sometimes 3.

If you guys are having problems after the machines are hot maybe try adding backers, i dont print back to back enough to say it will fix the problem but they are meant to keep the frames thermal expansion in check by making it tri metal. Sounds like a bunch of snake oil but the data shows they work .

I can make a coupon for the forum and even send a free set to anyone who volunteers to gather data and post back

 

[Laserbea4k43]

I was running a QGL the other night and saw one of my Z-joints were just sitting at the bottom of the rail!!
It turns out, I wasn't using an M5x20 like I was supposed to. Latest (cold) QGL took 2 tries.

Anyway, I HAVE been looking at backers. I'd be willing to go for the coupon AND/OR the free backer (if you tell me how to gather the needed data...)

-EDIT- I wonder how hard it would be to CNC titanium MGN12 Rails...