First Layer stil uneven after G29.

First Layer stil uneven after G29.

Kossel Clear,
Proximity Sensor,
RAMPS 1.4
Marlin 1.1.0-RC8
Aluminium bed.
Printer has worked before the upgrade and with manual leveling.

Hi printers,
finaly it got my proximity sensor with a fresh version of marlin working on my Kossel Clear.
G29 seems work just fine. It takes a lot of probes over the whole print bed.
I have written the G29 Z-0.8 command in Cura “Start Code” AFTER G28 to make sure it keeps the data.
The first layer ist still way too low in the corner of the X-tower. (In Pronterface the X- Y- Corner)
I have no idea where to start tweaking since every manual says that G29 is exactly there to avoid this problems.
And since a managed to upgrade to a contactless sensor there can be no mechanical problem with the Probe -Switch.
As far as i remember the Kossel Clear tented to have this problem before the upgrade. With the old “switch sensor” and even without ABL. It was not a big problem because i have printed smaller parts. And my printing was not that consistent without the P-Sensor anyway.
Overall G29 compensates that very well. Exept for the corner of the X-tower.
Yes the aluminium plate ist flat.
So i think the problem must be deeper.
Maybe the steppers or the stepper settings?

One of the M48 Results:
Finished!
Mean: 1.915115 Min: 1.915 Max: 1.915 Range: 0.000
Standard Deviation: 0.000000

regards from germany and sorry for my poor english.

I’d guess that your issue is the probe offset from the nozzle. There’s no way to properly compensate for that offset with a delta unless you can guarantee that your effector is parallel to your bed throughout the entire envelope.

Did you make your diagonal rods yourself? Does your heated bed heat evenly?

Can you wiggle yoir effector by hand easily?

Do you have the EEPROM option enabled in Marlin?

Also, which auto leveling option are you using? If not Bilinear, then try Bilinear.

@Taylor_Landry I try to understand what you are saying. Mine runs on G29 Z-0.8. Does that mean i have to adjust the sensor to run with no Z offset? Or does it mean the horizontal axis of the effector has to be fully on the axis of the bed?

@NathanielStenzel effector has no play. These are the rods that came with the printer. I have even switched the ones from the Z-Tower to the Y-Tower. No changes.

@Greg_Nutt EEPROM is enabled. Bilinear is used.

And you are using auto or manual leveling? As I understand, manual isn’t working correctly in RC8. Try the RC8 Bugfix which as I understand fixes it.

Gents, this isn’t a firmware issue. It’s a hardware issue. Unless you are certain that your effector is perfectly parallel to the bed throughout the entire build envelope, an offset Z probe will not work well on a Delta. It’s that simple.

Did a quick drawing to illustrate why…

This is an effector shown in the center of a bed.

The hot end is shown centered. It’s 15.5mm long, and is shown 1mm off the bed. The probe is offset 7.5mm to the left, and is 15mm long.
So, you can see the probe is 1.50mm away from the bed in this position.
missing/deleted image from Google+

This is the same effector out at the edge of the bed (120mm). If you have a 2degree tilt on your effector, you can see the effect.

Aside from the fact that your nozzle tip is not parallel to the bed which makes it slightly closer to the bed, the probe is now at 1.12mm from the bed - a full 0.38mm closer.

Now, imagine the other side of the bed had the opposite tilt. The probe would then be farther from the bed.
missing/deleted image from Google+

There’s no way to account for this offset properly.

So, unless you are 100% certain that your effector is perfectly parallel to the bed throughout the entire build envelope, then an offset Z probe is not the way to go.

@Taylor_Landry Thank you very much for your explanation!! I think you must be right. So everything must be exact parallel in the first place. And this is of course absurd. And proof for your thesis would be that the problem only accrues on the axis of the sensor and nozzle So what to do? Does it makes sense to level the effector with shims? Maybe switching to a sensor that uses the nozzle of the hot end to push against the bed? I have seen those on youtube. At the moment my workaround ist to raise the left side of bed with two pieces auf paper when G29 is working and pull them out right before the print starts. :slight_smile: Already used this method for two prints and the first layer was perfect. Again thanks. Good man.

The Kossel Pro by Open Beam had this same issue, and they did use shims to fix it, so that is an option - though not very elegant.

Personally, I would look at using your nozzle as the probe. The FSR kit from Ultibots is inexpensive, easy to implement, and we’ve gotten great results from them on deltas.
https://www.ultibots.com/fsr-kit/

What ist a “FSR kit” i have no Idea at all.
Right now i am scratching my head on how to keep my proximity sensor. I really like the concept. And it took my a lot or work get the thing to work. I have started from scratch wit a i thin 3 year newer version of Marlin and made a holder our of aluminium.
I will try to shim the effector and get the sensor as close as possible to the nozzle. A printed holder without the nuts will gain a view mm space I hope the heat of the hotend will not kill him :-/ Anyways what is a FSR kit? Hey btw you seem to know a view things about this stuff. What ist your background? (sorry curious.)

FSR = force sensitive resistor.

Basically, it is three small pads that sense force. You place them under your bed, and they can detect when the nozzle presses the bed.

You are not the first to try the offset probe on a delta. @Alex_Skoruppa has posted extensively about his custom delta build. He started with a Kossel Pro, and has since designed and built his own, beautiful machine - the Carbonoid.

He will also attest to the difficulty of that probe setup on a delta.

I work for MatterHackers. I get to try and test pretty much every printer, filament, and 3D printer accessory out there. :slight_smile:

@Taylor_Landry wow. ok. got it. I have a heated bed. Is it possible to mount them between the sheet of MDF wood and the frame of the printer to keep them away from the heat? have you seen solutions like that? So if i geht it right the bed has to sit vertical loose on the sensors while should not have any play horizontal?
Verry good advice anyway i will look into it! Tumbs up!
You do this for living. Uh! I tend to get completely sucked into things like that and have to force myself to take care of my everyday life. :slight_smile:

Yes, it is possible to do that.

You can see how we mounted them on the Kossel Pro here - http://forums.openbeamusa.com/threads/matterhackers-collection-of-3d-printed-upgrade-parts-azteeg-x5-fsrs-e3dv6-etc.2543/

There is a max load limit for them - which I think is around 5lbs. Not sure how thick your MDF is, but something to be aware of.

You can have the bed floating, allow it to run G29, and then lock it down if you are having trouble constraining it. You would likely just need to add appropriate Z offset to account for the change.

Also, depending on how you do it, you can store your G29 values so you don’t need to run leveling before every print if you don’t want to.

Ok. I can´t wrap my head around switching to FSR System nor to Repetier Firmware. As tempting as it is.
I need to make something proper to mount my new e3d hotend anyway. So i figured to put the hotend on something like a hinge that wings horizontal. On the end of the hinge would be something like the head of a screw an a litte spring that pushes the hinge together. The metal screw would trigger my beloved :slight_smile: proximity sensor.

  • The explained delta-problem should be solved.
  • The sensor is out of the way.
  • Propper mounting of the e3D hotend round 20mm higher in the effector.
  • No major changes in the firmware and electronics witch would result in new problems to solve.
  • Two settings 1. Z-probe offset = 0 And new Z-probe hight.