Trouble with larger prints sticking

I Have a working printing mini kossel from T3DP3D upgraded with duet8.5+Ir sensor for auto calibration + PanelDue touchscreen all running DCv1.09. Everything works and calibrates great, i think, SD after a few calibrations is (0.055)

My printer prints small prints in the center of the bed very nice however it struggles with larger prints. Particularly in one area of my bed where the nozzle does not quite squish the layer down to the bed to get a good enough stick. I use hair spray on a glass bed but this still does not catch and squish the filiment ( i can see this based on trying to print larger prints that take up the bed and continually having problems in the same area). I think this means my bed is not being properly compensated for during the autoleveling and maybe i could add additional probing points around this area in the .bed file??

I have tried to squish the layer using slicer by setting the first layer greater than my nozzle diameter based on other forums i have been reading which helped A LOT so did slowing my printer speeds down thanks for all the help from that forum. However i am still struggling with my larger prints and particularly in the same area of my bed. Has anyone else had this type of problem and have any advice on how to go about fixing it? (like possibly shimmy up that corner of my bed (with small cuts of index cards?) Let me know if anyone has any advice, small parts are fine but print multiple small parts or larger prints its still not optimized so any help is appreciated!

To home in on this solution and stay up to date. I have watched my autoleveling protocol and it is skipping the area i am having difficulties with... I will be trying to add a probe point in the .bed file.

It's a good idea to print a single-layer calibration spiral or calibration grid covering all areas of the bed, to check that you have a uniform Z=0 height. You can find models for these on Thingiverse.

You may find my bed.g file generator at [escher3d.com] useful. However, the bed probe pints don't need to be in a regular pattern. If you add any probe points manually, make sure that they are numbered consecutively from P0 upwards, and that only the last one has the S6 parameter.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

so im having problems with the spiral sticking to the bed even though i added the probe point i thought was missing. Is there some other way to compensate for this?

My guess is that the effector is tilting slightly as it translates in the XY plane. This causes the relative heights of the IR sensor and the nozzle to vary, which in turn affects the nozzle height at which the sensor triggers. So I suggest you measure the trigger height at different XY probe points. If you do find a variation, you can correct for it in the bed.g file, although it is better to eliminate the tilt as far as possible first.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

what command in the bed.g file would i edit? Or is this due to something mechanical in the build like tightening the effector arms on the carriages?

More specifically adding the H parameter to the G30 command just comes at the end of the existing code or do i need to put it in front of that?
Also is this H the offset (difference between the offsets at the points im having trouble with) or do i use the greater or smaller of the different offsets?

As I said before, you need to measure the trigger height at each probe point. If you have feeler gauges, you can also measure the height of the bottom of the sensor board above the bed when the nozzle touches the bed. If those heights also vary by similar amounts, that confirms that the problem is caused by the effector tilting.

The H parameter is used to correct for variations in trigger height that you can't eliminate ny improving your build. For each of the G30 commands in bed.g, the H parameter is the amount you need to add to the trigger height that you set in config.g to get the actual trigger height at that point.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I solve it with tons of hair spray, because large pieces with spikes, it doesnt matter how well your z heigh is calibrated, it can warp and unstick.
And use 2mm brim

I am currently working on compensating for the differences in trigger height i am measuring then working the results in the H parameter inside the G30 commands in my bed.g file. I am using the difference between my (Z trigger height set point) - (Actual measured trigger height at Probe points). Please correct if i am wrong

G30 P8 X36.8 Y32 Z-99999 H0.00 Is this the correct?

@PurpleSensation
I do use hairspray and plenty of it before a print and i still have this problem. Does adding "brim" help squish the first layer? I am not a slicer expert sorry if that is a newb question.

Quote
Ace67aod
I am currently working on compensating for the differences in trigger height i am measuring then working the results in the H parameter inside the G30 commands in my bed.g file. I am using the difference between my (Z trigger height set point) - (Actual measured trigger height at Probe points). Please correct if i am wrong

G30 P8 X36.8 Y32 Z-99999 H0.00 Is this the correct?

It's the other way round: H parameter = (actual trigger height at that point) - (trigger height set in G31 Z parameter).

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Using brim absolutely helps with this problem, i recomend this process when someone have to print a large piece, and much more important if this piece has spikey corners:

-Clean very well the base
-Spray it (a lot) before placing it on the printer
-Place it
-Calibrate z-height with a tipical 3 point method
-Print it with 2mm brim

First layer must be opaque! Not translucid and not too thick, you must see how the layer is crushed but still have visible thickness.

dc42 Thanks! ill swap it right after i remeasure i had to rotate the sensor so my additional print cooling fan could be mounted on the other side (so it wont crash into anything.

Purple ill try adding brim to my larger prints to try this out. I still have yet to truly optimize my slicer settings for "general printing" but first i gotta get everything level.

THanks for all the help

+1 to PurpleSensation's suggestions.

If I have a problem with a part sticking, then I:
* Clean the surface with acetone
* Add a 4mm brim to the part in slic3r


When the first perimeter purge is being put down, if this isn't the right height you will be able to see it. The lines should merge nicely into each other. If they are translucent (nozzle too low), or slightly not touching (nozzle too high) in places then you need to do a height adjustment (endstop adjustment on my delta).

So I have very little differences in my Z-height all under 0.10, however i have corrected for these anyhow and this still does not fix the problem i am having. I recently have installed PEI to my glass print top as well which really helped with first layer sticking (I painted one side black so it would not be transparent) I get a clean calibration from my auto calibration taking the deviation to 0.011 but for some reason I am still unable to get a level bed resulting in uneven first layer i can tell because some areas do not stick while others get squished too well.... I seem to have everything correct but i am having this same problem, all my end stop corrections are within 0.05 of each other as well so im not sure what the issue could be.

*Note, I am able to "eyeball" the Z-height where is does not squish and has too much squish and adjust the H parameters at those probe points accordingly which is tremendously painful and usually results in multiple extruder jams but i am able to get relativly acceptable results this way but I would like to have a automatic calibration i can rely on

Anyone have any ideas what would be going on??


Hardware:
T3DP3D mini kossel kit
Duet8.5 DC42 v1.09
Panel Due
Differential IR sensor z-probe
Black PEI heated print surface

Hi Ace67aod, to check the bed calibration you can do the following:

1. Copy file /sys/bed.g to "/macros/Check bed"

2. Edit file "Check bed" by removing the G28 command near the top and replacing the S6 parameter on the final G30 command by S-1.

If you run the Check Bed macro, it will probe the bed without homing or adjusting the delta parameters, and report the errors. I suggest the following:

1. Power up or reset the printer, then home it.

2. Run Check bed

3. Run G32

4. Run Check bed again

5. Home the printer

6. Run Check bed again

7. If the results of step (6) show low deviation, then move the head to each of the XY bed probe coordinates and use the paper test to see if the Z=0 height is accurate at that point.

Please post your results, and let me know exactly which firmware version you are using (should be 1.09m or 1.09r).

Thanks for your patience.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

After running the G32 the probe points of the check bed show low deviation and hold after homing and rechecking. Thanks for this little rundown, it saved me numerous hours of plugging in my XYZ probe coordinates one at a time to get these individual probe points. This learning curve is steep!

Results of my paper test at individual probe point positions. I also included my probe points in case there could be issues with them.

Z= Z height where paper snagged
G30 P0 X-75.10 Y-35.5 Z-0.50
G30 P1 X0 Y-76 Z0.10
G30 P2 X73.6 Y-30.5 Z0.20
G30 P3 X78 Y20 Z0.30
G30 P4 X0 Y67 Z0.0
G30 P5 X-78 Y20 Z-0.3
G30 P6 X-36.8 Y-21.25 Z-0.3
G30 P7 X36.8 Y-21.25 Z0.1
G30 P8 X36.8 Y32 Z0.2
G30 P9 X0 Y42.5 Z0.0
G30 P10 X0 Y0 Z0.0

Here is my paper test results. You can see here why i am having trouble.... for some reason my bed seems tilted maybe?

Thanks for the support!

Firmware I have not updated my firmware sense October 2015 i want to say 1.09m. Could this be my problem?

Did you make a note of the last set of readings from Check Bed before you checked the heights with the paper test? If there is a large discrepancy between them, then the probe trigger height is varying with position, probably due to effector tilt. Also, did the deviation reported by Check Bed agree reasonably well with the devirion reported by the preceding G32?

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

here is a snap shot of the protocol starting from power up (Note start from the bottom Connection Established)

M98 P/macros/Check bed.g
Bed probe heights: 0.004 0.016 0.030 -0.011 -0.008 -0.001 -0.061 -0.014 0.010 0.007 0.026, mean -0.000, deviation from mean 0.023

(printer homed)

M98 P/macros/Check bed.g
Bed probe heights: 0.004 0.016 0.018 -0.011 -0.008 -0.001 -0.061 -0.014 -0.003 0.007 0.026, mean -0.003, deviation from mean 0.022

G32
Calibrated 6 factors using 11 points, deviation before 0.160 after 0.020

M98 P/macros/Check bed.g
Bed probe heights: -0.004 -0.359 -0.135 -0.025 -0.184 -0.012 -0.163 -0.190 -0.058 -0.122 -0.142, mean -0.127, deviation from mean 0.098

Connection established!

You evidentry have varying trigger height with position, which is almost certainly due to the effector tilting as it moves. Possible causes include:

Play in the joints
Diagonal rods in a pair not quite the same length
Bearing spacing not quite the same at the top of a pair of rods as at the bottom (use shims or washers to adjust them)
Bearing lines not quite at 120 degrees to each other because the carriages are not parallel to the towers (check that you have the right number of washers between each carriage wheel bearing and the carriage truck body)

The first two are unlikely if you are using premade rods from T3P3.

Any trigger height variation you can't get rid of can be corrected in the bed.g file by putting H parameters on the G30 commands. For example, if at one of your probe points the trigger height is 0.1mm higher that at bed centre, add parameter H0.1 to the G30 command for that probe point.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

My question is that my nozzle is actually higher than this z height difference that is being measure by the probe ( as you can see by looking at the paper test at probe points)...so is the goal to have 0 deviation from z-trigger set height when i run the check bed?

When i adjust for this by adding the H parameter this makes things worse because i have to add a lot more than just the probe difference for the nozzle to be low enough to squish the first layer (or even be close enough to stick the plastic to the PEI) which then throws off the rest of the leveling

*If this is effector tilt would magnetic rods prove to be a solution?

Quote
Ace67aod
When i adjust for this by adding the H parameter this makes things worse because i have to add a lot more than just the probe difference for the nozzle to be low enough to squish the first layer (or even be close enough to stick the plastic to the PEI) which then throws off the rest of the leveling

I don't understand that. This is how I measure the trigger height at a given point:

1. Use the paper test to establish where Z=0 is, to the nearest 0.025mm.

2. Send G92 Z0 to tell the firmware that the head is at Z=0.

3. Command the nozzle up 5mm, then send G30 S-1. When the probe stops, read off the Z coordinate from the web interface, which is the trigger height at that point. Do this 2 or 3 times to make sure the result is consistent.

4. The H value for that point is the trigger height at that point minus the trigger height at bed centre.

Provided your machine moves to each point in as reproducible manner, I can't see any reason why you would need to further adjust the H parameters.

Quote
Ace67aod
*If this is effector tilt would magnetic rods prove to be a solution?

Only if the tilt is caused by play in the joints or binding joints.

One other thing to check is that the glass is flat. I suggest holding the edge of a metal ruler against it.

Edited 3 time(s). Last edit at 01/22/2016 03:03AM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I must be doing other things wrong because i have done all this and the result is usually a perfect skirt but then the inside of the print wont even stick i just get a ball of plastic at the nozzle. Could I not be snagging my paper correctly? how tight does the nozzle need to grab the paper? Would this even make a difference?

My belts are tight and my standard deviations for each probe points are low (they are consistent)


Compensating with the H parameter improved the outer adhesion but at the cost of the inner so i am at a loss...

My glass is flat but it is also fitted with PEI (painted black) stuck to the glass with 3M 468MP transfer tape

Quote
Ace67aod
I must be doing other things wrong because i have done all this and the result is usually a perfect skirt but then the inside of the print wont even stick i just get a ball of plastic at the nozzle. Could I not be snagging my paper correctly? how tight does the nozzle need to grab the paper? Would this even make a difference?

Sounds like you have made a mistake somewhere. After the last auto calibration, did you remember to copy the new delta radius to config.g along with the other parameters (or print immediately after recalibrating)? Or maybe you have some backlash or other problem that makes the Z height vary with the direction the head has just travelled in. I suggest you chart by checking that at each probe point, the head doesn't snag the paper at Z=0.1 but does at Z=0. In particular, check this at the centre of the bed.

Edited 1 time(s). Last edit at 01/30/2016 05:22PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

once i set Z0 with the paper test using G92 i probed that point using s30 then i moved to the next point and repeated (snagged paper then set zero and then probed) i did this 4 times then averaged the results and set the H parameter in the bed file. Then i start up the printer home and auto-calibrate usually 2 times to confirm consistent results. then i attempt to print. Should i reset the config file with the new M665, and M666 parameters it gives me after i add the H values?

After changing anything that affects the calibration of the machine, I recommend you run autocal about 3 times to check it is giving consistent results, then copy the values into the M665 and M666 commands in config.g. With this done, the calibration will be accurate or at least very close when you switch the printer on. For extra accuracy you can heat the bed and nozzle to operating temperature and run autocal just once immediately before you print. .

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Would more probe points help? i just cannot stick a print using the H parameter unless i over compensate to move the nozzle down enough to squish the print i am going to remeasure z-height and individual probe point trigger heights using the paper snag G92 set and probe (at that point)

So after i remeasure and compensate with the H parameter then i check zero after an auto-cal my paper snags at zero but HARDLY snags... So it seems almost as if this is still not enough to squish the first layer (usually only in the middle the skirt prints fine) is there something else i could be missing to make this happen?

Quote
Ace67aod
So after i remeasure and compensate with the H parameter then i check zero after an auto-cal my paper snags at zero but HARDLY snags... So it seems almost as if this is still not enough to squish the first layer (usually only in the middle the skirt prints fine) is there something else i could be missing to make this happen?

I've never understood the advice to set the height at which paper just snags as 0. This doesn't take into account the thickness of the sheet of paper.

80 gsm paper (your normal printer paper) is pretty close to 0.1mm thick.

Hence the advice I received when calibrating, to set the height at which paper just snags to 0.1mm. If you calibrate your paper snagging height to 0, then the printer head will be 0.1mm too high, which will be especially noticeable on the first layer. If you calibrate to 0.1mm, then your print head will be at the proper height.


I hope this helps.