effector tilting into bed on print

dc42

Your Y tower angle offset of more than 4 degrees sounds suspicious. If the 3 sides of your base and upper triangles are all of equal lengths, it should be much smaller than that.

Your endstop adjustments are also rather large, assuming you attempted to get all 3 endstop switches at more of less the same height (getting them equal to within 1mm isn't normally difficult).

What type of Z probe are you using?

TimVukman

When you mention y tower offset, are you referring to the -4.385 that was returned in M665?

This was an FLSun kit, so my verticals are all the same length(3) as are each of the cross members(9)

As I recall, the end stops were not too hard to adjust. Maybe I should redo them after all the moving I have been doing?

dc42

And what type of Z probe does it have?

TimVukman

The probe is a 3 pin switch that can be set to either normally open or normally closed.
The switch is on the side of the effector and is triggered by pushing the nozzle into the bed.

Is the Duet effector with carriages hard to install? I would like to stop pushing the nozzle into the bed to trigger it. It seems to really push on it around the outer diameter of the bed when all the rods are on an angle.

T3P3Tony

@timvukman The Smart effector has carriage adaptor that are sized to screw into "standard" linear rails that alot of the wheeled carriages are modelled on as well:

https://duet3d.dozuki.com/Wiki/Smart_effector_and_carriage_adapters_for_delta_printer#Section_Carriage_Adaptor_Dimensions

You will need to do some requiring to use it:
https://duet3d.dozuki.com/Wiki/Smart_effector_and_carriage_adapters_for_delta_printer#Section_Connectors_and_Wiring

dc42

@timvukman said in effector tilting into bed on print:

The probe is a 3 pin switch that can be set to either normally open or normally closed.
The switch is on the side of the effector and is triggered by pushing the nozzle into the bed.

I suggest you watch carefully as it probes the bed, to see if you can see any reason why the height at which the probe triggers may change with different XY positions.

TimVukman

@t3p3tony
Hi:
I have reviewed that documentation in quite a bit of detail. As I read it, it indicates that it would support rails or wheels. My current factory setup has three wheels spread over a fair amount of vertical carriage height. The measurements for the duet carriages indicate that the wheels would be vertically closer together which I feel would be more stable. I have lots of extra wheels, so I feel that adding a fourth wheel would be fairly straight forward.

I am not a huge fan of trying to install crimp connectors on small wires / sockets. I'd like to feel comfortable that these are large enough to work with. I'm not as good as I once was with these things.

TimVukman

@dc42
I would think that the height must change at various XY positions to some degree, by virtue of the diagonal rods following pendulum arcs. I thought that was why there was an H value in the G30 codes so that offsets could be put in to level the effector.

I think I am still confused between calibration and height adjustments. The documentation shows a beautiful light green flat surface for the mesh. I get something that looks more like a coffee cup with lots of blue. I think it is all indicated above.

I honestly think that my tilting problem was the nozzle contacting the bed because the plastic clip on the bottom rail between the X and Z columns was not sitting flat. On that basis, I think it is "fixed"

I guess I don't get what Auto Configuration means. It doesn't seem like it is automatic since it changed values and now put the nozzle off the bed diameter.

dc42

If the height map is cup shaped, then most likely your delta radius is out. But that wouldn't result in an error of 4 degrees in one of the tower positions. Perhaps you should try using a smaller probing radius, at least for now.

TimVukman

Ok
I started over from the beginning.

I home my printer. I took out the little screws from the top of the carriages. My switches click at pretty much the exact same time.

My Z shows as 245 in DWC
I move Z down until it traps a piece of paper.
My Z shows as 34 in DWC
I issue a G92 Z0 to set that as my 0 height
I issue G30 S-1 to get my trigger height of 1.17
I issue an M500 to save the settings.
I power off the printer to give it time think about what it's going to do when I power it back up.

Power Up
Home All
Z registers 245

I have adjust my height to 216 since when I home it, it drops by 5. Once there, the actual distance to the bed surface is 211. If I send my printer z to 0, it is perfectly aligned on a sheet of paper.

I thought I should be able to print from here, but I can't. The nozzle digs into the bed.

So, I guess I need to run auto config, even though that screwed up all my settings last time.

No idea what I should do next.

Obviously there are people who have working delta printers. Also obvious is that some of these working printers are also using these boards as well..

Unless someone has a suggestion I am going to leave it for a bit, until I get the duet effector / carriage kits.

adavidm

@dc42 his probing radius is 10mm according to that bed.g. that doesn't look right but I don't think going smaller is going to help....

I think there is confusion over Cartesian coordinates Vs the naming of the three towers as X,Y and Z and that is causing a mismatch between expectation and reality, both in understanding the G31 offsets and the general movement of the effector.

adavidm

@timvukman

Tim, a couple of things to make sure we have the same terms of reference. Please don't take these questions as being patronising, it's just vitally important we are talking about the same things.

are you aware that the X, Y and Z labels of the towers have nothing whatsoever to do with the movement of the effector in Cartesian (X,Y,Z) space when talking about delta printers?

1a) if the answer to 1 is "yes" can you describe how you have connected the towers in terms of their relative position? In standard configuration, a single tower, Z is at the back in the centre. X is at the front-left and Y is at the front-right. Anything else will cause problems when we are giving you advice.

the X and Y offsets in G31 relate to Cartesian coordinates and have nothing to do with the towers.
you've configured a radius of 10mm in the bed.g configurator, was this intentional? That radius should be as large as you can reasonably manage without colliding with towers.

It sounds like you are very close to having a working solution so please don't give up. The magnetic arm solution with smart effector is awesome but it's essentially only going to solve the G31 offset issue you have, IMO.

adavidm

dc42

@timvukman said in effector tilting into bed on print:

I issue G30 S-1 to get my trigger height of 1.17
I issue an M500 to save the settings.

M500 does not save the trigger height in recent firmware versions. You need to edit the G31 command in config.g, setting the Z parameter to 1.17.

TimVukman

@adavidm
You cannot offend me and there is no reason to be concerned about what you say.
It is most obvious to me that there is a lack of understanding on my part. I am disappointed in that, based on how many times I have gone through the documentation and how many you tube videos I have watched. Having said that, my goal is to learn.

I realize that I can't say that something is moving in and X or Y direction relative to how I know those Axis in a cartesian model. I am trying to use the X Y and Z columns when describing directions of effector heading, unless I am using Z in the context of Height where 0 is the surface of the bed with the nozzle resting on a piece of paper. I have my printer facing me on a desk where I am looking between the X and Y columns with the Z column centered in the rear.
1a. I had to switch my X and Y column motors and end stops to match the description that I was given here, but X is to my left face on, and Y is to my right and Z is at the back.
I haven't really paid much attention to those offsets. I see that 11 in the X reference, but I have no idea why it is there. My bed is 220mm in diameter. From the DWC, the position reports X0 and Y0 when homed. I measured across X in the cartesian directions and my nozzle is at 120mm. I would think it should be 110mm to center. My Y value measures as 120mm front to back. I would think it should also be at 110mm
I suspect you are viewing an older version of bed.g. I noticed that and went back to the configurator and changed it to 100, to make sure I had a band around the perimeter of the bed that would keep my effector from contacting any of the drive belts or bed screws.

I sincerely appreciate the assistance and I will follow any instructions given to help isolate, troubleshoot and resolve this. Inclusion of information as to the "what and why" something is expected as a result for any part of this are appreciated

Bed.g is now:
; bed.g file for RepRapFirmware, generated by Escher3D calculator
; 10 points, 6 factors, probing radius: 100, probe offset (0, 0)
G30 P0 X0.00 Y100.00 Z-99999 H0
G30 P1 X86.60 Y50.00 Z-99999 H0
G30 P2 X86.60 Y-50.00 Z-99999 H0
G30 P3 X0.00 Y-100.00 Z-99999 H0
G30 P4 X-86.60 Y-50.00 Z-99999 H0
G30 P5 X-86.60 Y50.00 Z-99999 H0
G30 P6 X0.00 Y50.00 Z-99999 H0
G30 P7 X43.30 Y-25.00 Z-99999 H0
G30 P8 X-43.30 Y-25.00 Z-99999 H0
G30 P9 X0 Y0 Z-99999 S6

TimVukman

@dc42
Thanks for the heads up on M500. I have manually entered the trigger height in the G31.

Danal

@timvukman said in effector tilting into bed on print:

I realize that I can't say that something is moving in and X or Y direction relative to how I know those Axis in a cartesian model.

You probably already get this, but just to be sure:

Actually, you can and should state cartesian X and Y directions, when referring to X or Y moves caused by G-Code and/or the movement buttons in the web interface. The tip of the nozzle absolutely DOES move in cartesian space. The firmware moves the carriages on the towers in varying amounts and directions to make this true.

Example: Assume the nozzle is a Z50 an X0Y0. Command a move to Y10. Two of the towers will move down, and one up, in different amounts, so that the nozzle moves to X0Y10Z50. The moves will be "timed" so that the Z and X planes never deviate during the move.

This is how Delta's work.

I am trying to use the X Y and Z columns when describing directions of effector heading, unless I am using Z in the context of Height where 0 is the surface of the bed with the nozzle resting on a piece of paper. I have my printer facing me on a desk where I am looking between the X and Y columns with the Z column centered in the rear.

This is all correct when referring to the TOWERS. To repeat with different phrasing: TOWER X is NOT in any way related to the way the nozzle moves when it moves in cartesian X. In fact, some Delta firmware refers to the towers as Alpha Beta Gamma just to avoid any possible confusion.

1a. I had to switch my X and Y column motors and end stops to match the description that I was given here, but X is to my left face on, and Y is to my right and Z is at the back.

Good. That will make all of us be "on the same page" when describing things.

I haven't really paid much attention to those offsets. I see that 11 in the X reference, but I have no idea why it is there.

This means your probe is 11mm in the (cartesian space) X direction from the tip of the nozzle.

My bed is 220mm in diameter. From the DWC, the position reports X0 and Y0 when homed. I measured across X in the cartesian directions and my nozzle is at 120mm. I would think it should be 110mm to center. My Y value measures as 120mm front to back. I would think it should also be at 110mm

Agreed.

Bed.g is now:
; bed.g file for RepRapFirmware, generated by Escher3D calculator
; 10 points, 6 factors, probing radius: 100, probe offset (0, 0)
G30 P0 X0.00 Y100.00 Z-99999 H0
G30 P1 X86.60 Y50.00 Z-99999 H0
G30 P2 X86.60 Y-50.00 Z-99999 H0
G30 P3 X0.00 Y-100.00 Z-99999 H0
G30 P4 X-86.60 Y-50.00 Z-99999 H0
G30 P5 X-86.60 Y50.00 Z-99999 H0
G30 P6 X0.00 Y50.00 Z-99999 H0
G30 P7 X43.30 Y-25.00 Z-99999 H0
G30 P8 X-43.30 Y-25.00 Z-99999 H0
G30 P9 X0 Y0 Z-99999 S6

This looks pretty good to me. It should result in a circle of 6 probes about 100 mm from center (which may be off a little from what you said about centering), then a circle of 3 more about 50 mm, and one in the center. (That could be off, I'm eyeballing the numbers). It will then calculate just about everything except the diagonal rod length because of the S6 in the last command. That's the way I do it as well (10 points total, calc with S6) .

So that all looks good.

What does it do? What is the response you see in the Gcode console? I see things like: (where the M500s are me writing it to config-override.g)

1:13:09 PM M500
1:12:13 PM G32
Calibrated 6 factors using 10 points, deviation before 0.043 after 0.033
1:11:20 PM M500
1:11:07 PM G32
Calibrated 6 factors using 10 points, deviation before 0.504 after 0.045

TimVukman

Ok. I am pleased that we got on to common terms for things. I am not convinced that X11 is correct, but I will deal with that a little later.

At this point, I have not run Auto Calibrate since resetting everything. I will run it.

I will run it in the morning. It blew up my system. My bad, I didn't home it first. Now things don't seem very happy in printer land.

I powered it back up, ran the G32. It seems to be moving outside of the bed, so I shut it down.

I will report in the morning

TimVukman

@timvukman
Hi:
I have reduced the radius to 95 which the printer seems happy with. I do have an issue where one of the diagonal rods contacts the housing for the Z switch on the effector. It's light contact at 95. At 100 it popped off the magnetic rod.

I ran auto config this morning with new values in bed.g based on the change in radius to 95

G32
Calibrated 6 factors using 10 points, deviation before 5.928 after 0.307

The before number sure looks out of whack, no?

dc42

I suggest you adjust one or more probe points to avoid that contact between the rod and the Z switch.

The 0.307mm deviation is rather large, however as the before-figure is to high, it may come down if you run 1 or 2 more cycles of auto calibration.

adavidm

@timvukman

Don't worry too much about that "before" value right now. That is based on the figures in your config.g and your config-override.g. We know they are some way out because you can't print at the moment.

Drop your radius to 90mm, you don't want any kind of contact. The reduced probing radius will not be an issue right now. EDIT - dc42 has suggested a better way by simply changing the coordinate of the problem point. If you are comfortable doing this then go ahead, my method should work if you are not feeling confident, however.

My suggestion is that you do the following:

Home all axes - either with a G28 or by clicking Home in the web interface (DWC)
Run calibration - either with a G32 or through the web interface.
WITHOUT HOMING - run G32 again.
Save to config-override.g by running M500

Once that's done, paste the complete content of config.g and config-override.g here, along with the results of the two G32 commands, above. You've changed quite a few things so worth getting an up-to-date look at your config.

It sounds like you are really close, so keep going!

adavidm

--edited for readability