Delta printer does not print the full height on the outer edges

TBS

@fcwilt: I will make a video later today in the evening.

dc42

I haven't yet worked out what might be causing that problem. Nobody else has reported it, and I don't see anything similar on my delta.

What happens when you home the printer? Your homedelta.g file homes the towers, then commands Z to move down by 5mm, then centres the effector. Do the carriages drop by 9cm during this process? If so, at what step?

Here's the code that limits the height, as at firmware version 2.02:

const float diagonalSquared = fsquare(coords[X_AXIS]) + fsquare(coords[Y_AXIS]);
...
if (applyM208Limits && coords[Z_AXIS] < reprap.GetPlatform().AxisMinimum(Z_AXIS))
{
	coords[Z_AXIS] = reprap.GetPlatform().AxisMinimum(Z_AXIS);
	limited = true;
}
else
{
	// Determine the maximum reachable height at this radius, in the worst case when the head is on a radius to a tower
	const float maxHeight = homedCarriageHeight - sqrtf(D2 - fsquare(radius - sqrtf(diagonalSquared)));
	if (coords[Z_AXIS] > maxHeight)
	{
		coords[Z_AXIS] = maxHeight;
		limited = true;
	}
}

Variable homedCarriageHeight is set to the height at which each carriage triggers the endstop switch, corrected by the M666 endstop adjustments. D2 is the square of the rod length.

As it happens, I've just replaced that block of code as part of extending the delta kinematics to support up to 6 towers all with different rod lengths. So I can provide a new binary for you to try. But I'd rather find out why the 2.02 firmware isn't working for you first.

bot

In the picture you post, is your effector centered while at its maximum height? Both in terms of XY position and real-world position relative to your towers.

TBS

Here is the announced video:
https://vimeo.com/313250184

Process:
The nozzle is homed with a height of 1024.5mm.
The nozzle moves in horizontal direction with G1 Y-350.
Than the printer shows a height of about 899mm.

But it is from the hardware arms a height of 990mm possible.
This height is never reached.
You can see that the arm is about 91mm under the maximum height (end stop).

Note:
If I execute G1 X-400 there are about 30mm more possible with the arms.
If I execute G1 X400 there are about 30mm more possible with the arms.
If I execute G1 Y400 there are about 5mm more possible with the arms.

TBS

@fcwilt: After reading the comment from dc42 I understood what you meant.
This is probably the internal code implementation.

@dc42: If you have the new implementation for the dynamic height calculation you can send them to me by e-mail to test it.
But you are right. The I2C error is much more important. I would like to start my 430 hours printjob

fcwilt

@tbs said in Delta printer does not print the full height on the outer edges:

You can see that the arm is about 91mm under the maximum height (end stop).

What arm does the video show?

Frederick

TBS

@fcwilt: The video shows the right arm on the front or the theoretical position is about X500 Y-500.

The left arm on the front is similar with 91mm.

fcwilt

Did you run the autocal feature when the homing file was as you posted?

Frederick

dc42

I've watched the video.

Can you confirm that when you home the printer, all 3 carriages go right up to the endstop switches, then drop about 5mm?
When you commanded Y to -350, did all 3 carriages drop? The video only shows one of them.

TBS

Yes, i can confirm that all three carriage trigger the end stops and move 5mm down.
The video is recorded in the front position.
The carriage in the front left moves about 91mm down.
The carriage in the front right moves about 91mm down.
The carriage in the background middle moves more than 300mm down. But I have not messured the exakt value for the background carriage, yet.

bot

Immediately after homing a delta printer, you can't freely move in X/Y without some moves causing a carriage to crash UP into an endstop.

This move shouldn't produce the behaviour, but perhaps the printer is moving down to a level that allows full X/Y movement without the carriage crashing up.

EG, if you gave the command G0 Y350, immediately after homing, the carriage would crash. Has the firmware been updated to prevent "illegal" moves of delta kinematics? IE ones that would put the printer at a height where the XY plane is not fully and freely accessible?

Danal

This may or may not have anything to do with your problem:

The dimensions given in the M665 in the config-override.g you posted seem odd.

M665 L1021.500 R527.601 H1030.482 B400.0 X-0.139 Y-0.828 Z0.000

Specifically, B (printable bed radius) of 400 with a R (Delta radius, measured horizontally, measured "joint center to joint center" of 527, this seems a bit off. The point at which a diagonal arm pivots on the carriage is 157mm beyond the edge of the bed? Normally this is a gap just large enough to run a belt...

Also, this makes your R (Delta Radius) more than half of your diagonal rod length. This would imply that the rods are at less than a 45 angle when everything is centered. Yet we can see in the video this is not true.

Again, may or may not be related to the problem...

Danal

As an aside to this:

If you have 1021mm arms, it should be easily possible to achieve a print radius of 445mm, Going all the way to 450 causes the rods to be below a 20 degree angle in their "worst case" position.

TBS

@fcwilt said in Delta printer does not print the full height on the outer edges:

Did you run the autocal feature when the homing file was as you posted?

Frederick

I have run the autoleveling routine and it was the same result.

@danal said in Delta printer does not print the full height on the outer edges:

As an aside to this:

If you have 1021mm arms, it should be easily possible to achieve a print radius of 445mm, Going all the way to 450 causes the rods to be below a 20 degree angle in their "worst case" position.

I have changed the R parameter to 450, but it is the same result.

fcwilt

@tbs said in Delta printer does not print the full height on the outer edges:

@fcwilt said in Delta printer does not print the full height on the outer edges:

Did you run the autocal feature when the homing file was as you posted?

Frederick

I have run the autoleveling routine and it was the same result.

But did you run delta calibration with the homing file as you posted?

That homing file has an error in it which can mess up the delta calibration procedure.

Danal

Please post a few photos of your bed & towers, with a tape measure or ruler. I am specifically interested in the physical bed size, and the offsets edge of bed to belt, and edge of bed to tower.

Like this:

TBS

@fcwilt said in Delta printer does not print the full height on the outer edges:

@tbs said in Delta printer does not print the full height on the outer edges:

@fcwilt said in Delta printer does not print the full height on the outer edges:

Did you run the autocal feature when the homing file was as you posted?

Frederick

I have run the autoleveling routine and it was the same result.

But did you run delta calibration with the homing file as you posted?

That homing file has an error in it which can mess up the delta calibration procedure.

Yes, the files that I have posted are the current leveling files. Where are the errors in the file?

@danal said in Delta printer does not print the full height on the outer edges:

Please post a few photos of your bed & towers, with a tape measure or ruler. I am specifically interested in the physical bed size, and the offsets edge of bed to belt, and edge of bed to tower.

fcwilt

@tbs said in Delta printer does not print the full height on the outer edges:

Yes, the files that I have posted are the current leveling files. Where are the errors in the file?

The last three lines where you move down and center. You don't want to do that part.

Just leave them up triggering the endstops - that is what the delta calibration procedure assumes for a starting position.

I had that same centering code and could not get my delta to print well. I finally noticed in the section of the duet help on delta calibration that I was not supposed to do that. I removed it and ran calibration and the printer then printed just fine.

I'm not saying this is directly related to the problem we have been discussing but it is something you should fix and then run calibration again.

Frederick

Danal

There really is 120MM+ gap belt to the edge of the bed. Maybe more like 140MM to the tower itself. So those M556 numbers are better than I thought.

As others have said, clean up the homing file, re-calibrate, and see...

dc42

I'll try to make time to reproduce this over the next few days.