Rostock V3 Bed issue

[mhackney]

Also, on my V3 I found that if I decreased the probing radius a bit I get much nicer calibration. I observed several early probes (with the original probe diameter) that the effector actually tilted a bit near my X tower. The results were basically the same as what I see for your Z tower. You can use David Crocker's bed file generator to generate the points automatically: http://escher3d.com/pages/wizards/wizardbed.php

Set number of peripheral points and halfway points to 3 and pick the probing radius you want. I used 135mm:

-116.910 Z-probe X1
-67.500 Z-probe Y1
116.910 Z-probe X2
-67.500 Z-probe Y2
0.000 Z-probe X3
135.000 Z-probe Y3

[javiercordero20]

Also, on my V3 I found that if I decreased the probing radius a bit I get much nicer calibration. I observed several early probes (with the original probe diameter) that the effector actually tilted a bit near my X tower. The results were basically the same as what I see for your Z tower. You can use David Crocker's bed file generator to generate the points automatically: http://escher3d.com/pages/wizards/wizardbed.php

Set number of peripheral pants and halfway points to 3 and pick the probing radius you want. I used 135mm:

-116.910 Z-probe X1
-67.500 Z-probe Y1
116.910 Z-probe X2
-67.500 Z-probe Y2
0.000 Z-probe X3
135.000 Z-probe Y3

Any other settings I should check besides those?

Could you also explain how I should use the generated G-code? should I just run it the same way I used the probing.Gcode provided by SeeMeCnC?

Capture.PNG (13.53 KiB) Viewed 15632 times

[mhackney]

You DON'T use the generated gcode, you use the X Y positions in it. "You can use David Crocker's bed file generator to generate the points automatically". I pulled out the points I use in my post so you can just use those.

[ComJak]

What exactly do you do with the points that are pulled from the bedfile generator?

[Th0mpy]

Hey everyone!

So, I don't have a V3 but I did upgrade my V2 with the new HE280 hot-end. So far I'm not getting anywhere near the quality I had been with the old setup. I have a few issues, but most seem correctable. The biggest issue I've seen so far, is with the automatic calibration. After running the automatic calibration I can see clearly that the end-stop offsets are being populated to a seemingly correct value, as well as the radius. If I then attempt to print a circular pattern the area opposite the Y tower is much lower than the rest. The other towers and corresponding opposite hypotenuse (i guess you could call it) all seem to be within an acceptable range.

So far, I've attempted to rule out mechanical issues. The towers are square to the plate. The belts are all tensioned properly and evenly. I've even ruled out the bowden tube and cable coming from above as binding on the arms. Also, the endstop switches are firmly mounted and solid.

I am calibrating cold. If I do not calibrate cold the hotend will crash into the plate OR will fail to complete the calibration and stop after it's probed the first tower.

Terribly frustrating, anyone have any good ideas? Is there something more which needs configuration to allow to calibrate while heated up? Could this be happening because I am not calibrating while heated?

[mhackney]

Do not calibrate while heating with the accelerometer probe. You already observed the results!

Is the bed level issue (sloped along the Z tower) consistent? One gent here had a belt looped over one of the braces and it caused a similar problem. A loose endstop (you checked) as well as a loose carriage trigger screw or even a loose pulley on the Y stepper or loose stepper could all be suspects.

[mhackney]

Oh, and watch the effector carefully when it probes out by the tower. I've actually seen mine rotate significantly, which is why I decreased the probing radius a bit.

[javiercordero20]

What exactly do you do with the points that are pulled from the bedfile generator?

Mr. Hackney could you please explain what we need to do with the points generated?

[Xenocrates]

What exactly do you do with the points that are pulled from the bedfile generator?

Mr. Hackney could you please explain what we need to do with the points generated?

What you do is you would edit the file that describes the bed probing points (bed.g on the duet). These points are inserted as different probe positions, and the rest of the bed.g file is assembled around them. As an example, I took the points, and have M561, G31 X0 y0, and G28 before the points, and then sometimes after the points a probe retract or similar would be needed.

[javiercordero20]

What exactly do you do with the points that are pulled from the bedfile generator?

Mr. Hackney could you please explain what we need to do with the points generated?

What you do is you would edit the file that describes the bed probing points (bed.g on the duet). These points are inserted as different probe positions, and the rest of the bed.g file is assembled around them. As an example, I took the points, and have M561, G31 X0 y0, and G28 before the points, and then sometimes after the points a probe retract or similar would be needed.

Hello,

Alright cool. What file would that be for us using the provided Rambo Board?

[mhackney]

You can edit them in the eeprom.

[Th0mpy]

Do not calibrate while heating with the accelerometer probe. You already observed the results!

Is the bed level issue (sloped along the Z tower) consistent? One gent here had a belt looped over one of the braces and it caused a similar problem. A loose endstop (you checked) as well as a loose carriage trigger screw or even a loose pulley on the Y stepper or loose stepper could all be suspects.

The slope seems consistent. What's weird is that when I had the old hotend I was getting really good prints, I didn't seem to have this issue after I manually calibrated the end-stop screws. After HOURS of manual calibration I even put a little locktite on the endstop adjustment screws to prevent them from wandering after it was set. Literally nothing else has changed except the firmware and hotend. So I don't think it's a pulley/loose stepper, but I will confirm tonight.

Oh, and watch the effector carefully when it probes out by the tower. I've actually seen mine rotate significantly, which is why I decreased the probing radius a bit.

What would be the cause for this? I might have noticed that but, again, i'll have to check when I get home. Which specific setting would adjust the radius; or is that simply the z-probing coordinates?

One final thing I forgot. After many hours last night, and a few beers, I decided to adjust the endstop screw out a ways. I did not runt the calibration after that and printed. It got better, not 100% exact, but I suspect that if I were to run the calibration again it would go back to the same thing. It does seem to point to a mechanical issue or a misconfiguration somewhere I just can't find it.

[mhackney]

The accelerometer probe contacts the bed with a lot of force. It was startling the first time I watched it! In my case, it looks like the ball cup joint separates a bit and this is worse when the arm is near vertical like it is with the default stock probing points. Bring them in a bit, the arm has a bit of angle and the ball does not pop out of the joint as frequently (I've had it happen once though).

Keep working on it, something will show up and you'll have that Aha moment!

[Th0mpy]

The accelerometer probe contacts the bed with a lot of force. It was startling the first time I watched it! In my case, it looks like the ball cup joint separates a bit and this is worse when the arm is near vertical like it is with the default stock probing points. Bring them in a bit, the arm has a bit of angle and the ball does not pop out of the joint as frequently (I've had it happen once though).

Keep working on it, something will show up and you'll have that Aha moment!

That is interesting!! I will have a look at that and report back.

I suppose I should get to work now, company time on a company computer, thanks for the feedback!

[mhackney]

This is what the eeprom settings look like in OctoPrint (if you install the plugin). You can also see/edit them in Repetier and MatterControl.

[Th0mpy]

The accelerometer probe contacts the bed with a lot of force. It was startling the first time I watched it! In my case, it looks like the ball cup joint separates a bit and this is worse when the arm is near vertical like it is with the default stock probing points. Bring them in a bit, the arm has a bit of angle and the ball does not pop out of the joint as frequently (I've had it happen once though).

Keep working on it, something will show up and you'll have that Aha moment!

That is interesting!! I will have a look at that and report back.

I suppose I should get to work now, company time on a company computer, thanks for the feedback!

Alright, I spent the better part of 4 hours banging away. I was able to fix one of my issues. I found out that the bowden tube did not get inserted all the way into the hotend. I had to make a small modification to the top nut, there was a bit of a burr which I had to trim in order to allow the tube to get inserted all the way. Anyway, you live and learn.

So, I pulled the probing points in a little bit. It didn't seem to make a difference. I checked and re-checked the pulleys, belts, and endstops. Any other suggestions? I am clean out of ideas.

[DeltaCon]

No suggestions to your problem, but just a small tip.

there was a bit of a burr which I had to trim in order to allow the tube to get inserted all the way.

I always use a pencil sharpener to "unsharp" the edges of the bowden tube a bit. That makes inserting it all the way a lot easier.

[mhackney]

+1 on DeltaCon's tip. I do the same or trim with a sharp hobby knife.

[mhackney]

Th0mpy, can you run calibration 3 times and after each, record the values saved to EEPROM so we can check for variability in the calibration process.

[Th0mpy]

So I was thinking of what could be the issue last night as I fell asleep. Is it possible that the radius of the Y arm is a bit shorter than the others? If that's the case could I set the delta radius correction for that tower?

[mhackney]

Possible but unlikely. I'd like to see if you are getting reasonable run-to-run reproducibility with the current setup.

[Th0mpy]

Th0mpy, can you run calibration 3 times and after each, record the values saved to EEPROM so we can check for variability in the calibration process.

Are you saying the endstop offset?

One thing I did last night which I thought was interesting... I used my iphone and did a slow-mo video of the hotend probing the glass, to see if there was any shifting of the platform. I ran the test about 4 times and while I was not specifically looking for changes in the offset it didn't seem to move about. I will do a better test tonight and focus on those measurements and report back.

[mhackney]

I'd like to see the results of all the calibration parameters:

x,y,z tower offsets
delta radius
z height

[Th0mpy]

I'd like to see the results of all the calibration parameters:

x,y,z tower offsets
delta radius
z height

Ok, here's what I got. What do you think?

Test1:

• X offset - 5
  Y offset - 3
  Z offset - 0
  Radius - 143.423
  Height - 369.7

Test2:

• X offset - 6
  Y offset - 4
  Z offset - 0
  Radius - 143.423
  Height - 369.7

Test3

• X offset - 6
  Y offset - 3
  Z offset - 0
  Radius - 143.437
  Height - 369.712

[mhackney]

Those are very consistent, which is good. Interestingly, I'm seeing a pattern with several of us having very similar "level" issues. I also ran Z height calibration a number of times in a row and it was dead on each time.

I am experiencing exactly the same thing you are also along the Y tower direction. I want to do this test: auto calibrate and test print to verify the tilt and then manually calibrate and compare the results.