Posts made by Jorge

I have found the problem. It was not the firmware, but my settings. As a test, I had set the jerk for the frequency measurement to 0. Now everything works again

Hi,
I wanted to test input sahping and had to update my firmware for this to 3.4.5. Before that I had a version running below 3.3. I'm not sure anymore, something with 2.x.
Since the upgrade, I now have a problem printing circles and arcs. Whether the radius is very large or very small, the printer starts to hum extremely, as if it stops for a millisecond between each segment. The print result looks accordingly. Input shaping is not activated.
Is it possible that this is due to the new firmware? Is there perhaps some function that recognizes curves and can lead to such problems?

Unfortunately the link is no more active. Neither to the V6 clone nor to the cyclops clone or the NF hotend.
But most tests were made with the V6 clone.
I already ordered a coated heatbreak from mapa refining here in Germany and a new bowden. I think the heatsink has no effect as long as the heatbreak and the nozzle are good. And I have a lot of nozzles here.

I printed PLA at temperatures down to non extrusion.

I have a hotend from aliexpress. There was no B Value mentioned.
For the calibration I removed the filament and the bowden. Then I put the temperature sensor of my multimeter from the top into the nozzle. In the nozzle was some molten filament left, so there was a very good thermal connection to the nozzle. Then I set 230 °C and tweaked the value until the printer and the multimeter showed exactly the same value. I assume, that I eliminated any thermal transitions and tolerances this way, as I measured the temperature of the molten filament and not the heatblock or so.

Thank you for that clue! You were right! I had to calibrate P1 and landet at B3980. The hotend was more than 20C to hot (at 230C)! But since I tested a lot of different temperatures, this didn't help for my problem

Finally I found the time to change the board and it did not help. Same blobs with Marlin. So it has nothing to do with the Duet or RepRap Firmware which means it must be a mechanical issue. I am kind of happy about that, because I love everything about the Duet. But now I have no idea how to solve the problem.

@danl said in Blobs at every start and end of line:

Could it be a bowden related mechanical issue, e.g. the curvature or a slight kink or dent of the tube, a loose tube/end fitting connection, slight play in the tube-to-extruder fitting......?

I assumed something like that, too. The bowden is about 50 cm long, which is longer, than I'm used to. That's why I try made a test with a very short bowden, allowing me only to print on one corner of the print area. Again no difference.
Watching the tube-to-extruder fitting while printing didn't bring the enlightment, too.

@dc42 said in Blobs at every start and end of line:

Blobs at the end of a line are typically caused by lack of pressure advance or insufficient retraction. Blobs at the start of a line are not normal. Are you sure that you don't have a positive "extra restart distance" or similar configured in your slicer?

Makes sense. But I even set ...

@jorge said in Blobs at every start and end of line:

... a negative restart distance until I got a fatal underextrusion at the beginning of a line but though I have a blob. So first a blob and then underextrusion for a few mm. This confuses me a lot.

@dc42 said in Blobs at every start and end of line:

But did you also set the jerk policy to 1 ?

Yes, no visible difference

I did not find the time to change the board yet but I did some coasting and restart distance tests. I set a high coasting distance and a negative restart distance until I got a fatal underextrusion at the beginning of a line but though I have a blob. So first a blob and then underextrusion for a few mm. This confuses me a lot.

I updated the firmware but unfortunately it makes no difference.
Next I will test it with a simple 8 bit board with Marlin. I hope it makes no difference because I really like the duet and the Dibond housing is milled for it...

@nightreaver said in Blobs at every start and end of line:

OT: You left your wifi password in there?

Thats an old password, so no worries
But I just deleted ist though.

@dc42 said in Blobs at every start and end of line:

IMO you are on the wrong track. Using a Bowden extruder, if you don't retract on layer change and you don't use pressure advance, you will get a blob unless the layer change happens very fast. Likewise if you don't retract before a travel move.
Are you absolutely certain that you were not using any retraction or pressure advance on your other printer? Retraction may be done by the slicer generating direct retraction commands, or by the slicer generating high-level commands, or perhaps by the firmware.

Yes, pressure advance is deactivated for sure. And the GCODE was the same file. So there is no retraction in the GCODE. I double-checked it in a text editor. I have Marlin running on the other printer. I never heard, that it is doing any retraction or so by firmware?

Here is one more test: I printed the same body again, once at the same speed as before (travel= 500mm/s, Infill = 60 mm/s, perimeter and testline = 45 mm/s), and once at 20% speed.

It's interesting, that the infill of the faster one is underextruded compared to the slower one, even though 60 mm/s is not so fast. And obviously the blob almost disappeared at the slower one. I conclude that the hotend must be not slippy enough and there must rise a big pressure in the bowden. With this pressure even the shortest stop is enough to build this blob. Maybe the hotend of the other printer is kind of slippier, so that less pressure grows when printing at 60 mm/s? Maybe I should try one more hotend or just an other nozzle? What do you think?

Thank you guys so much for your help! According to your posts I did some more tests but without success yet.

@change3d-gmbh said in Blobs at every start and end of line:

Did you test if enabling retraction reduces the amount of blobs or do you just assume that the blobs will get worse?
Because I could imagine retracting could prevent that too much plastic gets out of the nozzle after the print move finished and before the nozzle takes off to another location.

Now I did
Actually the retraction shifts the problem a bit from the starting point to the ending point. So the blob in the end of a line becomes even worse as the nozzle stays there for the time of retraction. The starting point becomes a bit better but far away from good. I tested this also with different retraction distances. The longer I retract, the worse is the blob in the end of a line.

@dc42 said in Blobs at every start and end of line:

Pressure advance reduces the need for retraction; but with a Bowden extruder, you will always need some retraction to avoid blobs around travel moves.

But the black part is printed with a bowden extruder and without retraction and pressure advance. I just don't get why both printers don't work the same way.

@deckingman said in Blobs at every start and end of line:

@jorge i think (and sincerely hope) that your quoted speeds are mm/minute because 7,000 mm/sec is really knocking on.

Sorry, my fault.

@deckingman said in Blobs at every start and end of line:

Anyway, assuming you mean mm/sec then the blobs you describe can usually be tamed with pressure advance - it works well for me anyway. I'm not sure about the correlation between RepRap firmware and Marlin though. I don't think Marlin "K" is the same as RRF M572. Comparing your extruder settings with mine (I use Bondtech BMGs but the gearing and steps per mm are very similar to the E3D Titans) I have lower accelerations (3,000 mm/min compared to your 9,000), but higher "Jerk" (I use 3600 compared to your 2400). My maximum speed setting is lower but that's irrelevant because it will never reach 3600 mm/minute let alone 9,000.
So higher extruder "jerk" will have an influence on pressure advance. Also, anything that alters the viscosity of the filament. I see that your corners have quite a bulge which might be an indication that the filament temperature is too high. That will cause blobs too. Have you tried lower temperatures and are you sure that your hot end thermistor has the correct values so that it is reading correctly?

I took over your acceleration and jerk and printed the test body once without pressure advance and once with k = 0.17 but the test body still looks the same.

The bulge of the corners confuses me, too. I decreased the temperature until it nearly didn't extrude at all. And indeed the corners became a bit better but the blob is still there.

@deckingman said in Blobs at every start and end of line:

I'm not sure that I understand your comments about not using retraction. It is normal to have to retract filament just prior to non-print moves. Pressure advance reduces the amount of retraction required but it doesn't eliminate the need for it completely.

From what I have gathered, retraction ist mostly important to avoid stringing and oozing when it comes to crossing open spaces. So if it's possible, I like to avoid retraction. So I usually set "Only retract when crossing open spaces" in Simplify3D and I deactivate "Force retraction between layers". With this settings a simple body like a block will be printed without any retraction. And for my other printers this works very well but the new one builds big blobs on every layer change.

So my current aim is to get the new printer to print those simple bodys as good as my other printers without retraction and actually without pressure advance. Only afterwards I wanted to calibrate retraction. Do you agree or am I on the wrong track?

@deckingman said in Blobs at every start and end of line:

It might be an idea to post the rest of your config.g file and the slicer settings you use.

sure
here the config.g, the S3D .fff profile and the S3D setting as images for those without S3D. Hope it is OK this way.
download

Hi guys,

my new coreXY prints not so bad but I have a big blob issue, which I don't know like this from any other of my printers.

First I thought it is an retraction issue but it is not as you will see in the following.

I have a bowden setup and I tested it with the geared Titan extruder and with a non geared extruder. And I tested different hotends, filaments and temperatures as well. All this has no significant effect.

To make sure my Titan extruder is fast enough I set pretty high values for E-speed, -acceleration and -jerk:
E-acceleration = 9000 mm/s²
E-jerk = 2400 mm/min
E-max-speed = 9000 mm/min

I can't here it loosing steps but I tried more conservative values as well with no big differences.

I assumed, that the nozzle stays too long at a place, so that oozing has time to build the blobs. So I tested different XY accelerations. Non print moves are always set to 7000 mm/s². Print moves are usually set to 2500 mm/s². And here the results for different print move accelerations:

As I said there is no retraction. The nozzle moves from the corner below right to the beginning of the line and instantly continues printing. I can not notice any delay. A high-speed camera would be interesting ...
You can see, that higher accelerations are helpful. But even at 7000 mm/s² I have a little blob. And I don't want to accelerate that fast, as I will have more ringing then.

So back to 2500 mm/s² and let's test pressure advance. I used the calibration pattern from the marlin firmware homepage and ended up with k = 0.35. But this value caused dramatic underextrusion. So I made a small test GCODE by hand. First I print a base and then I move from the corner below right to the middle and print a single line on it. Again no retraction activated:

The black part is the same GCODE printed on a widely similar machine but with marlin - no blob.
As you can see, the blob disappears at k = 0.18 but even at k = 0.15 I have underextruded areas. So I'm afraid pressure advance is no solution for my problem.

Now I can't think any further myself. The biggest difference between the printer of the black and grey part ist the electronics and the firmware. But I assume, that both translate the GCODE to the same movements without differences, right? I mean I can't imagine, that the duet includes any additional delays here!?

I know I could set my Slicer (S3D) to force retraction at almost every travel move and layer change. But this makes the problem even worse, as the nozzle stays even longer at every start- and endpoint of each line. Plus I want to understand why I have this problem only with this printer. I hope so much for any advice.

I could verify it. changing extruding speed from 0 allows no jerk. But when changing the speed from one direction to another, the jerk works fine.

Ah OK. I will write a little gcode file to test this.

I am pretty sure, that marlin behaves different. For my marlin printers I was able to drive single extrusion moves while increasing jerk by hand until I could hear the extruder loosing steps.

So Sorry, I always meant mm/min.
So the homing is much slower than specified.
M203 is set to 60.000 mm/min.

Where you able to make the stall detection trigger at all? I know it will unlikely work, but I would like to test it though. But I think I must do something wrong, as it never triggers, regardless of the threshold.

But the extruder jerk/acceleration/speed should not affect travel moves, right?
I just increased the extruder acceleration and jerk.
XY acceleration is still very low.
But XY jerk ist very high. But still the jerk seems to be ignored and the ramp starts from 0 mm/s.

I think my coreXY is what you mean?

It works very good and is a cheap solution though. But you are right, power loss would be a problem. Luckywise we never have this problem in Germany

Spoken from experience: Make sure to couple both sides mechanically! In a previous version I used two separate motors and had a big issue with homing repeatability.

But my understanding was, that the jerk sets an initial speed.
In the posted link it is not considered.
So if I set the jerk to 20 mm/s and I want to travel at 20 mm/s, there should be no ramp, because the travel already begins at 20 mm/s. Or am I wrong?

Hi again,

I have read, that the stall detection is not working so good on a coreXY as they have long belts and therefore soft current edges. But I wanted to give it a try though.

In my config.g stands:
M574 X1 Y1 S3

And my homing script is:
G91 ; relative positioning
G1 S1 X-247 F2400 ; move quickly to X axis endstop and stop there (first pass)
G1 X5 F6000 ; go back a few mm
G1 S1 X-247 F360 ; move slowly to X axis endstop once more (second pass)
G90 ; absolute positioning

The problem is, that no stall is detected. I tried to modify the configuration but I can not configure it to trigger at all. I thought, that there must be a threshold value, at which it triggers instantly. But even if I go down to
M915 P0 S-64
, no stall is detected.

One more thing: The homing speed is limited to something less than 600mm/s. But why? The specified 2400mm/min are never reached.