[Testers Needed] Inconsistent Extrusion  

Re: [Testers Needed] Inconsistent Extrusion

Sorry- but I don't believe it.

FWIW, you can believe whatever you want. Everybody has that option.

But when you call me a liar, I don't really feel the inclination to help you with your problem. I believe that you have a problem with your prints; why would you not believe me when I tell you that I don't have that problem?

Unlike some on these forums, I have been here for quite some time. I have been extremely critical of PR at times and highly defensive of them at others. They get some things very right; they also get some things very wrong.

All I do know is that they are working to solve your problem (as they should) and I am sure that as soon as they have a solution, it will be published as a matter of urgency.

Peter

Re: [Testers Needed] Inconsistent Extrusion

I did not call you a liar- I said I don't believe it. I believe this is a matter of perception- what some believe to be "great" and "perfect" is something entirely different than what is being discussed here. After those users post pictures of their prints, we see the same flaws and in most cases those users then agree with our assessment. This has happened time and time in this thread, as well as the github issue addressing it. Since you seem to be unwilling to post any pictures as we have asked, then we can only assume you fear we will once again point out the issue and that your prints are no better than others coming off a MK3.

This is not a personal issue- we're trying to get to the bottom of a problem and picture evidence is everything.

Re: [Testers Needed] Inconsistent Extrusion

@PJR: It would be great if you could provide a sample gcode and the STL file for the object where you think MK3 produce best results. I would like to compare it to my own current settings.
The whole current "inconsistent extrusion" discussion on all platforms lead me to dive deeper into FDM printing.

Re: [Testers Needed] Inconsistent Extrusion

> Has anyone (who speaks German) tried contacting one of Trinamic's FAEs? I'm sure they would have all kinds of great troubleshooting ideas.

Trinamic works closely with us.

Recently we asked them to review the Einsy board, and they just received a current factory built MK3 to play with.

Re: [Testers Needed] Inconsistent Extrusion

@PJR: It would be great if you could provide a sample gcode and the STL file for the object where you think MK3 produce best results. I would like to compare it to my own current settings.
The whole current "inconsistent extrusion" discussion on all platforms lead me to dive deeper into FDM printing.

Yeah, I can do that, but before you print any test object, I would suggest that you replace the cable ties holding the linear bearings at the rear of the extruder and ensure that the replacements are really tight. Please do this first and run one of your own test prints to see if there is any improvement.

Then (when I have the time, I would be happy to slice your model for you with KISS to see if there are any differences).

Unfortunately, the new forum layout is taking up a whole load of my time (checking posts, removing spam - I cannot batch delete 🙁 hence the many spams from a couple of users which remain here, banning spammers etc) and I don't have more than a couple of hours a day play with printing (I am also alpha testing KISS...).

I am firmly of the belief that this is a hardware/build issue and that slicing can only go a very little way to resolving this issue.

Peter

Re: [Testers Needed] Inconsistent Extrusion

> Has anyone (who speaks German) tried contacting one of Trinamic's FAEs? I'm sure they would have all kinds of great troubleshooting ideas.

Trinamic works closely with us.

Recently we asked them to review the Einsy board, and they just received a current factory built MK3 to play with.

That’s good to know, we would be interested to here any feedback from what they have to say, has there been any progress from PRUSA on this? I know you have a lot of people trying to resolve the problem now

Thanks

Re: [Testers Needed] Inconsistent Extrusion

The Prusa i3 MK3 issue nobody wants to talk about

What changed so dramatically from the mk2 to cause this issue? Is this hardware issues or firmware?

Re: [Testers Needed] Inconsistent Extrusion

Yeah, I can do that, but before you print any test object, I would suggest that you replace the cable ties holding the linear bearings at the rear of the extruder and ensure that the replacements are really tight. Please do this first and run one of your own test prints to see if there is any improvement.

Then (when I have the time, I would be happy to slice your model for you with KISS to see if there are any differences).

Unfortunately, the new forum layout is taking up a whole load of my time (checking posts, removing spam - I cannot batch delete 🙁 hence the many spams from a couple of users which remain here, banning spammers etc) and I don't have more than a couple of hours a day play with printing (I am also alpha testing KISS...).

I am firmly of the belief that this is a hardware/build issue and that slicing can only go a very little way to resolving this issue.

No worries, there is nothing urgent. If I compare my almost new MK3 (14 days print time) vs my old MK2 (over 200 days print time) then my MK3 prints look better. The 3.2.0 firmware with linear advance is pretty good and everything is tight and firm. I just had to lower the extrusion rate to 90% with 0.48 nozzle width.

Right now almost every bad looking print is put under "extrusion issue". So I agree with you that most of them are just hardware/build related. But there is also one thing which is being under the PR investigation right now. All the small irregularities which are barely visible on the MK2 are now more pronounced on the MK3. So there is definitely a print quality decrease. It's still pretty good but worse then the old setup. That's the reason why so many people are looking into that at the moment.

Re: [Testers Needed] Inconsistent Extrusion

The Prusa i3 MK3 issue nobody wants to talk about

What changed so dramatically from the mk2 to cause this issue? Is this hardware issues or firmware?

The person ultimately decides that the issue is Linear Advance in the text area. Well if he actually read the github thread he would have known that tests were done both with and withOUT LA and the "problem" continued.

Lots of talk in the video with no backup!

Re: [Testers Needed] Inconsistent Extrusion

The Prusa i3 MK3 issue nobody wants to talk about

What changed so dramatically from the mk2 to cause this issue? Is this hardware issues or firmware?

The person ultimately decides that the issue is Linear Advance in the text area. Well if he actually read the github thread he would have known that tests were done both with and withOUT LA and the "problem" continued.

Lots of talk in the video with no backup!

Youtube personality trying to be an authority on a subject as usual. Gotta get those subscribes. 🙂

@PJR,

I don't mean to call you a liar and mean this in the most polite way possible: We have yet to see a single MK3 print without this effect, we would be thrilled to see you do the test and show us one without it. We've tried to hunt one down. We've had at this point at least a dozen people who believed they didn't have the artifacts, but when guided through the test they clearly did each time. That said:

Can you print a 0.2mm single wall cube in a solid color like white or silver? Then take a pic of it with light shining from above?

Just to highlight how much the light matters: https://imgur.com/a/DH8ym this is the same print with and without flash. Flash makes these prints look flawless, but in real world conditions the effect is pretty easy to see as well, especially if you're moving the print around and changing its lighting conditions.

Re: [Testers Needed] Inconsistent Extrusion

Tried my 2 mk3's with the latest firmware today, upgrade, full reset, calibration etc, haven't been switched on for weeks because of this issue

No change, both still unusable 🙁

Re: [Testers Needed] Inconsistent Extrusion

@PJR,

I don't mean to call you a liar and mean this in the most polite way possible: We have yet to see a single MK3 print without this effect, we would be thrilled to see you do the test and show us one without it. We've tried to hunt one down. We've had at this point at least a dozen people who believed they didn't have the artifacts, but when guided through the test they clearly did each time. That said:

Can you print a 0.2mm single wall cube in a solid color like white or silver? Then take a pic of it with light shining from above?

Just to highlight how much the light matters: https://imgur.com/a/DH8ym this is the same print with and without flash. Flash makes these prints look flawless, but in real world conditions the effect is pretty easy to see as well, especially if you're moving the print around and changing its lighting conditions.

Well, I think I finally managed to get something that doesn't look right. I had to push the print speed to 80mm/s (usually 40) and reduce min layer time to 1sec (usually 5 sec).

This photo was taken without flash and the room light above and slightly in front of the prints.

One print is from the Mk3, the other is a Mk1->Mk2->Mk2S-> Mk2s+MMU (with Bondtech) upgrade, both PR Silver. Single wall (0.4mm extrusion width), no upper/lower surfaces, no infill.

Unfortunately, I only have room to work with 2 printers at at time, and I am currently doing a lot of alpha testing with my MMU printers.

Peter

Re: [Testers Needed] Inconsistent Extrusion

Well, I think I finally managed to get something that doesn't look right.

Left one doesn't look right. It has moire all over it, and looks wobbly. And we're still not seeing the 'light slightly from above' effect which Brigandier demonstrated.

Right one looks fine.

That's what we'd like on the MK3.

Re: [Testers Needed] Inconsistent Extrusion

Left one doesn't look right. It has moire all over it, and looks wobbly. And we're still not seeing the 'light slightly from above' effect which Brigandier demonstrated.

Right one looks fine.

That's what we'd like on the MK3.

Yeah well the light source was as described and was the worst I could get. And I tuned off the "Moire" calibration (was 1.07, printed "off") but the picture does not display what is seen on both prints - humungous "ringing" from left fading to right due to speed of print.

Peter

Re: [Testers Needed] Inconsistent Extrusion

@PJR,
.....

Well, I think I finally managed to get something that doesn't look right. I had to push the print speed to 80mm/s (usually 40) and reduce min layer time to 1sec (usually 5 sec).

This photo was taken without flash and the room light above and slightly in front of the prints.

One print is from the Mk3, the other is a Mk1->Mk2->Mk2S-> Mk2s+MMU (with Bondtech) upgrade, both PR Silver. Single wall (0.4mm extrusion width), no upper/lower surfaces, no infill.

IMG_20180520_224236.jpg

.......
Peter

The cropped right most piece also appears to show the issue...

Re: [Testers Needed] Inconsistent Extrusion

Yeah well the light source was as described and was the worst I could get. And I tuned off the "Moire" calibration (was 1.07, printed "off") but the picture does not display what is seen on both prints - humungous "ringing" from left fading to right due to speed of print.

Peter

Hi PJR,

It looks like you're in a room with a lot of ambient lighting. I can still see it in the print, but it's not casting a shadow so it may not be obvious. You do have a pretty nice print compared to others though, likely due to you not having other issues confusing the problem. Is it possible for you to try it under a single point light source (like a flashlight, single bulb, etc). The light needs to originate from a single point. This GIF from TynanTynan on Github demonstrates why:

As you can see, shining a single point light source down the face causes the over extruded layers to cast shadows onto the layers below them (rooms with lots of light from different angles tend to not have many shadows). Your image on the right looks nice, but I can still see some lines sticking out on it and I am confident if you try this with a single point light source you'll see it too.

Re: [Testers Needed] Inconsistent Extrusion

Okay, I have set up an oscilloscope and had time to do just one measurement of the voltage on one side of one of the phases.

All looks well, as far as my understanding. There's a ~23 kHz PWM voltage pushing a current through the coil. This current can be seen as oscillating the entire waveform (at about 1.25 Hz) , because of the voltage drop over the current sense resistor.

This was just while the printer was running the final "square" part of the Live Z calibration.

So nothing seems strange or unexpected about that, in my mind. I might have to look at the step signal pin instead. Maybe something is wrong with the IO voltages or noise between the MCU and the stepper driver, and this is causing the TMC to sometimes overextrude because of false flags. But it is very weird if that is only an issue towards the extruder driver and not the X, Y and Z drivers, which don't seem to be suffering from the same issue.

Re: [Testers Needed] Inconsistent Extrusion

Okay, I have set up an oscilloscope and had time to do just one measurement of the voltage on one side of one of the phases.

scope_1.png

All looks well, as far as my understanding. There's a ~23 kHz PWM voltage pushing a current through the coil. This current can be seen as oscillating the entire waveform (at about 1.25 Hz) , because of the voltage drop over the current sense resistor.

This was just while the printer was running the final "square" part of the Live Z calibration.

So nothing seems strange or unexpected about that, in my mind. I might have to look at the step signal pin instead. Maybe something is wrong with the IO voltages or noise between the MCU and the stepper driver, and this is causing the TMC to sometimes overextrude because of false flags. But it is very weird if that is only an issue towards the extruder driver and not the X, Y and Z drivers, which don't seem to be suffering from the same issue.

Not a scope guy, so may be asking the impossible here lol. Is it possible for you to find pauses? In the Github issue we're now exploring the possibility that this is micro stutter due to some process (like updating the LCD, see here: https://www.cnx-software.com/2018/02/03/he3d-k280-delta-3d-printer-review-part-2-upgrades-calibration-and-more-3d-prints/#micro-stutters - scroll down to micro stutters if it doesn't happen automatically).

Re: [Testers Needed] Inconsistent Extrusion

Okay, I have set up an oscilloscope and had time to do just one measurement of the voltage on one side of one of the phases.

scope_1.png

All looks well, as far as my understanding. There's a ~23 kHz PWM voltage pushing a current through the coil. This current can be seen as oscillating the entire waveform (at about 1.25 Hz) , because of the voltage drop over the current sense resistor.

This was just while the printer was running the final "square" part of the Live Z calibration.

So nothing seems strange or unexpected about that, in my mind. I might have to look at the step signal pin instead. Maybe something is wrong with the IO voltages or noise between the MCU and the stepper driver, and this is causing the TMC to sometimes overextrude because of false flags. But it is very weird if that is only an issue towards the extruder driver and not the X, Y and Z drivers, which don't seem to be suffering from the same issue.

My upgraded mK2 to MK2.5 has it now, like first print off it... there's no TMC drivers here. it didn't just prior, cause i printed the parts for it and it has none of the artifacts.... there's even moire now... so i think the moire in mk3 is probably from the same problem.... not the drivers.

Re: [Testers Needed] Inconsistent Extrusion

In the Github issue we're now exploring the possibility that this is micro stutter due to some process (like updating the LCD, see here: https://www.cnx-software.com/2018/02/03 ... o-stutters - scroll down to micro stutters if it doesn't happen automatically).

I think that what we would see is stretching and compression of the sinusoidal offset. Pauses would mean the extruder completely stops. But right now, I'm also a believer that the problem lies in CPU load/interrupts. I don't have much time to experiment with it this week, so I'm putting my trust in you other guys for now :-). I'm following both the discord and the github thread religiously