Sketchy, stuttered brim but otherwise good part.

Share your successful printing settings with others here. If you have any question about the parameters or printing skills of some filament, please find here.
User avatar
jetdillo
Posts: 148
Joined: Mon Jun 12, 2017 5:10 am

Sketchy, stuttered brim but otherwise good part.

Postby jetdillo » Tue Mar 06, 2018 8:28 am

I've noticed lately that the steppers will kind of stutter their way through a brim or skirt but the rest of the part is fine. Even the inside couple loops of the brim are drawn smoothly. With the first couple though, the print head travels smoothly down the straight part of the brim but steps and pauses going around the corners. This can cause small drips or blobs of filament to gather at the corners, causing the nozzle to bump over or tear at the side. Paradoxically, the print actually still turns out pretty great.

I'm using Ideamaker 3.10.1545 and don't recall this sort of behavior in previous versions I've used.
I'm wondering if there's a new setting in this version that I'm missing or if it's something else. I think this is bugging me because it's happening so early in the print and it's leaving artifacts behind as a result of the stalling at the corners.

My minimum print speed is 10mm/sec. Minimal layer time is 15 seconds. I'm wondering if there is a minimal speed beyond which Ideamaker can't effectively control the steppers.

Squenz
Posts: 140
Joined: Fri Dec 01, 2017 11:18 pm
Location: Stuttgart/Germany

Re: Sketchy, stuttered brim but otherwise good part.

Postby Squenz » Tue Mar 06, 2018 9:22 am

There's no setting you can change to get a better result, it's the nature of this ideamaker version and has been flagged as not good already by other users. viewtopic.php?f=3&t=6299&p=26920#p26920

Jetguy
Posts: 2700
Joined: Tue Mar 22, 2016 1:40 am
Location: In a van, down by the river

Re: Sketchy, stuttered brim but otherwise good part.

Postby Jetguy » Tue Mar 06, 2018 1:23 pm

This brings up a good question and debate.
First, let's talk about why the setting produces stuttering. The thought is that when Raise3D custom skirt makes the curve, there is no such thing as curves in current 3D printing gcode. A curve is just short segments like the part of a multisided polygon. Each of those segments becomes a line of gcode that must be read and followed. As such, it is imperative to understand the underlying architecture of the Raise3D printer and it's front panel display and separate motion control board connected by a USB cable but technically operating as a serial connection. That standard baud rate is 230400. This number was chosen most likely because long ago in 3D printing forums and especially involving Linux based systems, that is one of the higher defined standard baud rates. The key to remember is that the motion control bard is only an 8 bit microprocessor and it's running the printer. It's reading one by line the gcode commands streamed over the USB to serial connection since the actual 8 bit process does not natively support USB. When you get to fine short segments as seen in a curve, we always knew there was a theoretical choking limit here where the system will not keep up. When that happens, absolutely the telltale symptom is stuttering where the system appears to micro pause and this is very obvious in the sound and visible motion of the steppers. You can actually test this theory because there is a better way to print and that is via direct memory storage (SD card) by using a ReprapDiscount Full Graphics style LCD control panel attached directly to the motion control board. Since the 8 bit processor controls it's own reading of the directly memory containing the gcode file, this is much faster and more efficient than over USB to serial. So what does this all mean? Well, recently while poking around in the configuration of the LCD, In noticed a new line in the config file that was not there before that contains the baud rate that the LCD touch panel is attempting to use to connect to the motion controller. Again, it's known the default here is 230400 and remember that the other end (motion control board running Marlin firmware) has to run the same rate. This means in theory with custom motion control firmware and change the LCD Raisetouch config to match, it might be possible to up the rate to 2500000 VS 230400. I do not know, this theory needs testing, just something I recently discovered about new Raisetouch versions. 90% of the time we never talk about this limit since most gcode in theory never hits it and the user would be completely oblivious to this architecture of the system. However, in high performance printing situations where absolutely the max segment performance is required, printing from directly attached memory on the motion control board has always been recommended by the 3D printing community, but in this case, the stock configuration of the hardware does not support that option. I know this limit exists because when attempting a custom Raise3D Delta printer calculation using the standard Raise 3D touch panel, motion control board, and simply configuring the Marlin source code to Delta style printer, I absolutely could get the system to choke badly on even relatively standard segmented curves on a print due to the math overhead the 8 bit processor requires when doing delta style 3D printing. It was basically unusable. Again, as a test, I was determining could Raise 3D ever produce a Delta printer on the same basic electronics and touchscreen and that project showed me serious limits that are no normally seen by the N series Cartesian printers that have minimal math overhead.

zemlin
Posts: 380
Joined: Sat Oct 21, 2017 2:02 pm

Re: Sketchy, stuttered brim but otherwise good part.

Postby zemlin » Tue Mar 06, 2018 1:59 pm

I see that sometimes with S3D ... but it's a brim, so I don't care about it.

Jetguy
Posts: 2700
Joined: Tue Mar 22, 2016 1:40 am
Location: In a van, down by the river

Re: Sketchy, stuttered brim but otherwise good part.

Postby Jetguy » Tue Mar 06, 2018 3:55 pm

FWIW, I just bench tested the theory about being able to increase the baud rate between the two boards and firmware to 2500000 and that did not work. I know the Marlin side works, that's easy to test with just connecting to the PC. SSH was showing me that it's entirely possible that even though the text configuration has the setting, the Raise3D touch application and pcDuino in general do not seem to accept or use that rate change in the config file. This was confirmed by flashing the motion board back to OEM 230400 baud rate firmware while keeping the RaiseTouch config at 2500000, and that's a mismatch and yet it connects. Because the actual Raise 3D Touch UI application is a QT5 application, from the user perspective, we have limited ways we can edit and change functionality and settings.
Screenshot of config file.
baud rate change test.jpg


Code: Select all

login as: root
root@192.168.1.90's password:
root@imx6qdlsolo:~# stty < /dev/ttyACM0
speed 230400 baud; line = 0;
min = 0; time = 0;
-brkint -icrnl -imaxbel
-opost
-isig -icanon -iexten -echo
root@imx6qdlsolo:~# login as: root
Password:



Bottom line, changing baud rate is not yet a viable option. Also, FWIW, oldschool discussion on why 230400 is a "standard" baud rate when dealing with Linux and mega2560. https://github.com/MarlinFirmware/Marlin/issues/205

Jetguy
Posts: 2700
Joined: Tue Mar 22, 2016 1:40 am
Location: In a van, down by the river

Re: Sketchy, stuttered brim but otherwise good part.

Postby Jetguy » Tue Mar 06, 2018 4:02 pm

Can you share a confirmed gcode file that stutters for testing? Just want to make sure I test apples to apples.
If too big to share here, Dropbox or Google drive?

User avatar
jetdillo
Posts: 148
Joined: Mon Jun 12, 2017 5:10 am

Re: Sketchy, stuttered brim but otherwise good part.

Postby jetdillo » Tue Mar 06, 2018 5:40 pm

Squenz wrote:There's no setting you can change to get a better result, it's the nature of this ideamaker version and has been flagged as not good already by other users. viewtopic.php?f=3&t=6299&p=26920#p26920


Is there a way to backgrade ?
Because this version also has some UI update bugs on OSX High Sierra.

Jetguy
Posts: 2700
Joined: Tue Mar 22, 2016 1:40 am
Location: In a van, down by the river

Re: Sketchy, stuttered brim but otherwise good part.

Postby Jetguy » Tue Mar 06, 2018 5:47 pm

jetdillo wrote:
Squenz wrote:There's no setting you can change to get a better result, it's the nature of this ideamaker version and has been flagged as not good already by other users. viewtopic.php?f=3&t=6299&p=26920#p26920


Is there a way to backgrade ?
Because this version also has some UI update bugs on OSX High Sierra.


Ideamaker previous versions https://www.raise3d.com/pages/ideamaker-all-versions

Caveat, I don't know how this affects custom settings, just pointing out where you can get older versions if you no longer have it.

User avatar
Vicky@Raise3D
Posts: 4563
Joined: Fri Mar 25, 2016 3:54 am

Re: Sketchy, stuttered brim but otherwise good part.

Postby Vicky@Raise3D » Thu Mar 08, 2018 8:08 am

jetdillo wrote:I've noticed lately that the steppers will kind of stutter their way through a brim or skirt but the rest of the part is fine. Even the inside couple loops of the brim are drawn smoothly. With the first couple though, the print head travels smoothly down the straight part of the brim but steps and pauses going around the corners. This can cause small drips or blobs of filament to gather at the corners, causing the nozzle to bump over or tear at the side. Paradoxically, the print actually still turns out pretty great.

I'm using Ideamaker 3.10.1545 and don't recall this sort of behavior in previous versions I've used.
I'm wondering if there's a new setting in this version that I'm missing or if it's something else. I think this is bugging me because it's happening so early in the print and it's leaving artifacts behind as a result of the stalling at the corners.

My minimum print speed is 10mm/sec. Minimal layer time is 15 seconds. I'm wondering if there is a minimal speed beyond which Ideamaker can't effectively control the steppers.


Shall we ask the gcode and stl file for your model to do a test? We are trying to fix this.

User avatar
jetdillo
Posts: 148
Joined: Mon Jun 12, 2017 5:10 am

Re: Sketchy, stuttered brim but otherwise good part.

Postby jetdillo » Fri Mar 09, 2018 6:18 am

Vicky@Raise3D wrote:
jetdillo wrote:I've noticed lately that the steppers will kind of stutter their way through a brim or skirt but the rest of the part is fine. Even the inside couple loops of the brim are drawn smoothly. With the first couple though, the print head travels smoothly down the straight part of the brim but steps and pauses going around the corners. This can cause small drips or blobs of filament to gather at the corners, causing the nozzle to bump over or tear at the side. Paradoxically, the print actually still turns out pretty great.

I'm using Ideamaker 3.10.1545 and don't recall this sort of behavior in previous versions I've used.
I'm wondering if there's a new setting in this version that I'm missing or if it's something else. I think this is bugging me because it's happening so early in the print and it's leaving artifacts behind as a result of the stalling at the corners.

My minimum print speed is 10mm/sec. Minimal layer time is 15 seconds. I'm wondering if there is a minimal speed beyond which Ideamaker can't effectively control the steppers.


Shall we ask the gcode and stl file for your model to do a test? We are trying to fix this.


Sure. The forum doesn't seem to allow me to attach .gcode files here. Can I PM or email it to you ?

User avatar
Vicky@Raise3D
Posts: 4563
Joined: Fri Mar 25, 2016 3:54 am

Re: Sketchy, stuttered brim but otherwise good part.

Postby Vicky@Raise3D » Fri Mar 09, 2018 7:49 am

jetdillo wrote:
Vicky@Raise3D wrote:
jetdillo wrote:I've noticed lately that the steppers will kind of stutter their way through a brim or skirt but the rest of the part is fine. Even the inside couple loops of the brim are drawn smoothly. With the first couple though, the print head travels smoothly down the straight part of the brim but steps and pauses going around the corners. This can cause small drips or blobs of filament to gather at the corners, causing the nozzle to bump over or tear at the side. Paradoxically, the print actually still turns out pretty great.

I'm using Ideamaker 3.10.1545 and don't recall this sort of behavior in previous versions I've used.
I'm wondering if there's a new setting in this version that I'm missing or if it's something else. I think this is bugging me because it's happening so early in the print and it's leaving artifacts behind as a result of the stalling at the corners.

My minimum print speed is 10mm/sec. Minimal layer time is 15 seconds. I'm wondering if there is a minimal speed beyond which Ideamaker can't effectively control the steppers.


Shall we ask the gcode and stl file for your model to do a test? We are trying to fix this.


Sure. The forum doesn't seem to allow me to attach .gcode files here. Can I PM or email it to you ?


Of course. Just feel free to send PM to me.

User avatar
jetdillo
Posts: 148
Joined: Mon Jun 12, 2017 5:10 am

Re: Sketchy, stuttered brim but otherwise good part.

Postby jetdillo » Fri Mar 09, 2018 1:37 pm

Vicky@Raise3D wrote:
jetdillo wrote:
Vicky@Raise3D wrote:
Shall we ask the gcode and stl file for your model to do a test? We are trying to fix this.


Sure. The forum doesn't seem to allow me to attach .gcode files here. Can I PM or email it to you ?


Of course. Just feel free to send PM to me.


I uploaded it to my Google Drive. This link should work:
https://drive.google.com/file/d/1EIOJFw ... sp=sharing

Note how the brim has the "stuttering" or "scalloped" effect, but the first line inside that is smooth and round and follows a (visibly) proper curve.

20180305_172313.jpg


20180305_175531.jpg

Jetguy
Posts: 2700
Joined: Tue Mar 22, 2016 1:40 am
Location: In a van, down by the river

Re: Sketchy, stuttered brim but otherwise good part.

Postby Jetguy » Fri Mar 09, 2018 2:37 pm

Thanks for sharing that excellent example of a test file and the images of the problem. This will help greatly in trying to improve future firmware versions to see if we can get any better performance.

User avatar
Vicky@Raise3D
Posts: 4563
Joined: Fri Mar 25, 2016 3:54 am

Re: Sketchy, stuttered brim but otherwise good part.

Postby Vicky@Raise3D » Mon Mar 12, 2018 6:45 am

jetdillo wrote:
Vicky@Raise3D wrote:
jetdillo wrote:
Sure. The forum doesn't seem to allow me to attach .gcode files here. Can I PM or email it to you ?


Of course. Just feel free to send PM to me.


I uploaded it to my Google Drive. This link should work:
https://drive.google.com/file/d/1EIOJFw ... sp=sharing

Note how the brim has the "stuttering" or "scalloped" effect, but the first line inside that is smooth and round and follows a (visibly) proper curve.

20180305_172313.jpg


20180305_175531.jpg



Seems only .gcode file inside the download link? Or did I miss something?

User avatar
Vicky@Raise3D
Posts: 4563
Joined: Fri Mar 25, 2016 3:54 am

Re: Sketchy, stuttered brim but otherwise good part.

Postby Vicky@Raise3D » Mon Mar 12, 2018 7:33 am

jetdillo wrote:
Squenz wrote:There's no setting you can change to get a better result, it's the nature of this ideamaker version and has been flagged as not good already by other users. viewtopic.php?f=3&t=6299&p=26920#p26920


Is there a way to backgrade ?
Because this version also has some UI update bugs on OSX High Sierra.


Would you like to open a new thread or PM to me more details about the bugs of UI you meet?

Jetguy
Posts: 2700
Joined: Tue Mar 22, 2016 1:40 am
Location: In a van, down by the river

Re: Sketchy, stuttered brim but otherwise good part.

Postby Jetguy » Mon Mar 12, 2018 2:01 pm

Vicky@Raise3D wrote:
Seems only .gcode file inside the download link? Or did I miss something?


Yes, this is the example file that demonstrates a limitation of the USB to serial communications and motion control between the front panel Raise Touch panel pcDuino sending the gcode line by line to the Marlin motion control firmware. The skirt option with curved corners is creating finely segmented gcode that exposes this limitation and leads to motion stuttering because the buffer and transfer rate of all those tiny segments is being reached.

The reason this discussion and topic was created is that in one of the recent updates to Ideamaker, the slicing option to put rounded corners that generate these small segments happened.
I'm using Ideamaker 3.10.1545 and don't recall this sort of behavior in previous versions I've used.
I'm wondering if there's a new setting in this version that I'm missing or if it's something else. I think this is bugging me because it's happening so early in the print and it's leaving artifacts behind as a result of the stalling at the corners.


It's a double problem:
#1 there is a limit in the system and we know how to expose it- the sample gcode file will produce the problem on any N series printer and firmware because it's a system architecture limitation, not a firmware version or bug. It's fundamental to the printer design.
#2 It appears that Ideamaker 3.10.1545 produces gcode (the sample) that invokes this limitation on the skirt/brim.

The goal is to work on BOTH parts of the system.
Ideamaker should not create gcode that can invoke bad print quality or exceed performance limits of the system.
and
The Raise 3D touch application and baud rate chosen for the Marlin motion control firmware- there may be potential room for improvement to raise this performance limit of the architecture. By having a known gcode example that demonstrates the limitation, further investigation in the firmware side of this can happen.

Jetguy
Posts: 2700
Joined: Tue Mar 22, 2016 1:40 am
Location: In a van, down by the river

Re: Sketchy, stuttered brim but otherwise good part.

Postby Jetguy » Mon Mar 12, 2018 4:48 pm

And, just to prove the theory and what I'm talking about here, back to back comparison of a "normal" Raise 3D print function using the LCD Raisetouch to stream the gcode over USB to motion board VS directly connecting a standard off the shelf ReprapDiscount Graphical LCD module with SD card to the motion board and printing from SD card.

Here is the gcode test file uploaded to the printer using Ideamaker, then selecting print from local storage and the results are what we expect- stuttering on the extremely short segment round curves of the brim on the example gcode.
Stutter1.JPG

Stutter2.JPG


However, if we connect the ReprapDiscount graphical LCD module with SD card to the motion board directly, then we do not get the pronounced stutter printing the exact same file, now just locally accessed by the mega2560 on the motion board. This is a well known and documented function in the 3D printing world that printing from a directly attached SD card can improve small segment performance on the typical 8bit AVR processor based control board. This is because the processor can read the memory when it needs to without the overhead of communicating or buffering text gcode sent over a serial connection. it simply frees up the processor and the result is proven right here.
Sd card direct print.JPG


Again, the LCD module and SD card is connected via 2 short standard ribbon cables to the unused expansion plugs on the motion board, and the standard Raise 3D marlin firmware has always supported adding this LCD for troubleshooting and feature tuning reasons.
LCD ribbon cable to mainboard..JPG

LCD and SD card module.JPG

Jetguy
Posts: 2700
Joined: Tue Mar 22, 2016 1:40 am
Location: In a van, down by the river

Re: Sketchy, stuttered brim but otherwise good part.

Postby Jetguy » Mon Mar 12, 2018 4:49 pm

A few more result pictures from the SD card print since the Raise 3D forum limits posts to 5 pictures.
Attachments
SD card print 3.JPG
SD card print2.JPG

User avatar
jetdillo
Posts: 148
Joined: Mon Jun 12, 2017 5:10 am

Re: Sketchy, stuttered brim but otherwise good part.

Postby jetdillo » Mon Mar 12, 2018 8:14 pm

Jetguy wrote:
Vicky@Raise3D wrote:
Seems only .gcode file inside the download link? Or did I miss something?


Yes, this is the example file that demonstrates a limitation of the USB to serial communications and motion control between the front panel Raise Touch panel pcDuino sending the gcode line by line to the Marlin motion control firmware. The skirt option with curved corners is creating finely segmented gcode that exposes this limitation and leads to motion stuttering because the buffer and transfer rate of all those tiny segments is being reached.


So, this is a tiny-packets/buffer-overrun issue between the front-panel and the motion-controller ?
That makes sense, esp. if you're opening the UART once-per-gcode-statement.
That's going to be a lot of interrupts on the serial lines of both controllers and a lot of overhead from tiny messages that could be better concatenated into longer ones.
It's both interesting and not to see the near-real-time effects of inefficient communication like this.

User avatar
Vicky@Raise3D
Posts: 4563
Joined: Fri Mar 25, 2016 3:54 am

Re: Sketchy, stuttered brim but otherwise good part.

Postby Vicky@Raise3D » Wed Mar 14, 2018 8:18 am

Jetguy wrote:
Vicky@Raise3D wrote:
Seems only .gcode file inside the download link? Or did I miss something?


Yes, this is the example file that demonstrates a limitation of the USB to serial communications and motion control between the front panel Raise Touch panel pcDuino sending the gcode line by line to the Marlin motion control firmware. The skirt option with curved corners is creating finely segmented gcode that exposes this limitation and leads to motion stuttering because the buffer and transfer rate of all those tiny segments is being reached.

The reason this discussion and topic was created is that in one of the recent updates to Ideamaker, the slicing option to put rounded corners that generate these small segments happened.
I'm using Ideamaker 3.10.1545 and don't recall this sort of behavior in previous versions I've used.
I'm wondering if there's a new setting in this version that I'm missing or if it's something else. I think this is bugging me because it's happening so early in the print and it's leaving artifacts behind as a result of the stalling at the corners.


It's a double problem:
#1 there is a limit in the system and we know how to expose it- the sample gcode file will produce the problem on any N series printer and firmware because it's a system architecture limitation, not a firmware version or bug. It's fundamental to the printer design.
#2 It appears that Ideamaker 3.10.1545 produces gcode (the sample) that invokes this limitation on the skirt/brim.

The goal is to work on BOTH parts of the system.
Ideamaker should not create gcode that can invoke bad print quality or exceed performance limits of the system.
and
The Raise 3D touch application and baud rate chosen for the Marlin motion control firmware- there may be potential room for improvement to raise this performance limit of the architecture. By having a known gcode example that demonstrates the limitation, further investigation in the firmware side of this can happen.


We were asking for the .stl file was due to we were trying to debug the Brim short segments in ideaMaker and wanted to ask the origin .stl file to test whether our fix works or not.
We guess that the stuck problem is due to the serial port has limitations of data transfer in short time, and the root reason of this is the Brim has been sliced to be too many short segments. So we want to debug this problem as a slicing problem from ideaMaker.


Return to “Printing Settings”

Who is online

Users browsing this forum: No registered users and 4 guests