3D printing

just spent some couple of hours playing with prusa slicer. an interesting software, a bit of a mental shift after cura. UI design wise I liked it better.
the good part - z-seam issue with it is just non-existent. i could not see it in the printed model, it was 'just fine'. all smooth and no gaps.

the bad part - i may be a moron, of course, i never exclude that part, but, it completely refused to alter the wall thickness for me.
i figured it all (or most of it) out. it would be shown correctly in the prusa rendering, i can see the wall size increased there. then i print it, and the 3 sets of 0.5mm walls still measure up the same damn 0.81mm altogether.
kinda gave up at that point for now. it looked also like the extrusion was not quite right overall, the cube was breaking apart way too easily. but, it looked a bit better than one from cura.
with same speed settings cura prints at 8min, the prusa printed in 11-12 min, as it would be hiccupping on some odd add-on fillers between walls, even when all infill was switched off.
there was noticably less ghosting. cube form was also kinda bit better. no elephant foot at all, and all automatic.

i guess there is more of a learning curve hidden in there, but, no miracles, all in all - mostly same BS, just of a different color.

still debating, to drop or not $200 bucks on the paid slicer or not. just reprinted old models with that damn cura again and it all looks decent enough.
 
Last edited:
So, the trick with prusa seems to be not to alter defaults for extrusion width and just multiply the walls.
Anyway- here is how klipper on same settings does it. Prusa is a small gap one.

Also can be partially seen under extruded bottom.
A12CFD72-AFED-4E99-BCEB-D9F42338A0D8.jpeg
 
Finally getting back into this... I bought an Ender 3 Pro at the beginning of Covid. I printed a bunch of stuff then got busy. Over the holiday I went back to finish the bullet feeder I started back then. I've been catching up with everything that has changed in the last 2.5 years. I just put a glass bed and yellow springs on. I started to fiddle with Klipper and got really frustrated trying to install it on this ancient v1.1.4 8-bit board. Something about it not having a boot loader. After fiddling for a few hours I said screw it, I've been wanting a silent board anyway. Thanks to Amazon prime I just installed a v4.2.7 32-bit board: I should have made this upgrade a LONG time ago. It's 1000x quieter and less annoying!
 
Finally getting back into this... I bought an Ender 3 Pro at the beginning of Covid. I printed a bunch of stuff then got busy. Over the holiday I went back to finish the bullet feeder I started back then. I've been catching up with everything that has changed in the last 2.5 years. I just put a glass bed and yellow springs on. I started to fiddle with Klipper and got really frustrated trying to install it on this ancient v1.1.4 8-bit board. Something about it not having a boot loader. After fiddling for a few hours I said screw it, I've been wanting a silent board anyway. Thanks to Amazon prime I just installed a v4.2.7 32-bit board: I should have made this upgrade a LONG time ago. It's 1000x quieter and less annoying!
that was my grief with pro and older ender 3 versions - the power supply vertically mounted, the board, other stuff. the ender 3 v2 for now is the ideal donor chassis.
based on what i know now - it is what i did for a second printer - got a vanilla 3 v2, then got on it yellow springs done, the $25 second z axis screw, and swapped out extruder for a sprite pro and cr touch combo that was on sale together for $105.
i think now this combo was once on sale for as low as $95.

the 3 v2 i was able to get with some creativity for $108. therefore, for a tad less that $250 you can get a real life practical copy of the S1 Pro unit they sell for $500, it goes to 310C on extruder and i print it now from klipper with very good quality:

1673810701346.png


for klipper - i posted above - the T95 max box is great all in one solution, was also at $30 on sale. web camera was $8.39.
Amazon product ASIN B0B461KB7VView: https://www.amazon.com/dp/B0B461KB7V
 
paul73 I've been lurking in this thread for a while. I ordered the T95 box you recommended previously.

Are you running it off the internal drive or an SD card? I could never get it to install to the internal drive so I'm running it off an SD card.

Does yours backpower your mainboard/display? If I shutoff my printer the display stays on unless I disconnect the USB cable from the T95
 
paul73 I've been lurking in this thread for a while. I ordered the T95 box you recommended previously.

Are you running it off the internal drive or an SD card? I could never get it to install to the internal drive so I'm running it off an SD card.

Does yours backpower your mainboard/display? If I shutoff my printer the display stays on unless I disconnect the USB cable from the T95
i should have documented my steps on t95, my fault. when i had to do a second box it took some time to rediscover the america again, so, here it is.

first of all - yes, the usb power backfeed is there, so, just put printer and t95 power supply on the same powerstrip and kill the switch on the strip to turn off it all. easy enough.

now on t95 boxes. there is a swarm of them on amazon and ebay, but they are nothing alike inside, so, we will consider you got EXACTLY the one from the link i posted. it is very important.

if you did that, then you just had to follow step the step the guide i also posted:

now, of what is NOT in the guide. :)

so, you need first to get raspberry PI imager and flash the 2gb quadra image to the SD card.

i think card should not be larger than 32gb for all this to work, i used a 16gb card.

to be able to flash this box you need to make sure NOTHING is plugged in into it. only the power cord. if hdmi cord is connected - it will not flash.
you do not press nothing at all also, you unplug the power cord and all else, put in the card, then put back power cord alone.
then it remains black for 20-30 sec, then the red led should lit and stay lit for the flash duration, around 5 min i think or bit more - then it will turn blue.
then you need to unplug the power cord after that.

THEN.

this shit by default comes with the absolutely retarded graphic UI the name of which i always forget, looking for it now on the box. 'lightdm'.
you will need to disable it - 'systemctl stop lightdm','systemctl disable lightdm' - later.

so, after you flash it - if you connect ethernet cord to the box, you get zilch UNTIL you connect it to the monitor with the hdmi cord, mouse and keyboard, and login this shit into its quadra account via this lightdm interface. only AFTER that it initializes network interfaces and connects to the network. a complete idiocy.

ALSO - adding it now - if you just connect hdmi cord after it is already turned on and booted - it will also not work. you need to unplug power, connect hdmi cord to the box and monitor, then turn it on, only then it will produce video output. keyboard and mouse are better to be plugged in initially too. first box was sensitive to usb power drain - the keyboard i used first had some backlit LEDs on it, and i think it overdrained the usb, so if box blinks its blue led and then reboots suddenly - it means you drain too much from the USB port and provided power supply is insufficient. use a different keyboard.


frankly the ethernet cord is not needed, as after you plug power cord and see something on the monitor finally and then login into armbian - you just run 'passwd' first to alter 'quadra' user, then change root password, then run 'nmtui' to conenct to wifi.

THEN. :)

those boxes have a nasty glitch - armbian has a glitch, really, as every time you reboot it the wifi gets different mac address and it produces different dynamic IP.
you need to set it on a manual IP address. to do that on those boxes on that flash - you need to alter the file that will be created under
/etc/NetworkManager/system-connections
with the name of your wifi network after you run the 'nmtui' command. you just need to edit the [ipv4] section in it, do not do nothing else.
[ipv4]
address1=192.168.10.123/24,192.168.10.1
dns=8.8.8.8;8.8.4.4;
dns-search=
method=manual

after that disable service for lightdm, and it should be able to reboot and get connected to the network so you can manage it remotely.
install klipper as the guide says - it is super easy with:
./kiauh/kiauh.sh


i think only step they also skipped in the guide was the mjpg-streamer-service install, you need it for the camera. i got it done with 'snap' - also super easy way, so it handles it as:
/usr/bin/snap start mjpg-streamer.mjpg-streamer-service

forgot the steps, i think it was from here:

PS. if this thing does something else to you of what i did not cover above - post the details here.

PPS.
I knew i will forget it. for the camera - more steps are needed. and i think i had lost them , AGAIN.
you need to install the 'v412-ctl' tool and all required support files, so the command:
v4l2-ctl -d /dev/video1 --list-ctrls-menus
would give you a list of what your web cam can do.

ok, what does my command history say... all to be done from under the quadra user - not from root.

sudo apt-get install git -y
sudo apt install python3 python3-pip python3-dev python3-setuptools python3-venv git libyaml-dev build-essential libffi-dev libssl-dev
sudo apt-get install libjpeg8-dev imagemagick libv4l-dev
sudo ln -s /usr/include/linux/videodev2.h /usr/include/linux/videodev.h
sudo apt-get install cmake libjpeg8-dev
sudo apt install snapd
sudo snap install mjpg-streamer
/usr/bin/snap start mjpg-streamer.mjpg-streamer-service

then you need to alter the 'config' startup file for the streamer at
/var/snap/mjpg-streamer/current$

mine reads now as:
INPUTOPTS="input_uvc.so -r 1920x1080 -f 12 -d /dev/video1"
PORT="-p 8080"
DAEMON="true"

you generally need to add the device to it - the '-d /dev/video1' part. it will unpredictably come to /video1 or to /video0 - you need to use the v412-ctl utility to see where it is.
also you need to flip daemon option to 'true' so it would autostart upon reboot.
a correct camera output looks like:

root@inovato:~# v4l2-ctl -d /dev/video1 --list-ctrls-menus
brightness 0x00980900 (int) : min=-10 max=10 step=1 default=0 value=0
contrast 0x00980901 (int) : min=1 max=32 step=1 default=16 value=16
saturation 0x00980902 (int) : min=0 max=20 step=1 default=10 value=10
hue 0x00980903 (int) : min=-5 max=5 step=1 default=0 value=0
white_balance_temperature_auto 0x0098090c (bool) : default=1 value=1
gamma 0x00980910 (int) : min=100 max=200 step=1 default=180 value=180
power_line_frequency 0x00980918 (menu) : min=0 max=2 default=1 value=1
0: Disabled
1: 50 Hz
2: 60 Hz
white_balance_temperature 0x0098091a (int) : min=2800 max=6500 step=1850 default=6500 value=6500 flags=inactive
sharpness 0x0098091b (int) : min=0 max=10 step=1 default=7 value=7
exposure_auto 0x009a0901 (menu) : min=0 max=3 default=3 value=3
1: Manual Mode
3: Aperture Priority Mode
exposure_absolute 0x009a0902 (int) : min=156 max=5000 step=1 default=512 value=512 flags=inactive
root@inovato:~#


an incorrect one:
root@inovato:~# v4l2-ctl -d /dev/video0 --list-ctrls-menus

Codec Controls

h264_profile 0x00990a6b (menu) : min=0 max=4 default=2 value=2
0: Baseline
1: Constrained Baseline
2: Main
4: High
hevc_sequence_parameter_set 0x00990cf0 (unknown): type=120 flags=has-payload
hevc_picture_parameter_set 0x00990cf1 (unknown): type=121 flags=has-payload
hevc_slice_parameters 0x00990cf2 (unknown): type=122 flags=has-payload
hevc_scaling_matrix 0x00990cf3 (unknown): type=123 flags=has-payload
hevc_decode_parameters 0x00990cf4 (unknown): type=124 flags=has-payload
hevc_decode_mode 0x00990cf7 (menu) : min=0 max=0 default=0 value=0
0: Slice-Based
hevc_start_code 0x00990cf8 (menu) : min=0 max=0 default=0 value=0
0: No Start Code


it is NOT a camera device.
 
Last edited:
Thanks. I bought the exact one from Amazon that you linked to a few pages back. I read about issues the wifi and not keeping an address so I've been using it with ethernet (printer is located closer to switch). I found a youtube video (
View: https://www.youtube.com/watch?v=nVpnZeS4ogM
) and was using the q-tip in the audio port. I used an image from (armbian storage). I let it sit for a long time and no light came on. I may have had the ethernet plugged in, the guy on the video had it that way. Maybe that was my problem. I was able to boot up and configure but it is running off the SD card. Then I used kiauh to install klipper, moonraker, and mainsail. At some point I'll go back and install it cleanly from scratch. Thanks.
 
and was using the q-tip in the audio port. I used an image from (armbian storage). I let it sit for a long time and no light came on.
i was like 95% sure it was what you did, as it was known as a standard procedure to flash it - with a 20sec press into the audio port.
but here, nope, just insert card, plug in power and it would do it yourself, with that image. all other cords should be out including network.

if you can run it on ethernet wire - then the card pastera got is a better solution, but i needed specifically one with a wifi.
 
Im going to need a custom front sight made for my 1022 M1 project
A modified M1 garand front sight style
arguably it would be a better project for a 3d milling than printing, would probably be better to make it out of aluminum. even though a nylon should hold, but, to think of a sight, it probably would be beneficial to have it done out of metal, not any kind of plastic.

it really is going to be my next step, i feel, just need to research more out of those milling machines, that market does not seem to evolve as fast as 3d printers market does.
 
Im going to need a custom front sight made for my 1022 M1 project
A modified M1 garand front sight style
arguably it would be a better project for a 3d milling than printing, would probably be better to make it out of aluminum. even though a nylon should hold, but, to think of a sight, it probably would be beneficial to have it done out of metal, not any kind of plastic.

it really is going to be my next step, i feel, just need to research more out of those milling machines, that market does not seem to evolve as fast as 3d printers market does.
I'd design in a CAD app, then 3d print prototypes or even a temporary one in nylon. If you need stronger, can ultimately cast in metal (lost PLA casting).
CNC milling machines are expensive and space-consuming; for my occasional use, I prefer to just use the machine at the local makerspace.
 
Last edited:
arguably it would be a better project for a 3d milling than printing, would probably be better to make it out of aluminum. even though a nylon should hold, but, to think of a sight, it probably would be beneficial to have it done out of metal, not any kind of plastic.

it really is going to be my next step, i feel, just need to research more out of those milling machines, that market does not seem to evolve as fast as 3d printers market does.
Nah , the plastic would be just fine for this project
 
Updated my X5SA Pro to a Creality sprite extruder and klipper.
The t95 boxes are fine if you get one works right out of the box and doesn't overheat in the middle of a print

Libre LePotato (Pi 3b clone) runs fine fine for $35 and is easy to hook up I/O for power control and an accelerometer.

Haven't explored the edges of speeds but travels at 300mm/s and printing at 200+ with good quality.
Limit on print speed is a standard nozzle can only melt around 14mm^3/sec.

Did try out printed powder through expanders for 9mm - works great and they only take a couple of minutes to print. Time will tell how they last - if they don't, no worries since I have a lathe but knowing you have the right dimensions before cutting steel is always better.

If the plastic expanders seem to last I may offer custom sets with the expander and a seating die for different cast bullets (put 2-3 expanders in to make sure it's worth it)
 
Finally getting back into this... I bought an Ender 3 Pro at the beginning of Covid. I printed a bunch of stuff then got busy. Over the holiday I went back to finish the bullet feeder I started back then. I've been catching up with everything that has changed in the last 2.5 years. I just put a glass bed and yellow springs on. I started to fiddle with Klipper and got really frustrated trying to install it on this ancient v1.1.4 8-bit board. Something about it not having a boot loader. After fiddling for a few hours I said screw it, I've been wanting a silent board anyway. Thanks to Amazon prime I just installed a v4.2.7 32-bit board: I should have made this upgrade a LONG time ago. It's 1000x quieter and less annoying!
I can load a bootloader on that old board if you want
But I'd you have a 4.2.7 then the best use for the old board is a trash filler.
 
I can load a bootloader on that old board if you want
But I'd you have a 4.2.7 then the best use for the old board is a trash filler.

I played a bit with input shaper and pressure advance over the weekend. Even with the print and measure approach the results are impressive. ADXL345's and pico should arrive tomorrow.

The Libre LePotato looks like a nice option with the raspi availability issues. I may pick up to play with. I have three pi's (2-2's, 1-3b) in service for the other purposes.
 
I played a bit with input shaper and pressure advance over the weekend. Even with the print and measure approach the results are impressive. ADXL345's and pico should arrive tomorrow.

The Libre LePotato looks like a nice option with the raspi availability issues. I may pick up to play with. I have three pi's (2-2's, 1-3b) in service for the other purposes.
The print and measure process gave me good results but didn't play well with PA.
The calculated results for input shaping is giving much better results and PA isn't freaking out and causing gaps at layer change.

On a corexy printer, klipper really improved the speed/quality even with a 32 bit controller.

I haven't swapped my ender 3 over to klipper so I'm not sure how much of an improvement that will get me. Will likely just run Octoprint on the LePotato to run the Ender and add relays to control power to both printers.
 
Limit on print speed is a standard nozzle can only melt around 14mm^3/sec.
playing with pla+ - sunlu one seems to be going out just fine at 20mm^3/s at 240C. seems to be maxing out around 25, where is starts producing gaps in outer walls.
what material you see to be limiting at 14?
 
playing with pla+ - sunlu one seems to be going out just fine at 20mm^3/s at 240C. seems to be maxing out around 25, where is starts producing gaps in outer walls.
what material you see to be limiting at 14?
I don't turn the temperature up to get more speed so I'm running temps between 195-205 for PLA.
I haven't tried an extrusion test at 240 since bridging and support separation suffer at too high of a temperature.
 
I don't turn the temperature up to get more speed so I'm running temps between 195-205 for PLA.
I haven't tried an extrusion test at 240 since bridging and support separation suffer at too high of a temperature.
not pla - pla plus. which are also not made all equal.
i switched to pla plus 100%, and the sunlu kind is very temperature and fluidity flexible - prints pretty well from 250 down to 210.
and sunlu meta ended up to be a complete crap, it maxed out at 210 and just would not be printing with a stable good finish at different layers speed.

Here is this 240C print, max speed set to 120, and finish wise it is pretty even.
05809D2B-67C9-49AA-9266-7DBE66326E20.jpeg
 
not pla - pla plus. which are also not made all equal.
i switched to pla plus 100%, and the sunlu kind is very temperature and fluidity flexible - prints pretty well from 250 down to 210.
and sunlu meta ended up to be a complete crap, it maxed out at 210 and just would not be printing with a stable good finish at different layers speed.

Here is this 240C print, max speed set to 120, and finish wise it is pretty even.
View attachment 712975S.frw
I print mostly with 3d Solutech PLA since that's what works for my needs out of the box. For items that need different materials I have TPU, PETG, Nylon and ABS on hand.
I also don't normally print at layer heights over 50% of nozzle diameter in order to preserve visual quality and printed thread fit.
I understand that you are trying to commercialize your prints so you need to bias toward print speed - Things that I design/print for other people are usually cycling related so visual quality if paramount. Printing at 0.12mm layer heights or lower is not uncommon in order to get as close to injection molded surfaces as possible.
 
I print mostly with 3d Solutech PLA since that's what works for my needs out of the box. For items that need different materials I have TPU, PETG, Nylon and ABS on hand.
I also don't normally print at layer heights over 50% of nozzle diameter in order to preserve visual quality and printed thread fit.
I understand that you are trying to commercialize your prints so you need to bias toward print speed - Things that I design/print for other people are usually cycling related so visual quality if paramount. Printing at 0.12mm layer heights or lower is not uncommon in order to get as close to injection molded surfaces as possible.
Yep, 0.12 layer speeds are not sustainable for what I do there, but as mold model - sure, it makes perfect sense.

I print now at 0.24 and 0.3. Sample above is at 0.3.
0.24 visually is arguably more pleasant but it depends upon the item.
 
Yep, 0.12 layer speeds are not sustainable for what I do there, but as mold model - sure, it makes perfect sense.

I print now at 0.24 and 0.3. Sample above is at 0.3.
0.24 visually is arguably more pleasant but it depends upon the item.
Going with a 0.6mm nozzle and 50% layer height should give you the speed and better wall appearance as long as you aren't printing 0.4mm wide features (but those aren't going to reproduce properly at 75% nozzle height anyhow)
 
Going with a 0.6mm nozzle and 50% layer height should give you the speed and better wall appearance as long as you aren't printing 0.4mm wide features (but those aren't going to reproduce properly at 75% nozzle height anyhow)
i did try 0.6 nozzle, it did not make surfaces i liked and small holes i also need were all ruined. i used it at 0.3 height. i still have plenty of them, nozzles, but kinda lost any interest to even try again.
0.24 is really a best compromise for me now, with a 0.4 nozzle. sunlu pla+ filament flows very well and forms perfectly at 0.3 as well. its normal print temp window is from 210 to 235.
 
to be honest, i feel like i am losing interest to all that. ebay sales are extremely sporadic, sometimes it hits ok-ishly well, like up to 4 sales in a single day, and then it is just nothing, like now - 4 days nothing, and almost even no views. barely worth my time as is.
and i am fed up with usps shipping. it is just a nightmare.
 
to be honest, i feel like i am losing interest to all that. ebay sales are extremely sporadic, sometimes it hits ok-ishly well, like up to 4 sales in a single day, and then it is just nothing, like now - 4 days nothing, and almost even no views. barely worth my time as is.
and i am fed up with usps shipping. it is just a nightmare.
Until PayPal went insane, I made more in donations (tips) for publishing my designs than actually selling printed parts.
 
Hard to put to words but I need a M1 style front sight but modified to attach to lyman match style front sight base. It also needs to be a good amount taller so I can get down to at least a 50-75 yard zero.
front of the barrel will heat up considerably, so, it is a bitch to do, as it will need to wrap the base so close to the end of the barrel.
got to be nylon with a fastener, and some sort of a contraption to accept and hold steady an iron front sight something assembly, as to print an actual sight from plastic would be probably futile accuracy wise and prone to bending or breaking off.
i can be wrong about it, but, my gut feeling is - it needs to be a metal insert there.

so, is it something like this you want?

41eKqLr9Y9L._AC_.jpg
 
Last edited:
Back
Top Bottom