Page MenuHomeAleph Objects Inc

usb bezel v1.1
Closed, ResolvedPublic

Description

10x black abs

Event Timeline

west created this task.Apr 25 2019, 7:08 AM
EricNugent moved this task from Restricted Project Column to Restricted Project Column on the Restricted Project board.Apr 25 2019, 9:50 AM
EricNugent reassigned this task from EricNugent to zachah.May 2 2019, 7:09 AM
EricNugent added a subscriber: EricNugent.
zachah added a comment.May 2 2019, 9:05 AM

zachah added a comment.May 2 2019, 9:05 AM

sliced and sending over to cluster

Re-sliced without a brim because the brim removal seems to be the biggest concern here

zachah added a comment.May 3 2019, 2:25 PM

zachah added a comment.May 7 2019, 2:46 PM

Print without brim failed. One of the ten came loose from the bed and they all looked like they were printed too hot.

Sliced and printed with a lower temperature with more fan and re-printed without success.

Sliced again at a slightly higher temperature and less fan and will print again.

Re-sliced this time as one at a time and with the print speed a little slower

Trying another slice with a much slower speed. Without the brim the parts keep lifting off of the bed about an hour in{F3427073}

west added a comment.May 8 2019, 2:01 PM

do you have images of the failures?

zachah added a comment.May 8 2019, 3:21 PM

I just took some pictures of the failures for you. The group of three is where the temp was a bit too high and one of the small legs came loose during the print.

The little bits are where it failed earlier on. The small legs look like the point of failure, they keep coming loose before the print finishes

zachah added a comment.May 9 2019, 6:21 AM

Printing them one at a time and with less on the bed to see if that helps with the parts getting knocked over{F3431841}

west added a comment.May 9 2019, 7:18 AM

Have you tried the standard print profile? Thats the one ive had luck with. The changes I made to the standard 0.25 is increase the bed and hotend temp by 5 degrees. I'm also not printing in an enclosed area so those changes may not be necessary for cluster printing

zachah added a comment.May 9 2019, 8:51 AM

That may be why I'm having some troubles. I've been using the high detail profile. I've just sent one over sliced with the standard profile.

zachah added a comment.May 9 2019, 9:55 AM

Failed with standard settings. The temperature looks like it's too high or cooling is not high enough. While watching them print I could see the small legs lifting slightly on the top layers creating a lip for the nozzle to get caught on and knock the part loose. I've re-sliced with additional cooling to see if that can resolve that issue by solidifying the small parts into place faster.

zachah added a comment.May 9 2019, 2:06 PM

Got 4 separate successful prints with none failing on bed adhesion. Noticed the hole for flashing the firmware did not look good the lines did not connect fully.

zachah added a comment.May 9 2019, 2:09 PM

Noticed in the slice that the gap was visible in layer view, and saw I had disabled fill gaps. With that enabled the gap is filled in sliced and sent over with that enabled.

zachah added a comment.May 9 2019, 2:09 PM

and the gcode

zachah added a comment.May 9 2019, 4:54 PM

The last prints came out looking good

zachah moved this task from Restricted Project Column to Restricted Project Column on the Restricted Project board.May 9 2019, 4:54 PM

For the last print I did 5 at a time to reduce extra movements that may have been knocking the prints loose. Slicing and printing x10 at a time next to see if the changes permit.

mbloom added a subscriber: mbloom.May 10 2019, 7:06 AM

Here's the file that I worked on a few weeks ago. David said they looked good and I haven't had a failed print yet from them.

Gcode for 15 parts at .3mm layer height on Mini 2;

Profile for the print;

I've also started a print of them on F. Not all of my settings were saved, so the support xy distance might need to change. Let me know if .3mm is a good layer height. I don't know what settings we typically use for production prints.

The x10 gcode


Changed support xy distance from .2 to 0.

David just came by and told me to share the stl I was using. It's the stl in
shared-j/devel/lulzbot/TAZ/quiver/production_parts/printed_parts/USB_bezel/
I haven't changed it at all.

Last batch of 50 printed 10 at a time on each printer. Only 4 out of the 50 failed due to bed adhesion

Just finished printing this gcode on a Mini 2;

All of the pieces came out perfect. The print time was 2:44 for 15 pieces.

I also made a Mini 1 file but couldn't get a taz file working, but that's not really a big deal.

Mini 1 file;

Addressing the last log from @zachah "only 4 out of 50" fail is almost 10%. That is far too high!

west added a comment.May 13 2019, 7:23 AM

@zachah

that has the connecting layer at .8mm

tutley added a subscriber: tutley.EditedMay 13 2019, 7:43 AM

@david.hall this ticket is for a slice change, not a model correction.
All model changes MUST go through an MCO since the machine is in production.
Also we cannot be modeling parts in blender, all model changes MUST be done in FreeCAD.

@tutley . I apologize, but I am not sure what you are referring to. I only made a comment about a fail rate.

@david.hall sorry if im mistaken, it appears the model was changed by @mbloom in his test gcode

i just want to be sure procedure is being followed as it can introduce larger issues

@tutley @west I believe @mbloom utilized the cura support blocker feature to generate that gcode.

Part of the reason the brim method works is it joins multiple parts together at the base, and the reason why I chose that method in the first place. Built in supports will work but they need tabs to connect neighboring parts together if we are to print more than 1 on a bed.

@EricNugent Thanks for clearing that up!

The last batch of 5 printers printing x12 each did wonderfully. It looks like none of those 60 failed. I'll run another batch to make sure{F3459630}

Had a few batches fail on the last round just sent up another.

Added a small support brim similar to mBlooms gcode with Cura and all 5 batches of the last g-code were successful. Sending another batch to verify the success rate is repeatable.

None of the last 15 batches sent failed due to bed adhesion. I did notice a small layer shifting in a few of the parts from three of the batches though which is not ideal.

Running another gcode with the fans turned down to see if that eliminates the issue{F3463840}

Last gcode did not have any fails due to adhesion or layers shifting the less aggressive fan speed fixed the shifting

pushed gcode into testing branch

zachah moved this task from Restricted Project Column to Restricted Project Column on the Restricted Project board.May 22 2019, 1:16 PM
EricNugent closed this task as Resolved.Jun 10 2019, 6:22 AM