Failure to Print Two part statement (a question,

Failure to Print

Two part statement (a question, and a thought process)

It occurs to me that I have a frankly unacceptable level of failures. I Hope to correct this when filament from a completely different vendor arrives sometime this week. So, the question:

****How long have you been printing? What percentage of those prints are failures?

****What mechanisims could be put in place to reduce those failures?

I see the effort placed in auto-bed levelling, so there’s clearly a large amount of brainpower involved at the bleeding edges of the hobby (as presumably a printer in calibration will stay in calibration.)

It seems to be there needs to be a method to determine when filament stops flowing. Whether this is an optical process whereby nonintrusive marks are placed on the filament, then read after the extruder (Bowden), or something that monitors the power sent to the extruder over time (presumably, a non-flowing extruder’s power requirements are different than one operating properly).

I could see a two part camera based failsafe that ‘looks for daylight’ under the extruder then reacts if it sees something it shouldn’t…but then there also needs to be a recovery process that allows the printer to step back through the g-code and pick up where it leaves off.

I’ve had some small successes in recovering from failed prints, that process consists of: unblocking the nozzle, manuallg jog the nozzle down to the Z height of the failure, and manually searching for z=, reinitializing the variables, then kicking the process back off.

Failure occurs when the system looses calibration of the effector location (due to me taking too long), or thermal events prevent the filament from sticking to the cold part (but that was ABS, on a printer that really isn’t configured for ABS)

We’re working with some serisouly small tolerances, and there’s been a LOT of people looking at printing…is there no solution because it’s not an easy problem to solve, or am I standing on an edge that hasn’t actively been looked at? (I can’t believe the big players haven’t looked into it…especially if pre-marked filament would mean patents, proprietary cartridges and the marketing for ‘fail-proof prints’)

Most of my failures are ‘soft’, meaning that the print competed but not to the quality that I would want it to be. Those are much harder to fix. For instance, the top layer has holes and cavities in it. This could stem from a dozen different sources, from the slicing algorithm to the air moisture level near the printer, age of the filament, temperature, belt tension, really I could spend a day listing possible causes.

I recall someone [can’t remember who] was developing an active filament gauge [optical] a few months back. I would suggest that most printers [people] get through the mechanical s*** in the first 3 months, and then deal with print specific issues as @Shachar_Weis suggests. Material changes will continue to plague those of us that play with a variety… PLA, ABS, Nylon, Tglase and all the possible colours … the list of variables seems infinite some days !

What would you say your success rates are? (Where you’re completely happy with the output.)

I would say 70%, but my printer is a cobbled together DIY hack I designed myself, and therefor is riddled with design flaws that I am too lazy to fix.

Mine have been filament and bed issues. But ive tied the bed issues down its now filament problems.
Mainly due to some filaments absorbing moisture faster than expected and others still dry even though they are out for around the same amount of time.

3 years, estimated 98% success rate. Find the right temp for your filament, use a driving gear with a bearing pressing HARD on the other side, and make sure the filament spool will spin easy and not get kinked. Most of my failed prints were the result of the filament spool not spinning freely, but now that i have them mounted sideways on a lazy susan instead of spinning around a shaft horizontally in the middle of the spool its mostly making sure the filament doesn’t unspool too much and get wrapped around the bearing plate. Used to have issues with hotend but this one is chugging along fine. I’ve run about 20 pounds or more since my last cleaning and it hasn’t clogged yet. My bed gets unlevel when i try too hard to pry a part off, but if i watch the skirt going down i can tell where its off and fix it before the part actually starts. 3 point manual leveling of the bed is 100% better than 4 point leveling. Warp was an issue but i built a chamber around my printer so now thats a thing of the past. Also i only print ABS and my primary source is http://newimageplastics.com

Phoenix3dprinter has a rewind function they were building into their software. Kickstarter, not shipping yet (that I know). It could pick up where a print failed according to the video on their site, so someone is looking at it.

@Brian_Gudauskas rewind is useless unless the printer can detect failure automatically, stop the print and notify the operator. I don’t see any sensors that will allow them to do that.

Maybe, maybe not…if the printer could recalibrate, and the user jog it to the fail point, then it would be programattically simple to step through the G-code to get to that spot.

@Shachar_Weis I agree, just pointing out that someone was thinking about failed prints.

@Mike_Miller not easy to rewind to .1 or .2 height and nozzle width by eye if your doing a detailed print.

print rewind seems like a whole lot of work to salvage a few cents, or maybe a dollar’s worth of filament when its so much easier to pull the mess of the bed and restart. Making sure the printer is reliable should be a far more important goal than recovering from a failed print. Especially since recovery means watching the print. If the printer is so unreliable that you need it to be able to rewind a print, then you need to fix the printer. What is the point of a robot if you have to babysit it?

Reasonable point, and I’ll admit, my comments are based on my current situation with an unreliable robot.