Filament Empty Alarm

The place to discuss filament...
User avatar
pyronaught
Posts: 684
Joined: Mon Dec 01, 2014 8:24 pm

Filament Empty Alarm

Post by pyronaught » Sat Jun 03, 2017 5:15 pm

After having so many prints ruined from filament run-outs, including while I was even sitting in the same room and did not catch it in time, I finally got around to making a simple alarm that can be heard anywhere in the house when a spool runs empty. The alarm can of course be turned off so that it doesn't go off when you change out spools. This is a bolt-on replacement for the standard M2 filament guide, and runs on a 9v battery instead of the more expensive and lower voltage button batteries most people use for these alarms.
alarm1.jpg
alarm2.jpg
alarm3.jpg
alarm4.jpg
Experience is what you get when you didn't get what you wanted.

User avatar
insta
Posts: 2000
Joined: Tue Sep 16, 2014 3:59 am

Re: Filament Empty Alarm

Post by insta » Sat Jun 03, 2017 6:46 pm

I realize this is just an SPDT leaf switch with the filament holding it closed, and the beeper on the NC contact ... but you can use the same switch, and feed it to one of the MAX endstop wires, change some pins in configuration.h, and get firmware-supported filament changes on runout.
Custom 3D printing for you or your business -- quote [at] pingring.org

User avatar
pyronaught
Posts: 684
Joined: Mon Dec 01, 2014 8:24 pm

Re: Filament Empty Alarm

Post by pyronaught » Sat Jun 03, 2017 9:21 pm

That all sounds possible, but I'm not motivated enough to do it. I can catch 80% of my run-outs with just the buzzer and still have time to feed in a new spool. It's the overnight run-outs or away from home ones that will still get me. It has the same loudness as a smoke detector alarm, so it might actually be loud enough to wake someone up at night if not too far away and not behind closed doors.
Experience is what you get when you didn't get what you wanted.

User avatar
zemlin
Posts: 457
Joined: Thu Feb 04, 2016 9:32 pm
Location: Indianapolis, Indiana

Re: Filament Empty Alarm

Post by zemlin » Sun Jun 04, 2017 2:03 pm

If I think there's a risk of running out of filament, I weigh the spool before I start. I have saved an empty spool from each filament supplier I use for a tare reference. The weight estimates from S3D seem to be pretty much spot-on. I use 1.05 for ABS density and 1.25 for PETG. I've run prints from a spool that had just a few extra grams of material over the estimate - and there might be less that one turn of filament left on the spool at the end of the print.

steve220
Posts: 67
Joined: Tue Oct 11, 2016 2:32 pm

Re: Filament Empty Alarm

Post by steve220 » Sun Jun 04, 2017 3:59 pm

This looks good. I bought something similar off Amazon, but the alarm isn't loud enough for me. I'd like to make this.
Any chance you can post the STL files and part list?

User avatar
pyronaught
Posts: 684
Joined: Mon Dec 01, 2014 8:24 pm

Re: Filament Empty Alarm

Post by pyronaught » Sun Jun 04, 2017 4:29 pm

zemlin wrote:If I think there's a risk of running out of filament, I weigh the spool before I start. I have saved an empty spool from each filament supplier I use for a tare reference. The weight estimates from S3D seem to be pretty much spot-on. I use 1.05 for ABS density and 1.25 for PETG. I've run prints from a spool that had just a few extra grams of material over the estimate - and there might be less that one turn of filament left on the spool at the end of the print.
I'll often run prints knowing that the spool is going to run out with the plan of feeding in a new spool to help clear out all the partial spools I have laying around. This just seems easier than splicing partial spools together and then tediously winding one onto the other. The problem is I don't catch when the spool runs out half the time.

That's interesting that the S3D weight estimate is so accurate, given that their time estimates are way off. I've found that the actual print times are about double what S3D says they will be.
Experience is what you get when you didn't get what you wanted.

mathisyourfriend
Posts: 100
Joined: Sat Apr 19, 2014 8:34 pm

Re: Filament Empty Alarm

Post by mathisyourfriend » Sun Jun 04, 2017 11:02 pm

"That's interesting that the S3D weight estimate is so accurate, given that their time estimates are way off. I've found that the actual print times are about double what S3D says they will be."

I've been wanting to post about this. I've had S3D since the beginning (it saved me from RepG!!!). The S3D weight estimates are consistently within 2-3%. On my last print S3D estimated 220.3g and the actual weight with supports was 223.1g (cheap digital scale but should be close enough). The time estimate was 10' 30" but the actual time was 14' 05" (it took over 1/3 more time than S3D estimated).

The time error seems to get exponentially worse as volume increases. (More than the 8x one would expect when doubling the size of a print). I haven't kept good enough records to prove it since I'm not using Ella Guru for my living so printing time isn't as important to me as it is to some of You.

Also, something like a Hedgehog (basically a hemisphere with lots of spikes shaped like square pyramids and a cute face) increases the perimeter a lot but S3D doesn't seem to account for the time required. I've seen S3D underestimate times on Hedgehogs by 60% to 100%.

User avatar
zemlin
Posts: 457
Joined: Thu Feb 04, 2016 9:32 pm
Location: Indianapolis, Indiana

Re: Filament Empty Alarm

Post by zemlin » Mon Jun 05, 2017 2:07 am

pyronaught wrote:That's interesting that the S3D weight estimate is so accurate, given that their time estimates are way off. I've found that the actual print times are about double what S3D says they will be.
Calculating the amount of filament is infinitely easier than estimating time.
Material density and filament diameter are entered by the user.
The only other factor is the total amount of extruder travel, so add up all the extruder instructions and you're done.

Estimating time is an entirely different beast.
For every X, Y, and Z movement you have to take into consideration velocity, acceleration, and jerk. Different movement will happen at different speeds depending on how you have your underspeed settings entered. Acceleration and jerk are not factors in the UI of the software - those might or might not be included in the start script.

I just looked at the GCODE file of a batch of 14 parts I ran yesterday - was probably around a 24 hour build. There are 3,088,087 instances of "G1 X" in that file. Every single movement of the gantry needs to be evaluated for speed, acceleration, jerk, and perhaps other firmware-dependent factors, and there are interactions that come into play as well.

Because of the shear number of instructions that need to be evaluated, the smallest error in the estimate multiplied by the millions of instructions that make up a program will result in huge accumulated errors.

mathisyourfriend
Posts: 100
Joined: Sat Apr 19, 2014 8:34 pm

Re: Filament Empty Alarm

Post by mathisyourfriend » Tue Jun 06, 2017 1:25 am

I figured the weight estimate was much easier (as you pointed out it's simple math really). I've also looked at the sheer number of moves a print can need and it is amazing.

User avatar
sthone
Posts: 876
Joined: Tue Jul 08, 2014 8:25 pm
Location: Connecticut
Contact:

Re: Filament Empty Alarm

Post by sthone » Sat Jun 10, 2017 5:04 pm

I'm always missing swap outs too while in the same room (easily preoccupied) and I've been meaning to make one of these for awhile... Pyro finally inspired me.

I dug through my pieces and parts boxes and actually found some stuff that would work.
Runout001.jpg
Runout002.jpg
It definitely won't be as loud as Pyro's with this tiny buzzer but it should get my attention while I'm in the shop anyway.

Thanks for the inspiration Pyro. :D

-Steve
____________________________________________________
See my projects at https://www.theneverendingprojectslist.com

Post Reply