Do you splice rolls of filament and, if so, how do you do it?

Today, I ran into the problem every 3D printer (as in the human variety) has to face eventually: that is, I need to print something that will take more than 500 grams of filament but don’t have any PETG rolls with that much left.

I watched a number of YT videos showing different techniques, some of them quite ridiculous in the amount of time and effort involved. In the end, I decided to try the method described in this video: [U]How to join ✂ 3D printer filament (No special tools are required) - YouTube.
Basically, you cut a piece of teflon bowden tubing around 2" long and slide it over the filament end of 1 of the rolls to be joined. You then heat each filament end slightly and pull it gently to thin it out. You then trim the thinned out section to within about 2 mm of the regular part of the filament. Then you heat the 2 thinned ends at the same time, push them together and slide the bowden tubing over the new joint, rolling & sliding it back & forth until it cools. With luck you wind up with a pretty nice joint.

It took me a couple of tries, but I got it to work.

The only issue is having to hang on to the piece of bowden tube while you wind the filament from the smaller roll onto the larger one. I think I’m going to need to print a filament winder.

I am a little surprised that there doesn’t seem to be any discussion on this topic on the forums. It would seem like a natural topic to me, unless everyone just throws out the leftovers.

Update: by some miracle, I happened to be checking in on the current print when I noticed a slight change in reflection off the filament. I wondered if it was the spot where I welded the 2 spools together. I could feel a very slight bump in the filament, and I do mean slight. Since this was my 1st filament splice, I decided to hang around and see if it went through OK, and whether it made any difference to the print.

I can’t see any difference. And this splice is joining 2 different brands of black PETG: Hello3D and Overture. There is a very small difference in the sheen of the 2 brands, but I don’t care for this use.

Anyway, it looks like the splice worked.

I’ve looked at all of those, and others. What I’d really like is for Mosaic Mfg to use the splicing core for their Palette multi-material splicers to make a simple filament splicer to use up leftovers. I bet they could sell a ton of them.

yeah, the filament splicer within the palette is nice, but I guess still to expensive as you need most of the components.

You need the heater, the cutter, the control board, at least two steppers with a proper gear set, … I guess it is cheaper to build such thing on your own. Such thing would still be 200 euro or more.

Thanks for starting this thread. I do not splice filament but will try in the future! This looks easy to make: [URL=“Soldering Iron Powered Filament Joiner - Make Multicoloured Parts by QUBD_Broadlands - Thingiverse”]Soldering Iron Powered Filament Joiner - Make Multicoloured Parts by QUBD_Broadlands - Thingiverse


I was thinking, how about an Arduino based system? Take a small aluminum or brass block, drill a 1.78mm hole in it, plus a hole for a hotend heating resistor. Also attach a thermistor. Heck, just take a heater block, complete with resistor & thermistor, from a cheap 3D hotend. Connect the thermistor to the Arduino (to analog ports or PWM???). Connect the heater to an Arduino shield?? that can provide 24V in sufficient amperage. Connect 2 buttons to the Arduino. One button performs a reset of the Arduino. The other tells the Arduino to go into a mode waiting for user input. When the reset button is pressed the Arduino runs a program that sends current to the heater resistor for a specific time (to be determined), then automatically goes into a “get user input” mode, after turning on an LED. In fact, use red & green LEDs: red while heating, green after the timing loop is complete. Turn on both LEDs while the unit is heating up.

OK, that’s the dream. Unfortunately, I don’t know all that much about electronics; only some basic stuff. However, I have been following Paul McWhorter’s video series on Ardino programming, and I suspect this is doable. Perhaps some of you with a lot more knowledge of electronics can step in.

All told I’m thinking this could cost less than $100, which is really a good investment when you consider how much filament costs.

The really hard part, I think, would be getting a perfect filament-sized hole into the block. You can’t just drill a hole then cut the block in half, cause that would make the hole too small. Perhaps a rounded end mill could be used after the block is cut in half, but I don’t have a milling machine.

I swap my smaller rolls in when I have minor prints like name signs or other 30 minute prints. I also printed LED inserts. That way I mostly never end up with more than a few meter of filament.

At a price of 12 Euro per Kg that is next to nothing. These filament ends also are handy for printing calibration tubes and stuff.

Yeah, at $30/Kg, I pay around twice what you pay, so it matters more. Also, I don’t ever want to get caught with a ruined print because it ran out of filament.

Also, I found this:…er-controller/. Not surprisingly, somebody beat me to the punch. This guy has used an Arduino Nano to create a fusing conroller, at a pretty reasonable price. I intend to contact him to find out what he did for a ‘fusing block’.

I tried splicing bits and pieces, but I think a better way is to get a filament sensor. It depends on your printer and firmware, but most printers will see the filament out signal and perform a M600, move to a parking position, retract and wait for the reload.

How has this worked for you? I’ve been leery about it because I worry about how well the new filament will adhere to the previously layed down filament.

That would definitely be my solution though I haven t installed one yet. I probably will one of these days.

Filament runout sensors would not work for me. I perform longs prints over night, so even if there is a problem I won´ t notice. And the printer can scream as loud as he wants because, I won´t hear it.

I also don´t like it having the printers build plate heated up without actually printing for hours. Without the part would get loose and the print would fail anyway. Also you need to keep the motors running unless your homing is very precise.

On the other side you have the sensor dangling around with some filament inserted to keep it quiet or you need to fiddle filament through the extruders and the sensors before you can start printing. You also need a very good sensor to detect filament clogging or breaking.

I may be alone with this, but this adds to many additional reasons to fail a print that the situations where it would have helped are minor. Over the years I got quite good in guesstimating the filament on spools and if it is enough for a specific print.

To return to the topic: I go for print with the remains, pause printer by hand, swap filament to the next small amount I have, print that and once I am done, I would simply swap to the next full roll. That way you need only to be around for the first few hours or so. Of course that wouldn´t be that easy with having to filament roles with 200gram each on it.

I guess there would be some home made stand alone sensor would come in handy. You simply clip it on next to the roll and when it ends it just screams, while it falls off. Or with some 8266 it could send out an email or tell your smart home.

This would be more handy than an installed printer sensor you always need to cheat around.

Speaking of 200g, that’s just about what my situation was. I had 2 rolls of black PETG, each with around 270g left, but I needed to print a 4th deck umbrella bracket that requires 450g to 500g. The print takes around 60 hours, so there’s no way I’m hanging around. That’s why I spliced the 2 rolls together, to get somewhere around 540 grams. I used our digital kitchen scale to get the net weight of filament before starting the print. It turned out to be enough, with maybe 10m to 20m left. To me, that’s not a lot of filament to spare.

@Geit, I agree about having the motors running & bed heated for no good reason. I might find a sensor useful for when I’m not at the printer, but somewhere nearby, but I think I’d rather splice the rolls together, go by weight, and not worry about filament runout at all.

A few days ago, I let the filament run out while watching and the sensor never stopped the printer! I had to manually pause it. I’m guessing dust buildup, so I need to add that to the maintenance schedule. So far, I’ve had filament runout sensors cause more problems (false alarms) than saving a print, but printer reviewers love them.


R U telling me they used an optical sensor in the filament sensor you have? Seriously? Surely an old-fashioned micro switch would be a much better way to go.

I’m not 100% sure, but there are various types. Either way, it failed.


I’m betting you could, pretty easily, print a design that uses a micro switch.

They often to. In fact to do it right they needed a motion sensor as well. Also a filament runout sensor with an optical sensor sounds far better than just a sensor with a mechanical switch from a selling point.

However in a closed case and with a cleaning sponge on the IN side the optical sensor should not get dusty. On the other side even mechanical Endstops fail, when filament, you cut from the tip of a spool, is jumping away it may end up between you X, Y or Z endstop as well and blocking the mechanic. Sure it is rare, but it happens. Had a filament clip on the Z belt once. It got mangled under the GT2 pulley and spit out on the other side. The print got ruined as it caused a Z layer shift.

Shit happens and the more tech you add, the more could go wrong.

If it gets too annoying, I’ll just take it off. :frowning: