Dual Extrusion: Yet Another Design. Repetier 0.92.4

[gchristopher]

Others have posted perfectly good designs, but sometimes trying to come up with your own idea is the fun part.

So, here's the concept for a dual extruder setup. I'm trying to arrange two EZStruders so that the filament comes down through the center hole. The EZStruder positions are offset so that both filament travel identical paths relative to the center hole. It uses the holes already in place on the top plate.

[img]http://gchristopher.net/3dprinting/dual ... oncept.jpg[/img]

It features Totally-Professional-Looking SeeMeCNC-esque eyes.

I'm hesitating on a first print because I'm worried about the heat from the motors weakening or distorting the bracket.

[ZakRabbit]

That's what this is all about, finding something different! Looks great. It appears the motors will be end to end: if you're worried about heat, why not design a fan shroud that will actively cool both motors? We can all guess and give you our opinions, but you'll never actually know until you try it. Give it a shot and let us know how it worked out.

[gchristopher]

Thanks! The motor ends are actually spaced close-ish so a pair of attached cheap 40mm water cooling blocks will have a very short distance between them. I don't have a good way to bracket the cooling blocks on, but people suggested thermal epoxy, so I'll be trying that.

The eventual dream is to have a heated enclosure for the build area also, and water cooling the extruder motors seemed like a good way to spare them any extra heat from that, too.

[gchristopher]

The first attempt is printed and ready to install. There were a few minor design changes from the image above. The second EZstruder is shown in place.

[img]http://gchristopher.net/3dprinting/dual ... _print.jpg[/img]

The top section fits into the center hole in the top plate of the build area. There are two filament guide holes and one wire guide for extruder, hot end, and thermistor wires.

[img]http://gchristopher.net/3dprinting/dual ... p_view.jpg[/img]

I'm hoping the V-shaped design will add a lot of rigidity to the structure.

One major question will be whether heat from the extruder motors affects the printed mount pieces. I think the motors are topping out around 60C without cooling, so they should be okay. The pieces are also 1/2" thick so they can capture 6-32 nuts without breaking. Also, I need to take the whole machine apart enough to add the extra wire runs for the second extruder setup.

Next up is figuring out how to mount the two E3D V6 hot ends. There's a few nice designs out there, but most seem to hang the hot ends down from the effector platform, since two of the E3D hot ends don't fit through the stock effector.

Does anyone know how reasonable it is to print an effector plate? The STL is available and seems reasonable to print. I'm pretty confident of the dimensional accuracy of the prints I'm getting, so I think it'd be very close to the exact size. Will an ABS printed piece perform a lot worse than the injection molded stock effector?

[gchristopher]

Designing a dual tool platform was reasonably straightforward if you can also print a new effector plate. (With all the important dimensions stolen from the SeeMeCNC STL files.)

This one is intended for the Trick Laser arm attachment, where I can screw the threaded rod through the blocks, then screw the ball end attachments on either side.

[img]http://gchristopher.net/3dprinting/cust ... st_try.jpg[/img]

The major difference to the effector platform is a larger inner radius. This pushes the support columns out a bit. Since this is printed especially for these hot ends, it wasn't too much extra work to integrate an airflow shroud into the design of the effector platform and tool attachment piece. The effector platform has a solid plate keeping the fans from blowing directly onto the hot ends.

The little extra printed material on the support columns is intended to help with getting the hot ends exactly level. I figure I'll file down a tiny bit off the left or right to get them as flat as I can measure. Sort of a one-time-adjustable setup?

I wish I knew if this was an intrinsically stupid idea before I print, assemble and test it all! Does anyone know if I'm missing something basic here?

[ZakRabbit]

I really love the innovation! I'm just wondering why not go with E3D's Cyclops?

[gchristopher]

I really love the innovation! I'm just wondering why not go with E3D's Cyclops?

Thanks!

I think I bought the E3D V6's last year before the Cyclops/Chimera were for sale and the V6's are already here. Were I starting from scratch, I'd buy a Chimera, to permit printing materials at different temperatures.

[ZakRabbit]

I really love the innovation! I'm just wondering why not go with E3D's Cyclops?

Thanks!

I think I bought the E3D V6's last year before the Cyclops/Chimera were for sale and the V6's are already here. Were I starting from scratch, I'd buy a Chimera, to permit printing materials at different temperatures.

Makes perfect sense. Got the popcorn popping so's I can watch what the dinosaurs come up with for you!

[geneb]

I hate to break this to you ZakRabbit, but the dinosaurs are running the show. They just keep him around because he can type faster than they can.

g.

[ZakRabbit]

I hate to break this to you ZakRabbit, but the dinosaurs are running the show. They just keep him around because he can type faster than they can.

g.

Fixed hahahaha!

[gchristopher]

Everyone has found that there's just no way to cram two heater blocks through the opening in the stock effector platform. So one early goal of this design approach is to create an effector platform with enough interior space.

Aluminum would be the material of choice, but I don't have access to the machinery, so let's try to make a printable effector platform for the V2!

A few revisions down the line from the initial design above, here's the first version. It's printed with a E3D volcano, .8mm nozzle, and .6mm layers. (Btw, loving the volcano, my prints are faster than my bed heat time nowadays, need a 24V PSU, I guess!) The results are a little coarse-looking, but should be fine mechanically. The only part that needed any filing was the 6-32 screw holes to attach the tool holder, and that was intentional to keep close to 0 play.

Top View:
[img]http://gchristopher.net/3dprinting/new_ ... v1_top.jpg[/img]

Bottom View:
[img]http://gchristopher.net/3dprinting/new_ ... bottom.jpg[/img]

I ended up going with washers on 4-40 screws and captive nuts to hold the 3mm threaded rods in place. That was inspired by the Cheapskate design, and though it's a little more complicated, I like how there's no reaming of any holes involved, so there's no potential for the manual cleanup to impact mechanical precision. (In theory, the center of the threaded rod should be 33mm from the center of the platform, right? That's what I gathered from inspecting the STL file for the stock effector plaform.)

3mm rods are pictured, because I'm taking this opportunity to upgrade to TrickLaser carbon fiber rod arms.

Next up will be doing the arm upgrade, installing the E3D volcano on the matching tool platform. The upright screws to connect the tool platform have also been moved outward, to 32mm from center.

[gchristopher]

Wow, the Rostock sure is great for making precision parts!

This is with a still mostly-stock Rostock and the E3D volcano hot end. The target was 37mm for the block width on the new effector platform, and once I got the good calipers back, it's as close to that as I'm capable of measuring.

Tyrannosaurus Rex included for scale.
[img]http://gchristopher.net/3dprinting/precision_rex.jpg[/img]

One thing I learned here was that ABS shrinkage seems to be negligible for the "big" measurements on most pieces. (Or, at least, slic3r, any extrusion slop, and shrinkage conspired to make exactly the right size for this print.) Small interior holes still need some extra slop, but the print head on the Rostock is exactly where it claims to be.

[CaptainMerica]

I must say, your design is impressive. I have been working on a new platform for mine to hold a chimera hot end, i plan to eventually have it machined out of aluminum. I'm going to be using this with trick lasers 325mm cf arms but in a double hung setup where they attach. So far i have it so that the mounting block on the platform can be removed and leaves the platform with 2 dowels and 2 tapped holes to secure whatever future end effectors i may wish to put on it.

Noob question, but i suppose i will need a bigger power supply to run both hot ends? I want to do the 24v upgrade regardless, but just wondering how critical it is for hot ends.

Also, any opinions on nozzle distance below the platform? I realize the potential crashes if its too low and the part curls, but what I am wondering, is as far as accuracy, would having the nozzle as close to coplanar as the axi of the platform arm links increase the stability? I picture as the nozzle is offset further and further downwards, it creates a lever that amplifies any slop in the platform, as well as inertia that would possibly make it jump around as well. Or is my anxiety for machine design failures a little high for this?

[gchristopher]

That's what this is all about, finding something different! Looks great. It appears the motors will be end to end: if you're worried about heat, why not design a fan shroud that will actively cool both motors? We can all guess and give you our opinions, but you'll never actually know until you try it. Give it a shot and let us know how it worked out.

You were totally right! The ends are too close and the fan cooling blocks will collide. (The water cooling blocks would fit, but that'll be a ways in the future.) I'm reprinting the holder blocks to go back to the 45 degree mount angle and move them a little farther apart.

I have been working on a new platform for mine to hold a chimera hot end, i plan to eventually have it machined out of aluminum. I'm going to be using this with trick lasers 325mm cf arms but in a double hung setup where they attach. So far i have it so that the mounting block on the platform can be removed and leaves the platform with 2 dowels and 2 tapped holes to secure whatever future end effectors i may wish to put on it.

That should be great! Also, using the chimera is a more sensible option that I'd do if I was spending money on new hotends today.

Noob question, but i suppose i will need a bigger power supply to run both hot ends? I want to do the 24v upgrade regardless, but just wondering how critical it is for hot ends.

Um, the E3D V6 specifies a 25W power draw, so that's a relatively small portion of the overall power budget. I think the bed should draw around 10A at 12V. (120W) The RAMBo maxes out at 15A, (180W). For those numbers the stock power supply should still be fine with an extra hot end, though I'm a little fuzzier on how you estimate the motor power draw. If I had to guess, (and I'm a relatively newbie too; I take my design orders from stuffed dinosaurs) I bet the most likely result would be the bed being even more anemic.

Also, any opinions on nozzle distance below the platform? I realize the potential crashes if its too low and the part curls, but what I am wondering, is as far as accuracy, would having the nozzle as close to coplanar as the axi of the platform arm links increase the stability? I picture as the nozzle is offset further and further downwards, it creates a lever that amplifies any slop in the platform, as well as inertia that would possibly make it jump around as well. Or is my anxiety for machine design failures a little high for this?

I had very similar thoughts, though I can't point to either math or empirical testing or forum posts to back them up. From reading forum posts on other dual extruder attempts, the relative height of the head seemed to be a much more common problem moving the head around than positional accuracy. That seems like it would be less affected by hanging the heads below the platform. I suspect actual testing will be a lot more illuminating than analysis there.

Ultimately, I just wanted to take up the challenge of fitting dual extruders while preserving the aesthetic of having the heads near the plane of the arm junctions as possible. That irrational urge played more of a role than engineering analysis, because most decisions around here are made by a hyperactive Tyrannosaurus Rex.

[CaptainMerica]

I bet the most likely result would be the bed being even more anemic.

Yeah this is what I am gathering from other discussions.

Not to sidetrack but there's basically 2 ways of doing the 24v bed right?

(1) atx power supply for everything (minus bed) + (1) 24v power supply that strictly heats the bed through a relay controlled by rambo's 12v bed heating output.
OR
(1) 24v power supply for everything (after being regulated down to X-vdc for specific components

And past that there is back and forth about fitting everything in the platform or not to for looks or heat or portability etc. Is a large problem the fact that the stock PSU just doesn't have enough muscle for everything? What if, say, I have a crydon SSR on its way, and a pair of 930W power supplies (12v) from a server sitting in my shop. Could a person simply use a relay to supply voltage to the bed heater from a separate source, much like a 24v mod, but even at 12v, just to take the pressure off of the stock power supply? I realize that the bed is only going to use so much power regardless of how much power the supply has to offer, but if the stock power supply is close to browning-out or just simply cant supply the amps to the bed, would it be worthwhile to use a separate power supply just to heat the bed, if even only at 12v? (which could easily be swapped for 24v later too). I guess at a certain point i have to just see how it is first before i get the cart any further in front of the horse.

I'm comfortable daisy chaining power supplies and such. My home computer is a retired mining rig that has 4 video cards quad-crossfired and 2 power supplies to power it all. Pulls as much power from the wall as a microwave if you really load it up haha. So i already have a ton of hardware for frankensteining computers or similar.

back to dual extruder stuff.

Has anyone ran into problems with the weight of extra stuff effecting the vertical movement of the carriage? my goal is to be at or below the weight of the factory components. i started thinking about this when looking at designs that mounted the ezstruders directly to the platform. but regardless of that, and this may be a larger question for all delta printers. should the effector platform be suspended neutral to the motor? Rather than depending on the holding force of the stepper motor? i.e. elevators, giant counterweight that offsets the elevator so that the motor basically is just used to accelerate the mass of cargo up or down. Like, I'm imagining a pulley and weight system on each tower that holds up on the carriage just enough that it acts as if in zero-g. In Cartesian printers, each axis is supported by the bearings, or a lead screw on the Z axis. so naturally the x and y axis only have to accelerate the mass of the printer head back and forth. If the motors that run the position of the carriage are in an open loop (are they?) and don't have any absolute position encoders, would it be possible for the motor to have too much load and not make a step (therefore position=wrong)

I let my adhd run with this idea for a while and sketched up a few ideas in my handy dandy notebook of a delta printer using vertical ball screws to position the three carriages with encoders attached to the end of the shaft to give absolute feedback. Now you have a closed loop and (should) never have to worry about mid-print coordinate shifts/errors. I was thinking about this because if one were to make a delta printer beefy enough to instead hold a spindle motor (or a dremel) to make a desktop mill, you could get crazy with the programming, (way way above what i could ever imagine to do), and set it up so you could tip the platform and effectively make a 5+ axis machine. I realize there are some rotational degrees of freedom that would have to be addressed in the arm linkages that keep everything square with the world, but i haven't gotten that far yet.

I need to quit typing and putting off working on siding my house, so I will leave you on that note.

-Peter Muzzy

[gchristopher]

Not to sidetrack but there's basically 2 ways of doing the 24v bed right?

(1) atx power supply for everything (minus bed) + (1) 24v power supply that strictly heats the bed through a relay controlled by rambo's 12v bed heating output.
OR
(1) 24v power supply for everything (after being regulated down to X-vdc for specific components

And past that there is back and forth about fitting everything in the platform or not to for looks or heat or portability etc. Is a large problem the fact that the stock PSU just doesn't have enough muscle for everything? What if, say, I have a crydon SSR on its way, and a pair of 930W power supplies (12v) from a server sitting in my shop.

Well, I'm relative a beginner too, but from the forums, I gather that some people have better luck with 12V and the Onyx bed than others. I switched to a 1300W PSU and that roughly doubled my heat-up time from the stock power supply. (And improves slightly more with a cloth insulator on top.)

Adding a heat spreader greatly improved adhesion across the bed, but also slowed down bed heat times significantly. That's reasonable, since the whole bed is being heated more evenly. The power supply isn't even closed to being maxed out by this setup, so I know that wattage is no longer an issue and either the RAMBo is unable to drive enough power to the bed, or 12V just isn't enough. There's a small voltage drop across the bed terminals (down to 11.6 - 11.8) from the 12.05V provided by the power supply, but that's probably not too significant.

If you have good 12V power supplies lying around already, I suggest hooking one up and seeing if the bed will reliably hold the highest temperature you want. (105C for ABS for me is a magic number.)

Could a person simply use a relay to supply voltage to the bed heater from a separate source, much like a 24v mod, but even at 12v, just to take the pressure off of the stock power supply?

Sure. My experience suggests that would be identical to replacing the stock PSU with a better one. With the stock 450W power supply, most posts I've read indicate that either would improve bed heating times.

Has anyone ran into problems with the weight of extra stuff effecting the vertical movement of the carriage?

How do you even quantify "problems" in that area? I'd be really interested to know! I see words like "resonance" being thrown around and old posts where they print really wiggly lines and compare them?

If the motors that run the position of the carriage are in an open loop (are they?) and don't have any absolute position encoders, would it be possible for the motor to have too much load and not make a step (therefore position=wrong)

Yes. This happened last night when a print shrunk unevenly and levered a corner off the bed and into the air, where the print head collided with the print, got stuck, and then all subsequent layers were printed with an offset. Catastrophic failure.

[gchristopher]

Argh!

The EZStruder holders are revised, printed and looking good, but aside from the nice aesthetic of having the extruders located there, I'm reconsidering their location.

After all, the intent is to enclose the Rostock in a heated build chamber, so wouldn't it make more sense to put the EZStruder motors outside that enclosed area? The filament feed path would be slightly longer, but probably within reach of the stock bowden tube length.

Either inside the upper area or hanging off the side of the upper printer structure would be workable places for EZStruders.

[CaptainMerica]

im currently printing a redesigned dual ezstruder mount that allows for room for the heat sinks we ziptied onto the extruder stepper motor after noticing it was considerably hotter than the other motors. so i made mounting holes to mount a 120mm fan (low speed n quiet) to help pull across the fins. Ill put up some pictures tomorrow or whenever i get it thrown together.

[CaptainMerica]

Hello again,

Well i have my new platforms in hand, just need to get the cross holes popped in, and I wired up the hot end to the rambo board, following the manual and other posts regarding heat 0 heat 1 t0 t1 etc etc.

printed platform assembled (just for testing)

I'm to the point where I need to make the firmware changes, and I'm not sure if i am just out of it or what but I am simply chasing my tail.. well.. if i had one ya'know...

I have scoured the forums and found various threads with what appears to be a table of the lines that need to be modified, but where exactly do i find/edit/upload/flash or whatever the firmware file?

http://forum.seemecnc.com/viewtopic.php?f=36&t=7532" onclick="window.open(this.href);return false;

http://forum.seemecnc.com/viewtopic.php?f=77&t=7447" onclick="window.open(this.href);return false;

http://www.repetier.com/firmware/v091/" onclick="window.open(this.href);return false;

I'm usually stubborn and proud enough to attempt to figure stuff out myself before ever asking for help but I am just to worn out to care right now haha. (from remodeling my house all weekend)

oooh, also, got my 24v power supply, solid state relay, world of difference, 2 min heatup instead of half an hour, got the trick trucks and 325mm CF arms installed, and the heat spreader. prints have never been better. life is good. but now I have the aluminum platform and the stock hot end got plugged today because the filament had a knot in it (kinda) on the spool and it broke the filament and cooked it in the nozzle for 4-8 hours on high. so i figured why not swap to the new hot end if i have to take it apart anyways.

aluminum platform without cross holes yet.

If anyone can help me out with this I will be ever so grateful. Heck, i'll have 2 extra aluminum platforms if they all turn out good once machined

Thanks,
Pete

[gchristopher]

Wow, great progress!

I've been printing with the printed effector platform for a while now and it's working great, so I think I'm ready to print the dual tool holder section.

I'll be doing the same task of trying to figure out wiring and repetier configuration this weekend. I'll post results when I get something working. (Haha, I sound so confident. In reality, I'll fail, give up, and the more diligent of the two dinosaurs will do all the actual work.)

[Xenocrates]

Some things that may help. (I haven't done dual extruder mods yet, but plan to eventually. The planned timeframe is Soon™)

I would start with the firmware from seemeCNC, or if you've done any firmware tweaks, that. SMCNC github repo

Next up, go to the configuration.h file using the arduino IDE (there are instructions for downloading/configuring it in the manual) Find the line #define NUM_EXTRUDER and change the number after it to two.

Find the extruder offset settings, and set them to be the offset distance from the center point of the effector (this is apparently in number of steps, according to the info i've found so far, and it may be easier to set the offset between one and the other, rather than from the centerpoint, although this would likely put your prints slightly off center and make it difficult to print a near max size object) . Extruder0 is the first one, Extruder1 is the second one. Otherwise, you should be able to mostly clone the settings to extruder1 from extruder0. This is making the assumption that you are using the same motors, drive mechanism, ETC. From what i've seen of this thread, that should be true. The PID may be different, especially if you do something like have different cooling fans, or different heater block styles. Supposedly, you should be able to add an E# arguement to pick your extruder, E0 for the first one, and E1 for the second. It's not standard G-code by a mile, and I've never had a need to test it, but it should work. Your PID may be close between the two. I would not suggest attempting to just calibration on one, and copying that to the other. But that's for after the firmware is loaded.


"// =========================== Configuration for second extruder ========================" should be searchable for in your config, and that way you can find where to make your edits. I suggest copying out the extruder0 config, pasting it into notepad, and doing a find and replace to make it work for extruder 1. then procede to remove the lines between what I pasted above, and #define EXT1_EXTRUDER_COOLER_SPEED 255 including that last line to ensure you don't have duplicates.

Beyond that, I haven't got much to offer. I hope our multi-extrusion guru's have better help to offer than I. Mhackney in particular should be very familiar with this, and thus very helpful. But this should get you started, if it isn't completely off base.

[CaptainMerica]

Well, we have our chimera installed and working.... kind of..

After fighting with the thermisistors wanting to ground to the hot-blocks and therefore each other, and then breaking one of the heaters, it finally prints, for the most part...

We have only done a couple test prints, using only 1 extruder.

First roadblock, the unused nozzle dragging across the print. I think this is a print settings issue, in that the nozzles should be (theoretically) level. but the second one will sometimes snag a corner or piece that maybe is higher than should be because of temp or layer settings. like its swelling or something. almost wander if a layer fan might help? but would likely cause worse problems with my luck. delaminating, cracking, etc. We print almost exclusively ABS and flex. Side note, layer fan for flex? yay or neigh?

Like i said we only have done a few small prints so far, just wanted to check in.

[gchristopher]

Okay, here's a little bit back on the original topic!

The dual extruder holders are printed and installed. The big effort here was a substantial teardown to add t-slot trim and run all the extra wires to support dual extrusion. So far, so good! Here's the current revision of the holder, design view, and printed and installed:

[img]http://gchristopher.net/3dprinting/dual ... design.jpg[/img]

[img]http://gchristopher.net/3dprinting/dual ... t_look.jpg[/img]

I'm not sure I'll stick with them in this location in the long term, but I think those extruder holders definitely achieve the goals set out for them! After a couple more tweaks and some code cleanup, I'll be publishing those.

I need to either find a darker background or stop printing in black all the time. Also, they form a very solid frame; those suckers are some beefy rigid extruder holders!

Regression testing (making sure I didn't break anything) has started, and one test print turned out fine.

Next up will be installing the second hot end and working on firmware changes.

Here's the current design idea for the dual tool holder:

[img]http://gchristopher.net/3dprinting/dual ... design.jpg[/img]

The immediate problem is fan placement. In the pictured position, they'll get in the way of the lash straps on the arms. It'll either have to change to a single fan, or else offset the hot ends some to make space, or both.

The pictured effector platform was been printed and seems to be working great after several prints. It differs from the test print above by adding some support material around the M3 threaded rods to keep them from being bent by the lock nuts when they're tightened down.

So, as far as printer modifications go, things are going pretty well. The wiring is in place and at least didn't break anything. The new extruder holders and effector platform look good to go.

[gchristopher]

Well, it all works!

I called this one a class project for the 3d Printing and Robotics class. Here's the full writeup.

The STL files are posted to Thingiverse and the source code is on github:

Dual Extruder Holders on Thingiverse. Source files (OpenSCAD, stl, dxf) on github

Printed Effector Platform and Dual Tool Holder on Thingiverse. Source files on github

The biggest positive surprise (other than getting dual extrusion working at all) was how well the printed effector platform works. I've been using it for over 50 hours of printing and the print quality is as good as the stock effector platform.

The biggest downside is that even with the wider hole in the printed platform, it's just too tight a squeeze to fit two E3D V6 hot ends AND all their wiring down through it. Fortunately, the whole setup works as an undermount, too!

[img]http://gchristopher.net/3dprinting/dual ... _small.jpg[/img]

Of course, if it's an undermount, then I can just go back to the stock effector platform, so I don't think there's any great reason to use this tool mount design over others. However, the EZStruder holders turned out wonderfully and I would recommend them to others.

Here's a couple action shots:

Printing 2-color ABS:
[img]http://gchristopher.net/3dprinting/dual ... _small.jpg[/img]

Printing ABS with HIPS :
[img]http://gchristopher.net/3dprinting/dual ... _small.jpg[/img]

It's too bad the only limonene I can find locally was in a cleaning solution that didn't work to dissolve HIPS. So far I haven't found anyone that ships limonene to Alaska.

The other major happy point from this process was getting comfortable with Repetier 0.92.4 firmware. Two configuration files are posted to github. Thank you thank you thank you to those who pointed out the single line change that's still necessary to pins.h: jfettig post, Polygonhell. That line in pins.h was the ONLY change outside of configuration.h needed to get dual extrusion working with 0.92.4!

Configuration files for Repetier 0.92.4:
Stock Rostock + E3D Hotend + Bluetooth on Serial 1
Dual Extruder E3D V6 + Bluetooth

[gchristopher]

I finally found some limonene in the organic section of the grocery store! It worked great to dissolve the HIPS filament, but is pretty expensive. (Just over $1/oz.)

Here's a picture of the cube in a cube in a cube with the supports finally removed:

[img]http://gchristopher.net/3dprinting/cube_cube_cube_1.jpg[/img]

There's some interesting artifacts on the print. It looks like the ABS was stringing INTO the area supported by the HIPS, so when dissolved, there's little bits of filament. I can't be sure, though, because this print spent a little over two weeks immersed in limonene-related cleaning products in an attempt to dissolve the HIPS.

Still, for a very first print with dissolveable filament, that's a nice little result. Filing, heat and acetone post-processing are all options to play with for the print artifact issues.