I'm rerunning the test (Suspect sensor wasn't implanted well enough) and I'll let everyone know. (Currently, it's holding 180 pretty well, so I suspect the outcome will be different this time. I think it ran too hot last time.)

Edit: The average diameter improved significantly, so that's good. However, I was getting anywhere from 1.62-1.79 (+- .08). That was 180C with the 1.70mm hole. (I also noticed it was current limiting much throughout 180C, suggesting it was too hot for the size. I've noticed if the temperature is 'too hot', the motor tends to current limit for some unknown reason. At lower temperatures, it doesn't tend to limit the current.)

It seems increasing the hole diameter affected variation in a way I didn't anticipate. Given that there's a greater amount of hot plastic being extruded and constant cooling rate, it might be staying hotter longer thus stretching more, so I might need to bring down the temp to compensate, so I'll try out 170C.

(Granted, it could just be reading the temperature better now since I've implanted it deep and secured it, so 170C nominal on the PID now could be pretty much identical to 180C nominal before.)

Man, all this fine-tuning/testing is a pain in the butt, but I figure it's worth it in the end. (And, I'm convinced I'll solve this one just like I've solved every problem I've attempted before.)

Okay, testing complete. At 170C, it varied between 1.72-1.84. It seems my variation is completely screwed up (+- .06), regardless of temperature, and this low of temperature obviously results in a filament size a wee too large. Yeah, seems my nozzle is kind of fucked at this point, lol. Not sure if it's because the hole is "too large", or because it's not drilled through the back. (Wouldn't think that the nozzle being 85% screwed in would affect variation.)

Well, all this experimentation shows that it is possible to get usable filament at the stock settings.

Ideally you would remove the nozzle and the screen, then drill all the way through. What you've done isn't likely to hurt you by more than 0.01mm of variation, though. Because it isn't a smooth bore all the way through, the flow isn't as laminar (smooth).

Again, what you're seeing in terms of average filament diameter is not unexpected. There a wide range of setups - horizontal, vertical, ABS, PLA, other plastics. I have to make one nozzle size work for all the setups and polymers, and that means leaning towards undersize since that can be corrected in the slicer or with a drill bit. If I provided larger nozzles and someone was producing 1.80-1.85mm filament, you can't fix hotend jamming with a slicer and you can't make the hole any smaller.

That's what I was suspecting. It needs 3-4 feet for proper coiling.

Anyway, it currently looks appreciably thicker. For some reason, it seems like my PID is struggling to get the temp above 137C. (Nevermind, apparently the sensor was a little loose.)

EDIT: And the results are IN! At 175C, taking proper consistent measurements, the diameter was between 1.60-1.66. This is looking good.

(I'd say it's good for production, now, yeayea.)

On second thought, I probably should drill out the nozzle. Increase it by .10mm (Use the 1.70mm drill bit), and run at 180C.

Luckily for me, I have a dremel highspeed drill press for situations like this (I usually use it for PCB work), so drilling is a breeze.

EDIT:

Drilled out with the 1.70 mm drill bit as far as it could go in, several times, through the front of the nozzle. Ran the filastruder at 180C and it varied between 1.55-1.65mm. I also noticed that even though I specified 180C, it was actually running anywhere between 168C to 179C at any given point of time. I'm thinking 175C reduces variation likely because it's a more attainable goal for this machine. I'm not sure if there was an actual increase in extruded diameter, all things considered.

(It's quite possible that my machine is telling me 173C, but it's actually at 183C. Might explain the unexpectedly thin filament.)

Okay, just to clarify. When drilling the nozzle, is it through just the front where the filament comes out, or is it the front AND back? I didn't drill the back because the filter was cemented in place by plastic, and I didn't think it's necessary.

Anyhoo, I took extra steps to /really/ implant that sensor deep into the nozzle's face and secure it, and I'll rerun the test tomorrow. It seems puzzling that it was having a hard time keeping at least 175, but it was having no problems keeping 193 earlier when going for 195. And, the variation of +- .05mm seems like it'd be more typical of 190C than 173C.

Edit: Seeing the temperature chart at http://jcflowers1.iweb.bsu.edu/rlo/Filastruder.htm , I think I'm going to try 175C.

That temperature chart is for a Filabot, which measures the temperature in a different location.

The extra distance (Greater gravitational potential energy) and extra temperature (Greater thermal energy) -> Greater "power output" (Power = energy * time) -> Significantly more extruded. There's no discrepancy. But, the tradeoff was a much greater variation in diameter and much lower average diameter, which is unacceptable for a 3d printer (Variation in particular).

(Higher temp = greater variation since it takes less force to pull the filament when it's runnier at higher temps. Less runny = Lower variation.)

This "power" equation is nonsensical. Increased temperature increases flowrate because it decreases viscosity. Decreased viscosity also means decreased stiffness, which means greater filament diameter variation. Distance from the floor does not change flowrate, it only changes average diameter.

+/-0.04mm variation is better than a lot of commercially available filament.

No, the correct thing to do would be to drill out the nozzle. Moving the Filastruder closer to the ground will cause problems with the filament coiling. It needs to be at least 3-4 feet off the ground.

What is the LCD display showing for current draw? It is possible you set the limiter too conservatively.

It probably has nothing to do with "previously melted plastic high in the barrel that is now semi-solid", that plastic would have been pushed into the hot melt zone in just a few minutes. The more likely cause is that the hotter the plastic, the less drag there is on the motor due to aforementioned viscosity.

swbluto4gems wrote:

Btw, I am seeing somewhere around 18 inches/minute (At ~1.4mm) so the current feed rate seems upto par. It might've been (probably was) misaligned during the purge, but I added alignment marks so I'm sure it's feeding well now.

18 inches per minutes would be 100 grams (0.22 pounds) per hour. This doesn't mesh with your previous statement of 0.3 pounds in 9 hours.

Hey, I already solved it. 5 feet was too high, and 195C was too hot. Brought it down to 3.5 feet and 185C and it's now extruding in the 1.65mm range like normal.

Anyway, there's some key differences between the purge and what I was doing previously just now. I'll highlight them.

Purge: 3.5 feet off the ground, 180C, no colorant -> 1.65mm ave.
Just previous: 5 feet off the ground, 195C, shimmering pearl colorant -> 1.45mm ave.

The extra distance (Greater gravitational potential energy) and extra temperature (Greater thermal energy) -> Greater "power output" (Power = energy * time) -> Significantly more extruded. There's no discrepancy. But, the tradeoff was a much greater variation in diameter and much lower average diameter, which is unacceptable for a 3d printer (Variation in particular).

(Higher temp = greater variation since it takes less force to pull the filament when it's runnier at higher temps. Less runny = Lower variation.)

Anyway, thank you elmoret for a great product! Obviously I just need to figure out how to use it by trial and error.

EDIT:

Now, the filament is no longer touching the wall (Thanks to the guide I created), and it's coiling in a nice 2 ft. wide circle on the floor. I measured and at most points it's around 1.65mm but at a few other points in certain segments of the circle, it's more like 1.52mm. Now how am I going to get this variation down?

My setup looks like (Nozzle is 3.5 feet off the ground)

|||
|\
| \
|  \
|   \
|    \
|     \
|-----0                    (This is the metal loop I created to guide the filament)
|       |
|       |
|        \
|  ____\                 (Nice and organized coil)

EDIT: Nevermind, I just took a set of proper measurements along the entire coil (Stopped it, made sure the calipers were perfectly perpendicular and close to the tip), and it varied between 1.50-1.58mm. I guess the variation isn't too bad, but average diameter is a bit low.

Oh dear, I might just have to get the filastruder closer to the ground.

I guess I could just use 1.54 mm in the software. Seems like I'm getting close to something usable.

(I also noticed that since bringing it down to 185c from 195c, the current regulator is always on whereas it never was before turning it upto 195c. I'm suspecting there's some previously melted plastic high in the barrel that's now semi-solid at 185c, increasing the auger's resistance and engaging the current limit.)

Edit: Seeing the temperature chart at http://jcflowers1.iweb.bsu.edu/rlo/Filastruder.htm , I think I'm going to try 175C.

At 180c and 3.5 feet off the ground during the purge, the diameter measures anywhere between 1.65-1.69 mm.

At 195C and 5 feet off the ground with the "shimmering pearl" colorant added, I'm seeing anywhere between 1.3-1.45mm at the low end (80% of the extrusion cycle) and ~1.65mm at the highend (The other 20%) when the coil is resisting the extrusion.

Then I would say the problem here is the colorant, not the Filastruder or the position/orientation of the filament. Change one thing at a time, changing temperature and position and adding colorant all at once is too many variables. Since you got good tolerance during the purge, this "coil resisting the extrusion" theory doesn't make sense.

18 inches per minutes would be 100 grams (0.22 pounds) per hour. This doesn't mesh with your previous statement of 0.3 pounds in 9 hours.

The barrel/auger/motor alignment has nothing to do with extrusion speed, the only thing that will impact is barrel wear.

At 195C and 5 feet off the ground with the "shimmering pearl" colorant added, I'm seeing anywhere between 1.3-1.45mm at the low end (80% of the extrusion cycle) and ~1.65mm at the highend (The other 20%) when the coil is resisting the extrusion.

I'm going to try guide it through a tube about 1 foot after it passes the fan, I'll see how that affects things. I don't think it'll relieve the weight off the filament much since it's a slick metal tube, but I'll see.

Btw, I am seeing somewhere around 18 inches/minute (At ~1.4mm) so the current feed rate seems upto par. It might've been (probably was) misaligned during the purge, but I added alignment marks so I'm sure it's feeding well now.

My vertical feeding tube is this: http://www.thingiverse.com/thing:219463

I'm suspecting my alignment might be off by 5-10 degrees (Between the motor->auger->barrel), but I would assume the socket provides decoupling so that's not really an issue? Plus, the temp right next to the vertical feed is 80F (barely above room temperature), so I'd assume there's negligible barrel-rubbing losses due to possible misalignment.

Edit: The tube didn't work (Blocked it), but I did get to see the filament at a constant 1.68mm after it was blocked. The shimmering pearl looks like a pretty white color, it has that 'gleam' that a white pearl has.

Edit: Made some new adjustments. Brought the temperature down to 185C, the diameter went upto 1.55mm. I then removed the icechest it was standing on, which removed about 1.5 feet from it, so it's now 3.5 feet above the ground, and now it's consistently measuring between 1.62-1.67. Yep, turns out it was too high and too hot. Meh, I guess that's the way it works.

During the purge, what were your diameter results? In other words, with no colorant added.

"Guiding through piping" is pretty unlikely to work, the hot filament at the nozzle has practically zero stiffness so it can't push through anything. Again, the filament needs to be completely undisturbed on its way to the floor. If the filament touches the wall at any point, that is going to throw off the filament diameter.

1/3rd of a pound in 9 hours sounds very, very slow. Most folks report speeds around 5 times that. I'd guess either the melt filter nozzle is full of debris or (more likely) your vertical hopper isn't feeding pellets well.

Posting a few photos of your setup would be helpful.

I guess I'm probably too far off the floor. And, yeah, it is right next to the wall though it seems like the coil isn't pushing against the wall during its high mark. During its ordinary "the coil isn't pushing against the flow" cycle, it's registering in the 1.3-1.45mm range which is obviously too low so I need to provide some kind of counter force, or some kind of relief at 1-2 feet below the extruder's nozzle. I'm going to try guiding it through some piping or some such.

And, I did purge for 9 hours which only used 1/3 of the bag of included pellets. Was it supposed to use all of it in those 8-9 hours?

Did you extrude the included pellets with nothing else added as the purge step noted in the instructions? What were your results?

Is the Filastruder far enough off the wall that the filament never touches the wall?

What are you talking about when you say "rollers"? No Filastruder has rollers.

When the filament coil is pushing against the extrusion, I read 1.65mm, but when the filament coil is pulling the extrusion, it reads anything from 1.35-1.45mm diameter. It's clear to me that a consistent diameter requires a stable force exerted on the extrusion end, and I can see why the filawinder is designed the way it is.

Man, I don't want to give up my vertical bottle holder in changing over to a horizontal extruder with the rollers, it's so nice! lol

Well, I guess I just have to come up with a comparable 'push back' to when the coil is resisting the extrusion, I guess maybe some tubing might help here.

Higher temperature -> greater flow
Higher height -> greater downward force -> greater flow

Sure, the filament might be thinner, but I'm guessing I'm still extruding more grams/hour overall and that's all that matters (As well as consistency, which this thing seems to be pretty consistent.). Of course, I could get greater extrusion rates with a thicker extrusion diameter, but I'm guessing the possible % improvement is minimal. Drilling to 1.65 -> 3.1%, 1.70 - > 6.2% with possibility of clogging the nozzle, so realistically I could improve by 3.1%. Plus, not drilling is easier than drilling.

(Well, actually, extrusion rate is proportional to the area which is square to the diameter, so (1.031)^2 = 6.3% improvement with 3.1% increase in nozzle size (From 1.6 to 1.65). Still negligible.)

Well for starters, there's no real need to increase diameter. You can just set the actual diameter in your slicing software, which will take care of the rest.

That's music to my ears. Thank goodness S3D makes changing profiles a breeze.

The nozzles provided are slightly undersized as different materials exhibit differing amounts of die swell, and if the machine produces filament that is too large it will jam the hotend - you can't get around that with settings in the slicer. The nozzles are drilled to 1.60mm, you can drill them to any size you like. A hand drill is fine since the hole has already been made.