New i3V owner: A few beginner questions

[printbus]

I've only got about a month more experience than you, so bear that in mind. Note that while the overall print speed may be 100%, Slic3r applies about 10 different speeds factored off that, depending on what is being printed (perimeter, infill, etc.). If you haven't seen them, these Slic3r settings are visible in the Speed section of the Print Settings screen.

One thing that was enlightening for me was to feed the gcode through a viewer like gcode.ws. You probably can with other viewers too, but with gcode.ws, you can click on layer info and then walk through the print layer by layer, observing how different speeds will be applied to different parts of the print. I would compare the print speed being applied to the top solid layers with what is being used elsewhere, since that seems to be printing OK for you.

Reducing 3mm filament to the 0.40 mm tip is a pretty good reduction (3mm has almost 3x the surface area as 1.75mm filament) and it could be you're just not melting enough fast enough. You might need a higher temp (210 could be low anyway) and/or a slower "top solid infill" speed in Slic3r. Or maybe even a higher drive adjustment for the extruder stepper motor driver to maintain the pressure at the nozzle during high flow volumes.

EDIT: Unless it has changed since I archived the config files, the MakerFarm config file for 3mm PLA on the 0.4mm hex hot end will have Slic3r print the top solid layers at 100 mm/sec. It'll print perimeters at 75 mm/sec, and the external perimeter wall at 70% of the other perimeters, or 52.5mm/sec. So, the external perimeter walls you can see are printed at about half the speed of the top solid layers. That could be part of the reason why the rest of the print looks OK, at least on the outside.

EDIT #2: Well, there is a another filament cooling setting in Slic3r that will slow the print down if the time spent on a layer is less than a default of 30 seconds. That should be the case for at least the knob, so the above numbers may not apply. Just another reason why I like to look in a gcode viewer before I print to see what the printer will actually be doing. Regardless, the config file will still have the top solid infill printed faster than the perimeter shells.

[printbus]

gmay3, how did your knob in http://3dprintboard.com/showthread.p...umbwheel-Model print? Looking at mine, my top side has the same mesh look that DanSilov is battling here. You can even see it in the picture I posted.

EDIT: Looking at the gcode, that was the last item I printed with a bridge speed of 150mm/Sec as set in the Makerfarm Slic3r config files. That seems awfully fast to my inexperienced judgement. For my knob, slic3r has a bridge layer under the top layers, and perhaps the bridge didn't work very well and the top layers aren't properly supported. My recent prints are coming out solid top (now I'm using a bridge speed of 70mm/sec), but I've also been dinking around with ABS since that knob.

[Roxy]

EDIT: Looking at the gcode, that was the last item I printed with a bridge speed of 150mm/Sec. That seems awfully fast

Agreed. I know my printer could not run at that speed doing bridges.

[DanSilov]

Well, I had an interesting evening last night. First I stumbled upon this post (http://3dprintboard.com/showthread.p...inter-Extruder), where the video describes exactly the symptoms I was having with my prints. So I immediately decided to calibrate my extruder using that video and gmay3's post a reference.

I quickly found out that my extruder was a bit uderextruding (94mm instead of 100mm requested), and I used LCD menu to adjust E-steps. I decided to use the same G-code I printed before – a 24mm 20% infill cube. The result, well, dreadful.

Original settings (mesh at the top, not watertight, but very neat print otherwise):
IMG_8008.jpg

After calibration (a mess all around):
IMG_8011.jpg

I then reuploaded firmware using link provided in MakerFarm guide (BTW, should I try vanilla Marlin firmware, will it make any difference?), enabled EEPROM access just in case. Then I decided to try printing 20mm 100% infill cube, as per gmay3's post. I used adjusted E-steps, which I saved to EEPROP and I also set Extruder Multiplier in Slic3r to 1.1 just to see what will be the result.

Here's the result:
IMG_8013.jpgIMG_8015.jpg

I've tried two times, and both times the print started bending around corners, and then it lost adhesion (the first time this happened for me). At this point I gave up.

Looks like something is definitely off with extruder, since even Extruder Multiplier doesn't force it to extrude more material. I'm not really sure what to try now.

[gmay3]

printbus, I don't think I had the same problem with my knob but will take a look tonight and post a picture.

DanSilov, sorry you've been having a frustrating time with this. Sometimes this is how it goes with 3d printing . I would double check your calculation of the Esteps. It seems like your 24mm cube with 20% infill has less plastic (hence the caving in corner) after your changed the Esteps. Your initial commanded 94mm instead of 100mm for the Esteps cal should have had more plastic come through.

Ok so as far as the 20mm 100% infill box goes, like you said we need to get your print to stick before anything else because if the corners are bent up the print will not be accurate enough to show if your 20mm 100% infill is flat on top.

I'm assuming you are using the good ol' glass bed with hairspray method. I have had really good success with this so far as far as sticking to the bed goes. I found that in the beginning I was putting too little hairspray so try just a little more to see if this helps. (I'd recommend taking your glass off and spraying it on top of an old tee shirt, then let it dry for 5 mins or so)

I would recommend trying two things next.

1. Twist your Y bed gently clockwise and counterclockwise to see if the delrin idlers on the Y bed have loosened up. This happens for me every 20 prints or so and can mess up your bed leveling. Then re-level your bed so that the nozzle hits the z endstop at a height of about 70 to 80% of your desired layer thickness away from the glass at each corner of the bed. So for 0.2mm about .16mm.

2. I think I read that your were printing PLA at 220 or and 70 for the bed. I'm not a PLA master as I usually print in ABS but when I've printed PLA at 185 and and 60 for the bed temp I've had success.

Any way, I hope some of this helps but keep us posted with how its coming along!

[DanSilov]

gmay3, thank you for all the advice. I understood the 3D printing is not easy from day one, in fact I welcome the challenge

The problem with sticking to the glass is that it's something entirely new. I've printed quite a few things with PLA (bad quality, but still), and never ever I had any problems with adhesion even without using hairspray. In this case it looks like the print is shrinking into the center when it cools down. Maybe it's because it is 100% infill. Will be tinkering around some more today.

[printbus]

Bed adhesion suggestions beyond gmay3's points -

You might post a bottom view of one of the prints so we can see how your first layer is coming out. Have they improved from your initial post?

I learned to let the heat bed soak at temperature for a while - maybe 5 minutes before enabling a print. That likely helps the heat soak through the glass. My adhesion failures with PLA have frequently been when I've skipped this soak time.

Did you reverse the hot end fan like you thought you might? I face the fan airflow forward too, but I found the MakerFarm fan does create a draft over the forward side of the heat bed that was creating problems on least large prints. The forward side of the print would tend to lift and curl up. I initially just added a piece of painters tape along the bottom side of the fan to help deflect the airflow off the heat bed. I've also ditched the strong and overly loud fan Makerfarm provided for a slower, quieter one. It puts through less air, but it's till cooling the hot end while not creating as much draft over the heat bed. Finally, I also now let the heat bed soak with the Y-bed pushed all the way back. That allows any draft from the hot end fan to clear the heat bed.

[printbus]

For grins, and since I've seen the open mesh top on a print as well, I ran some comparison prints of that knob. These may not provide any firm answers for you, but there might be some info of value. I assume this is the knob you started with - http://www.thingiverse.com/thing:221920.

I ran three prints of the knob. The first on the left used Slic3r 0.9.9 and the unmodified MakerFarm config file for 0.4mm hexagon hot end, 1.75mm PLA. White doesn't photo well, but it's the only PLA I currently have. Although not as extensive as on your print, the top layers definitely show the mesh and are not "closed". Watching the print layer by layer, substantial droop was observed when printing bridges at 150mm/Sec as specified by the config file. One such layer is where the knob transitions from being open on the bottom to closed, and the other is just under the logo on the top. For the lower layer where bridge was used, it took several additional layers before the newly applied honeycomb infill seemed to not suffer from it. For the upper layer where bridge was applied, the drooping was substantial enough that the next solid layer also sagged, and the top and final layer therefore didn't have a substantial subsurface to extrude against - hence the stringy and open appearance. More extensive infill or a print cooling fan would probably help here.

The second in the middle used Slic3r 1.1.6 and a custom config file for 0.4mm hexagon hot end, 1.75mm ABS. Version 1.1.6 adds some config parameters beyond v0.9.9. I didn't know what would happen if I used the 0.9.9 config file with it, so when I started using it I recreated new config files for v1.1.6, initially leveraging settings MakerFarm had in the 0.9.9 files. Over time, I've been tweaking many of the print speeds. Right now most more closely follow what is suggested starting here - http://richrap.blogspot.com/2012/01/...tings-and.html than the MakerFarm values. Most significantly, the bridge print speed is considerably slower than 150mm/Sec. Beyond those changes, I now also typically use rectilinear infill at 25% (just works better for me than 20% honeycomb), first layer extrusion width of 250% (my ABS seems to do better with this), and top solid infill extrusion width of 125%. A crosshatch pattern is noticeable on the top layers, but the top is solid.

The third knob on the right also used Slic3r 1.1.6 and a config file identical to that for the ABS knob except that the bed and nozzle temps were set for PLA (70/225 per the MakerFarm suggested settings). The same crosshatch pattern shows up as on the ABS knob, but it is solid.

The underside of the knobs is interesting. I had previously noticed that v1.1.6 either extrudes a bit more and/or does better at going back and filling small missed areas. For the v0.9.9 print, you can see numerous gaps that carry up into the knob. In the v1.1.6 knobs, the printer was diligent at going back and touching up numerous spots on each layer. I can't speak to whether the difference is in v1.1.6 itself or in the config file differences. Likely because of these touchups, v1.1.6 takes longer to generate a gcode file, and prints take longer. I'm not implying that v1.1.6 is better - I think I have work to do in tweaking retraction, acceleration, and who knows what else related to all these touch up spots. Right now I tend to end up with a slightly excessive amount of plastic in areas where the touchups occur. I wanted to experiment with a 1.x.x version since support capability was revamped in 1.0.0, and 1.1.6 was the latest available.

All knobs printed with calibrated esteps/mm based on feeding 100mm filament length, but not fine tuned for single wall print thickness or variation in filament diameter.

[DanSilov]

printbus, thank you for such extensive comparison! I really hope to achieve something similar to the last one you printed, which looks quite solid. Also, thank you for the link to settings blog-post, will definitely help. Maybe you can also share your 1.1.6 config, the one you used for printing last knob?

Yesterday I've printed 4 different 20mm blocks with 100% infill with all kinds of settings, and results are still really bad. I didn't have any problems with adhesion, but that might be because I've flipped the fan again (yesterday I tried to print with the fan blowing air out of hot end). I've increased flow rate, reduced speed settings (pretty randomly, but to much lower values), all to no avail. Below are some results. Bottom of every block looks like the one one picture.

IMG_8020.jpgIMG_8021.jpg

I've even sliced and printed one block using Cura, which is a nice enough tool, but results were still bad (one side of block was 9mm, the other – 10mm).

It looks though I might have some problem with the extruder. I've gone through calibration procedure about 5 times yesterday, and I couldn't get the extruder to reliably extrude 100mm. Before it was underextruding, then after I adjusted E-steps it was overextruding by a lot. Then again underextruding and so on. When I assembled the extruder, it had some issues with guidler fitting correctly, it looked a bit overprinted, so I had to use a knife to take off some plastic. Also, my hobbed bolt doesn't that many teeth as the one on this picture. I managed to print all the parts for the extruder earlier. Although they have the same open mesh on top, they look good enough. I will try to reassemble the extruder with my parts and then run the calibration process to see if I get reliable results this time.

Also, did anyone measure and adjust stepper voltage? I checked my extruder stepper voltage after printer assembly, and it was 0.36v. It doesn't skip and turns quite nicely, so not sure if this might be a problem.

[printbus]

My extruder stepper driver is set to 0.35v.

The MakerFarm hobbed bolts don't have as many teeth as the one you linked to. I don't have a picture, but looking down into the hobbed bolt area I can see about five teeth. Regardless, as long as the teeth are consistent around the bolt I don't think the number of teeth shouldn't matter much.

Just trying to help with ideas, I would scrutinize everything from the filament spool to the print surface. These come to mind -

• Make sure the filament spool isn't catching on whatever is holding it
• Part of the guidler bearing is exposed on the left side. Mark it with a Sharpie or something and verify it seems to always rotate when the hobbed bolt does.
• Retract a section of filament and observe how the hobbed bolt teeth are digging into the filament. I don't have a picture to offer, but mine leaves a pretty rough serrated feel on the filament. Spacing between marks should be consistent. If the marks aren't very substantial, try increasing the tension screws more.
• ADDED: Looking down into the hobbed bolt area, do the channels between teeth on the bolt look clean? Does the cut area of the hobbed bolt appear to line up with the filament guide above it? (I futzed with finding the right spacers under the head of the hobbed bolt to improve the alignment; maybe on 3mm filament that's even more important)
• Check all the extruder hardware for tightness. I've had the motor start to get loose and the two extruder mounting bolts get loose. As mentioned earlier, make sure that (with the hot end cold), you get no mechanical play when you manually try to manipulate the hot end tip around.
• I used longer bolts on the motor so I could put large flat washers on the backside of the mounting plate. I have to be careful that the large gear is clearing the motor bolt heads on one side and also the delrin wheel bolts on the other. Rubbing on either side would lead to erratic extrusion.
• If you haven't already, consider tearing down the hot end to clean it.
• Do you have any other filament to try? Is the diameter of the filament you're using consistent? Unroll a length of it and run your fingers over it. Feel any bumps on it? I had a roll with small bumps that I just happened to notice one day by doing that. Some of the bumps were large enough to jam up at the top of the hex hot end, but others could have been sized right to struggle in passing through the body of the hot end.