Thin roofs on right sides of prints

[Samantha]

I've had a recurring issue lately with thin roofs specifically occurring on the right sides of my prints. It seems that one of the more recent Makerbot Desktop software updates have introduced this, as it was not something we saw for the first few months of operation of our machine (one of the first 5th gen Makerbots, yes yes, I know, we are outside the return zone). I've fiddled with  "roofThickness" and "roofLayerCount" but neither seem to make any difference. I thought possibly it was underextruding due to gunk build-up so I ran some of the eSUN cleaning filament through it, but also no luck.

Any ideas on what might be causing this? Even if it's an inherent flaw related to the design of the machine I'm just generally curious to know why it does this.

[Jetguy]

Temperature of extrusion is simply insanely hot - evidenced by the glossy finish on the printed parts.

The 5th gen has a very long melt zone inside the extruder and puts a ton of heat -beyond what is required to melt it, into the plastic.

As such, even with all the cooling on the plane from the ducts that blow at the print around the nozzle, the plastic stays goey/runny rather than returning solid so that the next layer doesn't have anythig to "build on top of".

 

Again, there are a couple of things going on here:

#1 is temp is too high

#2 not enough infill percentage for the hot roof layer to not droop down and thus build up a flat layer. Hot plastic cannot bridge.

#3 speed or edge overlap setting of roofs may not be the same in the software.

#4 order of infill VS perimiters may have changed

 

KNOWING the 5th gen nozzle has massive repeatability issues with returning to the exact same position after a retract may be contributing to the infill edge not lining up to perimiter to allow the plastic to bridge for a roof layer.

That coupled with the weak and simply incorrect extruder mounting leads the entire extruder to wobble like a pendulum below the gantry so speed can be just as much of a factor in nozzle positioning errors.

 

 

I'm sorry, but the 5th gen is junk. You can attempt to mask these massive mechanical flaws but probably are not going to be hugely successful.

Lowering temps should be the right answer but may lead to jams in an unmodified smart extruder. This is because the shoddy thermal design conducts sheat into the upper part of the filament path. What they found was that if they ran it hotter than normal and let it melt in places the filament should be solid it reduced jams, but then overheats the filament. Once again, they made the tradoff of works at all, VS works properly.

 

Then, the one thing you can improve is changing the spring inside the smart extruder to a stronger one helps ensure the nozle returns to the same down position each time after a retract. You cannot go crazy here because a stronger spring means the bed may deflect more trying to push up the nozzle during homing and bed leveling sequences. I've only tested on a mini that doesn't ever do bed leveling, only homing and that's at the back of the Z stage capable of huge force with minimal deflection.  Also, too strong of a spring could let the smart extruder be pushed off the magnets holding it in some cases.

This is a tradeoff/hack but worked well on my mini

[Samantha]

I'm still somewhat perplexed by this, so I've continued to test this. I've run the parts from 200C-250C, halved the speed, upped the layer time to allow the layers to cool more, increased the edge overlap settings, and my parts still turn out identical every time. I've got about a dozen of these identical blue parts and I see no effect in changing any setting in the output. It's to the point where I feel that perhaps these custom profiles are not even being used by the software, and it's defaulting to their own profiles instead. 

I'm not sure its the wiggle inherent to the nozzle as these gaps (not shown very well due to poor photo quality) are nearly identical and very consistent from print to print, and only occurring on one side of the print. I know the printer itself is crap, but this specific issue was not one we saw in the first few months of operation, leading me to think that it's some sort of software change that's occurred along the way.

Perhaps the next thing I'll try is the drop test....I'm sure at least it'll be therapeutic. 

[Dan Newman]

On 14/11/2014, 9:51 AM, Samantha wrote:
> I'm still somewhat perplexed by this, so I've continued to test this. I've
> run the parts from 200C-250C, halved the speed, upped the layer time to
> allow the layers to cool more, increased the edge overlap settings, and my
> parts still turn out identical every time. I've got about a dozen of these
> identical blue parts and I see no effect in changing any setting in the
> output. It's to the point where I feel that perhaps these custom profiles
> are not even being used by the software, and it's defaulting to their own
> profiles instead.

Wild speculation based upon past problems in this area with makerbot software:
how are you sending the prints to the Gen 5? Over USB, it saves them
locally I believe. To USB thumb drive you have a little more control. I
ask since in the past Makerbot has had bugs whereby you write a file
directly to SD card (their older printers) and their software would
not correctly write the new file and instead would append it to the
old one. Net, net, people would make profile changes, reprint, and
see no difference since the file always started with the same old
print data. (And then midway through had an "end print" command so
the new, appended print never happens.) I wouldn't put it past them
to have the same sort of bug all over again. So, if you're printing
from USB drive, have Desktop write the print file to your local hard
drive and then use your computer's drag and drop to move the new
file to the USB drive -- your computer's OS will hopefully function
correctly. If you're printing over USB, you can try changing the
model's name so that a different file name gets used on the bot's
local storage.

OTOH, it could simply be a different bug in Desktop and it really
is ignoring your profile changes.

Just a wild speculation from someone who doesn't own a Gen 5.

Dan

[Enginwiz]

On all top surfaces you can see through several layers of too thin infill strands.

This could be caused by underextrusion.

Did you already print a 20 x 20 x 20 mm solid calibration cube with 100% infill

to calibrate your extruder output factor on your 5th gen printer? Underextrusion

first shows up on the infill part of a print. Then solid cubes are not filled up on the

inside and in extreme cases you can see through a mesh of layers. In this case

you have to increase the extruder output factor in your print profile to push more plastic

through the nozzle during printing.