Welcome to the 5th gen disaster. Dan says everything a lot nicer than I am about it, but again, this just the straight truth.
First off, the whole smart extruder thing isn't. It should detect underextrusion conditions but never does anything about it. Besides that, it should never underextrude- hence poor top layer fills.
To really grasp the problem, understand that NO you can't edit a profile setting and "fix" the problem. The extruder jams internally and doesn't extrude the volume of plastic being commanded, but this jamming is impossible to predict. So telling the extruder to extrude more does nothing but cause more problems in places that the extrusion was fine. In a Replicator 2 or 2X, we'd call this problem extruder clicking because you could hear the extruder motor skipping steps. Because the 5th gen is so noisy, you can't hear the extruder over all the other noise. There is just no way software can predict when this thing fails to "do what it's told". I think they aren't using the sensors because they realized it's failing hundreds of times in every print. You'd never get a print finished because it would keep telling itself the extruder was failing.
Problem #2 is that because the extruder hangs WAAAAY down from the mechanics, and the smart extruder is only attached with magnets the system must have acceleration turned waaaaay down in the firmware (A setting you cannot touch anyway).
So, setting feedrate faster does nothing since the acceleration is so low, you never get to the feedrate you set. Again, this is a mechanical disaster. They compensated by slowing the machine down. You can attempt to fool yourself in telling it to print faster but it won't actually do it. Even if you did go faster the huge and heavy extruder hanging down bobs like a pendulum in an old grandfather clock. You get ringing and all kinds of other problems.
Then back to the other disaster of the smart extruder, Z axis homing. The idea was OK, the nozzle system can rise and fall in the extruder and to home, they sense when the nozzle hits the build plate and rises inside the extruder housing. That might be OK but nothing ensures the nozzle returns to the exact same position in X Y Z coordinates. The heater block is supposed to guide this with the cone shaped bottom and the stamped stainless cage that surrounds the hotend. But, because of cheap manufacturing, the entire stamped cage can move around slightly an the hot aluminum heater block against the stamped stainless cage kind of binds up too during operation. Now, to the lay person, they look at it an go "so what". Well, if the nozzle isn't rock solid to the entire extruder housing, and the enter extruder housing can shake and wiggle hanging down from the overhead gantry, your accuracy of nozzle positioning is just crap. I refuse to sugar coat this, it's just all insanely bad.
Prove this to yourself with the nozzle cold obviously. Simply push the nozzle side to side by hand and you can see in operation, it's going to slop around like a wet noodle. It only has to move a fraction of a mm for visible defects in the print to show up in vertical walls where each layer should be stacked on top of each other. Every time the filament retracts during a print, the nozzle rises up (which first off, negates some of the retraction distance inside the hotend and likely contributes to the stringing) and secondly, because the nozzle is not securely guides by some type of linear bearing system, there is zero chance of it returning to the same position each time extrusion begins. This is painfully obvious in prints that I have done.
You can even see it in your prints in the vertical walls.
So, given the mechanics are a HUGE limitation, and even if they did "fix" a new smart extruder, do you really ever think it's going to be "right" in comparison to known machines on the market? I certainly don't have that faith and I'm stuck with a MakerBot mini.
On Wednesday, October 15, 2014 9:04:39 PM UTC-4, 911Ducktail wrote:
I bought a 5th gen replicator on monday (in retrospect from reading here it wasnt a good idea I guess,) and ran through its calibration/bed leveling and test screw print