Problems with PRE-echo on models

[jsc]

I don't think your theory about acceleration control is correct. Stepper motors work by counting an absolute number of steps, regardless at the speed at which they are delivered. If the extruder were out of position and turning the corner early, every layer would be shifted. However, it might be the case that there is a mechanical lag in the system such that the extruder position is not quite to the corner when the other stepper kicks in to begin motion in the other axis, which might lead to asymmetric effects between X and Y axis because the mechanical properties of the axes are different.

Anyway, here's my print of your test object. There is some vibration/overshoot induced rippling but it is in line with what I have seen with vertically oriented holes on other models.

test.jpg (68.1 KiB) Viewed 10062 times

Here is an object to test for ghosting, with an example photo of its effects: http://www.thingiverse.com/thing:277394
And a test object I made a long time ago that shows similar artifacts: http://www.thingiverse.com/thing:247162
A particularly bad example on a MakerBot: https://robosavvy.com/forum/viewtopic.php?t=8149
And a general discussion of the problem: http://forums.reprap.org/read.php?1,214990

Try printing at 2400 mm/min (40 mm/s) and lowering accelerations by adding
M201 X800 Y800 ; lower default accelerations

to your startup gcode in the Scripts menu of Simplify3D.

[Tim]

If the extruder were out of position and turning the corner early, every layer would be shifted

I don't think that was what Denys meant. He didn't suggest that the extruder was out of position. What he's referring to, more or less, can be found in the firmware; see "planner.cpp", where there is the comment section "Basic definitions" at the top, and comments further down. It's apparent that it does do acceleration and braking calculations. It also appears that the acceleration and decelaration are "trapezoidal", meaning the speed ramps up, hits a plateau, then ramps down again. The points of the trapezoid would be discontinuities in the acceleration and could cause ripple both after and before a change in direction. So I think that Denys may have a valid point. Some of the ripple is certainly due to vibration, but some, especially the "pre-echo", may be built into the firmware (and would have a corresponding regular echo that may be making the ripple worse than it should be).

And I could be totally wrong. But it's a possibility.

(To edit what I said---the speed profile is trapezoidal, so there are discontinuities in accelaration itself, not the derivative.)

[insta]

I know one of the authors of TinyG, and he talks at length about this. Apparently in the newer versions of TinyG they are going for linear pop on their acceleration (5th order!), in an attempt to smooth out ringing and such.

[Dale Reed]

Hmm. I would think constant jerk (3rd derivative w.r.t. time) would be sufficient. (Think express elevators...) Our variable-speed drive systems guys at work deal with this all the time. I'll ask what they think... more later.

Dale

[insta]

They originally used constant jerk, but the math was actually harder. He said with linear pop the math got simpler since everything just became addition and subtraction. Smoother prints was just a happy side effect.

[DenysNazarenko]

Hi Guys,

I find out the reason of the problem ))

To do this we try -
- replace RAMBO with Arduino mega
- replace electric-motors
- rewrite firmware

Yesterday during programing of firmware I got clear answer in my mind. Andddddddddddddddddddddddddd
The problems is........... Nozzles. Yes. It's so simple.
The makergear Nozzles has the plateau around the hole for plastic. With this plateau extruder shift the plastic on corners.
Other company provide nozzles more acute with much small plateau.

I'm very happy that after 3 month of research we find out the reason and my father will tweak the problem with lathe and we fix this from our side.

But I'm very angry on MakerGear since they should find and fix this problem from own side.

And as I promised the print with cheap China printer

11.jpg (167.62 KiB) Viewed 9968 times

[Tim]

I'm doubtful that it is "so simple".

The flat area around the tip of the nozzle is there for a reason; it flattens out the filament. MakerGear designed the flat part to be a bit wider than the nozzle opening itself. I have noticed that it is wider than the flat area on my E3D nozzle. However, when I was using the E3D nozzle, I found no change in the echo ripples on the print. After using the E3D for a while, I concluded that I liked the MakerGear nozzle design better.

[rpollack]

Denis gave us a 1 star review on Amazon over this and he used your photo Jim!

[rpollack]

Denis - please post a photo your best print of this object on your M2. The photos you posted at the beginning of this thread show very poor print quality. I would like to see that you are able to at least get the same print quality as the other posts here. Did you mention anywhere that this was a kit and that you are in Ukraine (which adds to the support challenge).

Denis is now also demanding a full refund (minus return shipping) for a kit that was shipped in June.

[DenysNazarenko]

I'm doubtful that it is "so simple".

From my side it's very simple at last if this was a "feature" of the printer you can tell me 3 month ago I'm just change the extruder and don't waste the time on story listening about "Machine vibration".

Denis gave us a 1 star review on Amazon over this and he used your photo Jim!

Jim make wonderful photo but if you want I can write one more review but with my own photo of other models. I have the same problems on the same models.

Sorry guys but you make me almost crazy with this and i'm lost few customers with my low quality of the job. And waste ton of time and brain since I can't made half/work with poor results.