Some recent 0.35 nozzle prints

250 views
Skip to first unread message

orcinus

unread,
May 23, 2013, 3:24:45 PM5/23/13
to arco...@googlegroups.com
I've sent these to Laszlo in an e-mail discussion and he kindly asked me to post them here, so here goes...

Some things i've printed in the last two days, testing out a new brand of filament (colorFabb).
They're printed on an ORD Bot Hadron with an Arcol.hu 4.x (a mix and match of a few Arcol.hu 4 hotend versions) with a 0.35 mm nozzle.
Sliced with KISSlicer, printed running Repetier FW. Layer height 0.16ish.

https://dl.dropboxusercontent.com/u/1702513/_5230327.jpg
https://dl.dropboxusercontent.com/u/1702513/_5230329.jpg

I've noticed i've been getting a lot of "ringing" near sharp edges recently.
I guess it's finally time to switch from MXL to GT2 belts.

https://dl.dropboxusercontent.com/u/1702513/_5230336.jpg
https://dl.dropboxusercontent.com/u/1702513/_5230337.jpg
https://dl.dropboxusercontent.com/u/1702513/_5230339.jpg
https://dl.dropboxusercontent.com/u/1702513/_5230341.jpg

You can notice a gap in the layers in the lions right side. That's filament's fault. It did that about 3-4 times now, and i've narrowed it down to either bubbles in the filament or trapped moisture. I've actually stopped a print once in the middle of a botched perimeter and dug out the remains of the filament from the nozzle. There was a bubble in there, but, like i've said, i wasn't sure if it's a manufacturing defect or a steam bubble.

Anyways, been using Arcol.hu 4 for a little over a year now and don't plan on stopping any time soon :)

nop head

unread,
May 23, 2013, 3:35:09 PM5/23/13
to arco...@googlegroups.com
Nice prints. Why do you think GT2 belts will have less ringing than MXL?


--
If you want to send a message to this group, please send the email to arco...@googlegroups.com
---
You received this message because you are subscribed to the Google Groups "arcol_hu" group.
To unsubscribe from this group and stop receiving emails from it, send an email to arcol_hu+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

Ante Vukorepa

unread,
May 23, 2013, 3:38:35 PM5/23/13
to arco...@googlegroups.com
Less backlash, more "braking" action to their profile. And they were made for positioning, unlike MXL.
Plus, anecdotal evidence - i've heard it confirmed twice now.

-- 
Ante Vukorepa
Sent with Sparrow

Kyle Kenney

unread,
May 23, 2013, 4:01:25 PM5/23/13
to arco...@googlegroups.com
I can't beleive the quality, jesus, that titan is amazing....the lion less so, nice work! this just goes to show why its great to have an arcol :)

2013/5/23 Ante Vukorepa <o.or...@gmail.com>

Nicolas Arias

unread,
May 23, 2013, 4:37:32 PM5/23/13
to arcol_hu
wow, really nice prints, it would be great if you can tell more about
the slicing, temperatures, etc etc etc (i have samples from colorfabb
being delivered tomorrow).

have you fliped the belts, so the smooth part goes contacts the idler?
(not the teeth part)

Laszlo KREKACS

unread,
May 24, 2013, 4:47:42 AM5/24/13
to arco...@googlegroups.com
Hi,

On Thu, May 23, 2013 at 9:24 PM, orcinus <o.or...@gmail.com> wrote:
> I've sent these to Laszlo in an e-mail discussion and he kindly asked me to
> post them here, so here goes...

Thank you very much. I usually do not get any feedbacks of my hotend users,
so it is always nice to see an exception.

> I've noticed i've been getting a lot of "ringing" near sharp edges recently.
> I guess it's finally time to switch from MXL to GT2 belts.

I'm not that sure GT2 is the only cure for cancer:)
Read here more: http://wiki.arcol.hu/blog:arcol-t25

> Anyways, been using Arcol.hu 4 for a little over a year now and don't plan
> on stopping any time soon :)

That is really nice of you:)

I think it is time for a new monthly (june) competition.
I can offer a full pulley+ belt set as a first prize (for a machine).
Thats about 3meter of belt or so.

Best,
Laszlo

Nicolas Arias

unread,
May 24, 2013, 7:37:09 AM5/24/13
to arcol_hu
im in for a second competition!

Kyle Kenney

unread,
May 24, 2013, 7:40:05 AM5/24/13
to arco...@googlegroups.com
Laszlo, did you not get my email about the prototype? Also, I'm down for comp too!

2013/5/24 Nicolas Arias <nico...@gmail.com>

Ante Vukorepa

unread,
May 24, 2013, 10:13:10 PM5/24/13
to arco...@googlegroups.com
The quality of the lion is pretty much the same as the mech, but it's much smaller and photographed at a much shorter distance (as evidenced by the shallow DoF), so all the flaws are easier to see. That's disregarding the forementioned layer gap due to filament bubble, of course.

-- 
Ante Vukorepa
Sent with Sparrow

Ante Vukorepa

unread,
May 24, 2013, 10:22:48 PM5/24/13
to arco...@googlegroups.com
Not much to tell, really.

Sliced using KISSlicer, skin thickness was 0.3-ish, i believe, layer thickness 0.1684, seam hiding at 0.5, loop speed 30 mm/s, solid infill 40 mm/s, sparse infill 45 mm/s.
Support on, at the "rough" setting, sheathe z-roof at 3 mm, temperature 210 deg C for the mech and 190 for the lion, bed at 50 deg C. And that's about it :)

colorFabb's specs claim the filament should be extruded at 210 deg C, but it seems to me it extrudes much better at 190 deg C. At least the pale gold filament does. Shining silver is maybe a tiny bit cleaner at 200-210. They ship the filament in cardboard boxes, without *any* wrapping or desiccants. Granted, cardboard *is* a desiccant, but still… I don't like their packaging one bit. The filament ends up full of cardboard fluff too. Also, they ship via DPD, which is probably one of the worst, slowest and priciest courier services here (Croatia).

Re: flipping the belt - tried that, didn't have much luck with it.

Speaking of, i've installed the GT2 belts today, plus, i've isolated the steppers, bed and extruder with double sided sticky foam tape. Did some other tweaks (rubber tape on the printer feet), all in an effort to reduce vibration and - hence - the ringing. End result - the ringing is still there. Exact same frequency too. It's a bit attenuated, though, and the whole thing is MUCH quieter now. The layers are a bit more even now too, i presume due to reduced belt backlash.

But yeah, ringing still present and accounted for. The only thing left in the equation now are the steppers, i.e. their natural mechanical oscillation around the poles. I've got no other explanation left for it.

-- 
Ante Vukorepa
Sent with Sparrow

Ante Vukorepa

unread,
May 24, 2013, 10:23:38 PM5/24/13
to arco...@googlegroups.com
I'm not sure what the competition rules are, but i'm game :)
Also, those pulleys look pretty nice!

-- 
Ante Vukorepa
Sent with Sparrow

On Friday, 24. May 2013. at 10:47, Laszlo KREKACS wrote:

I'm not that sure GT2 is the only cure for cancer:)

I think it is time for a new monthly (june) competition.
I can offer a full pulley+ belt set as a first prize (for a machine).
Thats about 3meter of belt or so.

Best,
Laszlo

nop head

unread,
May 25, 2013, 3:24:44 AM5/25/13
to arco...@googlegroups.com
Steppers act like a spring because they exert torque proportional the displacement. That with together with the mass of the carriage forms a resonant system. I think the motors are far more "springy" than a tight belt.

Ante Vukorepa

unread,
May 25, 2013, 8:21:22 AM5/25/13
to arco...@googlegroups.com, arco...@googlegroups.com
Yeah, i'm aware of that.
The thing is, i didn't expect its resonant frequency to be quite as low as what i'm seeing (30 mm/s / 0.95±0.5 mm = 33 to 30 Hz), considering it's a rather rigid system.

It also exhibits two interesting properties:

1) it's *highly* repeatable (if i print an object twice and line the prints up, the phase and frequency of the ringing match perfectly)

2) it very noticeably favours one side of symmetrical objects along Y

Property #1 suggests it's something that's not positionally dependent (in the sense that position on the bed, homing variations etc. do not have an influence on the phase). #2 suggests it's triggered or at least exacerbated by some gcode event (most likely loop start or layer change), not necessarily (just) a surface feature.

I'll need to play with it more. I want to check what effect the following will have:

- speed
- low current microstepping
- current (does the spring constant change with stepper current?)
- microstepping (going down to 4x and back to 16x)
- enabling jitter (varying the loop starting points across layers)


Ante Vukorepa
----
Sent from my iPad

Matej Rozman

unread,
May 25, 2013, 9:05:11 AM5/25/13
to arco...@googlegroups.com
Hey all,

first post in this group so hello to everyone!

About ringing, If you dare try marlin FW, the ringing will be completly gone.
Its funny, I have a delta and I also noticed ringing when I changed to repetier and couldnt found the issue, changed back to marlin and everything back to normal.
You dont have to search machanical issue anymore, becaose there isnt any (probably), its a firmware isssue. Im supprized nobody noticed this before? Well at least havent heard of it....

Hope it helps,

all best,


2013/5/25 Ante Vukorepa <o.or...@gmail.com>



--
Matej Rozman

Robicomp d.o.o.
Koroska 37a
3320 Velenje

Office:  +38659022050
Fax:     +38659022051
Mobile: +38640188669
gotovi....@gmail.com
www.carbonautica.com

Ante Vukorepa

unread,
May 25, 2013, 9:14:29 AM5/25/13
to arco...@googlegroups.com
Wow, hm.

Now that you mention it, it *did* get worse around the time i updated to the latest Repetier.
It was there before, but not as noticeable. I've assumed it got worse due to my printer "shaking loose" lately.

Also, that would explain the perfect repeatability even after considerable hardware changes.

I'm gonna try and whip up a Marlin config quick and see what happens...

-- 
Ante Vukorepa
Sent with Sparrow

Ante Vukorepa

unread,
May 25, 2013, 10:29:15 AM5/25/13
to arco...@googlegroups.com
Nope. The ripples are still there.
And once again, they look *exactly* the same.

I'm beginning to think that - somehow - this is a slicer issue…

Either that or stepper/gantry resonance, as mentioned before, although i'm stumped as to how that would produce consistent, repeatable results every print, regardless of firmware, belts, vibration dampening measures etc.

-- 
Ante Vukorepa
Sent with Sparrow

nico...@gmail.com

unread,
May 25, 2013, 11:00:39 AM5/25/13
to arco...@googlegroups.com
Non concentric pulleys?


From: Ante Vukorepa <o.or...@gmail.com>
Date: Sat, 25 May 2013 16:29:15 +0200
Subject: Re: [arcol.hu] Some recent 0.35 nozzle prints

Ante Vukorepa

unread,
May 25, 2013, 11:34:16 AM5/25/13
to arco...@googlegroups.com
Nope. The phase would change with homing and centering shifts.
And the frequency is way too high for that.

I've tried changing the speed. It seems to have had an effect on the frequency of the ripples. It's slightly hard to tell, because the surface finish got worse with higher speeds. I'll need to do more testing later this evening or tomorrow...

-- 
Ante Vukorepa
Sent with Sparrow

nico...@gmail.com

unread,
May 25, 2013, 11:53:01 AM5/25/13
to arco...@googlegroups.com
Stepper drivers?

From: Ante Vukorepa <o.or...@gmail.com>
Date: Sat, 25 May 2013 17:34:16 +0200

Ante Vukorepa

unread,
May 25, 2013, 7:27:44 PM5/25/13
to arco...@googlegroups.com
Okay, latest batch of tests…

- Reducing the microstepping to 1/4 exacerbates the issue (but not as much as you'd expect).

- Enabling low current microstepping (ROSC tied to GND) makes the rippling worse (it gets sharper).

- Increasing the current and adding more foam padding to the idler plate (which, in my case, is a rather thin and bendy aluminium plate, under constant strain from the belt tension) helps a little. The ripple is now perhaps 1/3 reduced, but still present.

That's more or less it, i'm out of (possible fix) ideas :)

-- 
Ante Vukorepa
Sent with Sparrow

nop head

unread,
May 26, 2013, 8:38:41 AM5/26/13
to arco...@googlegroups.com
Perhaps printing something simpler like a cube would make it easier to diagnose. Are the ripples definitely from one axis, or are they flow rate variations in the extruder due to stepping?

I can't understand how low current micro stepping could do anything but make movement smoother. Especially as 1/4 step makes it worse.

Does increasing the current change the frequency? I would expect it to get higher and that might reduce the amplitude.

Ante Vukorepa

unread,
May 26, 2013, 9:15:37 AM5/26/13
to arco...@googlegroups.com, arco...@googlegroups.com
That's the thing, it doesn't show up on cubes, i.e. simple X-to-Y (or Y-to-X) cornering doesn't exhibit visible rippling. I'll have to whip up a test solid purpose-buikt for testing this... I'm thinking a cube with a dimple (small cylinder subtraction) in half of the faces and a vertical gash in the other half.

And yes, the low current microstepping effect doesn't make sense to me either. Especially because, inexplicably, it made printing noisier than it was (i'd expect less lost microsteps = less noise and less harmonics).

Increasing the current does indeed increase the frequency but also makes it more erratic (the ripples aren't "perfect" anymore, but result in phase variations every x layers), making them more noticable in some objects, as they result in a more mottled, orange-peel like surface (so it's a win some - lose some solution). Also, they make the print noisier as well. By ear, the noise always seems to increase around a particular frequency, which i assume is the resonant frequency of the machine or the axis.

Flow variations did occur to me, but the rippling mostly appears in the Y direction. Also, if it were flow variation, it wouldn't mimic surface features. 


Ante Vukorepa
----
Sent from my iPad

nop head

unread,
May 26, 2013, 9:29:47 AM5/26/13
to arco...@googlegroups.com
Which stepper drivers are you using and what voltage. Low current microstepping definitely reduces the harmonics for me. It makes the whine when stationary louder because ripple current is higher due to fast decay mode. That is far too high a frequency to show in the print.

Seems like it might be resonance on each motor step, rather than the impulse of starting and stopping. Perhaps there is something wrong with the microstepping. I assume you have seen my blog article and the video. Have you tried stepping very slowly and listening to the clicks to see if they are regular?

Ante Vukorepa

unread,
May 26, 2013, 10:51:26 AM5/26/13
to arco...@googlegroups.com
A4982, on Roy's (Panucatt's) carriers @24V. They are louder when stationary, as expected, and low current microstepping does reduce (but not eliminate) the "skip" at slow movements (i.e. the "rukka-rukka-rukka-rukka" buzz). That part is pretty much as you've described and as should be expected.

But for some reason, it behaves unexpectedly during fast moves. The harmonics i've mentioned are pretty much fixed in frequency regardless of the speed, they just become louder at certain speeds. My only theory so far is that noise i'm hearing during the print is the natural frequency of the machine and the missed microsteps somehow accidentally manage to "break up" the resonance, thus alleviating it. So when the low current microstepping is *on*, the effect disappears and i'm hearing the full effect of the resonance.

-- 
Ante Vukorepa
Sent with Sparrow

nop head

unread,
May 26, 2013, 11:13:11 AM5/26/13
to arco...@googlegroups.com
Have you tried 12V? It's hard to get the low current steps accurate with more than that. The chip regulates the peak current of a saw tooth ripple but the torque will be proportional to the average. That means there is an offset proportional to half the ripple current. On the low current steps it becomes significant and higher supply makes ripple worse.

When you say the harmonics are independent of speed do you mean audibly or spatially? If they are fixed spatially on the object, independent of speed than it is almost certainly a stepper motor problem. Possibly what is termed mid band resonance. Normally micro stepping mitigates that, but not if the waveform is badly distorted.

Laszlo KREKACS

unread,
May 26, 2013, 3:52:31 PM5/26/13
to arco...@googlegroups.com
Hi,

Without photos it is hard to give any advice. (ie. some side-by-side
comparison photo).
Feel free to start a page under
http://wiki.arcol.hu/blog:printing_reasonance or something more
meaningful title.

If you compile a test-object and you print it at different speed,
I try to print the same object for you for comparison.

But without reproducing the problem at my side, it is really hard to
advice anything.

So I'm waiting to your test object already printed by you at least 5
times with comparison photo.
Then I'm sure many others (myself for sure) will print to give you a comparison.

Best,
Laszlo

Kyle Kenney

unread,
May 26, 2013, 4:21:26 PM5/26/13
to arco...@googlegroups.com
I made up a 15mm cube, with a dimple, and a 1mm 10mm wide verticle strip, one of those on opposide sides, same as with the dimple...it's hollow aswell, so no need for infill issues.

I attached the file, lemmie know if you wanna do something different


2013/5/26 Laszlo KREKACS <li...@arcol.hu>
testsolid.stl

Kyle Kenney

unread,
May 26, 2013, 4:26:21 PM5/26/13
to arco...@googlegroups.com
I just realized after throwing the file through slic3r the dimples are right on the edges of the inside wall, so I updated it, here is the newer version, also I mean 1mm wide, and 10mm long gash *

Hope it helps!

2013/5/26 Kyle Kenney <dudl...@gmail.com>
testsolid.stl

Ante Vukorepa

unread,
May 27, 2013, 7:34:26 AM5/27/13
to arco...@googlegroups.com
Haven't tried with 12V, but i'm sure the results would be better, re: skipped microsteps. I doubt it'd have a noticeable impact on the ripple (print ripple, not ripple current), though.

Re: harmonics - i've meant audibly. Spatially, their frequency changes with speed as you'd expect if they were a fixed frequency in the time domain, which is why i'm pretty convinced it's mechanical resonance. The thing i don't understand is - why isolating the main components with silicone and foam padding doesn't affect the result much (even though it reduces the noise).

In any case, yesterday, while testing under higher jerk and acceleration in an attempt to exacerbate the effect on purpose, i've noticed the X stepper casing vibrate *noticeably* at what appeared to be somewhere within that 30 Hz range (no idea what the actual frequency is, of course, but i know a 50 Hz flicker when i see it and this seemed close). Considering the way the gantry stepper is mounted, it's no wonder it's doing that - it's a textbook example of a mass oscillating on a spring (4 bolts):


(left side of the gantry in the render)

Similar issue with the Y stepper and idlers. I've got a bunch of pre-cut 2x thicker replacement mounting plates that i never took the time to install, i'll see if those change anything. Also, probably will beef up the X stepper mount.

-- 
Ante Vukorepa
Sent with Sparrow

Ante Vukorepa

unread,
May 27, 2013, 7:35:01 AM5/27/13
to arco...@googlegroups.com
Sorry about straying off topic :)

I'll move to wiki after the next message.

-- 
Ante Vukorepa
Sent with Sparrow

Ante Vukorepa

unread,
May 27, 2013, 7:40:02 AM5/27/13
to arco...@googlegroups.com
Thanks for doing that! :)

I've made a slightly different version (attached).
This one is 20x20x20, has an X and Y shaped cut on corresponding faces and two dimples on each of the remaining faces.

The idea being that after you print it, you don't have to mark it in order to remember which face was which.

Also, half the object Z has sharp vertices, while the other half has fillets.
Should be useful for estimating how much impact the corners have on rippling.


-- 
Ante Vukorepa
Sent with Sparrow

Attachments:
- testsolid.stl

Ripple_test.STL

Kyle Kenney

unread,
May 27, 2013, 7:50:10 AM5/27/13
to arco...@googlegroups.com
I'll print that out and post my results :)

2013/5/27 Ante Vukorepa <o.or...@gmail.com>

Laszlo KREKACS

unread,
May 27, 2013, 9:26:19 AM5/27/13
to arco...@googlegroups.com
Hi,

On Mon, May 27, 2013 at 1:40 PM, Ante Vukorepa <o.or...@gmail.com> wrote:
> Thanks for doing that! :)

np
Can you please post your pictures also?

> Sorry about straying off topic :)
> I'll move to wiki after the next message.

It is not off-topic at all. It is printing related, feel free to discuss here.
So discussion here, summary on the wiki for future reference.

Updating the wiki page is the original conversation starter's job.
So that is you.
That is the "price" you pay in return of the others help.
(keeping the wiki up to date about your issue).
Others are also free to help in the wiki page, but it is your
responsibility in the end:)

So chit-chat here, summary on the wiki.

Best,
Laszlo

Ante Vukorepa

unread,
May 27, 2013, 10:24:26 AM5/27/13
to arco...@googlegroups.com
Yes, yes, i'm working on it :)

In the meantime, here are the first photos:


Now, remember, when photographing your own tests, the objective is to capture them in the *worst* light possible (i twisted these around for a while until i've found the lighting angle that makes the effect the most obvious, and i did it out in the sun - harsh light makes it more noticeable) :)

-- 
Ante Vukorepa
Sent with Sparrow

Ante Vukorepa

unread,
May 27, 2013, 10:27:15 AM5/27/13
to arco...@googlegroups.com

-- 
Ante Vukorepa
Sent with Sparrow

Ante Vukorepa

unread,
May 27, 2013, 10:29:07 AM5/27/13
to arco...@googlegroups.com
By the way, it would appear that all the changes i've made (which were mostly focused on the Y axis) seem to have had an effect after all, considering the photos now make it obvious the X axis is much much worse.

It used to be the other way around.

So maybe there is no magic bullet for this and it's a matter of tweaking and isolating and adjusting a lot of little things until the issue is attenuated enough.

-- 
Ante Vukorepa
Sent with Sparrow

Kyle Kenney

unread,
May 27, 2013, 11:05:16 AM5/27/13
to arco...@googlegroups.com
Mine is printing at this very moment, though I didn't realize you'd want such small layers (mines printing at .25) But yeah, I'll take some pictures aswell and load'em up when done.

2013/5/27 Ante Vukorepa <o.or...@gmail.com>

Laszlo KREKACS

unread,
May 27, 2013, 11:52:41 AM5/27/13
to arco...@googlegroups.com
Hi,

I have printed it in white, and it is almost impossible to do the photo right,
but I dont't have this issue.

However I have seen this issue on Stratasys machines besides on many
reprap printers.
Here are some pictures from haveblue.org:
http://haveblue.org/wordpress/wp-content/uploads/2012/06/bolt-hold-lugs-break.jpg
(on the left, the edge ripples)

http://haveblue.org/wordpress/wp-content/uploads/2012/06/rear-takedown-counterbore.jpg
(the hole under the drillbit ripples)

http://haveblue.org/wordpress/wp-content/uploads/2012/08/buffer-tower-gap.jpg
(bottom middle, the metal pin ripples)

http://haveblue.org/wordpress/wp-content/uploads/2012/08/extraction-failure.jpg
(bottom middle, the oval metal thingy)

Although it is almost unnoticable, if you don't know where to look.

I'm starting to believe it is first a mechanical reasonance caused
(exaggered) by the stepper motor
"stepping", causing the reasonance like peeling a stretched rubber
band (or guitar string).
Changing the microstepping or the firmware (ie. acceleration, edge
optimization),
makes the issue worse or less noticable.

That is my theory at least.

I don't want to talk you into my belt+pulleys
(http://wiki.arcol.hu/blog:arcol-t25), but
maybe it is worth a try.

Best,
Laszlo

Kyle Kenney

unread,
May 27, 2013, 11:58:42 AM5/27/13
to arco...@googlegroups.com
I printed mine..it seems I have a little echoing so to say, on the Y face...which means through the x-axis, while the X face has nothing...and the two faces with the dimples have no appearance of echoing/ringing at all...though, I printed in natural pla, and I'm find it difficult to take a picture of it, I'll try later tonight when the suns down if the lighting changes anything

2013/5/27 Laszlo KREKACS <li...@arcol.hu>

Ante Vukorepa

unread,
May 27, 2013, 12:10:21 PM5/27/13
to arco...@googlegroups.com
Your theory pretty much exactly matches mine (see the wiki page).

I believe ORD Bot's rigidity is its undoing in this case - it's *too* rigid and most of its construction elements are under relatively high tension, which makes it "ring" easily and makes dissipating the resonance a tough job.

I've got some TI DRV8825-based drivers on order (no idea when they'll arrive), so i'll test what difference (if any) 1/32 microstepping might have.
As soon as the paint dries, i'll install the beefed up stepper and idler brackets, see if that helps in any way.
And i need to figure out a different mount for the X gantry stepper (there's an alternative or two on Thingiverse) and see if that dampens down the X vibration i'm now seeing.

How much do you plan on charging for the belt+pulleys+idler bearings?

-- 
Ante Vukorepa
Sent with Sparrow

Ante Vukorepa

unread,
May 27, 2013, 12:13:18 PM5/27/13
to arco...@googlegroups.com
Forgot to mention that before - the choice of filament has a HUGE effect on the ripples being noticeable vs. not.
Opaque silver and black tend to make it the most obvious.

It's invisible on white and natural, as well as any semi-transparent color (including semi-transparent silvers and metallics).

Printing at big layer heights can hide the effect somewhat.
Same for other issues like irregular layers - they all tend to muddle the "rippling" and make it less visible or invisible.

-- 
Ante Vukorepa
Sent with Sparrow

Ante Vukorepa

unread,
May 27, 2013, 12:18:51 PM5/27/13
to arco...@googlegroups.com
Sorry for the email spam…

Just remembered one more thing.
Slicer can affect the "ringing" too. Well, not affect, rather disguise or amplify.

I can vary the amount of "ringing" in KISS by changing the oversampling resolution and crowning threshold.
Both of these controls can soften the high frequency details, which - unsurprisingly - disguises the problem.

Of course, that's at a cost of surface detail and corner sharpness, but i'm just trying to make a point - you probably won't run into this problem unless you're printing very sharp (high spatial frequency) detail on your objects.

-- 
Ante Vukorepa
Sent with Sparrow

Laszlo KREKACS

unread,
May 27, 2013, 12:41:18 PM5/27/13
to arco...@googlegroups.com
Hi,

> I've got some TI DRV8825-based drivers on order (no idea when they'll
> arrive), so i'll test what difference (if any) 1/32 microstepping might
> have.

Or you could try for experiment a soft motor coupler, like aquarium tubing,
to see if it changes the output.
Although it is only a workaround, so it is always better to fix the
root cause of issue.

Maybe the reasonance coming the lack of rigidity of the machine too.
(ie. too weak frame). But it is only a wide guess.

> How much do you plan on charging for the belt+pulleys+idler bearings?

I will determine the price this week. But in your case I can give you
a special discount.
If it does not fix your issue, I only charge you the half price, or
you can send me back (at your cost),
and I full refund.
(just remember reselling an already cut to length belt will be rather
difficult.)

Best,
Laszlo

nop head

unread,
May 27, 2013, 12:42:47 PM5/27/13
to arco...@googlegroups.com
What is crowning?

Ante Vukorepa

unread,
May 27, 2013, 12:46:09 PM5/27/13
to arco...@googlegroups.com

Matej Rozman

unread,
May 27, 2013, 12:53:37 PM5/27/13
to arco...@googlegroups.com
It just occured to me, i was thinking of something completly different when I was mentioning changing from repetier to marlin,
but now seing the pictures, this is a mechanical problem, and It almost looks like you have a low cost servo drive with poor loop.

There are two possible problems/solutions for this.

1. Your hotend is not attached rigid enough to your carriage.
Consequence is that when machine changes direction suddenly it will produce shaking of the hotend tip (dont know how to put this in different words). But you see what I mean when looking at the pictures.
Solution: lower your acceleration by half, (not jerk) and the effect will be half the size.

2. Your belts are not strong enough to couterforce the stoping action of the carriage/table.
Consequence wery similar to upper problem, but in this case carriage will be going back and forth rather than stop. I hope you know what I want to say.
Solution: lower your acceleration in FW, try to lower the weight of the axis making problems, try stiffer belts (from 6 to 9mm or maybe even more)

To all, stepper always stops exactly where its supposed to. So I think you can rule out voltage, current and all other electrical issues, this is all phisics.

Please do me a favour and lower your acceleration to 1000 or lower and try to print another one. Please leave jerk as is.

Id really love to see where this goes, since I am also interested in one of these ORDs...

BR, Matej


2013/5/27 Ante Vukorepa <o.or...@gmail.com>



--

László KREKÁCS

unread,
May 27, 2013, 12:54:14 PM5/27/13
to arco...@googlegroups.com
On Fri, May 24, 2013 at 1:40 PM, Kyle Kenney <dudl...@gmail.com> wrote:
> Laszlo, did you not get my email about the prototype? Also, I'm down for
> comp too!

Oh, seems like I forgot to reply to you.
Of course, I have read it, and I like the positive feedback.

I have not decided yout what path will I follow.

Best,
Laszlo

Ante Vukorepa

unread,
May 27, 2013, 1:03:40 PM5/27/13
to arco...@googlegroups.com
Yes, all of that was covered before.

Acceleration (and jerk - because jerk in Marlin and Repetier aren't true jerk, but just a speed limit below which there is no firmware-based acceleration, only the acceleration caused by the stepper's torque and stiction) affect the rippling and reduce it.

But that proves nor solves anything :)

It's akin to me saying "my car shakes rather violently if i go above 60 km/h, any ideas on how to dampen that" and you replying with "then don't go above 60 km/h" :)

And yes, it's a mechanical issue. Cartesian bots have lots of degrees of freedom and oscillation modes. Nothing new there.
And yes, the Arcol.hu, being a mass suspended on a thin SS tube is rather prone to oscillating. But those are all symptoms and contributing factors, not causes.

Regarding belts - that's been covered before (plus, the table and the carriage are about the same weight, or the table is a bit heavier. But Y is currently exhibiting less rippling than X. Stiffer belts might help somewhat, tho, as they'd "eat up" some of the vibrations, or they might change the k spring constant and raise the rippling frequency (which is also a solution, if raised high enough).


By the way, whenever someone says they never saw this effect on their printer / prints, i tend to not believe it. Every design out there is prone to oscillations. Perhaps deltas might be slightly less sensitive to them. It's just a question of finding the right test case, printing with the right filament (with surface texture and color that shows undulations well) and looking at it under the right light. I guarantee you you'll find the same artifact on *every* printer if you look or try hard enough, at least to some degree.

If all you do is print herringbone gears and mostly print in clear or white PLA, you'll quite probably never encounter this issue.

-- 
Ante Vukorepa
Sent with Sparrow

Ante Vukorepa

unread,
May 27, 2013, 1:06:10 PM5/27/13
to arco...@googlegroups.com
Oh, one more thing.

Switching from Repetier back to Marlin was a good idea for other reasons.
It turns out Repetier has some issues with buffer underflows (regardless of the setting) with highly detailed surfaces. It slows down to a crawl, jitters and blobs, creating a mottled, orange-peel surface, which actually obfuscates the ripples (and looks bad).

I've noticed that bug before with it, but wasn't bothered enough to go back to Marlin. Your recommendation made me finally do it :)

-- 
Ante Vukorepa
Sent with Sparrow

On Monday, 27. May 2013. at 18:53, Matej Rozman wrote:

Matej Rozman

unread,
May 27, 2013, 1:14:08 PM5/27/13
to arco...@googlegroups.com



2013/5/27 Ante Vukorepa <o.or...@gmail.com>

Yes, all of that was covered before.

Acceleration (and jerk - because jerk in Marlin and Repetier aren't true jerk, but just a speed limit below which there is no firmware-based acceleration, only the acceleration caused by the stepper's torque and stiction) affect the rippling and reduce it.

But that proves nor solves anything :)

It's akin to me saying "my car shakes rather violently if i go above 60 km/h, any ideas on how to dampen that" and you replying with "then don't go above 60 km/h" :)

I know, but then again, if the quality is the issue, go slower, its sometimes faster ;)

And yes, it's a mechanical issue. Cartesian bots have lots of degrees of freedom and oscillation modes. Nothing new there.
And yes, the Arcol.hu, being a mass suspended on a thin SS tube is rather prone to oscillating. But those are all symptoms and contributing factors, not causes.

Sorry, wrote it completly wrong...I dont know how to put it in english....

Regarding belts - that's been covered before (plus, the table and the carriage are about the same weight, or the table is a bit heavier. But Y is currently exhibiting less rippling than X. Stiffer belts might help somewhat, tho, as they'd "eat up" some of the vibrations, or they might change the k spring constant and raise the rippling frequency (which is also a solution, if raised high enough).

Aha, we are getting closer, this is one way to find out whats wrong with your setup, so you probably have Y axis lighter than X, this is one possibility, the other one is still your hotend mounting. You can check this pretty easily if you have a dial indicator laying around. Just put it somwhere near your hotend tip (heater or some flat spot) and engage motors so they will hold the machine steady, then try gently to put pressure to the tip from the corresponding side-180 degrees opposite to the dial indicator. Then rotate dial to 90 degrees and try again. Try to put the as much the same pressure as possible to the tip and see how much play you have.

Matej Rozman

unread,
May 27, 2013, 1:17:48 PM5/27/13
to arco...@googlegroups.com
Yes, I had those problems also, when I turned oversamling in Kiss up (well down), it had problems, and I reckon they were even bigger since delta makes more calculations.
Im also a happy marlin user now.



2013/5/27 Ante Vukorepa <o.or...@gmail.com>

nop head

unread,
May 27, 2013, 1:24:58 PM5/27/13
to arco...@googlegroups.com
The picture makes it clear it is a resonance causing ringing when one axis stops. The low current micro stepping might make it worse because there might be less electrical damping when the driver is in fast decay mode, compared to slow decay.

What you need is an anti-resonance driver but they are expensive. The old fashioned method is to attach a fluid damper that adds dynamic friction without adding static friction (which would increase backlash).

Ante Vukorepa

unread,
May 27, 2013, 2:01:43 PM5/27/13
to arco...@googlegroups.com
So mid-band resonance + resonance of the mechanical structure?

Yeah, i've looked at anti-resonance drivers and have seen the prices.
Fluid dampers i've seen mentioned in CNC context, but never saw anything applicable to small-scale machines.

I wonder if some kind of flexible coupling would help and how much negative impact on precision and repeatability it'd have.

-- 
Ante Vukorepa
Sent with Sparrow

Nicolas Arias

unread,
May 27, 2013, 2:04:11 PM5/27/13
to arcol_hu
maybe try another brand of steppers?

Laszlo KREKACS

unread,
May 27, 2013, 2:24:52 PM5/27/13
to arco...@googlegroups.com
Hi,

Would be nice to have an another machine to cross-check.
Fortunately I have an another one, just for issues like this.

Best,
Laszlo

nop head

unread,
May 27, 2013, 2:49:29 PM5/27/13
to arco...@googlegroups.com
>So mid-band resonance + resonance of the mechanical structure?

Mid band resonance is when the motor is moving. It speeds up and slows down such that it has a sinusoidal displacement from where it should be at any instant. If only one axis was moving then it would only be apparent in the width of the filament oscillating. When moving diagonally you would get a wavy line if the other axis was moving at constant velocity. That probably isn't the problem.

You have waves emanating from features where one axis has stopped and is ringing but the other is moving at constant velocity. The axis that is ringing is a mass spring resonance. If the spring element is the motor then an anti-resonant drive should help. If the frame is springy then probably not, or not as much.

I found a tiny fluid mass damper in an old disc drive, so small ones must exist, or have existed: http://hydraraptor.blogspot.co.uk/2009/04/unexpected-find.html
  

 

Ante Vukorepa

unread,
May 27, 2013, 4:33:06 PM5/27/13
to arco...@googlegroups.com
Hmmmmm…


Anyone tried anything similar (and by similar i don't mean cork or rubber gasket)?
Of course, these wouldn't fix the vibration transmitted down the belt, but it would at least kill the vibrations coming through the mount.

-- 
Ante Vukorepa
Sent with Sparrow

Chris Gibson

unread,
May 27, 2013, 4:40:53 PM5/27/13
to arco...@googlegroups.com
Funny you should mention those. I just got a bunch of the exact ones. I got them hoping they would make the printer quieter. I haven't tried them yet but I'm going to use them on the printer I'm building right now. When I have a test print done (probably in the next few of days) I will post pictures. Can you send me the file you are having difficulty with?

Ante Vukorepa

unread,
May 27, 2013, 4:44:53 PM5/27/13
to arco...@googlegroups.com
No particular file (or, rather, after debugging this, i've started seeing the rippling everywhere :)), but here's a test object:

Best noticed on black or silver PLA.

Did you order any extras by any chance?

-- 
Ante Vukorepa
Sent with Sparrow

Chris Gibson

unread,
May 27, 2013, 4:54:14 PM5/27/13
to arco...@googlegroups.com
I ordered 12. But I need them all for the printers I'm building right now. I could get some for you with my next order if you like. You do need to keep in mind that they move the stepper away from the plate by about 5mm so you need to make sure you have clearance and enough shaft length...

Ante Vukorepa

unread,
May 28, 2013, 2:48:44 PM5/28/13
to arco...@googlegroups.com
No worries, i've got in contact with Astrosyn in the meantime and they're willing to send some my way, despite Croatia not being listed on their web shop.

As a sidenote, i found two interesting things while googling in the off hours.

1) Vexta offers these: http://www.orientalmotor.com/products/accessories/clean-dampers-cmk.html They mount on double shaft steppers (on the "non-business end"). They also have an assortment of rather nice and expensive steppers. I wonder if they're worth their price...


Some more late-night developments from yesterday:
- i've mounted the extruder+hotend on a thick silicon pad, haven't tested the effect yet
- i've bought a bunch of appropriate sized (M4, M5) rubber washers and some M3 gaskets (couldn't find rubber washers in that size anywhere), i'll see if they help
- i've flicked a finger at the X stepper and realized that it continues to resonate when i do that, along with the X gantry - that's with it tight and completely solidly mounted; if i loosen the bolts a little, it doesn't resonate as badly, i'll see about printing a different mount for it, this one pretty much turns it into a tuning fork
- i've noticed a new artifact in the prints, which is unrelated to "regular" rippling - more about that below

So, yesterday, i did some slow 0.1mm prints to test something and noticed something new - one X side exhibits very regular banding. Then i went back to the test objects i've posted here and noticed the same thing in one of them:


(top right shot)

You might discount it as corner ripple or a belt artifact, but take a better look and you'll notice some interesting things:
- it's present on the filleted half as well
- it's only present on one side (the top left shot is opposite to it)
- its highly regular
- its frequency coincides with the frequency of the regular ripples
- the belt is flipped on the idler and the frequency of the artifact is nowhere close to belt pitch (it's approx. 2.5x higher)

I believe *that* is the natural resonance frequency of the X gantry. The reason it's so even and not caused by the edge is because it's continuously fed by the vibration of the stepper and has nowhere to dissipate. The only question is - why is it not present on the opposite side.


By the way, i showed the test prints to a few people and had to twist and turn them for a minute and then explain for a few more before they've even noticed the ripples and vibration artifacts. So i might be obsessing too much about this :)

-- 
Ante Vukorepa
Sent with Sparrow

nop head

unread,
May 28, 2013, 3:14:25 PM5/28/13