will1384's Ord Bot Hadron

Post your build logs here

Re: will1384's Ord Bot Hadron

Postby orcinus » Wed Apr 09, 2014 5:54 pm

will1384 wrote:I found this web page

http://wiki.arcol.hu/blog:printing-reasonance


That was actually my article :oops:
(made as a culmination of a lot of back-and-forth on the Arcol.hu mailing list with suggestions and comments from Laszlo, Nophead and others)

Important thing to note re: feet pads - just *any* feet pads won't work. The point is to have feet that are wobbly by design, the idea being that the *feet* actually absorb the vibration that would otherwise continue ringing through the ORD Bot frame. Wobbly is good. Multiple layers of padding made from different materials is also good (e.g. layer of cork, layer of silicone, layer of foam).

None of the solutions i've proposed is a magic bullet. They all add-up together and target different sources of vibration.
The only ones that are 100% certain to help are:

1. tightening the V-wheels (by A LOT - if you *think* they're tight, they're not tight - if you're having quite a bit of trouble moving the carriage by hand - they're tight)
2. loosening the belts (by a LOT)
3. reducing the acceleration

Obviously, each comes with its downsides. 1) will wear out or break your V-wheels and create bumps in them if left too long in one place. 2) will make your layers uneven. 3) will make your prints slower and corners duller. You need to find a compromise between them.

Another note - you will never, ever be completely vibration artifact free. If you think you are, i can guarantee you it'll rear its ugly head again if you try:
- printing hotter
- printing in PLA instead of ABS
- printing in a different kind of PLA (or a different color)
- printing faster or printing an object with sharper surface features

What you can do is mitigate it to a degree.
orcinus
 
Posts: 720
Joined: Tue Feb 14, 2012 4:03 am

Re: will1384's Ord Bot Hadron

Postby orcinus » Wed Apr 09, 2014 6:02 pm

NB: after that article, i've tried some other things.

1) dampers on steppers (e.g. http://www.cloudynights.com/ubbthreads/ ... G_1429.jpg)
2) stiffer belts (very stiff steel lined T2.5 belt)
3) 0.9 degree steppers with lower microstepping (1/4)
4) moving the X-carriage CoG further back

Neither of those turned out to be a magic bullet either (as expected).
They did reduce the artifacts a bit further (or increased the maximum allowable acceleration if you want to look at it that way), but haven't eliminated them.

One interesting thing i've noticed is that recently, after all the modifications, i've been having one direction of the prints almost perfectly vibration-free, but the other with vibration artifacts present. In other words, -X face of the test object is vibration free, while the +X has artifacts. +Y is vibration free, -Y has vibration artifacts.

After excluding all the other factors, i'm 90% sure this is due to cooling (i have a ducted fan blowing at one corner of the nozzle and a non-ducted one blowing from the opposite side and cooling the hotend as well), but i haven't tested that hypothesis yet.
orcinus
 
Posts: 720
Joined: Tue Feb 14, 2012 4:03 am

Re: will1384's Ord Bot Hadron

Postby orcinus » Wed Apr 09, 2014 6:07 pm

Oh, and one more thing.

The vibrations seen don't really have anything to do with wobbling.
They're the product of rigidity, not lack of rigidity.

If you think about it, the wavelength of the ripples is actually indicative of near-audible frequencies. That's not wobble, that's pure vibration.
And the reason we're experiencing it with the ORD Bot (especially after everything else gets tuned in) is because it's actually too rigid for its own good.

It rings. And there's nothing in the structure to absorb that either, as it's nearly all metal.

As far as vibration artifacts are concerned, machines with relatively "flimsy" vertices connected by plastic connectors (like Prusa & co.) are actually *better* at mitigating this.
orcinus
 
Posts: 720
Joined: Tue Feb 14, 2012 4:03 am

Re: will1384's Ord Bot Hadron

Postby flickerfly » Wed Apr 09, 2014 6:13 pm

That makes me wonder if it would be beneficial to fill the extrusion with hot glue or some better material that might absorb it?
flickerfly
 
Posts: 247
Joined: Sat Apr 27, 2013 10:04 pm
Location: Allentown, PA

Re: will1384's Ord Bot Hadron

Postby OliverK » Thu Apr 10, 2014 1:58 pm

Ran two quick builds of orcinus' Ripple Test: (tried to take picture with unflattering light to show the imperfections better).

vibro test.jpg


This machine has MXL not GT2 belts (not for long) and M8 Z rods (ACME rods will be installed as soon as the flexible shaft couplings arrive). Printer was run as-is - didn't tighten belts or wheels for the occasion, nor weigh down the extrusions*, plus the Slic3r settings weren't tuned either - note slight over extrusion.

Result was better than expected: there is over-travel just after the change of direction, but there are no vibration artefacts.There is also far less difference between the 40mm/s and 70mm/sec samples - looks like the lurch is not directly (linear) in proportion with the speed, but that will require further testing. I'd guess that the lurch is sum of all the little "gives" in the system... stepper, belt stretching, frame flexing, bed mount wobble, etc.


tinyg uses impressive speed ramping to minimise vibration... check this out:



*) flickerfly's idea of filling the (fixed) extrusions with something heavy is worth pursuing. Using a heavy material that has a bit of give, like sand or shot, ought to act as a massive dampener/energy sink.
User avatar
OliverK
 
Posts: 5
Joined: Fri May 17, 2013 10:14 am
Location: Sydney, Australia

Re: will1384's Ord Bot Hadron

Postby will1384 » Thu Apr 10, 2014 5:43 pm

I tried using large rubber washers on the extruder, tightening the belts up as much as possible, and using soft silicone pads under the Ord Bot feet.

The rubber washers I used, might be latex, they were rubbery and spongy.

Image

Image

Soft silicone pot holders I used as feet pads.

Image

With it setting on the feet pads like this, the Ord Bot would rock side to side when printing.

Image

I think "maybe" tightening the belts made things worse by a tad.

Image

I get the parts for the Bowden extruder setup today, I will post back if that helps any.

OliverK wrote:Try two things:

1. Mounting the (at times rapidly oscillating) build plate and PCB on springs, held in tension by four loose M3 screw and nut assemblies always struck me as introducing excess movement (due to jerk/lurch) into the system. Get rid of the springs and replace them with Nylon (PA) PCB spacers. Obviously they don't have the same adjustment range as a spring, but they compress (and bounce back) enough to allow perfect build surface leveling. For grosser adjustment they can be packed with M3 washers. This makes for a very rigid assembly. I expected the heat bed to warp due to being fully constrained, but nothing happened. Head crashes are taken care of by Barts' excellent Z-wobble fix.

2. Reduce moving mass as much as possible. The standard Y carriage assembly (carriage plate, heated bed, glass, V-groove wheel assemblies, etc.) weighs more than 900g. I made a Y carriage plate with lots of holes that's half the weight of the original. It doesn't just reduce vibration because there is less mass being thrown around (less jerk/jolt during changes of direction), but it's also gentler on the whole drive system. This would also let you drive things faster, or reduce the stepper power.


I will also try replacing the springs, I might also take a hole saw to carriage plate, thanks for the ideas everyone :D
will1384
 
Posts: 35
Joined: Sat Mar 15, 2014 10:56 pm

Re: will1384's Ord Bot Hadron

Postby flickerfly » Thu Apr 10, 2014 6:22 pm

One post I saw somewhere said that actually loosening the belts might help. Since you tightened yours and the vibration pattern got more rapid it makes me wonder. Also, is your X-axis stepper on metal standoffs? My kit from ATI came with long nylon standoffs. I wonder if that actually helps grab some vibration. I've noticed them vibrating recently so I think they may actually be a bit loose right now, but they definitely have a touch of give in them.

There is a thing on thingiverse for the X-axis mount that your could print that would probably be equivalent to my nylon spacers.

http://www.thingiverse.com/thing:43879
http://www.thingiverse.com/thing:106721
http://www.thingiverse.com/thing:75160
http://www.thingiverse.com/thing:137220
http://www.thingiverse.com/thing:154547

EDIT: adding links and removing search suggestion
EDIT2: Found a few more, thingiverse sucked me in
flickerfly
 
Posts: 247
Joined: Sat Apr 27, 2013 10:04 pm
Location: Allentown, PA

Re: will1384's Ord Bot Hadron

Postby orcinus » Thu Apr 10, 2014 8:39 pm

Re: tightening the belts - like i've said, it's the *opposite* that helps:

2. loosening the belts (by a LOT)


... but that comes at a price (uneven layers).

OliverK wrote:Ran two quick builds of orcinus' Ripple Test: (tried to take picture with unflattering light to show the imperfections better).


That looks very nice!
I'm curious about the material and texture, though - usually, the higher the temp and the glossier the print, the more visible the artifacts get.
Was that PLA or ABS?
orcinus
 
Posts: 720
Joined: Tue Feb 14, 2012 4:03 am

Re: will1384's Ord Bot Hadron

Postby orcinus » Thu Apr 10, 2014 8:43 pm

FYI, i've tried ditching the springs on the bed a while ago too.
Didn't really notice any difference.

And even though i have the Z axis floating free, i still don't feel comfortable with a fully fixed bed so i've gone back to screws.
Your mileage may vary, though.

Reducing the bed weight - now that i'd like to try.
orcinus
 
Posts: 720
Joined: Tue Feb 14, 2012 4:03 am

Re: will1384's Ord Bot Hadron

Postby will1384 » Thu Apr 10, 2014 10:56 pm

I cut the carriage plate with a jigsaw into an "X" shape, my carriage plate was 1/8th thick, from what I can remember I marked 1 and 3/8th inch from the corner on all sides to make my "X" pattern.

Image

Image

This gives me more travel also.

Image

I removed the bed springs, and used nylon washers, I did have to use extra metal washers on two ends of the hot bed to get it level.

The test print looks a lot better than it did :)

Image
will1384
 
Posts: 35
Joined: Sat Mar 15, 2014 10:56 pm

PreviousNext

Return to Build Logs

Who is online

Users browsing this forum: No registered users and 29 guests

cron