1 (edited by tealvince 2013-04-22 15:11:31)

Topic: Fixing banding (z-wobble) with fine threaded z-rod

I was thinking... since the frequency of banding on the Solidoodle matches the thread spacing of the z-rod (18 threads per inch), why not just use a z-rod with a finer thread?  Unlike other hardware solutions, which try to only minimize the "amplitude" of banding, using a finer thread would also change the frequency, instantly making it much less visible.  The finer thread would also eliminate almost all backlash.

I decided to incorporate a couple of other improvements at the same time, so I used a very thin stainless steel M3 z-rod to benefit from its flexibility ($3.01 + $6.31 shipping):
http://www.ebay.com/itm/300885751211

This rod had a thread spacing of 0.5 mm, almost 3 times finer than the default z-axis rod.

And attach it using a small 5mm-3mm flexible helical coupling ($4.20 + $1.50 shipping)
http://www.ebay.com/itm/220977869760

For the new z-nut, I used a long brass spacer (pack of 10 - $4 + $2 shipping)
http://www.ebay.com/itm/250955139328

I found that M3 components are much more affordable than the larger components typically in demand for 3D printing applications.

http://www.soliforum.com/misc.php?action=pun_attachment&item=1829

I re-tapped the inside of the spacer to make sure the thread ran all the way through.  If I had thought it out more fully before disassembling the z-carriage, I would have printed out some sort of holder adapter for the z-nut.  Since I didn't, I just tapped most of the outside of the brass spacer and threaded on a larger nut and sealed it with locktite.  With a short section of tubing, this nut fit snugly in the existing nut hugger.

http://www.soliforum.com/misc.php?action=pun_attachment&item=1833
Here is the assembled nut

http://www.soliforum.com/misc.php?action=pun_attachment&item=1831
And the printer with the installed z-rod

After reassembling, all I had to do was add some grease and increase the Z steps/mm value to 6400 (M92 Z6400).  I did find I needed to turn down the maximum z speed from 5mm/s to 3mm/s (M203 Z3) to keep it moving reliably, but since the bed doesn't move much during printing, that seemed fine.

http://www.soliforum.com/misc.php?action=pun_attachment&item=1832
And here are the results from printing with no software compensation enabled.   The object is a small octagonal cylinder printed at .1mm and viewed through a magnifying glass. The banding that was present before is now so slight and fine that it's negligible.

Post's attachments

CIMG1072.jpg 979.29 kb, 3 downloads since 2013-04-22 

CIMG1073.jpg 948.15 kb, 1 downloads since 2013-04-22 

CIMG1074.jpg
CIMG1074.jpg 858.06 kb, 1 downloads since 2013-04-22 

CIMG1078.jpg
CIMG1078.jpg 936.73 kb, 1 downloads since 2013-04-22 

You don't have the permssions to download the attachments of this post.

2

Re: Fixing banding (z-wobble) with fine threaded z-rod

very well made... it prove that you need either a very rigid ball screw or a thin flexible screw... so the standard middle size screw of most 3d printer is the worst choice...

3

Re: Fixing banding (z-wobble) with fine threaded z-rod

ysb wrote:

very well made... it prove that you need either a very rigid ball screw or a thin flexible screw... so the standard middle size screw of most 3d printer is the worst choice...

Yes!  I checked out the price of ball screws at one point and .. yikes!

The thin lead screw and coupling I got are an incredibly inexpensive upgrade (~$7 + shipping) using off the shelf hardware (if you printed a modified nut hugger).  I hope someone a Solidoodle sees this, as I imagine they could offer this in bulk for no more than they currently spend on the the heavier drilled-out leadscrew that they ship as standard.

4

Re: Fixing banding (z-wobble) with fine threaded z-rod

How bad was your banding before at the .1mm level?

5

Re: Fixing banding (z-wobble) with fine threaded z-rod

Jonathan wrote:

How bad was your banding before at the .1mm level?

It was not nearly as bad as some, but without an anti-backlash device it was fairly prominent.  With an anti-backlash device there was still some there that I used software compensation to address. With a thin rod,I don't think any amount of banding would look as bad as even slight banding at the coarser thread pitch, since the scale of the banding is automatically so much smaller.

6

Re: Fixing banding (z-wobble) with fine threaded z-rod

Nice take on sorting the problem.

As a suggestion/question could you not have put a thread on the brass spacer so it would have fitted in the standard bed nut and the just use a second lock nut to hold everything in place? It would save any rework to the bed and make it simple to revert if needed.

As for a good working solution though I'm off out to find some fine threaded rod and bits of brass smile

7

Re: Fixing banding (z-wobble) with fine threaded z-rod

I PM'd this thread to SDsupport lol.

This should be passed on their dev department.

No trees were harmed in the creation of this email, though some electrons were horribly inconvenienced.

8

Re: Fixing banding (z-wobble) with fine threaded z-rod

Zarni wrote:

Nice take on sorting the problem.

As a suggestion/question could you not have put a thread on the brass spacer so it would have fitted in the standard bed nut and the just use a second lock nut to hold everything in place? It would save any rework to the bed and make it simple to revert if needed.

As for a good working solution though I'm off out to find some fine threaded rod and bits of brass smile

The spacer I used is too small to be threaded to fit inside the standard nut, but yes, cutting one from scratch brass rod is probably the way to go.

After playing with it a few days, I've decided that the helical coupling gives a little more vertical movement than I'd like, and I can bounce the bed up and down a bit pushing on it with my hand.  I guess that's not unexpected, since the couplings aren't really meant to be used in this way.  I think I can improve it either by using a stiffer coupling, using an inflexible coupling (relying on the flexible rod alone), or pre-compressing the existing coupling  with a rubber band.  I'll report on what I find.

9 (edited by tealvince 2013-04-24 01:32:54)

Re: Fixing banding (z-wobble) with fine threaded z-rod

Ok, it turns out that fixing the compressibility of the helical coupling was really easy;  I just pushed in both the z-rod and motor shaft all the way into the coupling so they touched each other inside the coupling (preventing compression). 

I even tried to stretch the coupling a little bit before tightening it up.  Now it's rock solid in the z direction but still flexible.  If I ever take it apart again, I might round the bottom end of the z-rod for good measure, but I'm getting some of the nicest prints I've seen from my printer now.

10

Re: Fixing banding (z-wobble) with fine threaded z-rod

Just did this mod a couple of days ago. Chose this mod over the others based on the image that tealvince shows while viewing through a magnifying glass..... WOW!

it makes a MASSIVE difference.

getonit!

11

Re: Fixing banding (z-wobble) with fine threaded z-rod

saymo606, do you have any before and after print photos?

12

Re: Fixing banding (z-wobble) with fine threaded z-rod

not the best pic using an iphone...

Post's attachments

photo.JPG
photo.JPG 458.44 kb, 4 downloads since 2013-05-19 

You don't have the permssions to download the attachments of this post.

13

Re: Fixing banding (z-wobble) with fine threaded z-rod

tealvince wrote:

After reassembling, all I had to do was add some grease and increase the Z steps/mm value to 6400 (M92 Z6400).

How did you come out to 6400 steps/mm? I am going to use m5 rod with .8 pitch or 1.25 rotations per mm. I would assume the Z motor is 1.8 degree/step or 200 steps per rotation. 1.25x200 = 250 steps/mm

Is there a multiple that I am missing? For M3 I was calculating 400 steps/mm so 6400/400 =16
Does that mean I should be at 4000? 250*16

I will probably start at 4000 and calibrate from that point. I am just curious about how the math works out.

SD3 w/ mods:
Glass bed with QU-BD heat pad upgrade, threadless ballscrew w/ 8mm smooth rod, spectra line belt replacement, lawsy MK5 extruder, Lawsy replacement carriage, E3D hotend, Ramps 1.4 w/ reprap discount controller, DRV8825 drivers, 12v 30A PS, Acrylic case, Overkill Y-idlers, Filament alarm, Extruder fan + more.

14

Re: Fixing banding (z-wobble) with fine threaded z-rod

The stepper drivers are all set to 16x microstepping.

Don: Folger Tech 2020 Kossel Rev A + Borosilicate + Snow Effector
        Davinci 1.0 + Repetier : Filastruder
        SD3 + RAMPS + Lawsy Carriages + E3D + Borosilicate + ... : Cupcake

15

Re: Fixing banding (z-wobble) with fine threaded z-rod

Leghk wrote:

The stepper drivers are all set to 16x microstepping.

That makes sense now. :-) thanks for clearing that up.

SD3 w/ mods:
Glass bed with QU-BD heat pad upgrade, threadless ballscrew w/ 8mm smooth rod, spectra line belt replacement, lawsy MK5 extruder, Lawsy replacement carriage, E3D hotend, Ramps 1.4 w/ reprap discount controller, DRV8825 drivers, 12v 30A PS, Acrylic case, Overkill Y-idlers, Filament alarm, Extruder fan + more.

16

Re: Fixing banding (z-wobble) with fine threaded z-rod

I will be doing this mod in a couple of days. Could somebody clarify where to make the change for the steps? I remember words like firmware and gcode, but my small mind can only hold so much info : )

17

Re: Fixing banding (z-wobble) with fine threaded z-rod

machinistjoe wrote:

I will be doing this mod in a couple of days. Could somebody clarify where to make the change for the steps? I remember words like firmware and gcode, but my small mind can only hold so much info : )

Repetier-Host

This screen.
http://2n2r5.com/pictures/printer/rh_eeprom_settings.jpg

SD3 w/ mods:
Glass bed with QU-BD heat pad upgrade, threadless ballscrew w/ 8mm smooth rod, spectra line belt replacement, lawsy MK5 extruder, Lawsy replacement carriage, E3D hotend, Ramps 1.4 w/ reprap discount controller, DRV8825 drivers, 12v 30A PS, Acrylic case, Overkill Y-idlers, Filament alarm, Extruder fan + more.

18

Re: Fixing banding (z-wobble) with fine threaded z-rod

So why would you want the rod to flex?  Does that prevent some sort of binding?

I know this works, and you guys have proved it. I am probably going to do it as well.  But I look at that tiny lead screw holding up the bed and I just keep having this image in my head of the whole screw bending and flexing and making the bed move around when its raised up. 

What stops this?  Or is the M3 lead screw rather strong and I'm just underestimating it in my head?  I just wonder that if such a tiny lead screw is better then a big one, why does everyone use larger diameter screws?

19

Re: Fixing banding (z-wobble) with fine threaded z-rod

ITman496 wrote:

So why would you want the rod to flex?

Because if two things are not aligned perfectly, one must flex. Better the rod than the bed.

How do you think the ride of a car would be with concrete tires? Same deal here - with a stiff rod, misalignments and disturbances get transferred all the way to the bed.

ITman496 wrote:

I just wonder that if such a tiny lead screw is better then a big one, why does everyone use larger diameter screws?

Not much in the way of comprehensive engineering analysis has gone into hobbyist 3D printers.

20 (edited by TheBaron 2013-07-21 03:41:46)

Re: Fixing banding (z-wobble) with fine threaded z-rod

I work on professional cnc equipment for a living, and I think the main difference here is tolerance. The heads on our machines are all milled in one piece from the same machine in one go. All the bearing car rails and the mounts for the z axis bushings are cut at very high tolerances so a heavier ball screw (about 18" long) can be used. The end result is a very heavy screw with very little flex. The other result is a lot of binding should the screw not be lined up just right. the screw has to be slightly tightened in its bushing and then moved again and again, tightening it as you go just to make sure the screw doesn't get destroyed a month down the line.

On the older machines with a full Y axis screw, the screw is a couple inches thick and about six feet long. Rarely have any binding issues with those and there is plenty of flex. A proper flexing screw is a ratio of length to thickness.

Remember whenever you change out a screw to check your bind at the top and bottom of the screw. That's where it's most likely to give you trouble.

Edit: also keep in mind that you greatly reduce the power output and increase the friction the pointier the threads are. Acme theads give a lot more power and a lot less backlash than pointed threads. That said, the weights involved aren't too bad and the results speak for themselves.

don't even have my SD yet so I should probably calm down a bit.

Edit again: Also, acme threading at 40 turns per inch is about $50 for a 18" length. Might not be worth the extra drive power.

SD3 w/ RUMBA (8825s), merlin hot end (in pieces). Ender3 w/ silent board for PLA printing. Ender5 w/ silent board, e3d v6, new z lead screw, and glass bed. DiY Kossel w/ smoothieboard (in pieces). Vellman Vertex (in pieces)
Shapeoko2 router in process of being converted to laser engraver (in pieces)
Multicam 5000 series CNC Router w/ 11HP spindle, 5x10' table, and auto tool changer (in pieces)

21 (edited by tealvince 2013-07-21 05:15:45)

Re: Fixing banding (z-wobble) with fine threaded z-rod

I wondered if an M3 rod might be too thin myself, which is why I chose to try it in the first place; to push the boundaries of what was possible. I wanted the finest possible thread which in theory should have the best effect on eliminating banding, since the banding frequency has been found to match the z-rod thread pitch.  I was pleasantly surprised that an M3 is indeed plenty strong for my SD2's range of motion, though I did need to reduce the maximum bed z-movement rate a little to keep the motor from binding in the long movement at the start of a print.   This is because an M3's .5mm thread is so fine that the motor has to spin 3x faster to move the bed at the same speed, and if driven too fast the steppers seem to get out of sync with the signal driving them.  Fortunately, z-speed doesn't really matter during printing since the bed only moves one layer at a time.

Incidentally, I manually entered gcode commands in pronterface to change the steps per mm, lower the max speeed, and to save these settings to flash.  I don't remember them offhand, but I think I mentioned them in an earlier post.

Addendum:  I believe the thickness of the default threaded rod might be a carryover from other previous designs.  On other 3d printers, the threaded rod(s) also keep the z carriage aligned, but on a solidoodle there are two fixed zrods that do this, so the threaded drive rod can be thinner.

22

Re: Fixing banding (z-wobble) with fine threaded z-rod

Makes perfect sense. Thank you!

23

Re: Fixing banding (z-wobble) with fine threaded z-rod

How would one apply this to the aluminum Z carriage?  How do you even get the old bolt out?

24

Re: Fixing banding (z-wobble) with fine threaded z-rod

ITman496 wrote:

How would one apply this to the aluminum Z carriage?  How do you even get the old bolt out?

http://www.soliforum.com/topic/2990/bos … -consumer/

Also look for the slopnut, I included the sketchup files that you could easily make your own M5 nut retainer if you didn't want to use it as a slop nut.

SD3 w/ mods:
Glass bed with QU-BD heat pad upgrade, threadless ballscrew w/ 8mm smooth rod, spectra line belt replacement, lawsy MK5 extruder, Lawsy replacement carriage, E3D hotend, Ramps 1.4 w/ reprap discount controller, DRV8825 drivers, 12v 30A PS, Acrylic case, Overkill Y-idlers, Filament alarm, Extruder fan + more.

25

Re: Fixing banding (z-wobble) with fine threaded z-rod

Thank you!