my rod is just fine :°

seriously, i don't  have a problem, the z rod is fine, it's just one of those mods,I do just to make sure.

so wich one should i take?

i 'm metric base, imperial does not speak to me...

Imperial isn't hard to figure out unless you just want to swap the rod out. Search it on Google. I believe it's a 5/16-18

seriously, i don't  have a problem, the z rod is fine, it's just one of those mods,I do just to make sure.

so wich one should i take?

i 'm metric base, imperial does not speak to me...

metaldrgn wrote:

I would just ditch the 5/16"s rod and upgrade to an M5. You need at least 210mm of it. The well nuts are a bandaid fix and you'll still see the problem.

not true if your rod is straight. if your rod is bent obviously the wellnut will not help because the platform will flop around as the screw turns. it's amazing how bent the Z screw can be without binding.

True but from what I read it seems most if not all stock rods were deformed. Was yours?

I would just ditch the 5/16"s rod and upgrade to an M5. You need at least 210mm of it. The well nuts are a bandaid fix and you'll still see the problem.

not true if your rod is straight. if your rod is bent obviously the wellnut will not help because the platform will flop around as the screw turns. it's amazing how bent the Z screw can be without binding.

I did a serach on ebay, and found one in uk, but i don't know witch one to take

http://www.ebay.fr/itm/M3-M4-M5-M6-M8-M … 3386fc5b8e

can you point the one please

as for changing the rod, i'm a tiny bit overwhelm at doing it, i did change the carriage and the hot end and it definitely challenged my skill, and I'm gonna have to do it all again, so I rather not change it now.

if i change it I rather have something like this bed... ;°

http://bigbox-3d.com/features

overkill much  but I really like the idea...

http://www.thingiverse.com/thing:786086

You need to upgrade to an m5 rod first which should eliminate the wobble (if you get a decent quality rod).  It significantly reduces backlash. You just need a flared 8mm nut for the top and another one (I used a 10mm) for the bottom.

hi

I'm french, and obvioulsy in live in france and cannot find one of those? can anybody help me,  send one to me?

thanks. (I will pay for obviously)

have you searched eBay? many sellers will ship world wide. thing is, it would be cheaper for you to convert to a metric rod than it would be for me to ship you a single wellnut, most likely.

I'm french, and obvioulsy in live in france and cannot find one of those? can anybody help me,  send one to me?

thanks. (I will pay for obviously)

Good point, I'll keep an eye out - hadn't considered drift over a long space of time / including vibration from printing. Could cause issues if the print effectively stretches out...

Hahaha so true. Looks like to my mind I find the z-offset setting more confusing than gcoding it, but yes.   (maybe this is the same reason I struggle with arduino but get along ok with regular AVR/GCC, for example).

I read "BRO, DO YOU EVEN Z-LIFT?"

With a z-axis this troubled, I haven't been brave enough. Makes a good argument to ensure mechanical repeatability.

Just thinking about this though, if the z-lift always brings the bed towards the nozzle before each layer, then wouldn't consistent backlash in the drive (like I seem to be getting) be irrelevant? I'd just need to increase z-lift distance to compensate (if that really is a critical parameter!). Turning on z-lift would negate the need for compensation at the start too. Hmmm. I shall experiment some more...

chancrescolex wrote:

I replaced my anti-backlash solution with the same well nut you used, except not upside-down like you have pictured. The banding came back. I think it has to do with my z rod being super bent. I'm just going to replace it with an M5 rod.

The M5 rod should make a big difference to wobble, and allow for genuine metric layer heights haha.
Good luck with the modification, let us know if you have any trouble!

Talking about backlash in z, you can quite easily compensate for it in software without patching Marlin at all. The z axis only changes direction once during a print: when homing just before starting. My latest solution is to measure the backlash with a dial gauge, and then compensate in the start-gcode just before printing, like so:

G28 Z ; home Z
G1 Z0.23 F450 ; eat up the backlash
G92 Z0 ; reset z height to 0
;... print...

(yes, 0.23mm is enormous, have to investigate the motor bearings/mountings a bit more...)

I think this only works if you have a lot of friction on your Z axis to keep the threaded rod contacting the bottom face of the nut thread.  Otherwise, gravity will cause the bed&nut to drift back down and rest on the opposite face again.

...The z axis only changes direction once during a print: when homing just before starting.

Don't you ever use Z lift during retraction?  If not, you should look into it.

It can be very helpful in certain circumstances, but it assumes you can achieve accurate repeatability from your Z axis.

I replaced my anti-backlash solution with the same well nut you used, except not upside-down like you have pictured. The banding came back. I think it has to do with my z rod being super bent. I'm just going to replace it with an M5 rod.

The M5 rod should make a big difference to wobble, and allow for genuine metric layer heights haha.
Good luck with the modification, let us know if you have any trouble!

Talking about backlash in z, you can quite easily compensate for it in software without patching Marlin at all. The z axis only changes direction once during a print: when homing just before starting. My latest solution is to measure the backlash with a dial gauge, and then compensate in the start-gcode just before printing, like so:

G28 Z ; home Z
G1 Z0.23 F450 ; eat up the backlash
G92 Z0 ; reset z height to 0
;... print...

(yes, 0.23mm is enormous, have to investigate the motor bearings/mountings a bit more...)