Fixing banding (z-wobble) with fine threaded z-rod (Page 2 of 4)

If you are going to mod your Z then I would say go with the smaller threaded rod because then you are effectively shrinking the thread error giving less noticeable unevenness. As for bending it since we are printing with plastic the weight concerns are negligible... now if we were print with steel then there would be some concern

I figured you would say that... lol

Personally, I am getting the parts to replace the rod with an M6 thread. The bed is incredibly light, so I am not doing it for strength. The M3 can handle 58,000 PSI tensile strength. The reason why I am going with M6 is that it is cheaper to buy the threaded rod, well nut, and a 5mm-5mm coupler. Total cost is $10USD.

Also, the 5/16"-18 thread is wrong for this system. It does not evenly divide steps for mm layer heights. I will skip the math and tell you to use Josef Prusa calculator and you will find that the steps per mm for this thread is 2267.72, which we round off and put 2268 in the EEPROM. This leads to inaccurate steps when calculating .3mm, .2mm, and .1mm layer heights.

Now the M6 thread conveniently adds up to 3200 steps per mm. This factor will evenly with the layer heights we usually calibrate:

960 steps for 0.3mm
320 steps for 0.1mm
80 steps for 0.025mm

The M3 can double the resolution, but you honestly dont need it. Just like you dont need the extra strength of the M6 over the M3. But a 2.5mm to 5mm coupler is harder to locate.

Pardon my ignorance Hazer,
I've never dealt with this couplers, but should its diameter match the rod you're trying to fit in?
M6 rod means its diameter is 6mm, does it fit into a 5mm coupler?

Nick

Tealvince from your pics I cannot determined if your M3 rod is attached only to the stepper motor, as the stock rod, or if the rod is longer and held in place also to the top with some bearing.
Additionally, I found M3 rods made of A2 stainless steel, A4 stainless steel, galvanized steel and brass, the first the cheapest. Is it fine to use the A2 version?

Thanks
Nick

While I agree that the round-off does not cause considerable banding, I do believe that the calculations mess up the actual layer heights. So, for the sake of having 20 layers come out to a physical height of 6mm, using a metric thread will help in that regard.

As for the M3, you are only getting twice the resolution. As for banding, this is better in the regards that you get twice the number of threads engaging with the same thickness of nut. But I plan to use a 3/4"-1" long well nut to get enough engagement to get the same effect. I also plan to print out a special nut and tap it for an effective backlash, but that is a future thing that needs testing.

honestly, I have more of a problem with the sloppy bearings.  The whole aluminum assembly tilts up and down, side to side, and then the bed itself wobbles all around on the springs. Z-wobble does not seem to have nearly as much of an effect as when the nozzle bumps a high-point of extrusion and shifts the entire placement of the print. It makes me think that changing the leads screw to solve Z problems is like pissing on a brush fire.

Yeah.  You make a good point about the backlash.  That's why I tapped a 1" long brass standoff as the nut for my M3. 

Another reason I think an M3 might help, however, is that the banding necessarily gets reduced when the banding frequency approaches half the frequency of the layer heights. 

For instance, when printing at .3mm layers, the highest frequency completely-periodic banding it can reproduce is one that makes bands .6mm apart.  This is because the "nyquist" frequency is 1/2 the "sampling" frequency, which is equivalent to twice the period (layer height).  This makes sense because you can't create any finer bands than layers that alternate back and forth between too wide and too narrow.  Although, if it were possible to have a workable thread pitch of .1mm, I would have gone for that!

Have you thought about trying to find or make a m3 to m5/m8 standoff adapter with m3 on the inside threads and an OD of 5 or 8mm, I have seen round m3 standoff's but can't find any at the moment.... however there are plenty of m3 with a 5mm hex outside that could be turned or filed down to round.

Cutting the rod is simple enough with a hacksaw or cutoff abrasive wheel the main thing to avoid is a shear type cutter as this WILL bend the ends being cut... unfortunately this is the method most used by supply houses and thus is the reason why if you have things cut by them the ends will most likely be bent... just make sure and grind or file a chamfer on the ends when done cutting.

SD2 or SD3?

Haha epic fail to me!
Yes I have a SD3, didn't notice Tealvince has a SD2. We should ask our admins if we can put our printer model under the nickname, so no one gets confused

Now another problem, it seems everybody just stocks 200 or 300 mm precut threaded rods. Probably I'll buy the 300 mm one and then cut them by myself. Is it okay if I clamp it then cut with a Dremel and a 426 Fiberglass cut-off wheel?

Tealvince, or others who have completed this mod

This is a great mod. Can you give some step by step. I looked at Tealvince's initial pictures and i'm slightly confused on how the M3 rod is attached to the stepper motor, and I'm confused on how the printbed nut stays in the 5/18 nut spot. Thanks in advance for a step by step directions on how to do this mod.