Im having some problems with .1mm printing (Page 2 of 3)

Got this reply.

So I sent them a link to this thread.

This is what I've been arguing for a while now and did some tests to actually show that the inches per thread on my new rod are inconsistent.  The tpi count is still the same, so the overall dimensions are correct, but not each layer.  There are squished and stretched layers matching the threaded rod tpi.  It is possible that people are getting banding from different places, but unless the bands are helical, it seems there must be an issue with the thread tolerances in the nuts and or threaded rods.  Someone, I can't remember who, did a cool test with a laser pointer and that could certainly be done, but the dial indicator is the best way to tell the step sizes.  I stepped at .1 mm, and .3 mm in different areas of the bed to simulate a print for my testing.

I am anxiously waiting for what SD decides to do and am assuming this will be addressed before the SD3 comes out.  They have been actively communicating with us and actively searching for an answer as well.  It is a little bit of a wait and see game right now.

Did the banding move helically around the print? Meaning, even if it was horizonta,l were the bands moving around the print. I have a bent rod that actually provides band free prints. They aren't perfect, but much better than the new rod that is perfectly straight.

I was thinking of a way of illustrating what I strongly believe is happening and maybe a simple way of fixing it.

I hope my crude renderings come through.  When there is no lateral force on the rod the nut sits low and centered on the rod (image is turned 90 degrees):

+-----------+
|           |
|\    /\    /
| \  /  \  /|    ---> Top of SD
|/\\/  /\\/  /\
/| \  /| \  /| \
 |  \/ |  \/ |  \
 |  |  |  |  |  |
 |  |  |  |  |  |  rod
 |  |  |  |  |  |
 |  /\ |  /\ |  /
\| /  \| /  \| /
|\//\  \//\ |\/
| /  \  /  \|
|/    \/    \
|           |  nut (centered on rod)
+-----------+

Any sufficient lateral force caused by a bent or offset rod causes the nut to rise upward, which leads to irregular layer heights and horizontal bands at the same frequency as the rod threads:

 +-----------+
 | nut       |
 |\    /\   /|
 | \  /  \ / |    ---> Top of SD
 /\ \/ /\\/  /\
/| \  /| \  /| \
 |  \/ |  \/ |  \
 |  |  |  |  |  |
 |  |  |  |  |  |  rod | (lateral force)
 |  |  |  |  |  |      |
 |  /\ |  /\ |  /      V
\| //\\| //\\| /
 |//  \\//  \\/
 |/    \/    |
 |           |
 |           |  nut shifts --> along rod axis
 +-----------+

A spring-loaded nut preload will help keep the nut seated down in the base, but won't work if the lateral force is too great (rod is too bent).

Some have suggested using a rod with a flexible joint where it connects to the motor, and this sounds like it would solve the problem when used in combination with a spring preload of sufficient strength.

As a cheaper fix, however, I'd try loosening the motor mount so that the entire motor and rod assembly can pivot a little bit.  I don't have my SD in front of me, but if there are say, 4 screws holding it on, I'd try taking off two of them and leaving the other two snug with some rubber washers in between the motor and frame.  It might take a little testing to find how tight the connection should be, and perhaps some superglue or thread lock to keep the screws in the correct place afterwards, but worth a try if possible, I think.

I'm wondering if the reason you can't duplicate it in the shop is that you haven't shipped those printers.  Remember some of the early packing methods had broken pieces and z assemblies tipping etc.  That could easily bend a z-rod beyond what you are finding in your raw component bins.  Although it does seem to be more common lately which would be after the improved packaging.  Just a thought.

We never said we couldn't delicate it - just that a new rod wont necessarily fix the issue. We're afraid it's a little more complex than just putting a new one on.

I think it comes down to the way they build them.  I am sure they don't stock a lot of the parts and order them as needed and probably not from the same supplier.  I hope they at least inspect them for correct specifications and quality, rejecting parts that don't meet standards.

Very nice! If this works we will put it on Solidoodle Github, or package it with out software.

While we wont give away most of our research in to the topic, banding can be caused by interactions between the rod itself, the rods threads, the rods internal boring, the Z-motor, the rods connection to the Z-motor, and the Z-motors connection to the Solidoodle. The problem doesn't manifest from the same circumstances Solidoodle to Solidoodle. We've tried a few things with good results, but the issue (as always) is cost. There are definite tried and true solutions to this issue, but they are typically expensive and would take us away from our mission to provide the printer at a reasonable figure for consumers.

That said, given sufficient demand we could consider putting some more expensive solutions in our store as "upgrade" options. We've shied away from this since we do not want to give the impression that you need to get an upgrade to make the Solidoodle work correctly in the first place. What we'd really like is a solution that makes the $499 base model work without banding each and every time - we don't want to have to increase the price. That said, who knows - if there is sufficient demand we'd be open to offering wobble reducing solutions in the store.

Regards,
John

High quality lead screws do help, as well as integrated motor / leadscrew combos (expensive!)

We might offer an advanced lead screw in the store at some point.

Sadly no, we are working on it though.

Any news yet?