Kossel custom build problems with Z axis

So I'm not new to 3D printing or custom building, but I am new to the world of deltas. About 18 months ago I started purchasing parts to build a Kossel and finally got it printing just after Christmas, however I'm less than happy with its performance especially with money I've spent on it.

The frame is built using Robotdigg corners and extrusions, I also got the rails and carriages from Robotdigg.
The electronics are running at 24v on a Duet 0.6 with the smart effector using rod from Haydn.
Stepper motors a 0.9° JK42HM40-1684
E3D Titan extruder in a bowden setup, I am planning on a Nimble extruder but I don't want to purchase that until I've all the issues ironed out.

So now the main issue:

The only way I can describe it is Z wobble, which I believe is the incorrect term on a delta, the print quality is improving while setting up the profile but something has to be wrong. The image attached is of 18mm x 18mm x 100mm and as you can see isn't straight.

I've run out of things to try so I'm hoping I can be pointed in the right direction.

Thanks
Paul

Maybe a pulley is not centered on its shaft.

Just checked them can't see anything wrong them.

It might be over extrusion. Some of the excess filament is pushed out at the side. Try 0.9 extrusion multiplier and are you printing with fan?

I had this issue, and it was one of the magball that was not properly fastened (the effector was loose). you need to check every bolt... And what is the length of your bowden tube ? may be it's too long ?

Quote
Vigilant
It might be over extrusion. Some of the excess filament is pushed out at the side. Try 0.9 extrusion multiplier and are you printing with fan?

The line were slight over extrusion and yes I'm running a fan but I still have the wobble effect.

Quote
Archiclem
I had this issue, and it was one of the magball that was not properly fastened (the effector was loose). you need to check every bolt... And what is the length of your bowden tube ? may be it's too long ?

I've checked the magballs and hotend to make sure they are secure, found nothing there. My bowden tube is 600mm long.

A similar problem occurred before that was finally traced to the drive belts.
Skewed prints

HTH

Haydn apologized about the quality of his ball studs. They seem to be uneven. IDK exactly what he meant with that, but maybe it's worth the time to measure and eyeball them closely.
A ball that isn't round is the same as a rod changing length, depending on the mag-sleeves position. That would cause effector tilt in a reciprocating pattern.

Quote
ayudtee
A similar problem occurred before that was finally traced to the drive belts.
Skewed prints

HTH

Made an interesting read, may have to strip out the belts and give them a good looking over.

Edited 1 time(s). Last edit at 02/09/2018 04:51AM by Hozza.

Quote
o_lampe
Haydn apologized about the quality of his ball studs. They seem to be uneven. IDK exactly what he meant with that, but maybe it's worth the time to measure and eyeball them closely.
A ball that isn't round is the same as a rod changing length, depending on the mag-sleeves position. That would cause effector tilt in a reciprocating pattern.

Well got to be worth a look while it's in bits lol

Is this all Haydn ball studs, or only older ones? I was not aware of such an issue

Quote
Origamib
Is this all Haydn ball studs, or only older ones? I was not aware of such an issue

I bought mine a month ago and in his build instructions he suggested to use 3/8 ball bearings to measure final rod length.
I then heard from DjDemonD, that's because the ball studs weren't even. Does it mean they are potato shaped or are they not the same diameter? IDK

Hi guys please read this thread:

[www.duet3d.com]

The issue was not all ball studs were exactly the same size.

I do need to correct a couple of things though...
1) we could not accurately determine if they were spherical, if you know how to do this let me know
2) it seemed that whilst there was a spread of sizes of ball stud, when I measured them (as best I could) and my rods and then reassembled my kossel XL printer so that each rod+2 balls set was as close in length to each other, and each pair of arms was comprised of the two best-matching pairs of rods/balls, the effect on improving my calibration deviation was not measurable. Over the central 100mm of the bed it managed 0.004mm and then 0.004mm after swapping everything around.

So if this problem is bothering you (and you can test for yourself), my advice would be to buy a pack of high grade (grade 50) 3/8th bearings and bond them to your carriage and effector, instead of using the ball studs. Don't buy stainless its not magnetic enough.

The variance in the ball studs we are talking about in the thread above is so small, if you printed your effector rather than having one machined, or made from a precision drilled PCB, you will be attaching the "dodgy" ball studs to a much "dodgier" effector! Even with the smart effector the holes are drilled slightly larger than 3mm to assist assembly, how much effort are we going to to ensure the ball stud is centralised in those holes before tightening?

Edited 3 time(s). Last edit at 02/10/2018 06:12AM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

I took a handful of Delrin sleeves and some ball studs and tested continuity between ball and backside of the magnet. ( Yes, I finally got my 3 sets of mag-arm kits )
The result showed minor differences between them. Some had "solid" continuity when the ball stud and sleeve were straigth aligned, but no conductivity when angled.
Others had so_so continuity ( the multimeter made a screeching noise ) at several angles. To me the difference is academic.

BTW: My Smart Effectors holes are 2.9mm, which is a tight fit.

Hmm mine were not that tight at all.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions

Well I've still been unable to find the root cause of my issue, so bloody frustrating or infuriating not sure which yet

Is it worth the effort of modding the Titan and make it a flying extruder?

I'm also going to be changing the motor pulley's to 16 tooth from 20, I believe I've read something about step calculations being more accurate on 0.9° step, plus I've noticed my idlers are 16 tooth.

Flying extruder is worth it. Reducing bowden to 20cm basically returns performance to direct extruder levels. But you have to print a little slower as the inertia of the flying extruder will show as added ringing to some degree.

Idler and drive pulleys should be the same diameter. Twist your belt that goes over the idler so it runs smooth side to the idler unless the idler is toothed. Using 16t improves resolution but its quite a small effect.

Edited 1 time(s). Last edit at 02/10/2018 12:08PM by DjDemonD.

---

Simon Khoury

Co-founder of [www.precisionpiezo.co.uk] Accurate, repeatable, versatile Z-Probes
Published:Inventions