Lines in print. Stepper resolution problem?

Electronics, Components etc
chumm
Posts: 176
Joined: Thu Jun 15, 2017 11:24 am

Re: Lines in print. Stepper resolution problem?

Post by chumm » Mon Sep 18, 2017 1:58 am

FranJans wrote:
Sat Sep 16, 2017 7:27 pm
Here is a picture of the normal situation. The problem is friction.
Image
This is an improper way to mount a flexible coupler. There should be space between the motor shaft and spindle, they should never touch. This space allows the coupling to occur with no friction and allows the flex to do its job. With no gap, you are essentially turning a flexible coupler into a solid coupler as there is no room to flex.

Elastomer couplings take care of this gap for you by filling it with the elastomer, which is more secure than relying solely on set screws to float the spindle (though I've never had issues with properly tightened screws). The clamp set screw design as previously noted also helps with centering the shaft and spindle.
Image

kleingeist
Posts: 34
Joined: Wed May 24, 2017 11:51 am

Re: Lines in print. Stepper resolution problem?

Post by kleingeist » Mon Sep 18, 2017 10:08 am

Can´t really see it on the photos but is the leadscrew captured at the top?

FranJans
Posts: 41
Joined: Wed Jul 19, 2017 10:35 am

Re: Lines in print. Stepper resolution problem?

Post by FranJans » Mon Sep 18, 2017 5:02 pm

chumm wrote:
Mon Sep 18, 2017 1:58 am
This is an improper way to mount a flexible coupler. There should be space between the motor shaft and spindle, they should never touch. This space allows the coupling to occur with no friction and allows the flex to do its job. With no gap, you are essentially turning a flexible coupler into a solid coupler as there is no room to flex.
That makes some sense. However, having a gap is not possible with this printer. The weight of the printbed has to be supported by something. Currently the bed is resting on the motor shaft. Modifying this doesn't make any sense.
The connection I created is not comparable with a rigit coupling. The flexible coupling can still correct angle alignment errors which is essential. This is why this solution is working properly. Even if it had to correct some lateral movement, it would compensate it by simply tilting the flexible coupling a tiny bit.
kleingeist wrote:
Mon Sep 18, 2017 10:08 am
Can´t really see it on the photos but is the leadscrew captured at the top?
No it isn't. That would be really bad! ;)

I had some time to think. There are still many possible issues with this printer. However, I think the most likely cause for the lines is still the extruder. The extruder isn't perfect. Modifying/redisigning the extruder will take some time though :(

FranJans
Posts: 41
Joined: Wed Jul 19, 2017 10:35 am

Re: Lines in print. Stepper resolution problem?

Post by FranJans » Wed Sep 27, 2017 10:00 am

Update. I got him!
Someone actually suggested the problem all the way back in the topic but I didn't understand it at the time :oops: This was the post.
ronka wrote:
Sun Aug 06, 2017 2:08 pm
In tom's video about steppermotors he had special z motors without bounce in the spindles.
can you check te z motors to see if you can get movement in that direction?
This was the problem all this time. There is a axial spring inside the Z motor which is messing up the Z position! The spring is actually placed at the bottom of the Z motor which is as bad as it can be. I discovered this by accident. You need to press really hard on the axis to get this spring to move!
I disassembled the motor, replaced the spring washer with a solid washer leaving enough axial play and put the motor back together. I immediately had to turn down the Z axis acceleration by 50% after this modification due to massive banging noises. This proves how much this spring was actually doing all this time.

The difference in print quality is huge! The lines are almost completely disappeared. :D This is awesome :D
There are still little things visible in the print. I think this has something to do with bad retraction settings. I'm still running retraction settings from before all the extruder modifications. I'm currently testing new retraction settings. I think I going to settle for 0.3mm retraction which I find surprisingly low :D

kleingeist
Posts: 34
Joined: Wed May 24, 2017 11:51 am

Re: Lines in print. Stepper resolution problem?

Post by kleingeist » Tue Oct 10, 2017 12:28 am

Glad you improved on this :)

Pictures!

I'm still stuck with fine deviations. Might be a fdm thing :P

FranJans
Posts: 41
Joined: Wed Jul 19, 2017 10:35 am

Re: Lines in print. Stepper resolution problem?

Post by FranJans » Wed Oct 11, 2017 9:11 am

Unfortunately no pictures. The pyramid test print came out rather nice. But I'm afraid I was cheering to early. The benchy that I printed was marginally better than the one I printed before all these modifications. :( I was expecting a big improvement! The results of the benchy test are contradictory to the results of the pyramid test. This is odd.
Geometry plays a big part when I look closely at the benchy. Print inaccuracy always coincide with geometry changes. This brings me back to the start of this topic. XY inaccuracy.
chumm wrote:
Mon Jul 31, 2017 12:49 am
,2mm is pretty standard for 1/16th microstep 1.8 degree motors/drivers. Yes there's some play in the microsteps, but never more than half a microstep, so you're still accurate to about .00625mm, which I don't think you'd be able to see.
This is something I don't believe. I can turn the stepper motor quite a bit without skipping steps when power is applied to the motor. I cannot believe that his is only half a microstep. Common sense tells me that the actual accuracy under load is somewhere between half a step and a full step. This translates to a XY position error of 0.1 to 0.2mm. That would be a problem.
I will be measuring the inaccuracy in real life to validate this theory. This will take some time.

Vertigo
Posts: 22
Joined: Mon Jul 31, 2017 9:00 am

Re: Lines in print. Stepper resolution problem?

Post by Vertigo » Thu Oct 12, 2017 1:42 pm

FranJans wrote:
Wed Oct 11, 2017 9:11 am
This is something I don't believe. I can turn the stepper motor quite a bit without skipping steps when power is applied to the motor. I cannot believe that his is only half a microstep. Common sense tells me that the actual accuracy under load is somewhere between half a step and a full step. This translates to a XY position error of 0.1 to 0.2mm. That would be a problem.
I will be measuring the inaccuracy in real life to validate this theory. This will take some time.
You're quite right. Absolute position accuracy doesnt really improve with microstepping. Good read on the subject:
https://e2e.ti.com/blogs_/b/motordrivec ... -confusion

mbuc
Posts: 7
Joined: Tue Oct 10, 2017 12:08 am

Re: Lines in print. Stepper resolution problem?

Post by mbuc » Tue Oct 17, 2017 2:29 pm

FranJans, do you still have the bed clamping / "leveling" setup installed under the bed? That weighs quite a lot and may not be worth keeping. On my prodesk3d I gutted it and modified it and had some improvements to my prints. I also printed and secured some spacers underneath which prevent the actual bed's posts from having any play.

FranJans
Posts: 41
Joined: Wed Jul 19, 2017 10:35 am

Re: Lines in print. Stepper resolution problem?

Post by FranJans » Tue Oct 17, 2017 7:55 pm

That is a surprise. I didn't think there where that many people who own one of these printers. The printer is absolute garbage. The list of moronic design faults is insanely long. I redesigned almost everything. I hope you didn't buy one but got it for free. That is how I obtained the printer.

I didn't modify the bed clamping devise. I got the one with only 1 motor. I think there was a version with 3 motors. It works rather well, after some fine tuning of the system off course. The system is surprisingly consistent. But indeed kind of unnecessary. I wasn't planning on changing it. This would involve even more modifications. How did gutting this system improve your prints? The system itself feels very solid. It is heavy though.

Do you have your printer working properly? If so, what did you do with the mechanics?
I'm still using the original motors in the original place. This means I'm using very long belts. I did chose steel wire belts for added rigidity. Off course I use actually ball bearings on top and steel rods everywhere.

mbuc
Posts: 7
Joined: Tue Oct 10, 2017 12:08 am

Re: Lines in print. Stepper resolution problem?

Post by mbuc » Wed Oct 18, 2017 11:06 am

I did get it for free, fortunately. When I got it it had been hacked on a bit but nothing had really been changed, just a few things broken. The thing is a sad mixture of expensive parts and poor engineering.

Mine was indeed a single motor. I've attached pics of the bed with the covers off. I'll take some pics later of the way I have it now. The plastic fingers that gripped onto each bed leg were warped, preventing proper contact with the rotational piece (not sure what to call that). I eventually got the bed clamping system to work with the new electronics I put in but it wasn't worth keeping. Removing everything in the bottom half of the bed platform took quite a bit of weight off, smoothed out lines in my print a bit, and is much easier on the Z rails. I don't have a measurement of the current draw before vs after but I might have recorded it somewhere.

I ended up replacing the electronics with an Arduino Mega 2560 w/ a cheap RAMPS 1.4 board and A4988 stepper motor drivers. Setting up, tuning, and calibrating the Marlin firmware took a lot of research, trial and error, calculations, and time, but it was worth it I think.

I'm still using the same mechanical system that came with it, though I did tighten up the vertical belts on both X and Y. The Contitech belts that came on it are a pretty good brand and I am seeing pretty good but not great dimensional accuracy (last measured 100mm X and Y calibration piece averaged 99.69mm hot and 99.32mm after cooling). I'm now trying to determine how repeatable these numbers are - is it a steps-per-mm issue? Mechanical? I'm not sure yet.

Eventually I would like to replace the rods and bearings and carriage. I'd like to move the X and Y motors to the top of the frame and use gearing to increase the number of steps per rotation. I'm currently using the PVA support extruder that came with it as my hotend, but I'm about to order an official E3D V6. The mixer extruder was absolutely destroyed by the time it made its way to me. I'm also about to change how the PCB heater attached to the bed is affixed - the way it came is just awful.
Attachments
20171005_181119.jpg
20171005_181119.jpg (3 MiB) Viewed 512 times
20171005_181053.jpg
20171005_181053.jpg (2.59 MiB) Viewed 512 times

Post Reply

Who is online

Users browsing this forum: No registered users and 2 guests