printing at 1/32 steps on z axis.

Show off your prints, mods, etc.
firesped
Posts: 699
Joined: Mon Mar 21, 2016 9:23 pm
Location: Tucson, AZ

printing at 1/32 steps on z axis.

Postby firesped » Mon Jan 02, 2017 9:34 pm

printing off a model if 4 different scales. I have reconfigured the Z stepper driver to run at 1/32 steps.
IMG_0737.JPG
RL name: Michael Nolen
printers:
raise3D N2 kickstarter Early Bird
Trinus Deluxe (running smoothieware on Azteeg X5 GT board)
Monoprice Maker Select v2

eagle1xray
Posts: 10
Joined: Sun Dec 18, 2016 8:48 pm

Re: printing at 1/32 steps on z axis.

Postby eagle1xray » Tue Jan 03, 2017 12:24 pm

INTERESTING what were the results?!

lami3d
Posts: 220
Joined: Sat Oct 29, 2016 7:40 am
Location: Marseille - France

Re: printing at 1/32 steps on z axis.

Postby lami3d » Tue Jan 03, 2017 12:47 pm

Yes interesting analysis !
Let us know what are the conclusions

Phife
Posts: 59
Joined: Thu May 12, 2016 3:47 pm

Re: printing at 1/32 steps on z axis.

Postby Phife » Tue Jan 03, 2017 2:41 pm

I just noticed you dont appear to have any cooling fans for your hotends? Dont you get issues with filament rising up the nozzle throat tube and clogging as they get hot?

firesped
Posts: 699
Joined: Mon Mar 21, 2016 9:23 pm
Location: Tucson, AZ

Re: printing at 1/32 steps on z axis.

Postby firesped » Tue Jan 03, 2017 3:50 pm

There is a cooling fan. though, i do have the sides blocked off. hahahahaha. my cold end cooling fans are those white blobs on the sides. it is printed in ePC. I messed up that print and i needed to switch to another filament. but I have had issues with the filament before so I printed a temperature tower.

IMG_0738.JPG


a close up of it with better lighting. the print gets better below 204 C. below 184 also seems to have something else going on. so I think I will got for 195 on this filament

IMG_0739.JPG
RL name: Michael Nolen
printers:
raise3D N2 kickstarter Early Bird
Trinus Deluxe (running smoothieware on Azteeg X5 GT board)
Monoprice Maker Select v2

Phife
Posts: 59
Joined: Thu May 12, 2016 3:47 pm

Re: printing at 1/32 steps on z axis.

Postby Phife » Tue Jan 03, 2017 4:46 pm

Oh I see them now.. lol, I was like, holy cow, how is he running with no cold end fans

Jetguy
Posts: 1659
Joined: Tue Mar 22, 2016 1:40 am

Re: printing at 1/32 steps on z axis.

Postby Jetguy » Tue Jan 03, 2017 5:28 pm

Greater microsteps does not translate to higher resolution. Microsteps are stepped percentages of coil current betwen real actual pole alignments of the rotor to the field of the stator. This interpolation has zero guarantee of linearity, in fact, it's more or less proven to be non-linear and even within a known factor, there are individual manufacturing defects (rotor not centered, minor alignment of the stamped rotor and field parts, variations in the permanent magnet, non centered mounting of the rotor via the end bearings within the field, on and on) that lead to variations between any 2 microsteps. One of the biggest reasons for microstepping is noise reduction and has nothing to do with resolution.

http://hackaday.com/2016/08/29/how-accu ... ng-really/

Some terms and points of discussion:
Resolution= the smallest increment one can resolve or address. This is NOT a good indication of quality. It's just a number. Imagine hand drawing a 12 inch ruler on paper, then have 20 people all make one on their own. You take all those end to end and that's your position indicators for target. No 2 inch positions are alike and all have some error, Cumulative error might not be bad (people got 0 and 12 inch points correct), but in the middle of each ruler is variations from one to the next.
Repeatability= the concept of moving away from a known position, moving a random distance away, then commanding the original starting position and measuring the new position VS the original point.
Accuracy= the concept that every single resolvable position is equal and repeatable to a known measured standard. The idea is that no matter what 2 random points you select as target positions, they are both accurate to a known external reference standard and they are repeatable so that every time commanded, they are the same position "exactly". There is no periodic error or difference in error between a real step and any microstep. The problem is, this is not reality.

Steppers have the following errors:
#1 The error is load dependent. The magnetic alignment is like a spring and the load deflects the actual position from the target position. Higher current and greater torque (often from a longer rotor with more magnets and poles results in greater rating AKA tall stack motor so called because the rotor is stacks of rare earth magnets and stamped pole plates). The stronger the "spring", the less deflection with a given load.
#2 The only rated accuracy is FULL step. Anything else AKA microstepping is KNOWN error. You are puposely balancing a rotor somewhere between 2 known poles.
#3 Manufacturing tolerances. The end housing hold the bearings and thus the rotor and it's supposed to be perfectly centered in the field formed by the stator and coils. Defects in assembly, alignment, manufacturing, and even the very permanent magnets used could make variations such that throughout the rotation of the rotor, it's not that every step is the same mechanical distance.

But then you have to remember, the motor is only one half of the system, the driver is the other half.
The driver has to measure and control the current in the coil dynamically at speed as the motor is moving. Not all drivers are equal and if you just look at numbers like microstepping count, you can easily be fooled.
Shining example is DRV8825s, probably one of the biggest disasters I can remember since I started the hobby circa 2009.
DRV8825s have a fixed time in the chopping circuit to facilitate measuring current. As such, this means that with a 24V supply, there is point where the driver cannot properly control the coil current in the lower percentages of the microstep table.
http://cabristor.blogspot.com/2015/02/d ... steps.html

What I'm trying to get at is:
#1 if you want accuracy, 0.9 degree AKA 400 steps/rev VS standard 1.8 degree AKA 200 steps rev motors is better than increasing microstepping.
#2 In either case, increasing resolution increases work on the 8bit AVR controller that is the motion control board.
#3 Avoid DRV8825s like the plague. A4988, SD6128s, and the new A5984, as well as the TMC2100 generally are better. Toshiba chips are supposedly OK, but not ideal as "plug in" steppers into the board.
I cannot stress it enough, the stock A4988 used on Z and the extruders in the Raise 3D series is a perfectly acceptable and reasonable driver at 1/16th stepping.

Now granted, Z is the slowest moving axis in the printer. We never use the speed and it does not affect printing (but could get annoying to move full length at the start of a print if the bed is lowered). I did install a TMC2100 in stealthchop mode as my Z axis in my N2. It's insanely silent stock with an A4988, but completely silent with a TMC2100, but I run it at 1/16th stepping.

All I'm trying to convey is some research articles that show that moving to greater than 1/16th stepping (which is stock on the entire N series) and especially for Z axis, has very limited if any gains. Some mods could actually lead to increasing microstepping error and thus potentially show up as Z ribbing- example, using a DRV8825 to get 1/32 stepping since A4988 maxes out at 1/16th.
Again, just sharing this because DRV8825s are extremely common and cheap. Someone might get the bright idea to use one.

Also, we should note that Z is by far already the highest resolution in the printer stock.
X and Y axis is 80 steps per mm or conversely, each logical and thus addressable position is 1/80th of a mm
1/16th microstepping times 200 full step motor=3200 steps/rev divided by (20T 2mm pith pulley@ 40mm/rev)=80 steps per mm

If we back that out and say microstepping is assumed to have some error, then actual hard mechanical resolution is 200 steps/rev and 20T GT2 2mm pitch pulleys. 200/40= 5 steps per mm or each full step is 1/5th mm.

Z axis is a ballscrew like this http://www.power-transmissions.com/Prec ... et/sfu.htm
Basically, the same specs with 16mm diameter and 4mm lead per rev.
Well, let's just compare to X and since they have a 20T pulley and Gt2 belt, one rev is 40mm VS one rev of the Z leadscrew is 4mm or 10 times the mechanical resolution before the motor even comes into play.
Bascially, Even if you had zero microstepping (not recommended for any axis), Z is still 200 full steps/rev and 4mm/rev or 200/4=50 steps per mm. Half stepping gets you 100 steps per mm or the smallest resolution is 1/100th of a mm. Stock at 1/16th stepping is 1/800th of a mm, which in all reality is 1, yes just about 1 micron. I find it hard to believe the very threads of the ballscrew are even 1 micron accurate across the entire screw length.

Jetguy
Posts: 1659
Joined: Tue Mar 22, 2016 1:40 am

Re: printing at 1/32 steps on z axis.

Postby Jetguy » Tue Jan 03, 2017 5:37 pm

Again, if not clear, I'm not saying there is something "wrong" with experimenting nor specifically changing to a 1/32 microstepping on Z or any other axis. What I am trying to say is the difference you will see is more likely the very driver itself and any known non-linear current control and microstepping limitation of the driver brand and chip than any difference between stock 1/16th and 1/32. Further, the math shows no gain on realistic resolution to the Z axis no matter what you do. It's already more based on the very screw itself than anything going on with the driver or motor.

Again, nothing wrong with using different drivers. You might have your reasons why to try some other ones.
But a solid understanding of the "system" will help you see where the value and gains are and that in general, you are hard pressed to get more accurate than the stock A4988 without replacing the motor with a 0.9 degree unit. Even then, the very screws themselves probably have more error than the motor + driver.

firesped
Posts: 699
Joined: Mon Mar 21, 2016 9:23 pm
Location: Tucson, AZ

Re: printing at 1/32 steps on z axis.

Postby firesped » Tue Jan 03, 2017 10:23 pm

I have a SD6128 installed on my Z axis, actually I have them install in all five slots.
RL name: Michael Nolen
printers:
raise3D N2 kickstarter Early Bird
Trinus Deluxe (running smoothieware on Azteeg X5 GT board)
Monoprice Maker Select v2


Return to “User's Gallery”

Who is online

Users browsing this forum: No registered users and 4 guests