The below picture shows three parts with different layer heights. They were printed at the same time using Simplify3D. They correspond to the resolution settings in Simplify3D for Gigabot. From left to right: Highest (0.1510mm), Medium (0.2223mm) and Low (0.3175mm).
Full sized picture can be viewed here: https://www.dropbox.com/s/j73eqfr1h9zbmps/IMG_0574.JPG?dl=0
awesome thanks for sharing , this will save me time from doing my own experiments.
Here is a picture of the same part printed at 65 micron layer height. At this resolution the layers are not distinguishable to the naked eye. From my experience most applications I run across give diminishing returns when printing less than a 200 micron layer height. Very few parts will benifit from layer heights of 150 micron or less.
Full size pictures here: https://www.dropbox.com/s/s52sg78u7rhrusb/IMG_0575.JPG?dl=0
Hey Matthew, wondering if you could shed some light as to the value choices for layer height? I realize now that the layer height should be chosen according to full step increments Z axis moves. I am printing with an E3D volcano nozzle so getting to bigger layer heights but need a sound method of calculating accurate layer heights for S3D to use
Hi reptomatic,
Here is the formula I use for selecting layer height:
=B3/(C3*D3)*E3
This means that every full step of the Z axis motor moves the bed 0.0079 mm. Any multiple of this value will give best results.
Hi Matthew, does this formula vary depending on which Gigabot you have? i.e. does it apply to OGB?
Just the sort of information I was looking for. Much appreciated! Ah yes and I should add that I have a kickstarter Gigabot do those values hold true still?
Yup! The formula and values work for all Gigabots and OGB too.
Hey Matthew,
On the re3D home page it shows the minimum layer height to be .004" or .1016mm. What is the relationship between the .0079mm/step and the actual minimum step size? Would you recommend an actual minimum step size of 10x.0079, or .00311"? Or, are you just meaning that the layer height should be an integer multiple of .0079mm. In your opinion, what is the actual minimum layer thickness that the GB or OGB is capable of?
Thanks!
"What is the relationship between the .0079mm/step and the actual minimum step size? "
One full step of the stepper motor will move the bed 0.0079mm. The motor drivers are also 32x microstepping so that one microstep is 0.0002mm in the Z axis.
The best surface finish will be with a layer height that gives a full step of travel from the motor.
Minimum layer height I think will depend upon the plastic or composite fill. Are you printing primarily PLA?
Thanks Matthew. No, I almost never print with PLA. It just doesn't hold any interest for what I would like to accomplish. I would say I print mostly with PETG or ABS then TPU's and co-polymers, then followed by the Nylons.
I think, after going back and reading this thread, that the step size should be some integer multiple of .0079mm then. So instead of a .5mm layer height for a .6mm nozzle for instance, I would choose .4977 instead? Does that also mean that if I design a part from scratch, that the overall design height should also be a multiple of .0079 (assuming a constant step size in z)? For example instead of a 50 mm high part in z, I should design the part to be 49.77mm or 50.56mm high for best accuracy?
In other words, what really happens if I design a 50mm high part and I specify a .5mm layer height? Will I actually get a .5mm layer height out on the part, and the part will measure 50mm tall? If so, then does the micro-stepping make selecting an integer value of .0079 a mute point?
The reason I ask is because I've printed parts in the past with these parameters (in petg with a .6mm nozzle and a constant .5mm layer height) and while they are accurate in x and in y, they are not so accurate in z even though the first layer is also specified as .5mm. The height difference is more than what can be accounted for by the first layer height difference due to the shimmed nozzle offset from the bed.
Overall, I just want to make sure that I'm getting the best z accuracy possible from the machines.
Thanks!
Good questions.
If you slice a 10mm tall object Simplify3D will not make the top layer thicker or thinner than your layer height setting.
If your layer height setting is 0.3175mm Simplify3D will create gcode for 31 layers and the top layer will be @ 9.8425mm not 10mm. S3D will round up or down to the nearest full layer height. In some cases the part will be taller than the 3D model because S3D will round up and add another layer.
Thanks for sharing this conversation.
[Edit - seems I started my response before Matthew but didn't get it posted till after his response...]
https://re3d.zendesk.com/hc/en-us/community/posts/206825566-Calibrating-XYZ
There Matthew said that the "Second thing that affects overall height is the bed home position. The bed needs to be touching the nozzle to give you a true z height."
So if I follow that correctly, if you are chasing exactly 50.00 mm Z height, you would take your 0.4977 x 99 = 49.77 and try to adjust your nozzle height to 0.13 mm.
Then theoretically with a 50 mm high model, Simplify would call for 100 layers, resulting in a 50 mm high print. Right?
But that doesn't take into account the First Layer Height settings in Simplify3D... how does that factor in?
Thoughts?
Shawn, that makes sense to adjust the nozzle offset in the print height calculation. It will be interesting to attempt a print that includes a pre-determined measured offset of nozzle height (actually it would be an additional .23mm offset in the case above). If the filament properties allow such a high offset so that the first layer sticks, then the remaining error should be that of the calibration plus any shrinkage due to cooling.
Overall I was trying to understand how the microstepping comes in to play. Can I get away with a specified .5mm layer height for ease of use and allow microstepping to come in to play to accurately deliver that extra .0023 per layer, or do I set an integer multiple of .0079 and not invoke the microstepping at all.
Maybe not so important for a 50mm high part, but perhaps a bit more so in a 500 mm high part. I'll have to measure some larger parts and see what the overall error is.
Darrel, not to distract from this thread, but I noticed your comment about using .6 mm nozzles. I realize you are not using PLA, but I'd be interested in what parameters you are using, etc. given my current testing as noted here:
https://re3d.zendesk.com/hc/en-us/community/posts/209092543-Parameters-for-8-mm-Nozzles
As a follow up, I decided to design and print a large test cube to check the accuracy of my Gbot using the parameters that I normally use (.6mm Nozzle Diameter, .5mm Layer Height). The cube is 200mm per side, with a 4.5mm wall thickness. Windows in the cube walls allow for measurements to be made from the cube edges, in order to determine the progression of error along the respective axes.
Here's a picture of the test cube:
Here is a summary of the results that I got:
I would be glad to share my factory file and the .xls file in case you'd like to reproduce what I've done. If there is interest, I'll put the files up on Dropbox and post the link here:
https://www.dropbox.com/sh/4ypkn2fiasmrlr7/AAATouznAuqmuGIxZhibj-Tka?dl=0
Overall I'm happy with the results. There is a troubling offset (~.5mm in this case) between the requested z-height and the measured value which is the subject of this thread. Also, I have noticed in past square prints I've done that the x and y values are different even though the calibration steps/mm are the same; the same result is obtained in this print. As the print progresses in z, however, these numbers all become less than 200 mm, which I assume is due to shrinkage during cooling of the material.
I'd love to compare my results with those of others in the re3d community. If you print this cube, post your results here.
Cheers!
Darrel
Hey Darrel,
Thanks for sharing your print. Regarding the microstepping operation and how it applies to your situation. When you ask the Z axis to move 0.5mm the stepper motor moves 15.748 steps. If the stepper motor turns off (check your firmware setting in configuration_adv.h for " #define DEFAULT_STEPPER_DEACTIVE_TIME") during each layer there is a chance the stepper motor will snap back to the 15th step and not be able to hold the 15.748 step position. In your picture there also seems to be little grease on the threaded rods. Be sure to liberally apply lubrication to ensure the friction isn't limiting the Z travel. What tool did you use to perform the measurements?
Hey Matthew,
Fun things to do with printers, eh? Anyway thanks for the comments. I'll check the firmware settings and see if the motors time out. I know that when the machine is finished with a print there's a relatively long delay before the motors deactivate which I believe is set in the finish script, but I didn't know that could happen between layers. Thanks for pointing this out, I'll check into it.
So I went and grabbed the z-axis threaded rods, and got a lot of black stripes on my hand. It's been several months at least since I greased them, but I haven't heard any squeaks when I traverse the bed up after a long print. I snapped a picture of the grease I'm using from Autozone, iirc its a transparent amber color, and doesn't show up like the green or red stuff does. I think I'm ok there, but I'll grease them again jic.
Lastly, I made the cube as big as I could and still be able to measure it with the calipers I have. They are Mitutoyo 500-753-10 8" Vernier calipers...
Let's dial these printers in :)
D-
