netfabb material defs: what would you like?

[ddurant 0]

One other thought... In a blog post about slic3r was a comment about using integer numbers for the layer height.

533.3333 steps is a mm in Z... Our favorite layer height is 0.1mm which is 53.33333 steps... Obviously something the stepperFW can't do... 54 steps is far more convenient... But it translates to 0.10125mm layer height... Anyway to integrate this into your software? Sort of a up/down arrow that translates the integer steps into mm?

Hm.. Interesting idea..

Would it be better to have a value for the min possible layer height - 0.0010125mm or something - and have my code adjust to multiples of that or would it be better to have a few values for motor steps and screw pitch and such then figure out the value..

All those 3's lined up make me lean towards the later but I think the former is probably easier, once people come up with the right value for their machine.

[joergen 2]

One other thought... In a blog post about slic3r was a comment about using integer numbers for the layer height.

533.3333 steps is a mm in Z... Our favorite layer height is 0.1mm which is 53.33333 steps... Obviously something the stepperFW can't do... 54 steps is far more convenient... But it translates to 0.10125mm layer height... Anyway to integrate this into your software? Sort of a up/down arrow that translates the integer steps into mm?

Hm.. Interesting idea..

Would it be better to have a value for the min possible layer height - 0.0010125mm or something - and have my code adjust to multiples of that or would it be better to have a few values for motor steps and screw pitch and such then figure out the value..

All those 3's lined up make me lean towards the later but I think the former is probably easier, once people come up with the right value for their machine.

The smallest unit in Z is of course one step. there are 200*8/3 steps per mm on the UM, 533.33333333/mm

unfortunately, the UM marlin FW (or arduino) reports 533.00 back to repg... which means all this theorizing about nicer Z steps is a bit mute... which makes 0.1mm 53 steps (if marlin does indeed not screw more with the rounding)

[joergen 2]

Just wanted to say thank you for your tool... I played a bit around, and found the extrusion was just 10% too thick, so that is pretty close to good. finally a way to more or less use any combo in print profiles. so, thank you again. (now the art picking the right things for the right task come in).

[ddurant 0]

Just wanted to say thank you for your tool... I played a bit around, and found the extrusion was just 10% too thick, so that is pretty close to good. finally a way to more or less use any combo in print profiles. so, thank you again. (now the art picking the right things for the right task come in).

You're welcome - glad you find it useful!

The 10%, was it off by that much across all thread widths and feed rates or did you notice that some were closer than others? Did you just try one material or did you see it off with different spools?

I haven't had much time to print lately.. :(

[ddurant 0]

Finally starting to get a little more time to mess with printer stuff and actually printing tonight!

On my todo list for this stuff is to add a fudge-factor tweak so people can adjust for issues like Joergan's 10% off one, though I'm still unsure on if that should be a global thing or per-thread width or per-feed rate.. Any thoughts there would be appreciated.

Also is maybe a setting to force layer height to be a multiple of the minimum layer height the machine can do which, in theory, maybe, will help keep tiny layers consistant. Was there a consensus on that? Is it worth doing?

And, of course, any other suggesstions would be great..

[cyclone 0]

I think it should be a multiple of min layer height. But that makes it machine specific. Perhaps make it a preset value that can be changed?

[blue-bot 0]

This looks really great Dave.

I've just started developing a new set of profiles for Netfabb at a new layer height. So I will try out your program, hopefully it can save me many, many hours of work.

Would love to see this in Netfabb.

Thank you so much for your time! (& saving mine)

Paul

P.S. wish I had seen this earlier

[ddurant 0]

This looks really great Dave.

I've just started developing a new set of profiles for Netfabb at a new layer height. So I will try out your program, hopefully it can save me many, many hours of work.

Would love to see this in Netfabb.

Thank you so much for your time! (& saving mine)

Paul

P.S. wish I had seen this earlier

Heya Paul,

I've been wondering where you've been hiding!

I'd love to hear any feedback you have - I'd like to get it all settled and working as well as possible then send it off to Alexander.. I'm behind on a couple features but I think it's mostly working. Of course, now that I've finally got a little more time lately, I've just loaned my printer to the Boston Museum of Science for a few weeks, so I can't actually print (unless I want to go into Boston and print in front of a crowd..)

[michiel 0]

I tryed to remake the high quality netfabb file, but with higher speedsettings (pretty much doubled them).

When i load the profile and calculate how much time it will take to print it, it says that it takes longer then the original high quality setting which runs at only half the speed.

Is there something i missed or am i unable to change the speeds this way.

[ddurant 0]

I tryed to remake the high quality netfabb file, but with higher speedsettings (pretty much doubled them).

When i load the profile and calculate how much time it will take to print it, it says that it takes longer then the original high quality setting which runs at only half the speed.

Is there something i missed or am i unable to change the speeds this way.

I've never seen the print time estimate anything close to reality.. Did you try printing a small object to see how well your material defs actually took?

[michiel 0]

I tryed to remake the high quality netfabb file, but with higher speedsettings (pretty much doubled them).

When i load the profile and calculate how much time it will take to print it, it says that it takes longer then the original high quality setting which runs at only half the speed.

Is there something i missed or am i unable to change the speeds this way.

I've never seen the print time estimate anything close to reality.. Did you try printing a small object to see how well your material defs actually took?

I didn't use the profile yet to print something, i just noticed that the estimated build time is about 5-10% higher with the homemade profile compared to the normally slower ultimaker profile(using thesame object and the exact same positioning).

[cyclone 0]

Dave, do you have a Github or some repository for the latest of this tool?

[ddurant 0]

Dave, do you have a Github or some repository for the latest of this tool?

I was planning on publishing the source but I'm a little reluctant to do that now because it doesn't work the way it's supposed to work - I did some heinous stuff in order to make profiles it generates work with existing profiles. It's that whole M92 E14 vs M92 E865 thing. The real end-goal is to get netfabb to pick up the functionality, hopefully picking up non-E14 too.

Do you have a specific question? The math isn't hard - I'm just reluctant to set a precedent with doing things the wrong way.

[cyclone 0]

I guess I have a question, I just don't want it to turn into a whole "how to use netfabb" discussion.

So I've tried a few materials. I'm trying to generate a larger layer height profile for low res fast prints.

As well as wanting to play with lower layer heights to see what are the limits. But when I try to use

one material lets say the .3mm one I generated, I choose a print quality that is big-ish layer tweaked

like "Standard" or "Low" but the material and/or the print quality puts out way too much plastic.

Should I be able to just pick a material and go or what do you need to do to successfully use the materials?

I was wondering if I just had an old version of your tool but I've flipped through the messages here and I have the

last one you posted. I think I must be missing something. Do I need to calibrate the material?

[ddurant 0]

Off the top of my head, I'd guess one of two things..

One is that you might be telling my stuff to use a thread width that's too wide. I haven't really tested the limits of the stock nozzle but I'd guess that if you try to go under 0.35mm or over 0.75mm, you'll be outside the range of the nozzle. That 0.75mm is a guess - I think I've gone as high as about 0.60mm thread width. You (or somebody) will have to try doing material defs that have a larger and larger width until it just refuses to print well.

Two is that the netfabb build style you're using has an override that tells it to use more plastic than the material def says to use. I'm not sure where exactly these settings are but I remember seeing them in the build styles. If you're sure the material def looks correct, you might want to create a basic build style to test it with.

[joergen 2]

Off the top of my head, I'd guess one of two things..

One is that you might be telling my stuff to use a thread width that's too wide. I haven't really tested the limits of the stock nozzle but I'd guess that if you try to go under 0.35mm or over 0.75mm, you'll be outside the range of the nozzle. That 0.75mm is a guess - I think I've gone as high as about 0.60mm thread width. You (or somebody) will have to try doing material defs that have a larger and larger width until it just refuses to print well.

Two is that the netfabb build style you're using has an override that tells it to use more plastic than the material def says to use. I'm not sure where exactly these settings are but I remember seeing them in the build styles. If you're sure the material def looks correct, you might want to create a basic build style to test it with.

I have a gut feeling (but no definite proof yet) that there might be a bug/feature in netfabb with the generated profiles, when the build style is using half/double layers, netfabb might pick the wrong rpm (calculating the half layer as a full layer extrusion), and try to deposit twice as much pla as necessary... I have to go back, and try making a profile that is not using half/double layers and see if the bug still exists

[ddurant 0]

I have a gut feeling (but no definite proof yet) that there might be a bug/feature in netfabb with the generated profiles, when the build style is using half/double layers, netfabb might pick the wrong rpm (calculating the half layer as a full layer extrusion), and try to deposit twice as much pla as necessary... I have to go back, and try making a profile that is not using half/double layers and see if the bug still exists

If/when you have more on this, I'd very much like to hear about it.

Again, the whole idea of my program is to get this stuff nailed down and figured out then hassle netfabb into putting it into their product. A total success for me here would be to have this program just go away because it's all moved into netfabb.

[cyclone 0]

The only override I see in build styles is "width factor." I made a style that had all the width factors set to 100%

and set all layer types to "every layer", still way too much plastic. Not sure what I'm doing wrong.

[joergen 2]

The only override I see in build styles is "width factor." I made a style that had all the width factors set to 100%

and set all layer types to "every layer", still way too much plastic. Not sure what I'm doing wrong.

how much is "too much"? can you print Paul's calibration cylinder (one wall with the thread width that isn't working for you), and measure how much off it is, and report all the settings back to Dave?

[ddurant 0]

Seeing the first 100 or so lines of gcode would be good, too..

[ddurant 0]

New version back on the first post. Just a minor change to the config file for when 1/2- or double-height layers aren't used.

If you use this program but never disable those options, there's no need to get the new version.