netfabb material defs: what would you like?

[ddurant 0]

edit: latest file should always be here, in the first post of this thread..

I've been working on software to generate netfabb material definitions. You plug in some numbers and it spits out a .reprapcalibration file, which can be imported into netfabb as a new material definition.

The program isn't quite ready for prime time yet but the results, I think, are getting close.

So.. I could use some help testing this stuff.. If you use netfabb and want to try some new material definitions, post here with the specifics of what you'd like and I'll generate the profile & upload it here.

Here's a screenshot of what I'm typing numbers into:

netfabb material generator by ddurant123, on Flickr

If you'd like me to generate a material for you, please post:

- base layer height

- filament diameter (measured accurately!!)

- temperature

- travel feed rate

- min and max feed rates

- min and max thread widths

For the min/max ranges of feed rates and thread widths, the code will generate 5 values for each, even distributed between the min and max values. If you'd rather specify individual values (like "feed rates of 20, 30, 40, 100 and 150" instead of "feed rates between 20 and 80"), just say so.

Of course, if you try one (or more!) of these out, feedback on how it worked would be much appreciated..

[cyclone 0]

Does this do the calculations based on the extruder hole diameter and is that hardwired to .4mm or can it be adjusted like say changing the min thread thickenss?

With people reporting more clogged nozzles I'm thinking I may try to cut some parts to allow for a makergear hot end, and in that case, would probably like to set it up with two heads of differing sizes. Probably have to just switch the wiring back and forth to use one or the other until more support for multiple heads comes in, but I'd love for both sizes to work in Netfabb.

-b

[ddurant 0]

Does this do the calculations based on the extruder hole diameter and is that hardwired to .4mm or can it be adjusted like say changing the min thread thickenss?

Different nozzle sizes have different usable min/max thread widths, which you (or, rather, "I" for now) can just feed to the program but nozzle size doesn't really have as much impact on the extruder speeds as you might think so no, the code doesn't factor that in.

It's really all about volume - matching the volume of threads in the print with the correct amount of filament feedstock..

 

With people reporting more clogged nozzles I'm thinking I may try to cut some parts to allow for a makergear hot end, and in that case, would probably like to set it up with two heads of differing sizes. Probably have to just switch the wiring back and forth to use one or the other until more support for multiple heads comes in, but I'd love for both sizes to work in Netfabb.

I'm suspicious about the clogged nozzle stuff - I don't think anybody here has ever actually clogged a nozzle up unless they've run things waaaay too hot and baked all the bendy out of the filament. Most of the problems seem related to too much heat traveling up the hot end and forming a plug at the end of the bowden cable. To actually physically clog a nozzle, you need particules far bigger than dust or hair..

[jelleatprotospace 0]

A thick nozzle is good for speedy prints, which is good if you want to make particularly high prints...

We have a ultimaker++ here with a 0.8 mm nozzle which would be very nice to drive with netfabb gcode (skeinforge takes it sweet time with high models..), but we also have two 'regular' ultimakers that are driven mostly with skeinforge 35 because that is perceived as easiest. So we certainly could do with some material defs:

Ultimaker++

- base layer height 0.25 - 0.5

- filament diameter (measured accurately!!) black 2.85, white 2.88, pink 2.93

- temperature: honestly, I have no idea what the 'best' temperature is. Usually I drive it at 240, so I guess a range from 210-250 should do?

- travel feed rate: fast! (again, no idea about the actual values, I really dislike the silly percentage thing instead of mm/min)

- min and max feed rates: see above

- min and max thread widths: again, no idea, I think it should be a little wider to make single skinned objects a bit stronger.

plain Ultimaker:

Basically the same thing, but these ones should print at a lower height too.

Is your program shareable? I really dislike burdening you with a boring data-entry job.

[ddurant 0]

Is your program shareable? I really dislike burdening you with a boring data-entry job.

It will be!

PM me or post your email and I'll send you a copy..

edit: the reason it's not public now is that it's only been run on one PC and generated stuff for one machine - I was hoping to get a few definitions test first and maybe shake out some easy bugs before releasing it.

I also had a direction change yesterday from having it generate machine definitions which include start/end gcode and some other bits to just generating material definitions and there's a bunch of cleanup needed because of that. It still works fine (I think) but there are bits to the UI that don't make sense anymore..

[rencio7 0]

Ello..

I wish to try this program you've made ..

I use abs and a 0.35mm nozzle with 2.89mm filament though...what do you advise I change?

Oh yeh I also use a mendel with a wade extruder.

I've always wondered what esteps per mm I should have for netfabb 4.8..I'm using 14 at the mo.

I know it all depends on your setup.

I've just used trial and error to create my profiles..

Oh my email is rencio7 at gmail.com.

I really need to get a good profile it's really frustrating.

I'll post some prints if you want..I want to reduce layer size as much as possible.

cheers

Rencio out.

[ddurant 0]

Here's the program. Yes, it's windows-only (though likely works on mono) and there's no source yet.. If those are a problem, the offer to be a data entry monkey for people still stands - only takes a minute to type stuff in and get a profile.

The original version of this (that nobody else saw) generated a whole netfabb machine configuration file and setup the start/end gcode for you and did some other fancy things but there were some problems with doing that so I changed stuff around to just generate a material configuration. I did leave the start/end gcode stuff in there 'cause I think it's cool but you'll have to copy & paste the gcode it generates into netfabb manually, if you want to use it.

ONE BIG THING TO NOTE is that this isn't an end product - it's just a step towards where I think we want to be, which is having all this integrated into netfabb with algorithms that can generate the correct settings for a wide variety of scenarios. So, that said, please post here if you use this and what you think of it - Did it just totally suck? Did it work but always produce a bit too much/too little filament? Did narrow threads work better or worse than fat threads? These are just some off the top of my head - I'm looking for any data you can provide.

To install it, just create a directory for the program then unzip the exe to that dir. There's no separate installer program. There's nothing stored in the registry. There's no internet stuff needed. It's just a simple, stand-alone exe. To uninstall, just delete the directory that you created for it. It does create a .xml file but that just contains all the values you've typed in, so the program can continue with your last settings the next time you run it..

To run it, just enter whatever values you want then click the Save button. This will open a standard Window save-file dialog - type in the name of the config file you want and it will create the material config (".reprapcalibration") file.

To import into netfabb:

- click on Ultimaker in the context area, where it says Parts, Slices and Ultimaker

- on the lower right of the screen, click Manage Materials

- right-click on the window that pops up and do Import Material

- browse to the material config you made and click Open

- click the Close Window button

- back on the lower right of the screen, set Extruder 1 to be the material you just imported

Note that the "profile name" setting in my program doesn't do much - this used to set the name of the material in the machine config but you (apparently) can't set the material name when just generating a material config. Instead, netfabb uses the file name as the material name. You can always change the name of the material in netfabb Manage Materials, though..

Not sure what else to say.. I think the program's pretty self-explanitory to run but if something's not clear, feel free to ask..

Questions? Comments? Bugs? Bring 'em!

edit: old version deleted - see first post in thread for the latest..

[ddurant 0]

Oh yeh I also use a mendel with a wade extruder.

I've always wondered what esteps per mm I should have for netfabb 4.8..I'm using 14 at the mo.

That's going to be tricky..

The original version of the program was closer to real volumetric 5D, doing the math so that netfabb ended up generating G1 E values with the amount of raw filament to pull in. Doing that screwed up all the built-in material configs and would screw up any custom defs people made outside of my program. This is one of the main reasons I switch to doing just material configs.

The program really wants to be run on a machine with an extruder that behaves the same way as an ultimaker with esteps set to 14.. If you can get the esteps on your machine to print fairly happy with the built-in materials/build styles on the ultimaker version of netfabb, my program should work pretty well on it too..

Does that help?

[rencio7 0]

This program is excellent as I didn't know the ratios between settings so it should help loads.

I'll have a tweak, a little print, a think and get back to you

thanks

Rencio

[ddurant 0]

Now that I think about it more, if you can figure out the 'real' esteps value for your rig, the correct M92 E value for netfabb (until it moves up to real volumetric 5d) is probably 14 * esteps / 865.888.

So, if you come out to esteps = 700, you'd add "M92 E11.3179" to your startup code and the built-in netfabb profiles should work for you.

[ddurant 0]

Er.. Don't use rafts with material configs that this program generates yet.. They won't work well (at all?) but it's on the top of my list to fix. Probably tomorrow..

[rencio7 0]

sooo lemme try to understand :S its early here so I'll try to concentrate! :geek:

Actually my esteps is 40 not 14..I forgot!

I think that's close to volumetric on my setup..as I worked it out a while back for skeinforge..

So for example I wanted to transfer my settings into your formula I'd do this...?

(14 * 40) / 865.88800 = 0.646734913

40 being my esteps..

0.646734913 being my new esteps - I know its low but my 40 with default rpm rates the extruder speed was waaay to high so this number sort of makes sense...my 1st extruder (profile) rpm being 0.06 xD before (in netfabb) after trial and error..

I know M92 doesn't work for my gen3 (FiveD) I've got ramps 1.4 ready to go tomorrow if I have time or I will try teacup on the gen3..that supports M92..

I don't mind reuploading firmware each time to change esteps..is that what M92 is for just changing esteps and other steps before the print?

I found this post - by coincidence I am the thread starter! I should of read this earlier..

http://www.netfabb.com/forum.php?aid=825842962&pid=441

Will add more later on if I think of anything..

Rencio out.

[ddurant 0]

Nooooo..

Have you figured out the "real" correct value for esteps? See

http://reprap.org/wiki/Volumetric_Dimension_settings

.

That's the value you want to divide by 865.888 then multiply by 14. Should probably come out somewhere between 10 and 20..

[rencio7 0]

hmm..

I'm pretty sure my volumetric esteps was around 40 maybe I am mistaken...and I've changed back since I have switched to netfabb..

I used the sf40+ volutmetric calculator ...

http://calculator.josefprusa.cz/#ExtrusionSF40

... :oops:

Ahh is this 'real' esteps a different beast to this or are you using the formula based on 16th stepping?

Umm I use half-stepping gen3 drivers.. and a modded x and y

 

#define X_STEPS_PER_MM   21.21593291404612//10.124171898769282480213688728022#define X_STEPS_PER_INCH (X_STEPS_PER_MM*INCHES_TO_MM) // *RO#define INVERT_X_DIR 0#define Y_STEPS_PER_MM   19.765323876051237#define Y_STEPS_PER_INCH (Y_STEPS_PER_MM*INCHES_TO_MM) // *RO#define INVERT_Y_DIR 0#define Z_STEPS_PER_MM   868.1096681096683//859.4285714285716//782.08//681.0008084074373 //842.398*0.9115770282588879#define Z_STEPS_PER_INCH (Z_STEPS_PER_MM*INCHES_TO_MM) // *RO#define INVERT_Z_DIR 1                                            #define E0_STEPS_PER_MM   40

 

[ddurant 0]

Oh.. Er..

That still sounds pretty low to me but try it - worst that'll happen (I think) is that you'll get a really anemic print.

 

its early here

And late here! I'll check back tomorrow..

[rencio7 0]

Ok...I've got marlin setup with ramps 1.4...so smooth and quieter with 1/8 step!

I'm not using 1/16th because its overheating - probs due to the belt upgrade/mod..

I've tried repetier firmware too but I'm having a few teething problems..

Ok back to netfabb.

My extruder is using 1/16th stepping.

My volumetric was around 45 (before) so I multiplied that by 8 to give 360

(14 * 360) / 865.88800 = 5.82061421

 

#define DEFAULT_AXIS_STEPS_PER_UNIT   {84.86,79.06,3472.44,5.82061421}

 

So then I tried the default and generated configs...

(semi)Sucessful print of sophos head...needs a tweak though.

I haven't succesfully done a 0.15mm layer print before straight off so it's lookin good.

One thing which I'm noticing is that the extruder speed isn't changing though noticeably..or at all its like a constant rate and the feed rate is compensating.

On ultra (on the fill) it is very slow (feed rate) and on fine (round the edges) it is very fast (feed rate) .

Any tips on whether my max feedrate is limiting me ? x y z e

 

#define DEFAULT_MAX_FEEDRATE          {1500, 1500, 5, 70} 

 

Oh and some feedback on your netfabb program ddurant.

[*]The start and end code don't seem to work - you probably realise this though ..

Oh and a question...what does time before filling hatches do?

Have a happy new year reprappers <3

Edit - here are the pics of the rough print..

http://horobox.co.uk/u/Rencio7_1325440985.jpg

http://horobox.co.uk/u/Rencio7_1325409732.jpg

http://horobox.co.uk/u/Rencio7_1325382671.jpg

[ddurant 0]

Good progress!

Did you read this bit?

 

I did leave the start/end gcode stuff in there 'cause I think it's cool but you'll have to copy & paste the gcode it generates into netfabb manually, if you want to use it.

 

Have a happy new year reprappers <3

And to you!

[spunkhart 0]

One question before i try it out.. is the "Ultimaker standard bolt" in the Start Code section the new one?

http://wiki.ultimaker.com/File:NewBolt.jpg

That is what gave me the most problems.. tinkering with the Netfabb Settings...

some results..

http://wirsch.com/Ultiprints.JPG

(2 Prints stopped right in the middle when using the Netfabb direct print.. can't wait for the new version)

[ddurant 0]

One question before i try it out.. is the "Ultimaker standard bolt" in the Start Code section the new one?

http://wiki.ultimaker.com/File:NewBolt.jpg

 

It is indeed.. If there are people still using the old bolt, I can try to work out the math for that..

[spunkhart 0]

I gave it a try but when i import the settings, netfabb says in an error window: "0,05" is an invalid float".

So i opened the file with wordpad, and replaced all ',' with '.' and netfabb imported them..

but when i calculate a toolpath, then there are only points.. something is wrong there..

[ddurant 0]

"0,05" is an invalid float".

Do you normally use , for a decimal place or . ?

[ddurant 0]

I gave it a try but when i import the settings, netfabb says in an error window: "0,05" is an invalid float".

So i opened the file with wordpad, and replaced all ',' with '.' and netfabb imported them..

but when i calculate a toolpath, then there are only points.. something is wrong there..

Ok, I think I've fixed this..

Most recent version attached, which also builds the right info for traditional rafts and includes a couple other small bits. Please let me know if this works better!

Also, if you want to edit the netfabbconfig file, feel free but be very careful to not change the formatting and use a plain text editor like notepad instead of a rich text editor like wordpad.

edit: old version deleted - see first post in thread for the latest..

[rencio7 0]

I'm printing at 0.1mm layers now with half layer support too! I might be able to go even lower

I'm having no extruder jams too because of the very controlled feed rates so I can leave it for like 4 + hours without fear of fails, jams etc.

I've never managed this feat before on my mendel .

You are a great contributor to the reprap/ultimaker community ddurant .

I see from the netfabb forum they have offered to add your configurator to the actual netfabb program! Good luck with that

I've realised I know quite a little about the math behind the reprap (even after like 3+ years of doing it) :oops: and you've made me want to learn more to unlock more potential in the machine and enjoy it more with less frustration because of that.

My prints are getting better and better each time with tweaking of course- a little over extrusion though ..

It really shows the lower tolerances my printed parts sells custom mendel has compared to the cnc made ultimaker when you do macro shots! I need to replace the x axis smooth rods as they are very worn after 2+ years..

Still I'm happy

Oh I'm printing in black abs with an arcol 0.35mm..

I use this mod supposedly doubling accuracy..

http://www.thingiverse.com/thing:8852

hmm..

I get a lot less backlash because of this mod too!

You need to double the steps per mm when using this mod.

http://i.imgur.com/XZAIw.jpg

http://i.imgur.com/Mov7d.jpg

 

[jelleatprotospace 0]

This seems to work, but I have some trouble with the start-gcode not showing up?

Is there a different place where those can be defined too, or do I need to update netfabb?

[jelleatprotospace 0]

Not really a surprise, as there is no code in the calibration file to define the start and end gcode?