Cura RepRap calibration help needed

[origamimavin 0]

I built a Reprap based 3D printer a while back (longer than I'd like to admit), and haven't been able to get it to print anything even remotely acceptable. I'm back at making a solid effort to get it up and running, and just found the Cura 12.12 software, and right off the bat got the best print to date. I've still been having some problems, so I came here for some help.

Current build specs:

Prusa Mendel, Marlin, Ramps 1.4

X: 180mm

Y: 170mm

Z: 84mm

Nozzle: 0.35mm

Heated bed (typically at 60°C)

1.75mm Filament

Steps per E: 865.888

Nozzle distance from bed at Z-Home: .051mm(.002")

Material: PLA

Printing surface: glass

Fans: None

Best printing specs to date:

Layer height: .2mm

Wall thickness: .6mm

Bottom/Top thickness: .6mm

Fill density: 20%

Print speed: 25mm/s

Print temp: 195°C

Bed temp: 60°C

Filament diameter: 1.71mm

Packing density: 1.0

Fill type: Rectangular

Support type: all

Skirt lines: 5

Skirt distance: 4mm

Retraction minimum travel: 5mm

Retraction speed: 40mm/s

Retraction distance: 2.5mm

Extra length on start: 0

Bottom later speed: 15mm/s

Initial layer thickness: .26mm

Infill overlap: 25%

First layers seemed great

 

Started to get messy as it was built

 

Finished decent model

 

The model I tested ended up only having 1 or 2 layers of infill for the bulk of it, but the top flat section still ended up pretty rough. The taller sections had a lot of infill even though it's narrow, so the perimeter ended up pretty rough. The first couple layers seem to go on really well, but after that it seems to get really messy. how can I fix this? does it have to do with the speed and temperature? I typically print my PLA at 185°C, but for this print, i slowed down the printer a bit and bumped the temp up to 195°C, and it seemed to work a bit better, but added more stringing. Could this have to do with the layer height? I tried a print at .15mm (as opposed to my usual .2mm), and it sisn't come out much better. The first layer seems to have trouble bonding with the glass when I run it at 195°C, and works much better at 185°C, but havent tried anything hotter than that yet. The nozzle also tends to drip a bunch as it's heating up, so I didn't try pushing it any more. I also don't know exactly if the Steps per E is correct, and just went with the number that was in there. I'm currently playing with that number a bit as well to see if that's going to help anything.

I'm also trying to print some simple cubes to see how the fill is working, and it seems to be moving quickly and isn't laying down properly. I can post more pictures of this and more of my attempts later.

The other thing I'd really like to change with the software is to add a couple more perimeter layers around the outside to hopefully add some more quality and definition to the model. Is this done by changing the wall thickness? Does that add more layers if I increase it enough? Is there a Plugin that can be used to adjust this?

Does anyone know of some specs I should specifically change to try and get better results? Is there something I can try and post pictures or video of the results to better understand the problem that's happening? Any and all advice would be greatly appreciated. I've worked with Dimension printers, Z-Corp printers, and Objet printers before, and while I know I shouldn't expect results quite like those, I do know generally what to expect and know I'm not getting there.

Thanks in advance!

-David

 

[rbc1225 0]

Can you get decent prints out of Slic3r and Pronterface? If so I would mimic those settings and see how it comes out.

 

[owen 19]

It looks like it is too hot though the temperatures you quoted are toward the low end. It also looks like you're under extruding.

 

[origamimavin 0]

I haven't been able to get any good prints using other software, so I can't really compare it to much.

I can try slowing it down and dropping the temperature back down to 185 and report back, but I would like to be able to speed up the prints a bit more. How can I turn up the extrusion amount, would I change the steps per E, or fudge the filament size any?

[joergen 2]

I am almost positively sure that your e-steps/mm setting is utterly and completely wrong:

865.888 is what I measured on a ultimaker, V2 bolt, 3mm filament, in October 2011, and it since has become the UM standard. The new V3 bolt requires a value of 835 for the e-steps/mm.

you are using 1.75mm filament, with a completely different extruder setup, with unknown to me gearing.

the first thing you should do is move 100mm filament through the printer, and measure how many mm of filament actually moved through (not through the nozzle, only through the e-motor. remove the extruder from the nozzle if necessary, this test is supposed to run without load.

then calculate your e-steps with the following formula:

100mm / (your measured length) * (your current e steps / mm)

burn the new value into your firmware, and re-run the test: you should see 100mm moving through your system, or something very close to it. if, not, continue, until it does.

once you have set your e-steps properly in your firmware, measure the actual diameter (1.xxmm, 2 decimals are important), and enter this in cura as your filament diameter, and start slicing and printing.

 

[rbc1225 0]

Yeah I agree with Joergen especially since you have not had any decent prints any other time. It certainly sounds to me like you skipped the step of calibrating the machine. Take for instance on my setup. The steps are 375 due to the electronics I have. I would expect yours to be closer to 750 to 800 range since you are using ramps and can go down to a finer step than I can. Here is a link that should help quite a bit. http://richrap.blogspot.com/2012/01/slic3r-is-nicer-part-1-settings-and.html

 

[gr5 2,167]

195C seems very cold to me. Try 220C just for a test. Yes the nozzle oozes (but shoudn't drip) below 180C, but it looks to me that when it is printing the part, sometimes the loops in the outer wall are pulled too tight makeing it look ugly. At higher temps each layer will stick better to the previous and the plastic won't solidify before it's placed.

Look at your gcode. There may be a line near the top:

M92 EXXX.XXXXXX

If you have Marlin, this value will override your extruder steps/mm setting in your firmwayre.

 

[origamimavin 0]

Thanks for the replies, all! I've been making some progress...

The first thing I did was this:

I got the new 13.03 siftware and moved some of my specs over and tweaked some others. My last print had critical dimensions as follows:

layer height: .2mm

wall thicknes: 1mm

bottom/top thickness: .6

fill density: 20%

print speed: 20mm/s

print temp: 185°C

bed temp: 60°C

support: everywhere

support type: rectangular

filament diameter: 1.71mm

skirt line count: 5

skirt distance: 4

bottom layer speed: 10mm/s

initial layer thickness: .15mm

infill overlap: 25%

loops > infill > perimeter

support material: 30%

Steps per E: 865.888

Results are as follows:

First 5 layers or so turned out great!

 

Support material turned out the best I've seen so far

 

Finished on bed

 

Finished piece

It's still a little rough, but it turned out really nicely. Some of the overhanginf material got kinda messy, but I know that' to be expected. I just think it can get a bit better. Temperatures stayed low, which I kind of like, but the print was VERY slow, and I'd like to bump that up over time. Any more tips would be helpful as to what to tweak next!

After printing the above, I decided to look into the firmware stuff and did as follows:

Joergern and rbc1225, thanks for the steps per E advice! I'm looking into it now. Here is an image of my extruder motor:

Extruder gearbox pictured disconnected from extruder

 

I also don't know the gearing ratio of the motor. I was able to dig up the firmware codes I loaded up the machine with, but don't know exactly which I used (but I have a pretty good idea). It should be one of two or three, so even if I upload the wrong one, I can do a print or so with each and figure it out again. My calibration variables (float axis_steps_per_unit[] = {X, Y, Z,E}) of the different firmwares I tested have E values of 1292.52, 1380, 1200, or 1000. The one listed at 1000 is what I had labeled as my most recent iteration, so that's the one I went with, and loaded it back onto the machine (variables of {58.1818, 58.1818, 2600, 1000}). With these variables, I extruded 100mm of material and my measurements were pretty off. I had to head out and couldn't calculate what I needed to change it to, nor did I get a chance to change or try it, but I'm about to go do that now. Once I get the extruder moving properly, I should be in much better shape.

 

After I printed the model I have shown above, I compared it to a model that I printed on a professional machine (that I assume to be much more accurate than mine), and noticed mine was taller. I havent compared it to the 3d model yet to totally verify this, but I have a feeling my Z axis is off in the firmware as well, giving me incorrect layer heights, which could also compound the printing problems I've been seeing. Any advice or links on how to fix this? is it also a matter of moving the nozzle up a certain amount and seeing if the actual movements matched the programmed movements and use a similar equasion to fix this too?

 

Another question along those lines; when I home my Z-axis, i have adjusted it to sit about .015mm (.002") above the printing surface (using feeler guages to measure this at multiple locations). When the machine starts to print, does it print at that Z-home height, or does it move up any? Like if I have the initial layer thickness set to .15mm, does that mean the machine is expecting the Z-home to be .15mm off the bed, or does it move the nozzle .15mm above Z-home?

 

Outside of the X, Y, Z, E movements listed above, is there anything else I really need to calibrate that I may have missed to get the machine up and running properly?

 

gr5, I'll try a hotter temperature (and I'm guessing a faster speed to go with it?) once I get more of these numbers dialed back in. It does ooze and drip even at 195°C, but hopefully some speed with help make up for the mess. I'll post back about this as soon as I can.

 

Thank you all for your help, I've made more progress with this software and your help in 2 days than I have in a very long time of failed attemps through other methods.

 

[rbc1225 0]

Did you go through that calibration document I gave you the link to? It has different pictures for different problems and their fixes. I would bet that will answer a lot of your questions. It goes into temp considerations, speed considerations and slicer setting and such. I would take a few days if I were you and go throuh and make sure your machine is calibrated properly. If you do that you are sure to not only understand your machine much better but also have fewer questions.

 

[origamimavin 0]

Thanks, I'm working my way through it all now. I don't have internet at my apartment where the machine is, and can only make small posts like this one from my phone at home, so I tried to write as much as I could about how far I got when I was in front of a computer online. Just hoping someone might have some advise on top of that.

[gr5 2,167]

>Like if I have the initial layer thickness set to .15mm, does that mean

>the machine is expecting the Z-home to be .15mm off the bed, or does

>it move the nozzle .15mm above Z-home?

If your initial height is .15, Cura will home your machine - call that zero and then move z up by .15 when printing the first layer. Whatever the initial layer height is set to (.3 default?) it sets the z to that height when it starts printing.

 

[origamimavin 0]

I've spent the last couple days messing with the extruder speeds, and it seems like i'm back where I started. In Cura, I had the Steps per E at 865.888, and was told this was probably wrong. I opened Pronterface and did the extrusion calibration (now I see that some of my settings were off in this test, working on testing it again), but my new number was about 1354.5033, which seemed to print much worse. Immediately, it stopped monding with the bed, and once I got it to stick again and upon dialing those numbers back down through trial and error, I'm back at the original 865.888, but it still seems to be extruding too much at the beginning (globbing/ dragging) and too little at the end (spaghetti stinging). Since I have the correct setting in Cura, I'd like to use that to do a filament extrusion test, but I don't see a GCode sender in that software. After preparing a file in Cura and going to Print, I've clicked on the extrude button, but there's no control on how far it goes. What is this set to by default? in measuring it, at the 865.888 steps setting, it was extruding nearly exactly 2mm per click. Is this correct? I'd like to find some logical reason why that 865.888 number seems to work, or possibly dial it in closer rather than just using a number that seems to work decently.

The difference between extrusion at the start and the end made me though.. can I use a plugin to set a M92 code for the first X number of layers (like 1-3 or so), then change the steps per extrusion after that to complete the rest of the model? or is there some other settings I can change to get the first layer to extrude less while it keeps it up for the rest of the build?

rbc1225, I've read through that whole link and then some, as well as other sites, and can't seem to come up with a solution to my extrusion problems...

 

[Daid 306]

The default of 865.888 is actually the older Ultimaker default, and it ends up as global default due to a bug.

Did you increase the temperature already? Always start out at a too-high temperature like 220C, this might feel like a lot of heat, but it ensures that you extruder won't grind because of backpressure from the colder filament trying to be pushed trough the nozzle while only being slightly melted.

 

[rbc1225 0]

Okay tell ya what. I have all but abandoned printing PLA, mainly because I don't like the way it looks for my uses but I will see if I can get my machine dialed in using slic3r and PLA and send you config file. I have not done a lot with Cura yet, Otherwise I would send you those config files but at least this will get you a way to start printing something decent and you can work from there. It might take me a few days (and maybe not) as I have not printed enough PLA to get mine dialed in for it. If it turns out to be the same settings I use for ABS just at a lower temp then it won't take too long. Keep in mind mine are not perfect either as I just started this hobby a month or so ago but I have been told I have made very good progress in getting this Prusa to put out quality prints. Send me your email in a PM and we can go from there. How does that sound?

 

[origamimavin 0]

Sounds great, thanks a lot! I'll also try bumping the nozzle settings up to 220 and see what happens, but when I calibrated the extruder motor, the nozzle wasn't attached.

[rbc1225 0]

Sounds great, thanks a lot! I'll also try bumping the nozzle settings up to 220 and see what happens, but when I calibrated the extruder motor, the nozzle wasn't attached.

 

The nozzle needs to be attached as a whole unit to get an accurate measurement of how much filament will be going through it while it is attached and at the temp you plan to start printing at, or so that is how I understand it.

 

[origamimavin 0]

I'll run some more extrusion tests now.

[illuminarti 18]

If you are just talking about calibrating steps-per-e, no you don't want the nozzle attached... unless a) it's hot, and b) you're moving the filament at a very slow speed - the extruder isn't going to handle having 100mm of filament shoved at it all in one go. Best case, you'll get a lot of slipping, and the measurement will be way off. Worst case you'll break things.

The steps-per-e measurement is solely about how the extruder moves filament.... nothing to do with the interaction with the nozzle.

 

[rbc1225 0]

If you are just talking about calibrating steps-per-e, no you don't want the nozzle attached... unless a) it's hot, and B) you're moving the filament at a very slow speed - the extruder isn't going to handle having 100mm of filament shoved at it all in one go. Best case, you'll get a lot of slipping, and the measurement will be way off. Worst case you'll break things.

The steps-per-e measurement is solely about how the extruder moves filament.... nothing to do with the interaction with the nozzle.

 

Cool good to know, but yeah I thought people would know the nozzle had to be hot and yes your not going to force it through at 350mm per min either. I did mine with the nozzle hot and at a slow rate. But here is my first try at PLA using the same settings I use for ABS just temp down a bit to 220. To me it looks like it printed a little hot as one of the pieces looks melted on the hinge portion of the print. Turning it down to 210 and see what happens. We I don't see how to attach pics from my machine only a URL so I will send them to you email David.

 

[rbc1225 0]

I sent you my config files. My prints were best about 195C. This was also without a fan so I am in the process of mounting a fan to the head and most likely will help a bit from what I have heard. I took the time to figure out how the new gallery works so here is the pic I was trying to show earlier.

 

[origamimavin 0]

Thanks everyone, especially rbc1225 for the direct contact assistance! I seem to be getting much better prints now, but I'm noticing the first couple layers seem to be too wide. The first later seems to be laying down more material than needed and it's sticking out and up, so the next couple layers are a little messy, but it seems to correct itself once it starts outputting the sparse inner material. I tried manually changing the M92 command in the beginning to something less than my standard extrusion rate, then bumping it back up right before the start of later 3, and it seems to be working better, but still could use more work. Is there something else that might help this problem? Also, is there some plugin I could run where maybe it asks the starting steps per E as well as an adjusted steps per E and the layer number in which it is changed?