Introducing KISSlicer

Sun,

I've been using pronterface to upload gcode to my printer from kisslicer and have had no problems.

Can you toggle the axis with the manual controls?

Edited 2 time(s). Last edit at 02/11/2012 06:57PM by Andrew Diehl.

---

www.Fablicator.com

I've been printing using SFACT with Pronterface just fine, but when using gcode generated by KISSlicer nothing moves.

This is what is shown on command line

CRC mismatch:3
rs 3
Invalid line:4
rs 3
Invalid line:5



This is first 10 lines from actual gcode file

; KISSlicer
; version 1.0.5.32
; Built: Feb 8 2012, 11:23:26
;
; Saved: Tue Feb 14 20:44:39 2012
; 'Measurin.gcode'
;
; *** Printer Settings ***
;
; printer_name = Default
; extension = gcode

Maybe it doesn't like being fed comments?

Thanks

-Sun

Quote
zsunsun
Maybe it doesn't like being fed comments?

Remove them and see!

If this is really the case you can possibly turn off the comments in KISSlicer or use a script to automatically remove them.

In AWK this would be:

/^;/ { next }
{ print }

You could do this on the command line (or batch file) as follows:

AWK "/^;/ { next }{ print }" in.gcode > out.gcode

---

Bob Morrison
Wörth am Rhein, Germany
"Luke, use the source!"
BLOG - PHOTOS - Thingiverse

Pronterface does not send the comments as far as I am aware. It looks like the firmware is having issues, actually. Maybe not recognizing something?

What are the first few lines that aren't comments?

Or better yet, just attach the whole g-code file.

Edited 1 time(s). Last edit at 02/15/2012 04:22PM by Andrew Diehl.

---

www.Fablicator.com

Estimated duration (pessimistic): 75 layers, 00:16:52
T:42 B:35
Print Started at: 18:14:15
CRC mismatch:3
rs 3
Invalid line:4
T:41 B:34

I'm attaching entire gcode here

it's odd that printerface does not show preview of gcode generated by KISSlier.

Maybe it's fault from printerface?

Found the problem! Your reminder for what the the g-code does is not commented out in the final gcode. You need to put the semicolons in for the start/end/extruderchange alteration boxes otherwise you comments get sent to the firmware and cause issues.


(**** beginning of start.gcode ****)
(This is being used lol?)
(This file is for TOM with 1.75mm ABS in a MK7 Extruder)


should be:


;(**** beginning of start.gcode ****)
; (This is being used lol?)
; (**** end of start.gcode ****)

---

www.Fablicator.com

Ah, that could be it, on my Skeinforge setting I have it to remove all the comments (it recognizes both parenthesis and semicolon as comments).

I should update my files accordingly and try again tonight.

I'm really looking forward to see how well this prints on my printer as I'm more and more frustrated with fiddling with Skeinforge settings.

Also, could you add acceleration parameters for the printer firmware that supports it? It will improve my printing time and quality immensely and I refuse to go back to the days without one.

Thank you

Sun

The slicer should have nothing to do with acceleration. The firmware should do that automatically with whatever g-code it receives.

---

www.Fablicator.com

Oh, I meant I would like to be able to customize different acceleration / feed speed per different portion of the layer.

For example, I would like to have 20 mm/s feedspeed at 40mm/s^2 acceleration for perimeter while 80mm/s feedspeed at 300mm/s^2 acceleration for infills.

Since infill is less likely affected by feedspeed / acceleration it would be a great time saving measure.

Thanks

Sun

Hi, Sun.

You can set the speed for 3 different regions: Perimeter (I set this fairly slow), Solid Infill (I set this medium), and Sparse Infill (I set this fast). Loops actually interpolate (non-linear, though) between the perimeter speed and the solid speed.

I don't control the acceleration profile at all, partly because BfB printers don't support that...does RepRap? I sort of assumed the firmware is set to accelerate at the highest non-skipping acceleration possible.

Thanks for continuing to play with KISSlicer, and thanks Andrew for helping out. Since I know so little about the RepRap world I really appreciate everyone's patience!

thanks,
Jonathan

Hi Sun,

While some of the firmwares allow on the fly acceleration changes, they were never really meant to be adjusted automatically withing a print. The intention was to be able to easily calibrate the machine's max acceleration without having to constantly re-upload the firmware if an acceleration proved too extreme.

In my experience, the highest acceleration your machine can handle reliably is always the best. When the printer is accelerating, the flow rate does not perfectly follow the head velocity. Therefore, the longer (lower value) the acceleration, the more severe the flow rate discrepancy becomes.

Also, most machines are running accelerations many times faster than you posted. (I run 2500mm/s^2, many other Prusa's are running in the 1200 range) and the acceleration is a very small percentage of overall print times in the majority of cases.

What is your particular machine setup which requires such low accelerations?

---

www.Fablicator.com

Do you really run at 2.5 meters per second squared? my TOM will vibrate to the point of self destruction if I set acceleration that high!

Mainly I am looking to have a smooth acceleration somewhere from ~10mm/s to 90mm/s, which is my TOM's max feed speed it can handle until extruder stepper motor starts to skip.

I'm still tweaking to reduce running noise and vibrations of my TOM, as many already have posted it creates a loud-almost-musical noises while running and also makes unfriendly -dunk- noises at every fast corners.

"the longer (lower value) the acceleration, the more severe the flow rate discrepancy becomes."

I'm not quite sure about that statement, figuring out what flowspeed is required to have constant extrusion material per distance would be trivial calculation based on acceleration curve generated, no?

It still puzzles me that people are still doing guess and check with feedspeed and flowrate while much simpler way of calculating exists, granted absorbed moisture in ABS introduces error but it won't be far off.

Edit: sorry I'm tired and grumpy right now and means absolutely no offense to anyone, I'm just thinking out loud

Thanks

-Sun

Edit : ps. sort checkbox on KISSlicer 1.0.5.39 does not work, all it did for me is arrange everything from Z-A

Edited 1 time(s). Last edit at 02/22/2012 12:57AM by zsunsun.

I'm attaching truncated gcode file generated using KISSlicer

I was looking through gcode and found a few weird floating spot.

First, first layer of the print has different supporting pattern compared to the rest of the print making next layer of support material impossible to have strong adhesion, making them detach in the middle of the print. Also, this makes bottom look ugly

Another problem is when support material color is going from green to grey (in the Kisslicer layer view) I can see that my grey support layer is just laying in the thin air creating mess in the middle of the print if not knocking over other printed side.

Also, 90 degree rotating support makes it very difficult for removal, I would like to have support stacked neatly in one thin feature. After all, support material should be easier to remove there is no reason to give it a structural support by interlinking them.

Finally, it seems like what I put in 'advanced setting' window is halved or in fraction of in the gcode generated. I tried changing maximum extruder speed and minimum print time, dragging slider all the way left and right, etc, but didn't seem to affect actual feedrate. Am I doing something wrong here?

Overall I really like the gcode generated by KISSlicer. Path generated makes sense when I look at them, most of the times , and its fast slicing time is a big plus as I prefer rapid iteration of tweak and test. The whole steps flow quite nicely, Cadding -> slicing -> printing -> tweak -> Cad and or slicing -> print -> tweak . . . etc etc.

It's getting late and I'm not sure I'm making any sense at this point, and thanks for reading I really hope your software to be good as it can get.

Thanks

-Sun

Whisper quiet at 2.5m/s^2. Can run up to 6m/s^2 but short fills hit a resonance causing skips.

Anyway, 20 and 300mm/s^2 of acceleration take 5mm and 10mm to come up to speed for 20 and 80mm/s printing, which is a lot.

The old Tom's were never really meant to take 80mm/s, and to be honest the fill quality suffers on my machine at that speed. I'd try and keep things 20/40mm/s if not a little slower, and just keep pushing the acceleration as high as you can.

"the longer (lower value) the acceleration, the more severe the flow rate discrepancy becomes."

While the calculations tracking flow rate to extrusion rate are perfect and easy, the actual extrusion rate does not immediately match what the stepper is doing. The filament is essentially a plastic 'spring.' Even though the extruder stepper may slow down as feed rate decelerates, the plastic will continue at the previous rate for a bit leading to extra material at the end of threads. Things may behave a little differently than i'm used to at such slow accelerations, but I'm pretty sure the same effects are there.

I think the support material is getting an minor overhaul soon? I agree with the first support layer should be aligned with the others, though. I haven't done much with the support yet myself.

As for the adjusting the speed slider, your firmware may be limiting your speed, or your may RPM may be set low and that is capping the feed rate?

---

www.Fablicator.com

I printed 80 mm/s feedrate with SFACT till yesterday, so firmware is ruled out.

If you take a look at gcode I posted, RPM is nowhere near 280, which is max RPM I set.

I'm guessing here that maybe KISSlicer is limiting speed due to the size of the object, so I will try with something else later today.

And I would like to repeat that first layer, not only support layer but the entire one, is out of place.

Thanks

-Sun

Please attach the .stl and your config files. I'll try to reproduce on my end.

---

www.Fablicator.com

STL file attached

Thanks for assist.

-Sun

Can you also post you .ini files please? (sorry I called them config by accident before) So far it seems to have worked fine with my settings.

280 rpm seems really high. What kind of extruder do you have on your tom?

The gcode you sent before had the layer speed reduction taking effect, which is probably why you aren't noticing the speed slider working.

---

www.Fablicator.com

Here's my ini files.

Only reason I set my RPM that high is to see if it is the limiting factor in feedrate and I just left it at that.

Thanks

-Sun

Well, under extruders you don't have the mm per revolution of the extruder set. That will impact your max feed rate settings I believe. Still looking, likely more suggestions to come.

---

www.Fablicator.com

That didn't change anything other than RPM number in the comment of gcode file.

DId you try slicing the file I posted and see if first layer came out to be weird?

Thanks

-Sun

Other than the support changing directions (which will be fixed soon) it looks fine to me.

I'm kinda grasping at straws here now. Maybe post some screen shots of the first few problem layer paths?

---

www.Fablicator.com

Support changing direction, floating support and speed issue are the only problems I think I'm having.

And for the 'weird first layer problem' thing, it was corrupt STL file, I re made the model and all seems to be fine.

Sorry about the confusion.

Thanks

-Sun

Sun, Andrew had a great idea about the 1st support layer changing directions, so I'll try to have that in by the next release. Yep, some support does float (if I'm understanding correctly), especially when the supported area is very small relative to the support density...the only fix for that right now is to choose a support style that is more fine / dense. And regarding the speed issue, I'm not sure what's happening there. Let me see if any of my beta testers have a TOM and if they could provide a working baseline INI file collection.

I just posted version 1.0.6: KISSlicer download page. Here's the changelog:
2012 February 25 (v1.0.6)

Breakdown of estimated extrusion time by extruder and path type (in G-code)
Can optionally extrude loops from inside to outside
1st layer temperature only happens after the raft if the material is different from the raft
Changed the name "Wipe Pillar" to "Prime Pillar"
Changed the name "Bed Flatness" to "Bed Roughness".
Changed the math for solid infill
to remove overlap with loops
to account for material in short connecting pieces
NOTE: you will probably need to change your "Flow Tweak" in material settings (close to 1.0)
Allows a 0mm wipe (will still trigger destring on 5D firmware)
Added a "Scale by X" right-click menu option to each object tab
Added an "All Models Menu" (to scale, revert, or delete)
Added an option to *NOT* rotate the model on load or packing
Fixed grid raft thickness bug
On Drag-n-Drop to 3D pane, can now handle filenames with spaces in them under Linux & Mac (was changing space to %20)

Thanks!
Jonathan

So, I forgot to let you know about the 1.0.7 release, but 1.0.8 is here!

http://www.kisslicer.com/download.html

The combined changelog is on the downloads pages, and is too big to list here, but some highlights are:
* Support reworked, including gap, inflate, and some finer settings, interface support on internal support if needed, etc.
* Raft has inflation as well (to control oversize), and has a "Skirt" mode
* 5D firmware updates
--- don't output Z if unchanged
--- Added minimum jump length...no destring if the jump < threshold
--- Destring feedrate setting
--- Filenames are not truncated to 8 characters
--- Optional extruder axis (E A B C) for Mach3 controllers
* better Mac integration
* Bulk Key in case people want to use a bunch in a lab
* Path Color Key

Hope you enjoy!
Jonathan

Considering there are free alternatives you may consider lowering your costing, you would actually get more pro users this way. I personally use Slic3r, Cura, SFACT, and SF41 in that order to generate my g-code. If I didn't like the first g-code I would use one of the other generating programs.

I understand that you want to make money, (seems like everyone in 3D printing is trying to atleast) but at the same time you are competing against free open-source and very mature programs that have a strong foothold in the community.

Hi, WildBill.

Thanks for posting. As you point out, there are a lot of slicers out there, and most are Open Source, and since most of them sit on skeinforge in some way, that makes sense, Slic3r is the exception and is a _very_ cool initiative...I am glad that it is gaining followers. skeinforge can do anything (just not quickly) and is a good mature codebase. The other main slicer I see (which isn't free) is netfabb.

Regarding KISSlicer, the pricing model was based a lot on my philosophy and guesswork. [8^) When I first wrote KISSlicer it was because I had some very complex models and skeinforge (via Axon) was taking forever. When my slicer actually seemed to work pretty well (as a command line interface), I released it for free, and started planning on making a PRO version.

The very first step was deciding what feature set belonged in the PRO category, and I settled on this mantra: "The hobby user should have everything they need for free". This meant no slowing down the FREE version, no cutting features you would need for a great print, no time-limited trials to hook someone and then force them to pay, and no restriction on size of prints or on the commercial use of KISSlicer (since many hobby users make huge prints, and may want to sell them).

Then I needed a working definition of "hobby user", and I settled on "someone with a single-head printer". That way I figured I could support almost the entire RepRap community and most of the kit-BfB printer users for free. (Note that the free version of KISSlicer also lets you use any single head of your multi-head machine.) So, it's not open source (to protect the PRO portion of KISSlicer), but it IS free if you don't need multiple heads.

Pricing was tough...I had a bunch of data points at 0, and netfabb hanging out in the $200-800 range. So, when searching for the answer without knowing the question, I naturally turned to Douglas Adams, and thus the price was $42 [8^) It also was not too high (if you just bought a 2- or 3-head printer you paid what...$2k-4k?), and hopefully not so low that it looked worthless. That, plus the fact that most users don't need the PRO features is a combination that I hope will keep KISSlicer from being wildly pirated (though 'KISSlicer' just turned up as a search term on SerialBay... [8^/

Anyway, all that to say, I really hope that the free version of KISSlicer works for most users. I don't want to suck money from hobbiests. If a commercial user finds that KISSlicer PRO helps, great! Schools or bulk uers get a serious discount. I'm also thinking of trying to make a few mini-contests where I give licenses as prizes, but I don't have any good contest ideas yet. (Does anyone have some ideas?)

Thanks for listening to my manifesto! [8^)
Jonathan

I think your pricing is spot on. As the number of printers grow so does the number of people willing to pay for their tools. I don't think you need to use a gimmick like a contest to raise awareness. I think you have the best slicing quality at the moment and I think the only thing holding it back from more adoption are its BFB roots which are confusing to people used to a Reprap style slicer setup.

Thanks, Billy.

Good point about the BfB roots..that is quite true. The other two things I feel are holding back adoption are a lack of good settings for a broad variety of common printers / firmwares, and some pretty complex settings (at least for the new user). I just released KISSlicer version 1.0.9 so the next big push is to solve these issues, and return to the KISS principle as much as possible.

thanks,
Jonathan

I just released KISSlicer version 1.1.0. Lots and lots of changes, so I'll just post the link:

KISSlicer download page

(There are builds available for Win / Mac / Linux / Raspberry Pi / FreeBSD.)

thanks,
Jonathan