Best posts made by arhi

Lot of thread hijacking happening with digressions around FreeCAD so I believe having one topic with some general FreeCAD discussion might be beneficial for us all

Some general data for the first post on the topic

URL: https://www.freecadweb.org/

The releases download:
https://github.com/FreeCAD/FreeCAD/releases

All releases (including pre-release) are rather old, and since this is rather active project using "latest and greatest" is useful even if there are few bugs in the "nightly" version. The problem is that there is no system generating nightly builds so you have to build it yourself. Not super complicated task and requires all free tools but it can take a while, it is fully explained here (you need to follow steps exactly):
compile on windows
compile on linux
compile on macos

I'm compiling it myself for windows 10 64bit, but so far did not figure out how to make installer but if you have any previous version installed just unpacking my archive anywhere and running from there works. I will update this post from time to time when I compile new version I will be adding new builds as I make them on meganz so check for time ti time. File name is informative enough :).

Documentation is decent and there's a number of YT videos too

If you want to help the project out and speed up the maturity there are number of ways you can help out the project, from setting up the nightly build system (this would be awesome), writing documentation, translation, donation, writing code...

As discovered by ppl with better eyes than mine the dev pre-release placeholder that's old actually host nightly builds so no need to compile ourselves

p.s. I think the reasons for moving to FreeCAD are irrelevant so let's keep focus on FreeCAD here. Noone is forced to migrate and all other tools have their place..

is it worth learning freecad: https://youtu.be/udIBhVIy5MI

IMHO if you need to store babystep value you are doing something wrong. Baby stepping is there so you can "compensate" for the bad repeatability of your Z probe.

what do you mean by "many printers not accepting arc g-code", all of the mainstream ones do, marlin supports it, rrf supports it, smoothieware supports it ?! iirc (have not tried myself) also the redeem (replicate stuff) and klipper support it too ... so where dis "many printers" that does not support arc g-code come from?

Repeatability of the endstops is rather high, even for high quality spring microswithes, let alone good optical endstops. Now, I personally never printer PEEK, PEKK.. (I machined them few times) but considering price of that filament why would anyone use open loop driver for any axes?!?! Conversion to high precision silent servo's or closed loop steppers cost less than one failed print with those filaments.

@Kolbi step works. Unfortunately, there's no (yet) option to detect features like what solidworks know how to do but ...

FreeCAD can import both mesh and solids by default without any additional "workspaces"

imports step, iges, inventor, dwg, dxf, 3d studio, obj, brep, svg, ply, pov, stl, obj ....

can export everything you can think of probably .. the file format's are not a problem... but for solidworks -> freeCAD I'd say step is the way to go

IIRC it's around 200k steps/sec, with some round 100steps/mm number your minimal "make sense" segment of 0.1mm would be 10 steps, so just for stepping we're talking 20k segments/sec/ Everything else needs to be done too (parsing, pathfinding, buffering..).

I have not run a test with DUET but Smoothieware on comparable cortex M3 32bit LPC1769 (compared to duet2 cortex M4 SAM4E8E that's a bit better but for this purpose more/less same, or duet3 cortex m7 ATSAME70 that's a lot faster) and it can process 2610 "simple" gcodes/second, and that's trough USB (from sd card should be faster, here you have overhead of calculating CRC, ACK-ing every message...). With full calculation, trough USB, controlling laser drawing raster (so very small segments) the team tested successfully 1000 pixels / second if 1 gcode per pixel (it can go 5-10x faster if you compress more pixels into single G-code using non-standard G-code extension available in one of the forks)

if you switch to RRF3 there's a trick I'm using

before you run any probing you reconfigure treshold:

G31 P{sensors.probes[0].value[0] + 5}

it reads current value and sets treshold to +5

works like a charm (you might want to use +10 depending on what type of interference you have on orion)

@CTRDevelopments old TT printers have something similar made, not the 2 extruders, but something similar. With DUET you can setup push-pull as you got already explained as a mixing extruder. I think you need to setup firmware retraction there and it should work great. The issue can in theory happen if one of them skip steps and they get out of sync you can have some tension in the filament inside the tube. Anyhow, that part is already explained and @deckingman have ton of exp there so he can help you with any issues if you go that way.

I chimed in for the TT way.

Old TT UP printer (UP Plus 2 is different) had a push-pull configuration. On the hotend side you have standard direct drive, there is a ptfe tube pushed into extruder (not fixed!! so not like bowden, just pressed on to the extruder, if you remove the filament it would fall off), on the other side ptfe tube is attached to the "push" mechanism. Push mechanism is a DC motor with a rubber wheel pressed on with a bearing between which is your filament. So when DC motor run the filament is pushed from the spool and pulled into the ptfe tube. No speed control, motor is either full on or full off. Motor is controled by the microswitch, when motor pushes filament into the tube the tension lifts the tube from the microswitch and it turns off the motor, when the extruder pulls filament and there's "slack" that presses the microswitch and DC motor is turned on to grab more filament from the spool and push it towards extruder.

It supposedly work great, I never tried myself as I started with UP Plus2 that does not have that mechanism (supposedly new extruder was much stronger / using much stronger motor and did not need that system to feed in the filament, downside is that pla got impossible to print with it as motor was heating up so the heat break was too hot for the PLA, but UP is mainly ABS printer so noone cared).

it's this thing on the left

@JohnOCFII I was having similar issue and I assumed it might be 'cause I'm using too precise encoder (600step/rotation) but you have 2.4mm/pulse, that's 2 or 3 orders of magnitude less precision than mine so there's no way duet is not able to handle the pulse train (I believe I was maybe sending too many inpulses so some got eaten away) ...

I ended up configuring the sensor to use high span, minimum 20% and maximum 5000% and I reset the sensor on every layer (I added the script in slicer) and that solved my problems making it properly trigger if filament jams or snaps

@Infinitysnek your schematic was very close to the real thing ...

here's my take

@Corexy said in Mosquito hot end...yes or no?:

YES!

I've searched here and elsewhere regarding the Mosquito hot ends vs the E3D V6, and there's

I'm running e3dv6 (original) since they exist (both 3 and 1.75)
I'm running hexagon's also since they exist (both 3 and 1.75)
I'm running 1730 from Kai
I purchased M-clone (it is POS )
I have seen original Mosquito in action and fell in love
I'm using original Mosquito (short period only )

1730 rocks, it is the best contraption I tried before Mosquito, it is much better than e3dv6 and hexagon. Hexagon is also IMO better than e3dv6. The problem is the environment, e3dv6 has env. that covers almost all your needs, others don't so getting parts for 1730 or hexagon ain't that simple. Mosquito came hard, adopting e3dv6 style nozzles and hitting all the shops out there so almost anyone selling 3d printer stuff is selling mosquito too, they reached BMG and other serious extruder manufacturers to create ready to push combos.. they are creating their own environment and doing it properly so while I do like the 1730 nozzle style better for e.g. Mosquito is overall much better device (of course if you want to print clay or some of those other Kai's crazy filaments 1730 is a must as Mosquito just like e3dv6 don't cut it)

Mosquito comes as a breath of fresh air IMO both small footprint and ease of use. But I see v2 comes with some changes that confirm heat creep with PLA on v1, especially for slow prints or multi head prints. If you are going to be printing mostly PLA and PETG - any hotend will work. Mosquito will for sure get the job done (you wanna get the magnum if you wanna do fast & large extrusions) but the question is if it's worth the extra $$$ for features you do not need where 30$ hotend with a good original nozzle will do just as good.

I see you are already advised to go .5 and not .4 for the nozzle. I personally use .6 for 99% of prints, .6 allow you to easily print both .1mm layer and .5mm layer.. not to mention, you can do .1mm perimeters and .5mm infill .. You can't have super precise fine features in XY but do you need them and how often?

@JohnOCFII there's lot of things useful to have around 3d printers

• leds
• 1k resistors 1/8W
• 10k resistors 1/8W
• 1000uF 36V caps
• 3.3 - 5v and 5v - 3.3V level shifters
• switches
• connectors
• crimp terminals
• crimp plyers
• schmitt inverter circuits
• ssr's
• fet's
  ...
  ...
  so see what's cheap that you know how to utilize that you can get there and make that $8 irrelevant ... that's how I usually go from $9 bill to $90 bill and ton of stuff I might use in future that I never utilize

@appjaws said in Simplify 3D configuration:

I produce the gcode file from simplify3D and transfer it to the SD card in the Duet using DWC and then use DWC to run the print..

Why not just CURL the file from S3D directly on to duet ?

in s3d, scripts tab, Addition terminal commands for post processing:

c:\path\to\your\sendtorrf3.bat "[output_filepath]" 

and your sendtorrf3.bat:

@echo off

for %%a in (%1) do (
    set filepath=%%~dpa
    set filename=%%~na
    set extension=%%~xa
)   
set if=%filepath%%filename%%extension%

c:\path\to\your\curl\installation\curl-7.50.3-win64-mingw\bin\curl.exe -G --data-urlencode "gcode=M30%%20%%22%filename%%extension%%%22" http://ender5.local.lan/rr_gcode
c:\path\to\your\curl\installation\curl-7.50.3-win64-mingw\bin\curl.exe  --data-binary "@%if%" "http://ender5.local.lan/rr_upload?name=gcodes/%filename%%extension%"

replace ender5.local.lan with ip of your duet, set proper paths and you are golden. The only thing that does not work here properly is when there is a space in the filename and if the filename start with number

RC9, largest print on duet2ethernet for me ever (not longest nor heaviest, just physically biggest, almost full volume of ender5, but only 3 shells, .35mm layer, 3% infill, .6mm nozzle) ~13 hours no issues. I guess positive feedback should exist too

you can solve this rather easy by adding this

; layer_height={layer_height}
; Layer height={layer_height}
; layerHeight,{layer_height}
; layer_thickness_mm={layer_height}
; layerThickness={layer_height}

to your template/gcode/start gcode. Duet will pick up the data from comment. Actually each of them is enough but I added them all for good measure

@Corexy said in Mosquito hot end...yes or no?:

I only remember seeing the 1730 some time back and ignoring it. Didn't realize it was good, as you don't often hear about them.

It's cause Kai Parthi is not a marketing guy but engineer/inventor. He's the guy who invented wood filament for e.g. but you will almost never find his wood filament out there but other brands that took the process and are using and marketing it etc etc... Kai made 1730 because he had problems with all other hotends on the market with his ceramic filament as his ceramic filament is very brittle and 1730 works ok with brittle filaments (and all other ones too). I don't know if he pushed it to market I'm using proto version. The idea is that he's using a "nozzle" that's all integrated, basically a surgical steel tube brased (soldered) to the e3d type nozzle. and then you put the heater block at the begining ad a big cooler at the rest of it and it works and it works great .. this is my 3mm nozzle:

If you look at feltlay, polylay, porolay, woodlay, layfomm ... etc etc .. those are all filament made by him. Unfortunately his filaments (like his hotend) are rather expensive so not very popular, and he's not investing in marketing so few ppl know about them too...
https://www.matterhackers.com/store/c/Kai Parthy Lay Series
https://shop.3dfilaprint.com/kai-parthy-filaments-229-c.asp

I think it's a good idea to stay away from V1 of anything IMHO...there's always issues that will only turn up in real world long term use, which is why I've been asking here.

Don't get me wrong V1 Mosquito works like a charm. It is AWESOME!!!. @omni was reluctant to mount it as he was using original extruder for a while and was "waiting for everything else of upgrades to finish before he..." .. when he moved to Mosquito he was confused how good it was compared to original (ok, original is wanhao mk something bs so not hard to be better but again). .. and he printed a lot of PLA and PETG with it (I think he did not try anything else yet) so v1 Mosquito works like a charm for PLA and PETG (and other stuff too, but for other stuff even the shitty clone is perfect with ABS/ASA/HIPS/PA so ..)

While it will be mostly PLA and PETG, I will look to enclose this machine later and want the option there to print exotic materials as well

I'd go with Mosquito for sure then!!! BUT I'd immediately ask about water cooling option. They mentioned it in numerous videos I see but I did not see water cooling block for sales. If they have it I'd get it. It will work awesome immediately of course but will be crucial in enclosed printer with exotic filaments IMHO so better immediately go that route as that's something most ppl forget when they plan for printing in enclosure... ABS is ok, enough to close the printer and get there some 40C (you need 70 for big intricate parts but 40 will do for number of parts) and air cooling work ok-ish in 40C env, but if you wanna print PA you need serious temps and air cooling will not work

I must admit I've got a subconscious resistance against moving away from a 0.4mm nozzle...I'll just but an assortment of sizes and play around with them. After all, I'll be having that "easy one handed nozzle change".

Dunno, when I started over a decade ago my first nozzle was .5mm and since it was super hard to get anything excet PP and HDPE locally I had to print with them and I had a huge roll of recycled PP that was full of shit so I had to redrill that nozzle to .8mm. Since then I'm used to big nozzles and you can really do a lot with them. If you use multiple nozzles .4 is the worse, it's a nozzle you will most probably never use. My take on nozzles is

• I print drafts .6mm nozzle - it is fast, it is strong, it can easily go down to .1mm layer and up to over .5mm (no need to go over .5)
• I print large structural parts (parts for printer, parts for cnc, lathe, holders for bike, racks..) I go with .6 or .8mm nozzle - it is super strong, super fast
• I print figurines, detailed stuff etc .. I go with .3mm nozzle
• I need top details - .2 mm - but this is super rare, I need to drop from .3 if I need to print a business card for e.g.

so .6 is for me really the all arounder .. I was normally (till recently) going with dual hexagon and that was 3mm 0.6mm nozzle + 1.75mm 0.3mm nozzle .. that's kinda perfect fit for me, .6mm for infills and stuff, .3mm for outside perimeter, top/bottom..

@BlueDust said in Slicer speed: S3d is faster then ideaMaker?:

I am trying to learn how to use ideaMaker, and the first print, I found took over 9 hours. Same print in S3d took ~5.5 hours. Settings were not exactly the same, but still...

IM defines speeds for lot more features than S3D so you need to match them all ... the g-code and speed is more less same between then IMHO. Now there are options in IM like adaptive infill that will increase speed significantly on some parts (but slicing will take much longer). There's lot more features to tweak so it will take a while till you get it all.

Compare the g-code preview. Often with "seamingly same" settings the IM will actually generate 2 more perimeters than S3D making object print significantly slower.

My first question is around ideaMaker gcode files, and why I wasn't able to increase the speed using options on the Duet2. Even at 300% or 20% speed, didn't affect the print at all. Duet also couldn't tell what layer was currently being printed. I already set ideaMaker to RepRap, and now wondering if I should change it to Marlen to see if that helps any but don't know why it would.

Few things are important wrt this.

printer settings

here, the "change gcode e lenght.... is super important, if that is not checked IM will try to setup flow rate by using G-code to manipulate % and not by actually changing E values, this also interfere with how speed % later on work so you really want this checked

second, add this to your start g-code in IM and the layer detection will work:

; layer_height={layer_height}
; Layer height={layer_height}
; layerHeight,{layer_height}
; layer_thickness_mm={layer_height}
; layerThickness={layer_height}
; ... from here continue with your normal start code

RRF uses this info from the g-code to detect layer height, since IM normally do not add this comment to g-code RRF has no clue about the layer height, this way you add the info

I tried printing a test cube, and it just make a single square then large blobs that kept jamming the nozzle. Filament sensor kept stopping print. "Appeared" it just wouldn't go to the 2nd layer but couldn't tell as printer kept stopping.

dunno about filament sensor issue, I don't have that problem, what is your filament sensor configuration?

I myself have this in "Layer change gcode"

M98 P"filamentsensor.g"

and in filamentsensor.g on the duet I have

M591 D0 P7 C"^e0stop" L0.124 R50:5000 E3 S1 A1

so I reset the filament sensor configuration for every layer

Some huge speed modifiers:

@bot yeah well between freecad, arcwelder and slic3r I think something interesting open source can be born... the major issue is an investment of time... just look how much kicad progressed after cern added a single paid developer to work on it full time.. prusaslicer has prusa research behind them, cura has ultimaker, idemaker has raise3d.. dunno who's backing superslicer but if it's just someone free time thats always scary

@gorf26 said in Percision peizo probing help:

yes its a Orion, i tried 2 different one's , one was the older version, and i just hooked up the newer version it has to wire connectors one if you want to use analog and the other for digital ..

I am using both "old" setup and the orion.
Old ones work rather nicely on digital input with RRF but I could not make ORION work for the life of me. I fiddled for a month and finally contacted PP guy's and they told me to use ORION's ANALOG output with duet and so I did. If you look at your orion there is both digital and analog output. Digital output trigger is configured with that #$^(#%&^#_&^@$+^ trimmer on the board and in most cases just don't work. On the other hand analog signal is more less what's happening on the sensor itself. Route the analog signal to the zprobe.in input on the duet and you are golden. Two things to consider

1. it must be zprobe.in as it is the only analog input pin available, all the endstop pins are digital only
2. use a coax cable (microphone cable will work) to pass analog signal from the A output on the orion to the zprobe.in on the duet. If you use normal wire the surounding noise from the motors and motor drivers and heaters will affect repeatability

Use RRF3.1.1 and configure your analog proble "on the fly"

so, the main configuration of the probe goes like this:

M558 P1 C"!zprobe.in" R2 H6 F2400 T3600 A4 S0.03 K0 
G31 X0 Y0 Z-0.1 P250 

and in your homing script you do something like this:

homeall.g

M98 P"homex.g"
M98 P"homey.g"
M98 P"wipe.g"
M98 P"homez.g"

and homez.g

; HOME Z WITH ORION

M400 ; WAIT FOR ALL MOVES TO FINISH

T0
G90
G0 X0Y0 F3000           ; go to center of the bed, complain if X or Y not homed
G91                     ; relative positioning
;G0 Z5   F900  H2        ; drop bed for 5mm
G90                     ; absolute positioning
G4S3                    ; wait 3sec

G31 P{sensors.probes[0].value[0] + 2}

G30                     ; Probe the bed at the current XY position. 
                        ; When the probe is triggered, set the Z coordinate 
                        ; to the probe trigger height.

This will set the treshold just befor you start probing for Z

In order to do mesh stuff properly you want to make your bed.g to do something like this

; bed.g
; called to perform automatic bed compensation via G32

G29 S2 ; delete compensations
M561   ; clear any bed transform
G28    ; home
G4 S1
M400
G31 P{sensors.probes[0].value[0] + 2}
G29 S0 

and do G32 to do a mesh leveling instead of G29

I hope this helps.

now, with the precise piezo boards that do not have analog output I strongly recommend, for those who knows how to solder, replacing those @^$()&$^(&!@$##!&%^_+#^( trimmers with a proper multiturn pot as that will make setting a trigger point possible

pay attention to "connect infill lines endpoints" makes a part stronger but adds basically one more perimeter so if you turn that on, you can drop one perimeter compared to s3d for e.g.