Teacup Firmware replicatorG & Skeinforge

One thing you have not mentioned is which skeinforge are you using in RepG: standard, 31, 35, 39. they are all included in repg245 by Ben and all do things a little different from the looks of it.

Second thing is which skeinforge profile are you using, I am using 39 with a copy of the thingomatic profile which I have changed all the setting back to white/grey, turned off features in craft: chamber, fillet, activate homing, hop, lash, inset, limit, multiply, oozebane, raft, reversal, scale, splodge, tower, unpause, widen, and wipe and then changed the settings shown in the pictures here .

I have not tried printing anything as I needed to rebuild my hotend first, but I have sliced a 50mmx50mm cube and run the machine through 200-300 lines of code.

As for endstops I have not activated them in Teacup yet but the soft limits seem to work perfectly when jogging the axis from repG's control panel.

I am not sure if the linux version of Ben's RepG is different but I no longer have a Machines.xml. They are all in separate xml's in a folder called machines. I personally erased the thingomatic.xml and cupcake.xml to cleanup the drop down in repG.

I'm at work so i can't go through all the settings in Skeinforge that i currently have as of last nights printing. I also think that these are specific to each machine pretty much. the ones I know I have changed/ turned on are speed and temperature and carve.still tweaking flow rate and feed rate to get good outputs. i rebuilt my hot end and moved up to Skeinforge 39 after my last really good test. haven't gotten back to that level of quality yet. was playing with a 1.75mm extruder design so... I have new 3mm plastic now so I'm switching back.

the big thing is flowr ate and feed rate.

Hi guys

I am currently running with Skeinforge39 selected.

Yup spotted the machines.xml being separated out into separate files. I created my profile for Teacup 57600 in a separate file.

I haven't deleted the other profiles yet though so I will do this see if it helps any. At least I can rule out that it is mixing something up (however unlikely)

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Heres an excerpt of what skein forge and RepG24 are turning out ( that is wrong for my machine)

Quote

()
(**** beginning of start.txt ****)
(This file is for a MakerBot Thing-O-Matic)
(**** begin initilization commands ****)
G21 (set units to mm)
G90 (set positioning to absolute)
M108 S255 (set extruder speed to maximum)
M103 (Make sure extruder is off)
M104 S225 T0 (set extruder temperature)
M109 S125 T0 (set heated-build-platform temperature)
(**** end initilization commands ****)
(**** begin homing ****)
G162 Z F500 (home Z axis maximum)
G161 X Y F2500 (home XY axes minimum)
G92 Z80 ( ---=== Set Z axis maximum ===--- )
G92 X-57.5 Y-57 (set zero for X and Y)
(**** end homing ****)
(**** begin pre-wipe commands ****)
G1 X52 Y-57.0 Z10 F3300.0 (move to waiting position)
M6 T0 (wait for toolhead parts, nozzle, HBP, etc., to reach temperature)
M101 (Extruder on, forward)
G04 P5000 (Wait t/1000 seconds)
M103 (Extruder off)

This little bit is the amin culprit and everything follows down hill from there.

G162 Z F500 (home Z axis maximum)
G161 X Y F2500 (home XY axes minimum)
G92 Z80 ( ---=== Set Z axis maximum ===--- )
G92 X-57.5 Y-57 (set zero for X and Y)

Even after copying the machine profile for,m the wiki pages as shown.

Anyone got any ideas ??

I am stumped.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Deleting the other profiles didn't help at all unfortunately, worth a try though as I have about run out of answers for this one.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

I'm right there with ya, but I delete all of that. I believe that code comes from the start.gcode file defined in the preface window in skienforge. give that area a check. I have not gotten there yet. I just delete that gcode after slicing cuz I know I don't do any homing.
heres the 20mmCube I use. typically I set my temp manually in replicatorG to say 200, re-slice this file with the changes I want to try. set the temp to 235 for PLA which is about 225C in actuality, extrude about 100mm of plastic then manually home my print head and hit print. usually that scenario works good for me. I'm working towards getting rid of all these manual steps one at a time and still getting better quality print jobs.

gonna watch really bad Hollywood movies and decompress, plan to get back at it in the AM.
night all.


-Architect

Thanks for those. You are a star.

I have had a look through the G-code and will check up on where you suggested the homing instructions are coming from.

The G-code looks to increment the Z axis from (pretty much nearly zero) upwards. A definite improvement over my own attempts so far.

The X and Y axes look to be working around a reference point with +ve and -ve displacements though. But there has been no move away from home at the start....

Are you manually positioning your head and start point somewhere or have I just been stupid and blind and not seen the moves in the code.

Where are you starting your prints from ??

I should get a chance to try the code out day after tomorrow hopefully.

Again thanks for the input, is has been most useful.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Yep I am currently manually placing my print head at hone and hitting print. Didn't want to get any further in my physical buIld until I proved out teacup could do what we are asking. Last night I made my first replacement part using teacup. :-) seems without my Z endstops I broke a platform Y axis mounting bracket. My printer is now officially self healing. Which btw, is how I am now promoting reprap. It's not so much self replicating as a true power but I own no other tech other then my 2 cats that when hurt can fix themselves. This is astounding to me. Think about it, if you broke the coffee maker could it fix itself? No, so i gotta go buy a new one!! Anyways, you'll probably hear more from me on this thought as it develops in my cranium feed by beer and such... Hehehhe

thought id share a pict of my LCD with the joystick attached.. so getto but it works :-)

controlled by a arduino nano.
btw, if anyone has suggestions for screen info to show I'd appreciate the input.

@aka47 If you copied your skeinforge profile using repg the new profile you created uses the old profiles prefs.

I started with the sf39 thingomatic profile and copied it in repG but it was still using the thingomatic preferences found @ /replicatorg-00245/skein_engines/skeinforge-39/skeinforge_application/prefs/SF39-Thingomatic-baseline/alterations/

You will see a start.gcode and end.gcode as well as others. These are what inject the homing code and other annoyances into the Gcode.

Either make a copy of the entire profile folder, rename it and start over in repG and change the files content to suit your machine or just change them in the thingomatic profile and never use a thingomatic with it.

Hope this helps.

aka47 Wrote:
-------------------------------------------------------
> The X and Y axes look to be working
> around a reference point with +ve and -ve
> displacements though. But there has been
> no move away from home at the start....
>
> Are you manually positioning your head
> and start point somewhere or have I just
> been stupid and blind and not seen the moves
> in the code.

Manual homing followed by negative coordinates will break on teacup's soft limits (X_MIN etc) unless you disable them.

Without the soft limits however, this should work fine as long as you manually home to the center of your table.

Architect Wrote:
-------------------------------------------------------
> thought id share a pict of my LCD with the
> joystick attached.. so getto but it works :-)
>
> controlled by a arduino nano.
> btw, if anyone has suggestions for screen info to
> show I'd appreciate the input.

* Layer x of y (based on Z(layer 3)-Z(layer 2), and Z(end), remember layer 1 is usually a different height to others. Alternately, just count delta Z != 0)
* extruder temperature
* bed temperature
* % complete (based on extruded volume vs total extrusion volume. don't use layers because higher layers are frequently smaller than lower ones)
* ETA (% complete vs elapsed time plus cool time per remaining layer if any)
* number of bad checksums, bad line numbers, etc
* % bad lines (n bad vs n good so far)

Quite a few of these require reading the whole file before starting, might be an idea to encourage your slicer to put this info in the file header perhaps in some specially formatted comments or something. I've seen layer indexes in skeinforge output before, but I think it's usually removed before being sent to the printer.

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

Triffid_Hunter Wrote:
-------------------------------------------------------

> Manual homing followed by negative coordinates
> will break on teacup's soft limits (X_MIN etc)
> unless you disable them.
>
> Without the soft limits however, this should work
> fine as long as you manually home to the center of
> your table.

I set my limits as such:

#define X_MIN -95.0 // 0 at center of 190
#define X_MAX 95.0

#define Y_MIN -70.0 // 0 at center of 140
#define Y_MAX 70.0

#define Z_MIN 0.0 // 0 at platform
#define Z_MAX 150.0

Which seems to solve the negative number limits for the soft limits and makes centring the platform easier too.

Ewww! Negative space?
I was never able to keep that straight when I was coding postscript. :-) I get the idea but couldn't keep the numbers right in my head. Lol
Triffid awesome ideas. Percentage done I have yet to see correctly. Good challenge. Skeinforge does compute stats at the end of a build could put that into the file. But maybe once current abilities are reached.
Found something interesting tOnight. In skeinforge 39 (still trying to convert from 31) the flowrate doesn't seem to effect my e stepper motor. No matter what I set it at. In 31 I set it to about 25 and it works well. But now I am wondering, if it s controlled at all but skeinforge or the firmware. I need to look further into this, currently my e stepper per mm is 36 which may be laying down a little more Plastic then nessisary.
@ak47, got my pla temp down to 195. Had to completely rebuild and simPlify my cold and hot ends but the temps are much better. I'm going with a makergear hybrid cold end and a kinda modified nophead resistor/makerbot looking heater core. Made my own nozzle from a brass barrel from makergear and a blank nozzle they sell. Seems to be working well. So far. I drilled it out to .35 mm but the pla extracted comes out at .65mm interesting

I'm thinking for now a status window. kinda sum up those idea triffid had but without the calculations.
here's what I would have.

Extruder temp: xxx c / xxx f
Extruder set temp xxx c (what temp I'm suppose to be at)
Bed temp: xxx c / xxx f
Bed set temp: xxx c (what temp I'm suppose to be at)
current layer height : XXX.XX mm
Line number : (N value)
Build time : XXX min.
Volume extruded (?) : XXX mm(2) or XXX mm extruded? (total sum of E)

Guys

Awesome.

Architect & Sublime.

Thanks for the pointers, I am using skeinforge39 and found it creates the new profiles in ~/.replicatorg/sf_39_profiles

~/ being the default home directory (sorry had to explain for windoze folks)

I have hacked the two gcode files and now get something that looks more sensible so hoping to try this out tomorow sometime.

Sublime

Tweaking the origin to default to the center of the X & Y build space in the firmware is a good idea. Having reviewed the G-Code that was now being generated I was wondering how to coerse Skeinforge into generating positive numbers etc.

I am a big fan of making the machinery and firmware do as much of its own setup as pos. Purely because us humans are a touch less reliable and end up breaking things by forgetting.

I will adjust this and go from there.

Questions: is there any mileage in coding this up as the default with Build Space Dims in the config.h such that it is a runner out of the box for repg24?? Or is this a poor idea because it breaks something else ??

I actually like the soft limits very much. Particularly as the endstops don't actually enforce electrical move ceases at the end of an axis. Merely request it from the firmware.

Good news about the Extruding temps too.

Thanks guys

Cheers

aka47

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Quote

Questions: is there any mileage in coding this up as the default with Build Space Dims in the config.h such that it is a runner out of the box for repg24?? Or is this a poor idea because it breaks something else ??

Not to disappoint you, but with at least 5 host applications and at least a dozen different mechanical setups a "runner out of the box" for more than one person is plain impossible. If you have a specific setup and want to share it with others, provide them your config.h.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Sure I agree.

However, having the build origin centered in the X & Y Axes of the Build Space and calculated from the build envelope dimensions may be arguably applicable to more of the permutations than perhaps not having it so.

RepG just happens to be one of them.

I would certainly be concerned about going a route which made sense for only one permutation at the expense of all the others.

Can we list the permutations perhaps that this would categorically not work for ??

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Hi

Skeinforge Multiply centres the job. Set the number of cells to 1 and in can be used to produce a single part.

---

Regards

Ian

OK more testing done.

Sublime

I tried changing the max and min stuff as suggested and it didn't work unless I manually homed the extruder tot he centre (which defeated the purpose a little)

emt

However setting the multiply option in Skeinforge with a cell size of 1 did work as the head now went to the center of the bed before starting (I dialed in 100 by 100 for my 200 by 200 bed) and the motion looked to be about right.

Problems that remain then are:-

1. Although Motion is about right now, Extrusion is'nt.

Definite improvements on the journey towards a printing machine but slow.

RepG/Skeinforge 39 is generating Extrusion code as follows.

M108 255.0 (set extruder speed this looks to be a max for non stepper extruder)
M101 (to turn the extrusion on)
.
.
.
Moves as G1 instructions without any E axis parameters
.
.
.
M103 (to turn the extrusion off)

I am currently using a M01 to pause and have to hit a key after I have left it to warm up a bit.

2. M6 T0 is not causing the necessary wait until the head is up to temp. It just breezes straight through as if it weren't there.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

If I compare this with the code that Architect uploaded for comparison.

G1 X8.89 Y-8.89 Z0.15 F1800.0 E17.78

Architect's code clearly shows the E values for each G1

So I think RepG245 & Skeinforge39 are not passing information between them that tells Skeinforge to generate 5D G-Code.

Consequently no extrusion.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

I might have found it I activated Dimension in the profile and I now get something with E parameters.

Off to do some more testing.

---

Necessity hopefully becomes the absentee parent of successfully invented children.

Yep,
dimension activates the 5D E code.
my current problem is that flow rate in SPEED does not seem to change anything in regards to how fast my Extruder stepper runs...
so Other then changing E_MM_PER_STEP in the firmware I cant control how much or how fast my extruder stepper runs...
grrr.
Almost got a working whistle to print out last night. but it is putting way to much plastic down and filled in the mouthpiece, as well as kept the ball intact inside. Does any one know of a good forum I could use to start trolling for examples and ask for help in tweaking Skeinforge?

Hey everybody I am getting good results with Teacup and will post some instructions here as soon as possible. I'm a busy for the next little while because my wife just had laser eye surgery and I have to be her eyes.

I have posted some pics here and instructions will follow.
[geometricobjectdepositiontool.blogspot.com]

Edited 1 time(s). Last edit at 04/15/2011 03:33PM by Sublime.

aka47 Wrote:
-------------------------------------------------------
> 2. M6 T0 is not causing the necessary wait until
> the head is up to temp. It just breezes straight
> through as if it weren't there.

I wasn't aware that it was supposed to, I thought that's what M109 was for. If you check gcode_process.c, M101, M109 and M116 wait for temps to stabilise near their targets, using enqueue(NULL) which inserts a wait for temps "move" into the queue.

Sublime Wrote:
-------------------------------------------------------
> Hey everybody I am getting good results with
> Teacup and will post some instructions here
> as soon as possible.

Those prints look great! Thanks for posting

---

-----------------------------------------------
Wooden Mendel
Teacup Firmware

Sublime, great looking output. maybe you can help me. seems I can either get good solid walls or good solid tops but mot both. hopefully this pict will show what i mean. when i get everything dialed in to where my layer thickness, flow and feed rate are acceptable, closing off the tops becomes too thin and the plastic sags and does not span over the top. I'm not sure if this is what is considered bridges but I've tried adjusting everything I've read about bridging and nothing seems to help. anyone have any suggestions? or even hints as too what might be causing this?

Is that abs? I'm using PLA

Is the cube hollow or solid?

How many surface layers are you using?

Are your bridge rate multiples in the speed panel of skeinforge set to 1.0 each as a starting point and both activate speed and add flow rate active?

One thing I was doing without even knowing was, I wasn't using the temperature settings correct. I had been using the manual control panel to preheat the nozzle and then starting the build assuming it would use the settings from skeinforge, but I had "RAFT" turned off and it needs to be on and set to 0, so it was using my pre-heat temp for the entire build.

To print single wall items like the tornado turn off infill but for hollow items you need to set the infill ratio to 0 but have it active, this should give you multiple outer layers which can be set in the infill properties under solid surface thickness. I haven't printed much with bridges and what I have had sags in the first two or three layers.

I started by turning off or setting to default almost all the features and have been adding them as needed. I find clicking on the "?" and reading the web page for every setting has helped to make sense of the jumble of settings that are only there to compensate for the lack of feature in the older firmwares. Teacup having acceleration and E_start/stop does not need the any retraction in dimension tab and oozebane may not be needed either. I am not using the limit settings either, I chose to adjust them in the firmware every time I made an adjustment.

I decided to try and fix a leak in my hot end and ended up damaging the teflon, since then I have made four new heater barrels but none that work like the leaky one I used to make those prints. So it may even be hardware related causing your issues.

To everyone else following this thread, I had a lot of connection issues with Ben's repG, but he has managed to find the issue causing the connection problems between teacup and repG. I haven't tested printing with it yet because of my hotend issues but it now connects every time I press connect or start repG while plugged in. I encourage the rest of you to download his latest version and see if it helps you too.

Edited 3 time(s). Last edit at 04/16/2011 07:32PM by Sublime.

yep, I've been using Ben's fixed repG since he released it. very stable (unless I forget to turn on my electronics which are powered separately from my arduino)

Those are printed with PLA at around 200 - 205 C from top to bottom. my surface layers are set to 3 perhaps I should raise that? and lower the temp for final layer?

bridge rates have been tweaked considerable as I was hoping they would fix the problem. I was wondering what happened to my temp settings... Thanks for the catch. I do the same thing with my temp right now. but at one point I did have different temps going and then for some reason they started disappearing from my gcode. didn't realize turning off Raft caused all the M104 commands to go away thats ridiculous.

Ben pushed a fix last night for communication errors I was having on Linux and others were having on OSX that related to timing errors.

Did you increase the speed and flow for the bridging or lower it? I would think increasing the speed to say 1.4 and the rate by maybe 1.3 should stretch the filament while the higher flow will lower the temperature. But I am learning along with you right now, so I'm just guessing.

Really? Hmmm, I didn't get the email. I haven't been having any communication errors for the last couple weeks.
I'll have to do a pull thanks.