Retract Z or not not?

( ANSWER TO THE PROBLEM IS POSTED in post # 20)

Do most people retract Z to stop strings or just retract the filament?

I have been printing for a couple months now and have my prusa tuned in very well. The only problem i havent totally solved yet is getting rid of the strings it leaves when jumping from the end run of one spot to the beginning of another. Using slic3r i have it set to retract the filament which helps alot but it still doesnt completely get rid of all the strings. When i set it to retract Z it would lift Z but when Z went back down to start the next run it was ramming the nozzle into the print, maybe i need to retract Z more to get out of the slop in the axis but i had no luck retracting Z. Also if i retract the filament too far it seems to either take a while to start getting filament to start coming out again or it puts out to much filament and leaves a blob.

I am using slic3r, have 1.75mm PLA, a .35 nozzle, i print at 175C and aroud 80mm/sec typically. Using a Makergear extruder with a direct drive geared motor.

Edited 2 time(s). Last edit at 05/29/2012 02:59PM by GITRDUN.

I only lift Z when printing tall thin parts to stop them being knocked over. It doesn't have any impact on strings. If anything it would make them worse as the time to do the move is longer, so more time to ooze.

Do you live in a humid country? I think ooze depends on moisture absorption. The room I run my machines in gets very hot so it is dry all year round and I never have any problem with strings.

---

[www.hydraraptor.blogspot.com]

Like nophead I don't lift Z unless the parts are tall and thin. I use a geared extruder so I don't know if this will help you, but try increasing the speed of the filament retraction instead of the distance. In my experience with Makergear hotends 1mm retraction is far enough, and I set the speed of retraction as fast as I can get on that printer.

Aha, I hadnt really thought of increasing the speed of the filament retraction. Its set at 30mm/sec now, ill set it at 60 and see what i get. Thanks for the tips.

GITRDUN Wrote:
-------------------------------------------------------
> Aha, I hadnt really thought of increasing the
> speed of the filament retraction. Its set at
> 30mm/sec now, ill set it at 60 and see what i get.
> Thanks for the tips.

It's most likely limited in the firmware to below 30 already. Most extruders can not retract much faster than 30 if the acceleration is fast enough that it actually gets up to speed in retraction distance. I would set the max speed in the firmware to 30 and turn up the retraction acceleration until the extruder skips steps on retraction and restart. Also a shorter melt zone in the hotend can really help, adding a fan to blow across the thermal barrier is the simplest way of doing this.

---

FFF Settings Calculator	Gcode post processors	Geometric Object Deposition Tool Blog
Tantillus.org	Mini Printable Lathe	How NOT to install a Pololu driver

I set the retract speed to 60mm in slic3r and it didnt haelp at all. Ill check my firmware and see whats up there. I have a fan blowing on the hot end already, it made a huge difference also in stopping the extruder jams.

Still no luck. I found in the firmware where it had the E limited to 5mm/sec max , i set it to 60mm. Then resliced with it set at 60 and got the same results. I then started playing with the extrusion multiplier and extra amount to extrude after retract and nothing is helping at all. The problem to me seems to be that the nozzle is extruding and when it wants to jump to another part or different location it rapids straight there but drags a tiny bit of the extruded filament off the top of its last location with it. The strings arent thick, maybe a hairs thickness but its a pain to have to clean up every print afterwards.

I'd play with the temperature.
Having said that I have yet to completely eliminate strings and blobs.
Skeinforge with the comb plugin enabled tends to produce less strings than Slic3r.

I tried everything i can think of including dropping temps til it skips around and cant fix it. Just now i was watching the extruder gear and it does not retract every time when it should. When printing perimeters it rarely retracts before moving to the next part. But when printing infill it retracts every time. i think this is what the problem is. Not sure if its an issue with my slic3r or something else i need to tweak.

What value do you have set for 'minimum travel after retraction'? FWIW I have mine set to 0.2mm.

I had it set for 2mm. I thought that was only effective if you were retracting Z axis. Since my Z retract is set to 0 i assumed that was not used.

Minimum travel is used to determine if retraction is used on a move. If the none printing move is less than the minimum travel, no retraction is used.
IME 2mm is a pretty good number.

Having PLA too hot, so it comes out as a liquid rather than a filament, can cause tiny hairs all over the place.

---

[www.hydraraptor.blogspot.com]

You hit the nail on the head, but Polygon called it earlier. I went stupid low on temp as it was printing and the print actually got better and the hairs disapeared entirely. I ended up at 148C. At 135C it started jamming the nozzle. Earlier i tried going down on temp and it started to spit and miss in spots when starting to print. So this time i set the software to run the first layer at 160C and then step down to 148C after that and it works great. I dont know why my temp is so far off the norm from what everyone else uses but hey if it works it works. My sensor must just be way off.

Thanks for the help fellas.

I have just started playing with Marlin and found the Epcos 100K table was 25C too hot and had a bad entry in it causing a jump of 5C around 200. That prevented it settling on the target temperature if it was close to 200.

Coupled with where people measure the temperature it is no wonder the results don't match. 25C too hot would damage the extruder if I ran at say my normal 250C for ABS. I tested it lower first luckily.

---

[www.hydraraptor.blogspot.com]

Ive never tried printing at these temps as it always caused problems anywhere below 160C. Lesson learned, finally. I think the trick was to let the nozzle get good and hot before trying to print and then let it settle down to the printing temp. But as you state im sure its not actually a correct reading, i have a laser temp meter i will check it with tomorrow and see what temp its actually running at.

My temps seem to fluctuate around +/- 8 to 10C from my target temp but other than that i dont see anything wrong with the Sprinter firmware, and that may not be the fault of the firmware either but i know there a few temp settings in there i am going to play with. It acts like it takes to long to heat back up when it shuts off the heating.

The only way to get an accurate temperature calibration is to put a thermocouple inside the heater barrel. That is where the plastic spends most of its time. Obviously that is much easier before you extrude any plastic.

Laser IR thermometers usually have too big an aperture and are not calibrated for the typical emissivity of shiny metals, which is low.

---

[www.hydraraptor.blogspot.com]

I've never gotten what I'd consider an accurate reading from a laser thermometer on a hot end, as nophead mentions the measurement area is too big.
A multimeter with a thermocouple will get you pretty close, I usually measure at the point the hot end meets the heater, it's probably off by as much as 5 degrees, but close enough.
Using repetier on Ramps my Hotend temperature will fluctuate less than 2 degrees once it gets hot.
PLA is a little difficult to get right, I recently had exactly the opposite issue, I was extruding at 185 (210 for the first layer), and getting occasional misses in long prints, I had to increase temperature almost 20 degrees to get rid of the misses. I think a lot of it has to do with Hotend designs and how the heat is distributed when it stabilizes.
One of these days I'll do a comparison of the popular hotends with PLA.

nophead Wrote:
-------------------------------------------------------
> I have just started playing with Marlin and found
> the Epcos 100K table was 25C too hot and had a bad
> entry in it causing a jump of 5C around 200. That
> prevented it settling on the target temperature if
> it was close to 200.

Nice catch Chris - Anyone using stock Marlin and EPCOS 100k thermistor should change the temp table in thermistortables.h

I think that's the standard Makerbot thermistor?

The bad section is here under thermistor #6 - (about line number 224)

{62*OVERSAMPLENR, 210},
{73*OVERSAMPLENR, 205},
{72*OVERSAMPLENR, 200},
{94*OVERSAMPLENR, 190},
{102*OVERSAMPLENR, 185},
{116*OVERSAMPLENR, 170},
{143*OVERSAMPLENR, 160},

The whole table sounds like it's out of calibration, but at least change the highlighted line from 72 to about 81, that's going to help when trying to print with PLA.

EDIT - And why I clicked on the post - I have been using lift Z for quite a few months now, I'm using it more and more with all sorts of objects, if you are moving fast enough and have reasonable speed on your Z axis (8mm/sec+) it really helps print quality, I have yet to see it decrease print quality.

Edited 1 time(s). Last edit at 05/28/2012 02:23PM by richrap.

---

[richrap.blogspot.com]

UPDATE :

Thought i should post this for reference to anyone else with stringy problems.

After dropping my temp the strings had completely disapeared but in doing so i started having problems with small cavitys showing up in the prints due to either extruder slipping or plastic just extruding inconsistently. I then went back up to 180C on my temps and the problems went away, but i also have no strings at all. What actually stopped the strings was going into the firmware (sprinter for me) and setting the max speed of the extruder motor to 30mm/sec vs 5mm/sec as it was origionaly set at. I changed that at the same time as going down on temps aparently. So my mistake, but the problem is completely resolved finally. So the extruder was being limited to a retract speed of 5mm/sec when trying to retract filament before moving to a new coordinate and just wasnt able to retract fast engouh.

did you mean that you change the code where its written
(line 149) #define _MAX_FEEDRATE {400, 400, 2, 45} // (mm/sec) (X, Y, Z, E)
on the "Configuration.h" file on sprinter, because as you see its written that E is 45mm/sec
can you write where you made the change.
thanks.

RichRap and Chris - I have an EPCOS and tried changing my Marlin thermistor as you suggested and it wrecks my extruder temperature readings. I'd like to know how the thermistor tables were derived originally. Then I might know why Chris has supposedly made a good catch, albeit without any posted code or verification. Then I might also know why I should follow the recommended code patch. Right now...none of these facts are available...so this thread has solved nothing for me and only created a needless firmware change and revert. Have fun printing!


GITRDUN Wrote:
-------------------------------------------------------
> ( ANSWER TO THE PROBLEM IS POSTED in post # 20)
>
> Do most people retract Z to stop strings or just
> retract the filament?
>
> I have been printing for a couple months now and
> have my prusa tuned in very well. The only problem
> i havent totally solved yet is getting rid of the
> strings it leaves when jumping from the end run of
> one spot to the beginning of another. Using slic3r
> i have it set to retract the filament which helps
> alot but it still doesnt completely get rid of all
> the strings. When i set it to retract Z it would
> lift Z but when Z went back down to start the next
> run it was ramming the nozzle into the print,
> maybe i need to retract Z more to get out of the
> slop in the axis but i had no luck retracting Z.
> Also if i retract the filament too far it seems to
> either take a while to start getting filament to
> start coming out again or it puts out to much
> filament and leaves a blob.
>
> I am using slic3r, have 1.75mm PLA, a .35 nozzle,
> i print at 175C and aroud 80mm/sec typically.
> Using a Makergear extruder with a direct drive
> geared motor.

This is what my temperature tables look like. I didn't publish them because I think they are so bodged I must have the wrong thermistors selected or something. I can't submit them as a patch as it doesn't make sense they are so far out. The discontinuity at 200/205 got fixed on github, but not in the same way I fixed it. I changed the 205 line to 66, but it is actually 228C for me.

#if (THERMISTORHEATER_0 == 6) || (THERMISTORHEATER_1 == 6) || (THERMISTORHEATER_2 == 6) || (THERMISTORBED == 6) // 100k Epcos thermistor
const short temptable_6[][2] PROGMEM = {
   {28*OVERSAMPLENR, 250 + 30},
   {31*OVERSAMPLENR, 245 + 29},
   {35*OVERSAMPLENR, 240 + 28},
   {39*OVERSAMPLENR, 235 + 27},
   {42*OVERSAMPLENR, 230 + 26},
   {44*OVERSAMPLENR, 225 + 25},
   {49*OVERSAMPLENR, 220 + 25},
   {53*OVERSAMPLENR, 215 + 25},
   {62*OVERSAMPLENR, 210 + 22},
   {66*OVERSAMPLENR, 205 + 23},
   {72*OVERSAMPLENR, 200 + 23},
   {94*OVERSAMPLENR, 190 + 19},
   {102*OVERSAMPLENR, 185 + 19},
   {116*OVERSAMPLENR, 170 + 28},
   {143*OVERSAMPLENR, 160 + 26},
   {183*OVERSAMPLENR, 150 + 16},
   {223*OVERSAMPLENR, 140},
   {270*OVERSAMPLENR, 130},
   {318*OVERSAMPLENR, 120},
   {383*OVERSAMPLENR, 110},
   {413*OVERSAMPLENR, 105},
   {439*OVERSAMPLENR, 100},
   {484*OVERSAMPLENR, 95},
   {513*OVERSAMPLENR, 90},
   {607*OVERSAMPLENR, 80},
   {664*OVERSAMPLENR, 70},
   {781*OVERSAMPLENR, 60},
   {810*OVERSAMPLENR, 55},
   {849*OVERSAMPLENR, 50},
   {914*OVERSAMPLENR, 45},
   {914*OVERSAMPLENR, 40},
   {935*OVERSAMPLENR, 35},
   {954*OVERSAMPLENR, 30},
   {970*OVERSAMPLENR, 25},
   {978*OVERSAMPLENR, 22},
   {1008*OVERSAMPLENR, 3}
};
#endif

#if (THERMISTORHEATER_0 == 7) || (THERMISTORHEATER_1 == 7) || (THERMISTORHEATER_2 == 7) || (THERMISTORBED == 7) // 100k Honeywell 135-104LAG-J01
const short temptable_7[][2] PROGMEM = {
   {46*OVERSAMPLENR, 270-52},
   {50*OVERSAMPLENR, 265-52},
   {54*OVERSAMPLENR, 260-52},
   {58*OVERSAMPLENR, 255-52},
   {62*OVERSAMPLENR, 250-52},
   {67*OVERSAMPLENR, 245-52},
   {72*OVERSAMPLENR, 240-52},
   {79*OVERSAMPLENR, 235-52},
   {85*OVERSAMPLENR, 230-52},
   {91*OVERSAMPLENR, 225-52},
   {99*OVERSAMPLENR, 220-52},
   {107*OVERSAMPLENR, 215-52},
   {116*OVERSAMPLENR, 210-52},
   {126*OVERSAMPLENR, 205-52},
   {136*OVERSAMPLENR, 200-52},
   {149*OVERSAMPLENR, 195-52},
   {160*OVERSAMPLENR, 190-52},
   {175*OVERSAMPLENR, 185-52},
   {191*OVERSAMPLENR, 180-52},
   {209*OVERSAMPLENR, 175-52},
   {224*OVERSAMPLENR, 170-52},
   {246*OVERSAMPLENR, 165-52},
   {267*OVERSAMPLENR, 160 -52 },
   {293*OVERSAMPLENR, 155-50},
   {316*OVERSAMPLENR, 150 - 49},
   {340*OVERSAMPLENR, 145 - 49},
   {364*OVERSAMPLENR, 140 - 44},
   {396*OVERSAMPLENR, 135 - 42},
   {425*OVERSAMPLENR, 130 - 40},
   {460*OVERSAMPLENR, 125-38},
   {489*OVERSAMPLENR, 120 - 36},
   {526*OVERSAMPLENR, 115-34},
   {558*OVERSAMPLENR, 110 - 32},
   {591*OVERSAMPLENR, 105-31},
   {628*OVERSAMPLENR, 100 -29},
   {660*OVERSAMPLENR, 95 - 26},
   {696*OVERSAMPLENR, 90 - 24},
   {733*OVERSAMPLENR, 85 - 22},
   {761*OVERSAMPLENR, 80 - 20},
   {794*OVERSAMPLENR, 75 - 19},
   {819*OVERSAMPLENR, 70 - 17},
   {847*OVERSAMPLENR, 65 - 16},
   {870*OVERSAMPLENR, 60 - 15},
   {892*OVERSAMPLENR, 55 -12},
   {911*OVERSAMPLENR, 50 -11},
   {929*OVERSAMPLENR, 45},
   {944*OVERSAMPLENR, 40},
   {959*OVERSAMPLENR, 35},
   {971*OVERSAMPLENR, 30},
   {981*OVERSAMPLENR, 25},
   {989*OVERSAMPLENR, 20},
   {994*OVERSAMPLENR, 15},
   {1001*OVERSAMPLENR, 10},
   {1005*OVERSAMPLENR, 5}
};
#endif

I also have no idea how the tables where made and there is no information to identify which Epcos thermistor it applies to or what type the "100K" refers to. Type 7 actually specifies the thermistor, but that was also way out for me when used for the bed.

I set the target to each of the temperatures in the table and then measured the actual temperature inside the barrel with a thermocouple and added the offsets shown. I only did the range I am interested in so there are big discontinuities.

Given R25 and beta a table can be constructed using the maths here: [hydraraptor.blogspot.co.uk] There is a Python script kicking about that does that. I don't use the datasheet beta because it is isn't accurate enough and doesn't cover the temperature range we use. I normally calculate it from two temperature measurements and I use my own software which doesn't need a table in the firmware because the maths is done on the host.

Edited 1 time(s). Last edit at 07/06/2012 05:23AM by nophead.

---

[www.hydraraptor.blogspot.com]

I've also found using the beta value from the datasheet gives very inaccurate results, as much as 40% out at 300deg. The beta formula is only good for the specifed range, typically 25-85 deg.

If I need to use a formula, I would use the Steinhart-Hart equation. Admittedly it requires a bit more math, but I borrowed a spreadsheet from a colleague But with 3 data points, it calculates S-H coefficients then any value can be calculated to a small percent error.

If possible I get the resistance values from the datasheet to generate a lookup table.

Edited 1 time(s). Last edit at 07/07/2012 07:54PM by bobc.

Here's a thought, not sure if it's worth anything or not:

With different hotends, thermistors, and materials, temps are mostly relative. Is there a way to calibrate or get the ideal temp for your machine by some observation or test? For example (not the answer), you start at 160 and try to extrude - nothing comes out, you go to 165 and it will begin to extrude, and so the ideal temp is 30 more than that. Again, this is not the answer but an example. There must be something that's observable that will help determine the ideal temp. I know there are better brains than mine that could help make this less of a trial and error thing.