Extruder Setup Help

Hi,

I'm having trouble setting up my extruder for Mendel. Configuration is Adrian's extruder - Nema 17, V2.2 controller.

Using Java RepRap host, whatever I adjust in the preferences area the stepper motor turns at the same speed, which is slightly too fast and results in slippage on the feedstock and loss of extrusion. If I turn up the temperature it will keep up OK but the extruded thread is not good enough for anything useful. If I tighten the pinch, then it leads to wedging in the barrel.

Without the motor connected I can turn the gear by hand and it extrudes very well if I keep the speed down, if I go faster I can feel it loose grip on the feedstock as the resistance increases. On the console if I change the "Extrude Speed xx mm/min" value, it alters the time that the extruder runs when I press "Extrude", not the speed. Reverse works OK and runs for the same time at the same speed.

I've been through the manual and tweaking guides on the website and mad a few changes where appropriate but nothing seems to alter the extrude speed. If I can slow the extruder motor down just a little then I'm sure I get it working well enough.

Any help would be much appreciated.

Thanks

Kirtlesog.

Quote
kirtlesog
On the console if I change the "Extrude Speed xx mm/min" value, it alters the time that the extruder runs when I press "Extrude", not the speed. Reverse works OK and runs for the same time at the same speed.

YES, that is all this setting does!
It does NOT change the values used for creating the gcode.

What you need to change is the Extruder0_ExtrusionSpeed.
Do the same for Extruder 1 and 2 while you are at it!

For my Mendel 1950 mm/minute seams to be a good value.
Then load your STL file(s) and generate the gcode file then you then load and print.

---

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

Thanks Bob,

The issue for me seems to be before I get to generating any gcode for parts. I'll try what you suggest with a lower extrusion speed, at present I loose extrusion during the barrel purge, it also seems faster than most of the videos I've seen.

Cheers

Kirltesog

Are you sure you have your E_STEPS_PER_MM set correctly in the firmware?

---

[www.hydraraptor.blogspot.com]

To be honest, no I'm not sure.

After you put me straight on my last error, I realised that the motherboard defines the stepping rate of the extruder and therefore this frequency could be defined in the sanguino firmware, and I should look at this. I didn't see a frequency but if it works this out from steps per mm then that will probably do it. This means recompiling and loading the new firmware, not a problem but a very strange way of doing things.

Cheers

Kirlesog.

Yeah I agree totally it is a very strange way of doing things especially because you have to change it every time you change your filament or your layer height, the latter of which I change several times a day. It is all a consequence of the decision to interpret the gcode on the micro rather than the PC.

---

[www.hydraraptor.blogspot.com]

Quote

It is all a consequence of the decision to interpret the gcode on the micro rather than the PC.

If I understand you correctly, removing the direct connection between XYZ-robot distance and extrusion length would solve this issue. Instead of G1 X10 E10 you'd send G1 X10 E12.5 or G1 X10 E6.8 to extrude 10 millimeters along X, depending on a factor to be set in the host software.

Another solution would be to send some S (spindle speed on CNC mills) command to set this factor at runtime. Firmware would have to be prepared for this, of course.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Yes the problem with sending gcode is that it is all in SI units, so all the conversion is done in the firmware.

I interpret the g-code on the PC and do all the bot specific conversions there, so I never need to change my firmware, even if I change the gear ratio of the extruder. The extruder just gets a binary packet with a number of steps to go and a timer value to use in the interrupt. All the constants that describe my bot, the extruder and the plastic are on the PC, not in the firmware.

---

[www.hydraraptor.blogspot.com]

Hi

Skeinforge has the ability to alter the E value as a proportion of the axis movement.

Controlled by feed rate and flow rate in Speed.

If they are the same flow is 1: 1

If you increase flow rate to 10% greater than feed rate you get 10% more extrusion and so on.

(This may be only in the later versions [fabmetheus.crsndoo.com])

---

Regards

Ian

I reduced the E_STEPS_PER_MM as an experiment and it slowed the extruder down well enough, but it meant that there wasn't enough extrusion compared to XY movement.

Reading all your comments it seems that I need to slow the whole machine down until I can get better force applied to the feedstock without slippage or too much deformation. (I shouldn't need to slow down the Z movement).

Skeinforge sounds good but I wanted to get a basic machine going before getting into the whole subject of software and firmware. I had thought (maybe naively) that the reprap host would give easy control of a reprap machine. Please let me know if Skeinforge is "better". The reprap host seems to be tied up with a lot of legacy information which just confuses the whole issue.

Considering the control system generally, I'm at something of a loss to understand why the extruder board is needed at all: the Atmega 644 is more than capable of doing everything. (but that's for another time).

Cheers

Kirtlesog.

Quote
kirtlesog
Considering the control system generally, I'm at something of a loss to understand why the extruder board is needed at all: the Atmega 644 is more than capable of doing everything. (but that's for another time).

Exactly, that is why the new Pololu based electronics have one Mega board and four Pololu stepper controllers.
No extruder controller is needed and you have a 'real' 1/16 step stepper controller instead of the funny dual bridge.

---

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

Quote
rhmorrison
Exactly, that is why the new Pololu based electronics have one Mega board and four Pololu stepper controllers.
No extruder controller is needed and you have a 'real' 1/16 step stepper controller instead of the funny dual bridge.

Yeah, or one Atmega 644 and four L298's could do the lot with some decent software, including microstepping. But, as I said, that's for another time, first I just want to get printing!

Cheers

Kirtlesog

The reason for having one board for the axis control and a separate board for the tool heads is so you can can have many and varied tool-heads.

The extruder controller is designed for a DC motor based extruder, so is totally not suited to controlling a stepper. What reprap should have done is move to something like this http://hydraraptor.blogspot.com/2009/08/time-for-new-extruder-controller.html, but instead has lurched to a 4 axis board which will look silly when we have two or three extruders.

---

[www.hydraraptor.blogspot.com]

Quote
Nophead
The reason for having one board for the axis control and a separate board for the tool heads is so you can can have many and varied tool-heads.

Good point. Looking at your circuit - it's altogether better. I should have read more before starting!

Cheers

Kirtlesog