Missing steps a lot...

[hradec]

I'm trying this Marlin_v1 branch, and it seems like my machine is missing steps a lot... Sometimes it's like it completely forget to move in one direction, or move just half of what it should.

Am I missing something important in my Configuration.h that should fix this behaviour or is this some sort of bug?

thanks for any help!

-H

[bkubicek]

there are three main settings relevant to lost steps:
you might want to change the acceleration to something suitable to your machine, probably in the start.gcode.

M204 - Set default acceleration: S normal moves T filament only moves (M204 S3000 T7000) im mm/sec^2 also sets minimum segment time in ms (B20000) to prevent buffer underruns and M20 minimum feedrate

Then you could also lower your travel feedrate and the maximum feedrates:
M203 - Set maximum feedrate that your machine can sustain (M203 X200 Y200 Z300 E10000) in mm/sec

it it happens only at corners lower the xyjerk:
M205 - X= maximum xy jerk, Z=maximum Z jerk

if its because too much material is extruded, lower your e_steps_per_mm
M92 E

Please be aware that all this values are lost on reset.

[hradec]

I tried all of then but M204... I was using M201 for acceleration... is it
different from M204?

also, what happens when a buffer underrun happen? by the sound of it, that
may be my problem...

I known for a fact that its not the stepper that is getting stuck for
something, like too much plastic, since when that happens you can hear that
very characteristic sound of the stepper trying to move but can't... In my
case now, the stepper is moving fine, no weird sound or anything and it
just miss some movements... I can't even notice that something is wrong
unless I keep watching the printing...

So, it does look like something in software...

-H
blog http://hradec.blogspot.com/ home http://hradec.com/

On Mon, Nov 21, 2011 at 2:33 PM, Bernhard Kubicek <
reply@reply.github.com

wrote:

there are three main settings relevant to lost steps:
you might want to change the acceleration to something suitable to your
machine, probably in the start.gcode.

M204 - Set default acceleration: S normal moves T filament only moves
(M204 S3000 T7000) im mm/sec^2 also sets minimum segment time in ms
(B20000) to prevent buffer underruns and M20 minimum feedrate

Then you could also lower your travel feedrate and the maximum feedrates:
M203 - Set maximum feedrate that your machine can sustain (M203 X200 Y200
Z300 E10000) in mm/sec

it it happens only at corners lower the xyjerk:
M205 - X= maximum xy jerk, Z=maximum Z jerk

if its because too much material is extruded, lower your e_steps_per_mm
M92 E

Please be aware that all this values are lost on reset.

---

Reply to this email directly or view it on GitHub:
https://github.com/ErikZalm/Marlin/issues/13#issuecomment-2822350

[ErikZalm]

Hello,

Can you send me the configuration.h ?

Best regards,

Erik van der Zalm

----- Original Message -----
From: "Hradec" reply@reply.github.com
To: "ErikZalm" erik@vdzalm.eu
Sent: Monday, November 21, 2011 11:10:38 PM
Subject: [Marlin] Missing steps a lot... (#13)

I'm trying this Marlin_v1 branch, and it seems like my machine is missing steps a lot... Sometimes it's like it completely forget to move in one direction, or move just half of what it should.

Am I missing something important in my Configuration.h that should fix this behaviour or is this some sort of bug?

thanks for any help!

-H

---

Reply to this email directly or view it on GitHub:
https://github.com/ErikZalm/Marlin/issues/13

[hradec]

sure...

this is my header in each gcode file netfabb generates:

G28
M92 X71.28 Y71.60 Z3236.25 E30.00
M203 X200 Y200 Z100 E3333
M204 S1000.00 T7000.00
M205 S0.00 T0.00 B20000 X0.33 Z0.01
G92 E0
M106

I don't have my configuration.h file here with me right now, but I'll send
it over in a few hours...

cheers...

-H
blog http://hradec.blogspot.com/ home http://hradec.com/

On Tue, Nov 22, 2011 at 2:40 AM, ErikZalm <
reply@reply.github.com

wrote:

Hello,

Can you send me the configuration.h ?

Best regards,

Erik van der Zalm

----- Original Message -----
From: "Hradec" reply@reply.github.com
To: "ErikZalm" erik@vdzalm.eu
Sent: Monday, November 21, 2011 11:10:38 PM
Subject: [Marlin] Missing steps a lot... (#13)

I'm trying this Marlin_v1 branch, and it seems like my machine is missing
steps a lot... Sometimes it's like it completely forget to move in one
direction, or move just half of what it should.

Am I missing something important in my Configuration.h that should fix
this behaviour or is this some sort of bug?

thanks for any help!

-H

---

Reply to this email directly or view it on GitHub:
https://github.com/ErikZalm/Marlin/issues/13

---

Reply to this email directly or view it on GitHub:
https://github.com/ErikZalm/Marlin/issues/13#issuecomment-2832855

[ErikZalm]

Can you test without the M200 lines?

Those are not used by me and could trigger a bug / configuration problem.

You can make the default in the configuration.h

On Nov 22, 2011, at 7:09 PM, Hradec wrote:

sure...

this is my header in each gcode file netfabb generates:

G28
M92 X71.28 Y71.60 Z3236.25 E30.00
M203 X200 Y200 Z100 E3333
M204 S1000.00 T7000.00
M205 S0.00 T0.00 B20000 X0.33 Z0.01
G92 E0
M106

I don't have my configuration.h file here with me right now, but I'll send
it over in a few hours...

cheers...

-H
blog http://hradec.blogspot.com/ home http://hradec.com/

On Tue, Nov 22, 2011 at 2:40 AM, ErikZalm <
reply@reply.github.com

wrote:

Hello,

Can you send me the configuration.h ?

Best regards,

Erik van der Zalm

----- Original Message -----
From: "Hradec" reply@reply.github.com
To: "ErikZalm" erik@vdzalm.eu
Sent: Monday, November 21, 2011 11:10:38 PM
Subject: [Marlin] Missing steps a lot... (#13)

I'm trying this Marlin_v1 branch, and it seems like my machine is missing
steps a lot... Sometimes it's like it completely forget to move in one
direction, or move just half of what it should.

Am I missing something important in my Configuration.h that should fix
this behaviour or is this some sort of bug?

thanks for any help!

-H

---

Reply to this email directly or view it on GitHub:
https://github.com/ErikZalm/Marlin/issues/13

---

Reply to this email directly or view it on GitHub:
https://github.com/ErikZalm/Marlin/issues/13#issuecomment-2832855

---

Reply to this email directly or view it on GitHub:
https://github.com/ErikZalm/Marlin/issues/13#issuecomment-2837874

[bkubicek]

hradec:
@M92: the z value looks very strange, maybe you misplaced the decimal point?
@M203: 200 mm/sec is quite high, but Z100 is insane. maybe you lost steps in Z, and hence had overdeposition and were hitting the already printed things? Z12 is the max on my printer.
@m204:looks allright
@m205: x and Z are very very low for you, usually you want x to be around 10-40, and z between 1-5.

[ErikZalm]

hradec,

When you miss steps. Do you see an endstop message in the host software?
You probably have false endstop triggers.
I will improve the FW to limit the effect of false endstop triggers. But if you have false endstop triggers you should increase the distance between the endstop cables and the motor cables. And even better use shielded cables.

[ErikZalm]

hradec,

I just pushed a new version that is less susceptible to endstop noise.

Can you try this one and report the results?

[cdaringe]

I used to skip all of the time as well, @hradec . my fix was ~reducing #define DEFAULT_XYJERK & #define DEFAULT_ZJERK @ the bottom of the configuration.h. lowering max accel vals seemed to contribute too. unsure which had greater impact.

#define DEFAULT_MAX_ACCELERATION {200,200,100,10000}
#define DEFAULT_XYJERK 10.0 // (mm/sec)
#define DEFAULT_ZJERK 1 // (mm/sec)

[ErikZalm]

If you must lower the acceleration and XY_jerk to these low values you have to check a few things.

1. Are the axis running smooth?
2. Increase the motor current
3. Check for endstop-errors in the pronterface window.
4. Check your motor. (Are the motors ok for reprap applications)

If you have a relative normal machine then the acceleration of 2000 and xy_jerk of 20 should not be a problem.

[cdaringe]

Hi Erik:

I appreciate the input! :)

1. Axis. Smooth as anyone could ask for. However, with decent belt tension, it is a bit tough to spin the motor spindle.
2. Motor current. I could look into that but am unsure on the ramifications to motor health. I am using pololu's. I really don't think this is the solution based on the fact that most prints are healthy
3. No endstop errors.
4. Yes, decent stationary torque and recommend motors on the wiki

Skipped steps are not the status quo. Rather, they tend to show up on thin features when the x or y axis oscillates back and forth & high speeds/frequencies. I believe this is when i occasionally drop a step. I wrote a brute force algorithm in the sprinter experimental github to force the head to stay stationary for x milliseconds on a direction change to alleviate such 'jerk.' I was excited to see that someone was trying to put nophead's frequency reduction algorithm in Marlin--that's exactly what i wanted!

However, i may bring my accel back up & keep my jerk low.

[bkubicek]

if you miss a steps at high frequcency, I think it less a jerk-velocity issue, but your acceleration could be too high.

[bkubicek]

is this still an issue? If so, please reopen the issue.

[cdaringe]

nah it's closed. i ended up w/ a jerk val of ~15, and i did reduce my accel some as you recommended. shes puttin' out some good lookin prints!