Backlash compensation

RAyWB · April 2016

Hi,
tried backlash compensation on threaded rod drive machine fired by due and have a question:
when getting to the turnover point where direction changes machine seems to stop shortly.
clearly visible when printing circles.
if i set backlash to 0 it has normal behavior. Is there a possibility to speed up ?

stopping position is visible in printed part :-(

Repetier · April 2016

You see backlash logic in motion.cpp line 206-242.

As you see all it does is adding a second move on direction change. That also needs 2 free moves to place so it reduces buffer space for travel distance. So all I can think of is that your max. buffer size should be increased to at least 32 to accomodate for extra moves without penalty for path planner, which is especially important in circles which I guess is what you tested.

RAyWB · May 2016

buffer size is 40 , also tried 80.
i do the test with circles , thats right, but on turnover point machine nearly stops.
if i set backlash values to 0 it runs absolutely smooth.
as soon as i put a value (in my case i need 0.005mm to 0.009) it slows down extremely just at turnover point.
machine step size is 0.00166 mm , so i need 3 to 5 steps.

is there a change of feedrate in the added moves?
any idea?

Repetier · May 2016

No feedrate stays unchanged.

Does it also happen in a simple case with 2 lines? Then you could add

#define DEBUG_QUEUE_MOVE

and enable echo before running. It will out to log the moves and speeds it computes for this. You might need 1-2 pre moves to fill buffer.

RAyWB · May 2016

seems to happen illogical
using cambam gcode with G3 circles everything is ok also circles by G1 lines.

here´s a snippet from gcode where that slow down happens(extruder is commented, feedrate set to 2400mm/min) :

Gcode is generated by slic3r , also Cura or S3D makes not a difference

any value in backlash slows down in turnover point, set values to 0 it runs smooth.
may be you can check it out , its really strange

G1 X291.517 Y113.106 F2400;E6.02030 F3600.000
G1 X291.795 Y114.315 ;E6.08249
G1 X291.994 Y115.535 ;E6.14451
G1 X292.113 Y116.763 ;E6.20640
G1 X292.153 Y118.000 ;E6.26846
G1 X292.113 Y119.237 ;E6.33051
G1 X291.994 Y120.465 ;E6.39240
G1 X291.795 Y121.685 ;E6.45443
G1 X291.517 Y122.894 ;E6.51661
G1 X291.165 Y124.072 ;E6.57830
G1 X290.735 Y125.233 ;E6.64039
G1 X290.228 Y126.367 ;E6.70270
G1 X289.657 Y127.452 ;E6.76420
G1 X289.011 Y128.510 ;E6.82635
G1 X288.299 Y129.520 ;E6.88834
G1 X287.524 Y130.483 ;E6.95032
G1 X286.681 Y131.401 ;E7.01285
G1 X286.401 Y131.681 ;E7.03272
G1 X285.479 Y132.528 ;E7.09548
G1 X284.528 Y133.294 ;E7.15677
G1 X283.510 Y134.011 ;E7.21922
G1 X282.452 Y134.657 ;E7.28137
G1 X281.367 Y135.228 ;E7.34286
G1 X280.233 Y135.735 ;E7.40518
G1 X279.072 Y136.165 ;E7.46726
G1 X277.894 Y136.518 ;E7.52896
G1 X276.685 Y136.795 ;E7.59114
G1 X275.465 Y136.994 ;E7.65317
G1 X274.237 Y137.113 ;E7.71506
G1 X273.000 Y137.153 ;E7.77711
G1 X271.763 Y137.113 ;E7.83917
G1 X270.535 Y136.994 ;E7.90106
G1 X269.315 Y136.795 ;E7.96308
G1 X268.106 Y136.518 ;E8.02526
G1 X266.928 Y136.165 ;E8.08696
G1 X265.767 Y135.735 ;E8.14905
G1 X264.633 Y135.228 ;E8.21136
G1 X263.548 Y134.657 ;E8.27286
G1 X262.490 Y134.011 ;E8.33501
G1 X261.472 Y133.294 ;E8.39745
G1 X260.521 Y132.528 ;E8.45875
G1 X259.599 Y131.681 ;E8.52150
G1 X259.319 Y131.401 ;E8.54138
G1 X258.476 Y130.483 ;E8.60390
G1 X257.701 Y129.520 ;E8.66588
G1 X256.989 Y128.510 ;E8.72787
G1 X256.343 Y127.452 ;E8.79002
G1 X255.772 Y126.367 ;E8.85152
G1 X255.265 Y125.233 ;E8.91383
G1 X254.835 Y124.072 ;E8.97592
G1 X254.482 Y122.894 ;E9.03761
G1 X254.205 Y121.685 ;E9.09980
G1 X254.006 Y120.465 ;E9.16182
G1 X253.887 Y119.237 ;E9.22371
G1 X253.847 Y118.000 ;E9.28576
G1 X253.887 Y116.763 ;E9.34782
G1 X254.006 Y115.535 ;E9.40971
G1 X254.205 Y114.315 ;E9.47173
G1 X254.482 Y113.106 ;E9.53392
G1 X254.835 Y111.928 ;E9.59561
G1 X255.265 Y110.767 ;E9.65770
G1 X255.772 Y109.633 ;E9.72001
G1 X256.343 Y108.548 ;E9.78151
G1 X256.989 Y107.490 ;E9.84366
G1 X257.701 Y106.480 ;E9.90565
G1 X258.476 Y105.517 ;E9.96763
G1 X259.319 Y104.599 ;E10.03015
G1 X259.599 Y104.319 ;E10.05003
G1 X260.521 Y103.472 ;E10.11278
G1 X261.472 Y102.706 ;E10.17408
G1 X262.490 Y101.989 ;E10.23652
G1 X263.548 Y101.343 ;E10.29868
G1 X264.633 Y100.772 ;E10.36017
G1 X265.767 Y100.265 ;E10.42248
G1 X266.928 Y99.835 ;E10.48457
G1 X268.106 Y99.482 ;E10.54627
G1 X269.315 Y99.205 ;E10.60845
G1 X270.535 Y99.006 ;E10.67044
G1 X271.765 Y98.887 ;E10.73243
G1 X273.000 Y98.847 ;E10.79444
G1 X273.407 Y98.859 ;E10.81483
G1 X274.235 Y98.887 ;E10.85640
G1 X275.465 Y99.006 ;E10.91840
G1 X276.685 Y99.205 ;E10.98039
G1 X277.890 Y99.482 ;E11.04238
G1 X279.074 Y99.836 ;E11.10436
G1 X280.238 Y100.267 ;E11.16663
G1 X280.609 Y100.428 ;E11.18689
G1 X281.359 Y100.768 ;E11.22820
G1 X282.455 Y101.344 ;E11.29033
G1 X283.517 Y101.993 ;E11.35273
G1 X283.850 Y102.220 ;E11.37292
G1 X284.517 Y102.698 ;E11.41410
G1 X285.483 Y103.476 ;E11.47629
G1 X286.401 Y104.319 ;E11.53881
G1 X286.681 Y104.599 ;E11.55869
G1 X287.528 Y105.521 ;E11.62145
G1 X288.048 Y106.155 ;E11.66258
G1 X288.294 Y106.472 ;E11.68275
G1 X289.011 Y107.490 ;E11.74519
G1 X289.657 Y108.548 ;E11.80734
G1 X290.045 Y109.273 ;E11.84860
G1 X290.228 Y109.633 ;E11.86884
G1 X290.735 Y110.767 ;E11.93115
G1 X291.165 Y111.928 ;E11.99324
G1 X291.387 Y112.660 ;E12.03162
;G1 X291.806 Y112.598 F6000.000
G1 X291.923 Y113.001 F2400;E12.05265 F3600.000
G1 X292.207 Y114.234 ;E12.11610
G1 X292.410 Y115.481 ;E12.17947
G1 X292.532 Y116.737 ;E12.24276
G1 X292.572 Y118.000 ;E12.30614
G1 X292.532 Y119.263 ;E12.36953
G1 X292.410 Y120.519 ;E12.43282
G1 X292.207 Y121.766 ;E12.49618
G1 X291.923 Y122.999 ;E12.55964
G1 X291.562 Y124.206 ;E12.62282
G1 X291.123 Y125.391 ;E12.68622
G1 X290.606 Y126.548 ;E12.74975
G1 X290.021 Y127.661 ;E12.81281
G1 X289.361 Y128.740 ;E12.87625
G1 X288.634 Y129.772 ;E12.93959
G1 X287.842 Y130.756 ;E13.00293
G1 X286.984 Y131.691 ;E13.06659
G1 X286.691 Y131.984 ;E13.08736
G1 X285.754 Y132.844 ;E13.15115
G1 X284.777 Y133.631 ;E13.21409
G1 X283.740 Y134.361 ;E13.27770
G1 X282.661 Y135.021 ;E13.34114
G1 X281.548 Y135.606 ;E13.40420
G1 X280.391 Y136.123 ;E13.46773
G1 X279.206 Y136.562 ;E13.53113
G1 X277.999 Y136.923 ;E13.59431
G1 X276.766 Y137.207 ;E13.65777
G1 X275.519 Y137.410 ;E13.72113
G1 X274.263 Y137.532 ;E13.78442
G1 X273.000 Y137.572 ;E13.84781
G1 X271.737 Y137.532 ;E13.91119
G1 X270.481 Y137.410 ;E13.97448
G1 X269.234 Y137.207 ;E14.03785
G1 X268.001 Y136.923 ;E14.10131
G1 X266.794 Y136.562 ;E14.16448
G1 X265.609 Y136.123 ;E14.22788
G1 X264.452 Y135.606 ;E14.29142
G1 X263.339 Y135.021 ;E14.35448
G1 X262.260 Y134.361 ;E14.41791
G1 X261.223 Y133.631 ;E14.48152
G1 X260.246 Y132.844 ;E14.54446
G1 X259.309 Y131.984 ;E14.60825
G1 X259.016 Y131.691 ;E14.62903
G1 X258.158 Y130.756 ;E14.69268
G1 X257.366 Y129.772 ;E14.75602
G1 X256.639 Y128.740 ;E14.81936
G1 X255.979 Y127.661 ;E14.88280
G1 X255.394 Y126.548 ;E14.94586
G1 X254.877 Y125.391 ;E15.00939
G1 X254.438 Y124.206 ;E15.07280
G1 X254.077 Y122.999 ;E15.13597
G1 X253.793 Y121.766 ;E15.19943
G1 X253.590 Y120.519 ;E15.26280
G1 X253.468 Y119.263 ;E15.32609
G1 X253.428 Y118.000 ;E15.38947
G1 X253.468 Y116.737 ;E15.45285
G1 X253.590 Y115.481 ;E15.51614
G1 X253.793 Y114.234 ;E15.57951
G1 X254.077 Y113.001 ;E15.64297
G1 X254.438 Y111.794 ;E15.70614
G1 X254.877 Y110.609 ;E15.76955
G1 X255.394 Y109.452 ;E15.83308
G1 X255.979 Y108.339 ;E15.89614
G1 X256.639 Y107.260 ;E15.95958
G1 X257.366 Y106.228 ;E16.02292
G1 X258.158 Y105.244 ;E16.08626
G1 X259.016 Y104.309 ;E16.14991
G1 X259.309 Y104.016 ;E16.17069
G1 X260.246 Y103.156 ;E16.23448
G1 X261.223 Y102.369 ;E16.29742
G1 X262.260 Y101.639 ;E16.36103
G1 X263.339 Y100.979 ;E16.42447
G1 X264.452 Y100.394 ;E16.48753
G1 X265.609 Y99.877 ;E16.55106
G1 X266.794 Y99.438 ;E16.61446
G1 X268.001 Y99.077 ;E16.67764
G1 X269.234 Y98.793 ;E16.74110
G1 X270.481 Y98.590 ;E16.80444
G1 X271.738 Y98.468 ;E16.86779
G1 X273.000 Y98.428 ;E16.93115
G1 X273.420 Y98.441 ;E16.95222
G1 X274.262 Y98.468 ;E16.99447
G1 X275.519 Y98.590 ;E17.05783
G1 X276.766 Y98.793 ;E17.12117
G1 X277.997 Y99.076 ;E17.18452
G1 X279.207 Y99.438 ;E17.24786
G1 X280.394 Y99.879 ;E17.31137
G1 X280.778 Y100.045 ;E17.33237
G1 X281.543 Y100.391 ;E17.37447
G1 X282.662 Y100.980 ;E17.43790
G1 X283.744 Y101.641 ;E17.50148
G1 X284.089 Y101.876 ;E17.52244
G1 X284.771 Y102.364 ;E17.56447
G1 X285.756 Y103.158 ;E17.62793
G1 X286.691 Y104.016 ;E17.69158
G1 X286.984 Y104.309 ;E17.71236
G1 X287.844 Y105.246 ;E17.77615
G1 X288.375 Y105.893 ;E17.81815
G1 X288.631 Y106.223 ;E17.83910
G1 X289.361 Y107.260 ;E17.90270
G1 X290.021 Y108.339 ;E17.96614
G1 X290.417 Y109.079 ;E18.00822
G1 X290.606 Y109.452 ;E18.02920
G1 X291.123 Y110.609 ;E18.09274
G1 X291.562 Y111.794 ;E18.15614
G1 X291.789 Y112.541 ;E18.19528

Repetier · May 2016

Can't say what is so special on that gcode. Looks like 2 circles. Difference might be the direction switch is like a triangle, but I do not find that really special.

There might be a chance that you have

REDUCE_ON_SMALL_SEGMENTS

defined somewhere. That will reduce speed for direction changes based on direction AND segment length. Smaller = Slower, and since backlash is small that might be the reason for a slowdown.

RAyWB · May 2016

didn´t define that, only found it in motion.cpp
#ifdef REDUCE_ON_SMALL_SEGMENTS

what about jerk? i have jerk 7.0 cannot go higher because my moving table´s weight is around 12Kg.
if i increase jerk it´s like an earthquake...

another question regarding circles :

tried to mill some round pockets diameter 6mm with 3mm flute using G2 , worked but segments are clearly visible.

i found :
#define MM_PER_ARC_SEGMENT 1
#define MM_PER_ARC_SEGMENT_BIG 3

thats ok for printing but not good enough for milling

are float numbers allowed?
so may be 0.1 or smaller?

Repetier · May 2016

Yes, the number gets converted to float anyway (in case it was integer), that is no problem.

Regarding jerk you might have seen the experimental defines

#define ALTERNATIVE_JERK

should give better behavior. REDUCE_ON_SMALL_SEGMENTS is for cases where you switch direction very fast having a small move between/at end that could cause bigger real jerks if direction changes are too fast in a row.

But the slowdown is then still a mystery. No luck on finding a short motion list you could debug with output created using echo as described above?

RAyWB · May 2016

Ok,thanks.

I´ll try to reduce the Gcode above to a part of just one circle to see if that still happens.

Hope i´ll find the time to investigate next week.

another finding regarding G-code handling (killed some flutes yesterday):

Cambam adds some single command line for spindle Speed ,

for example just like this:

G1 .....

S1000

G0 Z3

G0 X....

So the line S1000 is a unknown command but the following G0 Z3 is not executed

is there a Chance to catch this ?

I think of something like

void Commands::processSCode(GCode *com) {

switch ( com->S) .....

or just adding a if(com->hasS()) to processGCode before switch ( com->G)

what would allow to set Spindle Speed or laser intensity without a move.

should be not hard to implement.

(easiest would be just to do nothing when detecting the S-line :-) )

Repetier · May 2016

S1000 gets ignored. Just run a test

S1000

M117 Hi

on Repetier-Host with this result:

10:11:02.865 : N41 M451*22

10:11:02.865 : N42 S1000*26

10:11:02.865 : Printjob finished at 19.05.2016 10:11

10:11:02.865 : Printing Time: 0s

10:11:02.865 : Lines Send: 5

10:11:02.890 : N43 M117 Hi*18

10:11:02.893 : N44 M104 T0 S0*17

10:11:02.893 : N45 M104 T1 S0*17

10:11:02.893 : N46 M84*45

10:11:02.975 : Disabled volumetric extrusion for extruder 0

10:11:02.975 : Disabled volumetric extrusion for extruder 1

10:11:02.976 : PrinterMode:FFF

10:11:02.981 : Unknown command:N42 S1000

10:11:03.158 : Successfully send push message.

As you see firmware detected it as unknown command and next line was send.

What did you use for sending it? Maybe the host/firmware combination makes a difference.

There is no branch to handle S only commands. We only support G/M/T as command identifier. Everything else gets ignored.

RAyWB · May 2016

sent it by Repetier host,

Printer mode CNC, and yes, the S1000 was ignored .

My Problem is , that the move following the S1000 was not executed and this is reproducable.

didn´t check it with a M117, but i´ll try that

will post code and log this evening

RAyWB · May 2016

Ok, skipping commands just happens in ASCII protocol.

see snippet from LOG using Ascii protocol:

17:55:20.874 : N24 M117 ETE 3m 04s *111
17:55:20.874 : N25 M105 *48
17:55:20.874 : N26 G1 Y155.5 *111
17:55:20.874 : N27 G1 X294 *122
17:55:20.874 : N28 G1 Y144.5 *97
17:55:20.874 : N29 G1 X280.4 *107
17:55:20.874 : ok 24
17:55:20.874 : N30 S1000 *63
17:55:20.874 : N31 G0 Z3 *114
17:55:21.342 : ok 25
17:55:21.342 : ok 26
17:55:21.342 : ok 27
17:55:21.342 : ok 28
17:55:21.342 : N32 M117 ETE 3m 00s *108
17:55:21.342 : N33 G0 X182.7 Y150 *46
17:55:21.342 : N34 G0 Z1 *117
17:55:21.342 : ok 29
17:55:21.342 : Error:Format error
17:55:21.342 : Resend:30
17:55:21.358 : ok
17:55:21.358 : skip 31
17:55:21.358 : ok
17:55:21.358 : Resend: N30 S1000 *63
17:55:21.358 : Resend: N31 G0 Z3 *114
17:55:21.358 : Resend: N32 M117 ETE 3m 00s *108
17:55:21.358 : Resend: N33 G0 X182.7 Y150 *46
17:55:21.358 : Resend: N34 G0 Z1 *117
17:55:21.358 : N35 G1 Z-0.4 F300 *22
17:55:21.358 : skip 32
17:55:21.358 : ok
17:55:21.358 : skip 33
17:55:21.358 : ok
17:55:21.358 : skip 34
17:55:21.358 : ok

with Repetier protocol it´s OK

RAyWB · May 2016

but with repetier protocol i get :

22:19:07.496 : Error:Checksum required when switching back to ASCII protocol.

when using buttons assigned to

case UI_ACTION_X_ZERO:
            if(!allowMoves) return UI_ACTION_X_ZERO;
            GCode::executeFString(PSTR("G92 X0\n"));
            break;

Repetier · May 2016

I see I'm in compatibility vs. error hell.

I made more strict tests to detect errenous injection of wrong gcode throug communication errors. And now they are firing back. Will check it.

RAyWB · May 2016

i added a handling for S commands, works in repetier protocol,
but still get format error in ascii protocol...
cannot see the trap, ist there something wrong in ascii parsing?

void Commands::processSCode(GCode *com) {
uint32_t codenum; //throw away variable

codenum=(com->S) ;
{
#if defined(SUPPORT_LASER) && SUPPORT_LASER
if (Printer::mode == PRINTER_MODE_LASER)
{
LaserDriver::intensity = constrain(codenum, 0, 255);
Com::printFLN(PSTR("Laserint:"), (int)LaserDriver::intensity);
}
#endif // defined
#if defined(SUPPORT_CNC) && SUPPORT_CNC
if (Printer::mode == PRINTER_MODE_CNC)
{
Com::printFLN(PSTR("Spindle s "),(int)codenum);
}
#endif // defined
}
}

RAyWB · May 2016

back to original problem with reducing speed with backlash compensation:
seems to come from jerk , no improvement with alternative jerk
as it doesn´t appear with circles generated by Cam program i guess problem is located in
STL-file with too big segments.
i´ll try to change my model to smaller segments in CAD and see what happens

here´s the log:

10:17:33.216 : N912 G1 X291.517 Y113.106 F2400 *110
10:17:33.216 : N913 G1 X291.795 Y114.315 *0
10:17:33.216 : N914 G1 X291.994 Y115.535 *13
10:17:33.216 : N915 G1 X292.113 Y116.763 *10
10:17:33.216 : N916 G1 X292.153 Y118 *31
10:17:33.216 : N917 G1 X292.113 Y119.237 *3
10:17:33.216 : N918 G1 X291.994 Y120.465 *3
10:17:33.216 : N919 G1 X291.795 Y121.685 *0
10:17:33.231 : ID:537350200
10:17:33.231 : N920 M104 T0 S0 *10
10:17:33.231 : vStart/End:417/417
10:17:33.231 : accel/decel steps:1/1/3
10:17:33.231 : st./end speed:0.7/0.7
10:17:33.231 : Flags:146
10:17:33.247 : joinFlags:5
10:17:33.247 : ID:537350200
10:17:33.247 : Delta 0 3 0 0
10:17:33.247 : Dir:36
10:17:33.247 : Flags:18
10:17:33.247 : fullSpeed:0.70
10:17:33.247 : vMax:417
10:17:33.247 : Acceleration:8.96
10:17:33.263 : Acceleration Prim:537600
10:17:33.263 : Remaining steps:3
10:17:33.263 : LimitInterval:50350
10:17:33.278 : Move distance on the XYZ space:0.00
10:17:33.278 : Commanded feedrate:40.00
10:17:33.294 : Constant full speed move time:151050.00
10:17:33.294 : ID:537339140
10:17:33.294 : vStart/End:1921/1921
10:17:33.294 : accel/decel steps:876/876/6589
10:17:33.309 : st./end speed:2.5/2.5
10:17:33.309 : Flags:146
10:17:33.309 : joinFlags:5
10:17:33.309 : ID:537339140
10:17:33.309 : Delta 148 6589 0 0
10:17:33.309 : Dir:60
10:17:33.309 : Flags:18
10:17:33.309 : fullSpeed:40.00
10:17:33.325 : vMax:30746
10:17:33.325 : Acceleration:12007.04
10:17:33.325 : Acceleration Prim:537600
10:17:33.325 : Remaining steps:6589
10:17:33.341 : LimitInterval:683
10:17:33.341 : Move distance on the XYZ space:8.58
10:17:33.341 : Commanded feedrate:40.00
10:17:33.356 : Constant full speed move time:4506555.00
10:17:33.356 : Echo:N912 G1 X291.52 Y113.11 F2400.00
10:17:33.356 : N921 M104 T1 S0 *10
10:17:33.356 : N922 M140 S0 *76
10:17:33.356 : ID:537339280
10:17:33.372 : vStart/End:167/167
10:17:33.372 : accel/decel steps:1/1/3
10:17:33.372 : st./end speed:0.4/0.4
10:17:33.372 : Flags:146
10:17:33.387 : joinFlags:5
10:17:33.387 : ID:537339280
10:17:33.387 : Delta 2 3 0 0
10:17:33.387 : Dir:55
10:17:33.387 : Flags:18
10:17:33.387 : fullSpeed:0.40
10:17:33.387 : vMax:168
10:17:33.403 : Acceleration:17.92
10:17:33.403 : Acceleration Prim:537600
10:17:33.403 : Remaining steps:3
10:17:33.403 : LimitInterval:124381
10:17:33.419 : Move distance on the XYZ space:0.01
10:17:33.419 : Commanded feedrate:40.00
10:17:33.419 : Constant full speed move time:373143.84
10:17:33.419 : ID:537339420
10:17:33.434 : vStart/End:1872/1872
10:17:33.434 : accel/decel steps:464/464/929
10:17:33.434 : st./end speed:2.5/2.5
10:17:33.434 : Flags:146
10:17:33.450 : joinFlags:5
10:17:33.450 : ID:537339420
10:17:33.450 : Delta 148 929 0 0
10:17:33.450 : Dir:63
10:17:33.450 : Flags:18
10:17:33.465 : fullSpeed:40.00
10:17:33.465 : vMax:29957
10:17:33.465 : Acceleration:1781.72
10:17:33.465 : Acceleration Prim:537600
10:17:33.465 : Remaining steps:929
10:17:33.481 : LimitInterval:701
10:17:33.481 : Move distance on the XYZ space:1.24
10:17:33.481 : Commanded feedrate:40.00
10:17:33.497 : Constant full speed move time:651555.31
10:17:33.497 : Echo:N913 G1 X291.80 Y114.31
10:17:33.497 : N923 M84 *55
10:17:33.497 : N924 M105 *8
10:17:33.497 : ID:537339420 MJ:40.00
10:17:33.497 : ID:537339420
10:17:33.512 : vStart/End:1872/29957
10:17:33.512 : accel/decel steps:832/1/929
10:17:33.512 : st./end speed:2.5/40.0
10:17:33.512 : Flags:146
10:17:33.528 : joinFlags:7
10:17:33.528 : ID:537339560
10:17:33.528 : vStart/End:30345/1896
10:17:33.528 : accel/decel steps:1/854/937
10:17:33.543 : st./end speed:40.0/2.5
10:17:33.543 : Flags:146
10:17:33.543 : joinFlags:5
10:17:33.543 : ID:537339560
10:17:33.543 : Delta 106 937 0 0
10:17:33.543 : Dir:63
10:17:33.559 : Flags:18
10:17:33.559 : fullSpeed:40.00
10:17:33.559 : vMax:30346
10:17:33.559 : Acceleration:1751.87
10:17:33.559 : Acceleration Prim:537600
10:17:33.575 : Remaining steps:937
10:17:33.575 : LimitInterval:692
10:17:33.575 : Move distance on the XYZ space:1.24
10:17:33.575 : Commanded feedrate:40.00
10:17:33.590 : Constant full speed move time:648970.68
10:17:33.590 : Echo:N914 G1 X291.99 Y115.54
10:17:33.699 : ID:537339700
10:17:33.699 : vStart/End:1913/1913
10:17:33.715 : accel/decel steps:471/471/943
10:17:33.715 : st./end speed:2.5/2.5
10:17:33.715 : Flags:146
10:17:33.715 : joinFlags:5
10:17:33.715 : ID:537339700
10:17:33.731 : Delta 64 943 0 0
10:17:33.731 : Dir:63
10:17:33.731 : Flags:18
10:17:33.731 : fullSpeed:40.00
10:17:33.731 : vMax:30612
10:17:33.731 : Acceleration:1733.28
10:17:33.746 : Acceleration Prim:537600
10:17:33.746 : Remaining steps:943
10:17:33.746 : LimitInterval:686
10:17:33.746 : Move distance on the XYZ space:1.23
10:17:33.762 : Commanded feedrate:40.00
10:17:33.762 : Constant full speed move time:647700.12
10:17:33.762 : Echo:N915 G1 X292.11 Y116.76
10:17:33.777 : ID:537339840
10:17:33.777 : vStart/End:1921/1921
10:17:33.777 : accel/decel steps:474/474/950
10:17:33.793 : st./end speed:2.5/2.5
10:17:33.793 : Flags:146
10:17:33.793 : joinFlags:5
10:17:33.793 : ID:537339840
10:17:33.793 : Delta 21 950 0 0
10:17:33.793 : Dir:63
10:17:33.793 : Flags:18
10:17:33.809 : fullSpeed:40.00
10:17:33.809 : vMax:30746
10:17:33.809 : Acceleration:1731.14
10:17:33.809 : Acceleration Prim:537600
10:17:33.824 : Remaining steps:950
10:17:33.824 : LimitInterval:683
10:17:33.824 : Move distance on the XYZ space:1.24
10:17:33.840 : Commanded feedrate:40.00
10:17:33.840 : Constant full speed move time:649743.06
10:17:33.840 : Echo:N916 G1 X292.15 Y118.00
10:17:33.840 : ID:537339980
10:17:33.855 : vStart/End:112/112
10:17:33.855 : accel/decel steps:0/0/2
10:17:33.855 : st./end speed:0.3/0.3
10:17:33.855 : Flags:146
10:17:33.871 : joinFlags:5
10:17:33.871 : ID:537339980
10:17:33.871 : Delta 2 0 0 0
10:17:33.871 : Dir:22
10:17:33.871 : Flags:18
10:17:33.871 : fullSpeed:0.28
10:17:33.871 : vMax:112
10:17:33.887 : Acceleration:9.33
10:17:33.887 : Acceleration Prim:373333
10:17:33.887 : Remaining steps:2
10:17:33.887 : LimitInterval:186843
10:17:33.902 : Move distance on the XYZ space:0.00
10:17:33.902 : Commanded feedrate:40.00
10:17:33.902 : Constant full speed move time:373687.50
10:17:33.918 : ID:537340120
10:17:33.918 : vStart/End:1921/1921
10:17:33.918 : accel/decel steps:474/474/950
10:17:33.918 : st./end speed:2.5/2.5
10:17:33.918 : Flags:146
10:17:33.933 : joinFlags:5
10:17:33.933 : ID:537340120
10:17:33.933 : Delta 21 950 0 0
10:17:33.933 : Dir:62
10:17:33.933 : Flags:18
10:17:33.949 : fullSpeed:40.00
10:17:33.949 : vMax:30746
10:17:33.949 : Acceleration:1731.14
10:17:33.949 : Acceleration Prim:537600
10:17:33.949 : Remaining steps:950
10:17:33.965 : LimitInterval:683
10:17:33.965 : Move distance on the XYZ space:1.24
10:17:33.965 : Commanded feedrate:40.00
10:17:33.980 : Constant full speed move time:649743.06
10:17:33.980 : Echo:N917 G1 X292.11 Y119.24
10:17:33.980 : ID:537340260
10:17:33.996 : vStart/End:1913/1913
10:17:33.996 : accel/decel steps:471/471/943
10:17:33.996 : st./end speed:2.5/2.5
10:17:33.996 : Flags:146
10:17:33.996 : joinFlags:5
10:17:33.996 : ID:537340260
10:17:34.011 : Delta 64 943 0 0
10:17:34.011 : Dir:62
10:17:34.011 : Flags:18
10:17:34.011 : fullSpeed:40.00
10:17:34.011 : vMax:30612
10:17:34.027 : Acceleration:1733.28
10:17:34.027 : Acceleration Prim:537600
10:17:34.027 : Remaining steps:943
10:17:34.027 : LimitInterval:686
10:17:34.043 : Move distance on the XYZ space:1.23
10:17:34.043 : Commanded feedrate:40.00
10:17:34.043 : Constant full speed move time:647700.12
10:17:34.058 : Echo:N918 G1 X291.99 Y120.46
10:17:34.058 : ID:537340400
10:17:34.058 : vStart/End:1896/1896
10:17:34.074 : accel/decel steps:468/468/937
10:17:34.074 : st./end speed:2.5/2.5
10:17:34.074 : Flags:146
10:17:34.074 : joinFlags:5
10:17:34.074 : ID:537340400
10:17:34.074 : Delta 106 937 0 0
10:17:34.074 : Dir:62
10:17:34.089 : Flags:18
10:17:34.089 : fullSpeed:40.00
10:17:34.089 : vMax:30346
10:17:34.089 : Acceleration:1751.87
10:17:34.089 : Acceleration Prim:537600
10:17:34.105 : Remaining steps:937
10:17:34.105 : LimitInterval:692
10:17:34.105 : Move distance on the XYZ space:1.24
10:17:34.121 : Commanded feedrate:40.00
10:17:34.121 : Constant full speed move time:648970.68
10:17:34.136 : Echo:N919 G1 X291.80 Y121.68

RAyWB · May 2016

update : Stl changed to very small segments, still same behavior

Speed changed to st./end speed:0.3/0.3 in turnover points

Repetier · May 2016

Any idea what 0.3 is? If I'm right your jerk is 5 resulting in 2.5 mm start/end speed.

Commanded speed is is 40mm/s so no need to go lower. safeSpeed would be 2.5 so if anything puts it down it must be computeMaxJunctionSpeed() function returning lower. So adding a print here with id might show if that is the case and then it should be easy to find the formula causing it.

I have just released a new development version where I fixed the communication error you had plus some other improvements. It also simplifies now minimum speed handling. Most of the code has been removed as I now check in bresenham for it allowing nearly any speed you want. Was a problem with CNC/laser sometimes that minimum speed forced higher speeds as wanted.

RAyWB · May 2016

0.3 in my setup is z-jerk, will change it to see if value in log also changes.
also will try the development release tomorrow.

RAyWB · May 2016

youre right, happens in computeMaxJunctionSpeed()
i added some prints, found it happens here:

   float jerk = sqrt(dx * dx + dy * dy) * lengthFactor;
    Com::printFLN(PSTR("jerk "),(int)(jerk*100));
#endif // ALTERNATIVE_JERK
#endif // DELTA
    if(jerk > Printer::maxJerk) {
        {factor = Printer::maxJerk / jerk; // always < 1.0!
        Com::printFLN(PSTR("jerk>pJerk "),(int)(factor*100));}
        if(factor * maxJoinSpeed * 2.0 < Printer::maxJerk)
        {factor = Printer::maxJerk / (2.0 * maxJoinSpeed);
       Com::printFLN(PSTR(" factor<pjerk "),(int)(factor*100));}
    }
......
   previous->maxJunctionSpeed = maxJoinSpeed * factor; // set speed limit
    Com::printFLN(PSTR("speed limit"),previous->maxJunctionSpeed);

so Log says :
23:50:47.729 : jerk 257 //2.57
23:50:47.729 : N56 G1 X265.767 Y135.735 *59
23:50:47.729 : speed limit40.00
23:50:47.729 : ok 49
23:50:47.729 : N57 G1 X264.633 Y135.228 *50
23:50:47.729 : jerk 254 //2.54
23:50:47.729 : speed limit40.00
23:50:47.729 : ok 50
23:50:47.729 : jerk 258 //2.58
23:50:47.745 : N58 G1 X263.548 Y134.657 *56
23:50:47.745 : speed limit40.00
23:50:47.745 : ok 51
23:50:47.745 : jerk 5563 //55.63
23:50:47.745 : jerk>pJerk 8 //0.08
23:50:47.745 : N59 G1 X262.49 Y134.011 *8
23:50:47.745 : speed limit3.59
23:50:47.745 : jerk 5748 //57.48
23:50:47.745 : jerk>pJerk 8 //0.08
23:50:47.760 : speed limit3.48
23:50:47.760 : ok 52
23:50:47.760 : jerk 258 //2.58
23:50:47.760 : N60 G1 X261.472 Y133.294 *53
23:50:47.760 : speed limit40.00
23:50:47.760 : ok 53
23:50:47.760 : N61 G1 X260.521 Y132.528 *51

values are *100 correct ones in comments

so it seems backlash compensation causes high jerk values .
do you have any idea how to solve this?

Repetier · May 2016

The only way I see is if you have reduce on small segments enabled

#ifdef REDUCE_ON_SMALL_SEGMENTS
	if(previous->distance < MAX_JERK_DISTANCE)
	lengthFactor = static_cast<float>(MAX_JERK_DISTANCE * MAX_JERK_DISTANCE) / (previous->distance * previous->distance);
	#endif

That can lead to high lengthFactors. While some reduction is wanted I guess this is simply too much. But assuming your correction is 0.02mm and MAX_JERK_DISTANCE 0.6 (not sure what it is) your factor becomes 900 which is a quite big and causes a big reduction as you see. Witout it jerk would be quite low. Only thing is how to reduce correctly without loosing the reason for it. So smaller segments should reduce jerk but limited somehow.

RAyWB · May 2016

ok, but REDUCE_ON_SMALL_SEGMENTS is for shure not active.
anyway, as backlash on my machine is just 5 micron i´ll disable correction for printing.
for milling and laser i´ll activate it.

does backlash compensation generate separate moves instead or add the backlash steps to existing move?

Repetier · May 2016

Could you write out dx,dy and lengthFactor as well? If you did not have it on it must give lengthFactor = 1, but then I wonder what dx/dy contain that sqrt of them is such a large number.

Backlash adds a additional move before the move changing direction. In your example it was just 2 steps in x direction.

RAyWB · May 2016

Ok,
changed prints:

   float dx = current->speedX - previous->speedX;
    float dy = current->speedY - previous->speedY;
    float jerk = sqrt(dx * dx + dy * dy) * lengthFactor;
     Com::printFLN(PSTR("dx "),(int)(dx*100));
     Com::printFLN(PSTR("dy "),(int)(dy*100));
     Com::printFLN(PSTR("jerk "),(int)(jerk*100));
     Com::printFLN(PSTR("LF "),(int)(lengthFactor*100));
#endif // ALTERNATIVE_JERK
#endif // DELTA
    if(jerk > Printer::maxJerk) {
        {factor = Printer::maxJerk / jerk; // always < 1.0!
        Com::printFLN(PSTR("Printermax "),(int)(Printer::maxJerk*100) );
        Com::printFLN(PSTR("jerk>pJerk "),(int)(factor*100));}
        if(factor * maxJoinSpeed * 2.0 < Printer::maxJerk)
        {factor = Printer::maxJerk / (2.0 * maxJoinSpeed);
       Com::printFLN(PSTR(" factor<pjerk "),(int)(factor*100));}
    }
.......
previous->maxJunctionSpeed = maxJoinSpeed * factor; // set speed limit
    Com::printFLN(PSTR("speed limit"),previous->maxJunctionSpeed);

as before factors are *100

so you see lengthFactor is 1.00 everyhere.
i wonder about that huge dx/dy numbers.

22:14:42.034 : N427 G1 X291.517 Y113.106 F2400 *101
22:14:42.034 : N428 G1 X291.795 Y114.315 *5
22:14:42.034 : N429 G1 X291.994 Y115.535 *14
22:14:42.034 : N430 G1 X292.113 Y116.763 *0
22:14:42.034 : N431 G1 X292.153 Y118 *23
22:14:42.034 : N432 G1 X292.113 Y119.237 *9
22:14:42.034 : N433 G1 X291.994 Y120.465 *7
22:14:42.034 : N434 G1 X291.795 Y121.685 *2
22:14:42.034 : ok 427
22:14:42.034 : ok 428
22:14:42.034 : dx -2835
22:14:42.034 : N435 G1 X291.517 Y122.894 *6
22:14:42.034 : N436 G1 X291.165 Y124.072 *2
22:14:42.034 : dy 2454 //24.54
22:14:42.034 : jerk 3750 //37.50
22:14:42.034 : LF 100 //1.00
22:14:42.050 : Printermax 500 //5.00
22:14:42.050 : jerk>pJerk 13 //0.13
22:14:42.050 : speed limit5.33
22:14:42.050 : ok 429
22:14:42.050 : dx -251 //-2.51
22:14:42.050 : dy 49 //0.49
22:14:42.050 : N437 G1 X290.735 Y125.233 *7
22:14:42.065 : jerk 256 //2.56
22:14:42.065 : LF 100 //1.00
22:14:42.065 : speed limit40.00
22:14:42.065 : ok 430
22:14:42.065 : N438 G1 X290.228 Y126.367 *2
22:14:42.065 : dx -254 //-2.54
22:14:42.065 : dy 33 //0.33
22:14:42.065 : jerk 256 //2.56
22:14:42.065 : LF 100 //1.00
22:14:42.081 : speed limit40.00
22:14:42.081 : ok 431
22:14:42.081 : dx -261 //-2.61
22:14:42.081 : N439 G1 X289.657 Y127.452 *7
22:14:42.081 : dy 16 //0.16
22:14:42.081 : jerk 262 //2.62
22:14:42.081 : LF 100 //1.00
22:14:42.081 : speed limit40.00
22:14:42.081 : ok 432
22:14:42.081 : N440 G1 X289.011 Y128.51 *53
22:14:42.081 : dx -44 //-.44
22:14:42.081 : dy -3997 //-39.97
22:14:42.096 : jerk 3998 //39.98
22:14:42.096 : LF 100 //1.00
22:14:42.096 : Printermax 500 //5.00
22:14:42.096 : jerk>pJerk 12 //0.12
22:14:42.096 : factor<pjerk 300 //3.00
22:14:42.096 : speed limit2.50
22:14:42.096 : dx -210 //-2.10
22:14:42.112 : dy 3997 //39.97
22:14:42.112 : jerk 4003 //40.03
22:14:42.112 : LF 100 //1.00
22:14:42.112 : Printermax 500 //5.00
22:14:42.112 : jerk>pJerk 12 //0.12
22:14:42.112 : factor<pjerk 300 //3.00
22:14:42.112 : speed limit2.50
22:14:42.128 : ok 433
22:14:42.128 : dx -261 //-2.61
22:14:42.128 : dy -16 //-0.16
22:14:42.128 : jerk 262 //2.62
22:14:42.128 : LF 100 //1.00
22:14:42.128 : speed limit40.00
22:14:42.128 : ok 434
22:14:42.128 : N441 M117 ETE 00s *34
22:14:42.143 : dx -254
22:14:42.143 : dy -33
22:14:42.143 : jerk 256
22:14:42.143 : LF 100
22:14:42.143 : speed limit40.00
22:14:42.143 : ok 435
22:14:42.143 : dx -251
22:14:42.143 : dy -49
22:14:42.143 : N442 G1 X288.299 Y129.52 *54
22:14:42.143 : jerk 256
22:14:42.143 : LF 100
22:14:42.159 : speed limit40.00
22:14:42.159 : ok 436
22:14:42.159 : N443 G1 X287.524 Y130.483 *9
22:14:42.159 : dx -253
22:14:42.159 : dy -66
22:14:42.159 : jerk 261
22:14:42.159 : LF 100
22:14:42.159 : speed limit40.00
22:14:42.159 : ok 437
22:14:42.159 : dx -239
22:14:42.159 : N444 G1 X286.681 Y131.401 *8
22:14:42.174 : dy -80
22:14:42.174 : jerk 252
22:14:42.174 : LF 100
22:14:42.174 : speed limit40.00
22:14:42.174 : ok 438
22:14:42.174 : N445 G1 X286.401 Y131.681 *9
22:14:42.174 : dx -248
22:14:42.174 : dy -101
22:14:42.174 : jerk 268
22:14:42.174 : LF 100
22:14:42.190 : speed limit40.00
22:14:42.190 : ok 439
22:14:42.190 : dx -225
22:14:42.190 : N446 G1 X285.479 Y132.528 *5
22:14:42.190 : dy -109
22:14:42.190 : jerk 250
22:14:42.190 : LF 100
22:14:42.190 : speed limit40.00
22:14:42.190 : ok 440
22:14:42.190 : dx -224
22:14:42.190 : dy -127
22:14:42.190 : N447 G1 X284.528 Y133.294 *1
22:14:42.206 : jerk 258
22:14:42.206 : LF 100
22:14:42.206 : speed limit40.00
22:14:42.206 : ok 441
22:14:42.206 : N448 G1 X283.51 Y134.011 *58
22:14:42.206 : N449 G1 X282.452 Y134.657 *9
22:14:42.315 : ok 442
22:14:42.315 : N450 G1 X281.367 Y135.228 *14
22:14:42.315 : dx -221
22:14:42.315 : dy -145
22:14:42.315 : jerk 265
22:14:42.315 : LF 100
22:14:42.330 : speed limit40.00
22:14:42.330 : ok 443
22:14:42.330 : dx -198
22:14:42.330 : N451 G1 X280.233 Y135.735 *7
22:14:42.330 : dy -149
22:14:42.330 : jerk 248
22:14:42.330 : LF 100
22:14:42.330 : speed limit40.00
22:14:42.330 : ok 444
22:14:42.330 : dx -197
22:14:42.330 : dy -169
22:14:42.330 : N452 G1 X279.072 Y136.165 *5
22:14:42.346 : jerk 260
22:14:42.346 : LF 100
22:14:42.346 : speed limit40.00
22:14:42.346 : ok 445
22:14:42.346 : N453 G1 X277.894 Y136.518 *4
22:14:42.346 : dx -44
22:14:42.346 : dy -212
22:14:42.346 : jerk 217
22:14:42.346 : LF 100
22:14:42.346 : speed limit38.78
22:14:42.362 : ok 446
22:14:42.362 : dx -197
22:14:42.362 : N454 G1 X276.685 Y136.795 *11
22:14:42.362 : dy -29
22:14:42.362 : jerk 199
22:14:42.362 : LF 100
22:14:42.362 : speed limit38.78
22:14:42.362 : ok 447
22:14:42.362 : N455 G1 X275.465 Y136.994 *10
22:14:42.362 : dx -169
22:14:42.362 : dy -197
22:14:42.377 : jerk 260
22:14:42.377 : LF 100
22:14:42.377 : speed limit40.00
22:14:42.377 : ok 448
22:14:42.377 : dx -156
22:14:42.377 : N456 G1 X274.237 Y137.113 *15
22:14:42.377 : dy -207
22:14:42.377 : jerk 260
22:14:42.393 : LF 100
22:14:42.393 : speed limit40.00
22:14:42.393 : ok 449
22:14:42.393 : dx -140
22:14:42.393 : dy -213
22:14:42.393 : N457 G1 X273 Y137.153 *21
22:14:42.393 : jerk 255
22:14:42.393 : LF 100
22:14:42.393 : speed limit40.00
22:14:42.393 : ok 450
22:14:42.393 : N458 G1 X271.763 Y137.113 *0
22:14:42.408 : dx -129
22:14:42.408 : dy -227
22:14:42.408 : jerk 261
22:14:42.408 : LF 100
22:14:42.408 : speed limit40.00
22:14:42.408 : ok 451
22:14:42.408 : dx -110
22:14:42.408 : N459 G1 X270.535 Y136.994 *7
22:14:42.408 : dy -227
22:14:42.408 : jerk 252
22:14:42.408 : LF 100
22:14:42.424 : speed limit40.00
22:14:42.424 : ok 452
22:14:42.424 : dx -99
22:14:42.424 : dy -242
22:14:42.424 : N460 G1 X269.315 Y136.795 *14
22:14:42.424 : jerk 262
22:14:42.424 : LF 100
22:14:42.424 : speed limit40.00
22:14:42.533 : ok 453
22:14:42.533 : dx -81
22:14:42.533 : N461 G1 X268.106 Y136.518 *9
22:14:42.533 : dy -242
22:14:42.549 : jerk 255
22:14:42.549 : LF 100
22:14:42.549 : speed limit40.00
22:14:42.549 : ok 454
22:14:42.549 : N462 G1 X266.928 Y136.165 *14
22:14:42.549 : dx -67
22:14:42.549 : dy -254
22:14:42.549 : jerk 262
22:14:42.549 : LF 100
22:14:42.564 : speed limit40.00
22:14:42.564 : ok 455
22:14:42.564 : dx -49
22:14:42.564 : Druck fertig gestellt am 23.05.2016 22:14
22:14:42.564 : Druckzeit: 0s
22:14:42.564 : Gesendete Zeilen: 36
22:14:42.564 : dy -252
22:14:42.564 : jerk 257
22:14:42.564 : LF 100
22:14:42.564 : speed limit40.00
22:14:42.564 : N463 G1 X265.767 Y135.735 *9
22:14:42.564 : ok 456
22:14:42.564 : dx -32
22:14:42.564 : dy -252
22:14:42.564 : N464 M104 T0 S0 *7
22:14:42.580 : jerk 254
22:14:42.580 : LF 100
22:14:42.580 : speed limit40.00
22:14:42.580 : ok 457
22:14:42.580 : dx -16
22:14:42.580 : dy -257
22:14:42.580 : jerk 258
22:14:42.580 : N465 M104 T1 S0 *7
22:14:42.580 : N466 M140 S0 *65
22:14:42.580 : LF 100
22:14:42.580 : speed limit40.00
22:14:42.596 : ok 458
22:14:42.596 : dx 3997
22:14:42.596 : N467 M84 *58
22:14:42.596 : dy 3869
22:14:42.596 : jerk 5563
22:14:42.596 : LF 100
22:14:42.596 : Printermax 500
22:14:42.596 : jerk>pJerk 8
22:14:42.596 : speed limit3.59
22:14:42.611 : dx -3997
22:14:42.611 : dy -4130
22:14:42.611 : jerk 5748
22:14:42.611 : LF 100
22:14:42.611 : Printermax 500
22:14:42.611 : jerk>pJerk 8
22:14:42.611 : speed limit3.48
22:14:42.611 : ok 459
22:14:42.627 : dx 16
22:14:42.627 : dy -257
22:14:42.627 : jerk 258
22:14:42.627 : LF 100
22:14:42.627 : speed limit40.00
22:14:42.627 : ok 460
22:14:42.627 : dx 32
22:14:42.627 : dy -252
22:14:42.642 : jerk 254
22:14:42.642 : LF 100
22:14:42.642 : speed limit40.00
22:14:42.642 : ok 461
22:14:42.642 : dx 49
22:14:42.642 : dy -252
22:14:42.642 : jerk 257
22:14:42.642 : LF 100
22:14:42.658 : speed limit40.00
22:14:42.658 : ok 462
22:14:42.658 : dx 67
22:14:42.658 : dy -254
22:14:42.658 : jerk 262
22:14:42.658 : LF 100
22:14:42.658 : speed limit40.00
22:14:42.658 : ok 463
22:14:42.658 : dx 81
22:14:42.658 : dy -242
22:14:42.674 : jerk 255
22:14:42.674 : LF 100
22:14:42.674 : speed limit40.00

Repetier · May 2016

Ok, we com ecloser. The huge dx/dy values are surely the reason for the problem. Looks like it is assuming a 45° move - no move - 45° move from the speed values. That is what wonders me most as I assume you are testing a circle part. Can you post the code tested and your steps per mm for x/y and movement speed (I assume 60 or 40 from the values). Maybe replaying the critical move while reading the code will help me seing where the speed difference comes from.

RAyWB · May 2016

ok, my machines axis steps are not symmetrical for test i set them the same value but still slows down in turnover point.

steps per mm y 768
steps per mm x 533.333
feedrate 40 mm/s

testing code is part of a circle :

G1 X291.517 Y113.106 F2400
G1 X291.795 Y114.315
G1 X291.994 Y115.535
G1 X292.113 Y116.763
G1 X292.153 Y118.000
G1 X292.113 Y119.237
G1 X291.994 Y120.465
G1 X291.795 Y121.685
G1 X291.517 Y122.894
G1 X291.165 Y124.072
G1 X290.735 Y125.233
G1 X290.228 Y126.367
G1 X289.657 Y127.452
G1 X289.011 Y128.510
G1 X288.299 Y129.520
G1 X287.524 Y130.483
G1 X286.681 Y131.401
G1 X286.401 Y131.681
G1 X285.479 Y132.528
G1 X284.528 Y133.294
G1 X283.510 Y134.011
G1 X282.452 Y134.657
G1 X281.367 Y135.228
G1 X280.233 Y135.735
G1 X279.072 Y136.165
G1 X277.894 Y136.518
G1 X276.685 Y136.795
G1 X275.465 Y136.994
G1 X274.237 Y137.113
G1 X273.000 Y137.153
G1 X271.763 Y137.113
G1 X270.535 Y136.994
G1 X269.315 Y136.795
G1 X268.106 Y136.518
G1 X266.928 Y136.165
G1 X265.767 Y135.735

also happens with circle from cam program :

; Made using CamBam - http://www.cambam.info
; Ohne Namen 5/23/2016 11:02:34 AM

G0 X51.0 Y26.0
G1 X50.99 Y26.707 F2400
G1 X50.96 Y27.4135
G1 X50.91 Y28.1188
G1 X50.8402 Y28.8225
G1 X50.7504 Y29.5239
G1 X50.6408 Y30.2224
G1 X50.5116 Y30.9176
G1 X50.3627 Y31.6089
G1 X50.1943 Y32.2957
G1 X50.0066 Y32.9774
G1 X49.7996 Y33.6535
G1 X49.5737 Y34.3236
G1 X49.3288 Y34.9869
G1 X49.0653 Y35.6431
G1 X48.7834 Y36.2916
G1 X48.4832 Y36.9318
G1 X48.1651 Y37.5633
G1 X47.8292 Y38.1855
G1 X47.4758 Y38.798
G1 X47.1053 Y39.4003
G1 X46.7179 Y39.9918
G1 X46.3139 Y40.5721
G1 X45.8936 Y41.1408
G1 X45.4575 Y41.6974
G1 X45.0057 Y42.2414
G1 X44.5388 Y42.7724
G1 X44.057 Y43.29
G1 X43.5608 Y43.7937
G1 X43.0506 Y44.2833
G1 X42.5267 Y44.7582
G1 X41.9896 Y45.2181
G1 X41.4396 Y45.6626
G1 X40.8774 Y46.0914
G1 X40.3032 Y46.5041
G1 X39.7176 Y46.9004
G1 X39.121 Y47.28
G1 X38.514 Y47.6426
G1 X37.8969 Y47.9878
G1 X37.2703 Y48.3155
G1 X36.6346 Y48.6253
G1 X35.9905 Y48.917
G1 X35.3384 Y49.1904
G1 X34.6788 Y49.4452
G1 X34.0123 Y49.6813
G1 X33.3393 Y49.8984
G1 X32.6605 Y50.0964
G1 X31.9763 Y50.2752
G1 X31.2874 Y50.4345
G1 X30.5943 Y50.5742
G1 X29.8974 Y50.6943
G1 X29.1975 Y50.7947
G1 X28.495 Y50.8752
G1 X27.7905 Y50.9358
G1 X27.0845 Y50.9765
G1 X26.3777 Y50.9971
G1 X25.6706 Y50.9978
G1 X24.9638 Y50.9785
G1 X24.2578 Y50.9392
G1 X23.5531 Y50.88
G1 X22.8505 Y50.8008
G1 X22.1503 Y50.7018
G1 X21.4533 Y50.5831
G1 X20.7598 Y50.4446
G1 X20.0706 Y50.2867
G1 X19.3861 Y50.1093
G1 X18.7069 Y49.9126
G1 X18.0336 Y49.6967
G1 X17.3666 Y49.462
G1 X16.7065 Y49.2084
G1 X16.0538 Y48.9363
G1 X15.4091 Y48.6458
G1 X14.7729 Y48.3372
G1 X14.1457 Y48.0108
G1 X13.5279 Y47.6667
G1 X13.5 Y47.6506
G1 X12.8927 Y47.2885
G1 X12.2959 Y46.9092
G1 X11.71 Y46.5133
G1 X11.1356 Y46.101
G1 X10.573 Y45.6725
G1 X10.0228 Y45.2284
G1 X9.4854 Y44.7688
G1 X8.9612 Y44.2943
G1 X8.4506 Y43.8051
G1 X7.9541 Y43.3016
G1 X7.472 Y42.7843
G1 X7.0047 Y42.2536
G1 X6.5527 Y41.7099
G1 X6.1161 Y41.1536
G1 X5.6955 Y40.5852
G1 X5.2912 Y40.0051
G1 X4.9034 Y39.4139
G1 X4.5324 Y38.8119
G1 X4.1787 Y38.1996
G1 X3.8424 Y37.5776
G1 X3.5238 Y36.9463
G1 X3.2232 Y36.3063
G1 X2.9409 Y35.658
G1 X2.677 Y35.002
G1 X2.4317 Y34.3388
G1 X2.2053 Y33.6689
G1 X1.9979 Y32.9929
G1 X1.8098 Y32.3113
G1 X1.6409 Y31.6246
G1 X1.4916 Y30.9334
G1 X1.3619 Y30.2383
G1 X1.2519 Y29.5398
G1 X1.1617 Y28.8385
G1 X1.0913 Y28.1349
G1 X1.0409 Y27.4296
G1 X1.0105 Y26.7231
G1 X1.0 Y26.0161
G1 X1.0095 Y25.3091
G1 X1.0391 Y24.6026
G1 X1.0886 Y23.8972
G1 X1.158 Y23.1935
G1 X1.2473 Y22.4921
G1 X1.3564 Y21.7934
G1 X1.4853 Y21.0982
G1 X1.6337 Y20.4068
G1 X1.8016 Y19.7199
G1 X1.9889 Y19.0381
G1 X2.1954 Y18.3618
G1 X2.421 Y17.6916
G1 X2.6654 Y17.0281
G1 X2.9284 Y16.3718
G1 X3.21 Y15.7231
G1 X3.5097 Y15.0827
G1 X3.8275 Y14.451
G1 X4.163 Y13.8285
G1 X4.5159 Y13.2158
G1 X4.8861 Y12.6133
G1 X5.2731 Y12.0216
G1 X5.6767 Y11.441
G1 X6.0966 Y10.872
G1 X6.5324 Y10.3152
G1 X6.9838 Y9.7709
G1 X7.4504 Y9.2396
G1 X7.9318 Y8.7217
G1 X8.4277 Y8.2176
G1 X8.9376 Y7.7277
G1 X9.4612 Y7.2525
G1 X9.9981 Y6.7922
G1 X10.5477 Y6.3474
G1 X11.1097 Y5.9182
G1 X11.6836 Y5.5051
G1 X12.2689 Y5.1084
G1 X12.8653 Y4.7285
G1 X13.4721 Y4.3655
G1 X13.5 Y4.3494
G1 X14.1173 Y4.0045
G1 X14.7441 Y3.6772
G1 X15.38 Y3.3678
G1 X16.0243 Y3.0765
G1 X16.6766 Y2.8036
G1 X17.3363 Y2.5492
G1 X18.003 Y2.3135
G1 X18.6761 Y2.0969
G1 X19.355 Y1.8993
G1 X20.0393 Y1.721
G1 X20.7283 Y1.5621
G1 X21.4216 Y1.4228
G1 X22.1185 Y1.3032
G1 X22.8185 Y1.2033
G1 X23.5211 Y1.1232
G1 X24.2256 Y1.063
G1 X24.9316 Y1.0228
G1 X25.6384 Y1.0026
G1 X26.3455 Y1.0024
G1 X27.0523 Y1.0222
G1 X27.7583 Y1.0619
G1 X28.4629 Y1.1216
G1 X29.1655 Y1.2012
G1 X29.8656 Y1.3007
G1 X30.5626 Y1.4199
G1 X31.2559 Y1.5587
G1 X31.945 Y1.7172
G1 X32.6294 Y1.895
G1 X33.3085 Y2.0921
G1 X33.9817 Y2.3084
G1 X34.6486 Y2.5436
G1 X35.3085 Y2.7976
G1 X35.961 Y3.0701
G1 X36.6055 Y3.361
G1 X37.2415 Y3.67
G1 X37.8685 Y3.9969
G1 X38.486 Y4.3413
G1 X39.0936 Y4.7031
G1 X39.6907 Y5.0819
G1 X40.2768 Y5.4775
G1 X40.8515 Y5.8895
G1 X41.4143 Y6.3175
G1 X41.9648 Y6.7613
G1 X42.5025 Y7.2205
G1 X43.027 Y7.6948
G1 X43.5379 Y8.1836
G1 X44.0347 Y8.6868
G1 X44.5172 Y9.2037
G1 X44.9848 Y9.7341
G1 X45.4372 Y10.2776
G1 X45.8741 Y10.8336
G1 X46.2951 Y11.4017
G1 X46.6998 Y11.9815
G1 X47.088 Y12.5725
G1 X47.4593 Y13.1743
G1 X47.8134 Y13.7863
G1 X48.1501 Y14.4081
G1 X48.4691 Y15.0392
G1 X48.7701 Y15.6791
G1 X49.0529 Y16.3272
G1 X49.3172 Y16.983
G1 X49.5629 Y17.646
G1 X49.7898 Y18.3158
G1 X49.9976 Y18.9917
G1 X50.1862 Y19.6732
G1 X50.3554 Y20.3597
G1 X50.5052 Y21.0508
G1 X50.6354 Y21.7458
G1 X50.7458 Y22.4442
G1 X50.8365 Y23.1455
G1 X50.9073 Y23.849
G1 X50.9582 Y24.5543
G1 X50.9891 Y25.2607
G1 X51.0 Y25.9678
G1 Y26.0

Repetier · May 2016

Ok, I now understand it and the logic is correct. In the circle you have only small angle changes but at backlash position this changes. There you have this

G1 X275.465 Y136.994

G1 X274.237 Y137.113 dir -1,228/0,119

G1 X273.000 Y137.153 dir -1,237/0,04

Here backlash move comes 0/-0,01 or whatever you have

dVx around 40mm/s as it stops, dVy -40 as we invert speed direction to downmove

G1 X271.763 Y137.113 dir -1,237/-0,04

G1 X270.535 Y136.994

So in this example of you you go left and up until the apex. There you switch from nearly left to pure down move = 90° direction change and then continue nearly pure left again having again 90° direction change.

Since the backlash move is not distinguishable the move planner sees a zig zag and reduces speed accordingly. When it does so it does not know that the direction is very short and it will continue old direction next move. With jerk 5 this speed approximates then to 5/sqrt(2). That is fully correct also completely unwanted.

But how could such a thing be detected correctly. If you move along a square you also get 90° change and that is fully correct here to slow down as the BIG move is really 90°. So the stupid solution always works and safes print but looses sometimes quality for unnecessary slowdowns.

RAyWB · May 2016

Ok, thanks.
so I´ll disable backlash in printing mode because error produced by slow down is much higher than my backlash.
(i have short bowden extruder).
for milling and laser i keep it because there is no change result because of precision and of coarse slower
feedrates

thank you for your effort!!

lukie80 · October 2016

Sorry for digging out this old post.

Is it possible with low effort to inject backlash compensation moves after the planner? Or is it possible to mark them as "ignore" for the planner? Could this approach eliminate the described issue because backlash-compensating moves are ideally load-free or in other words they do not move the carriage?

Repetier · October 2016

Every move needs to go through planner as the planner computes important parameter. A marker would be possible I guess, but that makes planning very complicated as you always have to check for that extra case.

aiz · May 2018

Hi,

I know I'm coming back to an old topic.

Would it be possible to set different jerk, accelerations or feedrate for the small movement of the backlash compensation?

I think that the problem is that the printer movement is almost stopped, not because the jerk is too small but because the initial speed of that segment (jerk/2) is too small.

For example, I'm printing a circle at 50 mm/s and when I set jerk at 100 (I know is too high) it prints without stopping at the point of the backlash compensation.

In my understanding, the backlash movement could be done very quickly because it is not really moving any mass. It is unloading the belts and loading them again in the opposite direction, but it is not really moving the extruder. That's why I understand that the backlash movement could be done at very high rates, much higher than normally used for printing, and this could be a solution for this topic.

I've tried to add a local change of jerk at motion.cpp at the "#if ENABLE_BACKLASH_COMPENSATION" (line 258 and line 412), first increasing the jerk to a very high value and leaving it as the original jerk at the #endif, but it didn't work.

The modification was the following:

#if ENABLE_BACKLASH_COMPENSATION
    if((p->isXYZMove()) && ((p->dir & XYZ_DIRPOS) ^ (Printer::backlashDir & XYZ_DIRPOS)) & (Printer::backlashDir >> 3)) { // We need to compensate backlash, add a move

GCode::executeFString(PSTR("M207 X100")); //change of jerk when doing a backlash compensation

        waitForXFreeLines(2);
        uint8_t wpos2 = linesWritePos + 1;
        if(wpos2 >= PRINTLINE_CACHE_SIZE) wpos2 = 0;
        PrintLine *p2 = &lines[wpos2];
        memcpy(p2, p, sizeof(PrintLine)); // Move current data to p2
        uint8_t changed = (p->dir & XYZ_DIRPOS) ^ (Printer::backlashDir & XYZ_DIRPOS);
        float back_diff[4]; // Axis movement in mm
        back_diff[E_AXIS] = 0;
        back_diff[X_AXIS] = (changed & 1 ? (p->isXPositiveMove() ? Printer::backlashX : -Printer::backlashX) : 0);
        back_diff[Y_AXIS] = (changed & 2 ? (p->isYPositiveMove() ? Printer::backlashY : -Printer::backlashY) : 0);
        back_diff[Z_AXIS] = (changed & 4 ? (p->isZPositiveMove() ? Printer::backlashZ : -Printer::backlashZ) : 0);
        p->dir &= XYZ_DIRPOS; // x,y and z are already correct
        for(uint8_t i = 0; i < 4; i++) {
            float f = back_diff[i] * Printer::axisStepsPerMM[i];
            p->delta[i] = abs((long)f);
            if(p->delta[i]) p->dir |= XSTEP << i;
        }
        //Define variables that are needed for the Bresenham algorithm. Please note that Z is not currently included in the Bresenham algorithm.
        if(p->delta[Y_AXIS] > p->delta[X_AXIS] && p->delta[Y_AXIS] > p->delta[Z_AXIS]) p->primaryAxis = Y_AXIS;
        else if (p->delta[X_AXIS] > p->delta[Z_AXIS] ) p->primaryAxis = X_AXIS;
        else p->primaryAxis = Z_AXIS;
        p->stepsRemaining = p->delta[p->primaryAxis];
        //Feedrate calc based on XYZ travel distance
        xydist2 = back_diff[X_AXIS] * back_diff[X_AXIS] + back_diff[Y_AXIS] * back_diff[Y_AXIS];
        if(p->isZMove())
            p->distance = sqrt(xydist2 + back_diff[Z_AXIS] * back_diff[Z_AXIS]);
        else
            p->distance = sqrt(xydist2);
        // 56 seems to be xstep|ystep|e_posdir which just seems odd
        Printer::backlashDir = (Printer::backlashDir & 56) | (p2->dir & XYZ_DIRPOS);
        p->calculateMove(back_diff, pathOptimize, p->primaryAxis);
        p = p2; // use saved instance for the real move

GCode::executeFString(PSTR("M207 X10")); //back to original jerk after the backlash compensation
    }
#endif

Any thoughts?

Thank you in advance

Backlash compensation

Comments