I would tend to disagree. An ideal TP should always obey constraints, and this is possible
even with a zero P value (motion simply stops at the corner) and enough lookahead.
The TP slows enough on non-differentiable features that it can both obey constraint
bounds and deviate less than P from the programmed path.

If the TP does not obey constraints, some external and likely less controllable
factor will determine the bounds.
 

I made this work around that seems to be robust.

if cur_task_mode_mon == 1 and cur_interp_state_mon == 1 and cur_queue_mon == 0:
cmd.abort()

M30 is actually read into the buffer and executed, but the inerp has already reset before updating the line number que. That is a bug in Linuxcnc and breaks the standard function of M30 in RS274.

I may simply modify my lcnc to fix the issue. But for now, the above works and should only activate if the yask_mode is manual, which M30 does actually set on M30 read, the task_state is idle, which it should be, and the queue is empty. Which is the planner que.

@ruediger123

Think of it like a train on tracks. Your train has a strict rule: it must never jerk the passengers more than a certain amount. It approaches a sharp 45° bend in the track. The only way to respect the jerk rule through that bend is to slow down enough before it. Fine so far.

But now imagine the track is divided into fixed 1 meter segments and the train can only change speed at the boundary between segments, never in between. The bend does not fall exactly on a segment boundary. So no matter how carefully the train slows down, the moment it crosses the bend it is mid-segment, and the direction change happens in one step. That step is a spike, regardless of how slow the train is going.

Your 1ms cycle time is those fixed segments. The corner of your octagon is the bend that never lands exactly on a boundary. The jerk spike is that one unavoidable step.

The only way to eliminate it completely is to stop the train at the bend, let it come to a full standstill, change direction, then start again. That respects every rule. But it is slow.

Everything else is a compromise: the spike exists, it is small, and a good drive smooths it out internally before it reaches the wheels. For most machines that is perfectly acceptable. But if you need the command stream itself to be clean, a full stop at sharp corners is the only option within your parameters.

@ruediger123
For your exact parameters: 1ms cycle, 50 m/s³ jerk, 5 m/s² acceleration, P0.01 on a 45° corner, the only cornering speed that mathematically fits within the jerk limit on a 1ms grid is zero. Any non-zero cornering speed produces a direction change that cannot be absorbed within the jerk limit given the grid resolution. This is not a tuning issue, it is a fundamental property of discrete-time trajectory execution on a fixed clock.
In practice the planner currently exceeds the limit only slightly, and for most servo drives this is acceptable because the drive interpolates internally at a much higher rate and the mechanical system never sees the spike. What HALscope shows is the command stream quantized at 1ms, not what the motor actually does.
That said, an option to automatically fall back to a complete stop when the jerk-constrained cornering speed rounds down to effectively zero could be useful, giving integrators a way to get mathematically clean HALscope traces at the cost of cycle time on sharp corners. Worth considering as a TP option.

Tommy,

The replies demo what is actually happening in hopes I may get some guidance to fix it. My question is above them.

The attached screenshot of a 20 tooth pocket shows a larger view of the program operation.  It runs by drawing nested figures inside each other by shortening (or lengthening) each side of the figure while drawing it.  The shortening (or lengthening) values are recorded in a table as they are encountered  when the figure is first executed. A multiple of the values is used for each execution of the figure and the values remain constant for the entire pocketing operation. The table starts at memory location 100 and the operators manual says that location 31 to 5000 are "user defined" so we can have lots of sides. Notice that the pocketing operation evolves such that the shown 40 sided figure produces a circular (or 20 sided) area that needs to be Pocketed to complete the entire pocket operation. I doubt that commercially available packages (eg autocad) can duplicate this activity easily since they would require an exhausting specification of each nested figure before executing it.  G-code's ability to process subroutines allows us to have the figure execute itself here for pocketing purposes. It also gives us the ability to hack individual sides and thereby make really ugly figures. 

That's all the information in the PDF.

sudo dmesg -wH
[Apr 4 07:43] EtherCAT 0: Master thread exited.
[  +0.000016] EtherCAT 0: Starting EtherCAT-IDLE thread.
[  +0.000038] EtherCAT ERROR 0-3: Failed to receive AL state datagram: Datagram initialized.
[  +0.000017] EtherCAT 0: Releasing master...
[  +0.000001] EtherCAT 0: Released.
[  +0.024165] EtherCAT ERROR 0-0: AL status message 0x001A: "Synchronization error".
[  +0.001309] EtherCAT 0-0: Acknowledged state SAFEOP.
[  +0.014483] EtherCAT ERROR 0-1: AL status message 0x001A: "Synchronization error".
[  +0.001578] EtherCAT 0-1: Acknowledged state SAFEOP.
[  +0.010458] EtherCAT ERROR 0-2: AL status message 0x001A: "Synchronization error".
[  +0.000839] EtherCAT 0-2: Acknowledged state SAFEOP.
[  +0.011212] EtherCAT ERROR 0-3: AL status message 0x002D: "No Sync Error".
[  +0.000528] EtherCAT 0-3: Acknowledged state SAFEOP.
[  +0.007163] EtherCAT 0: Slave states on main device: PREOP.
[  +0.783938] EtherCAT WARNING: Datagram 000000007d6dd556 (master-fsm) was SKIPPED 1 time.
[ +12.203295] EtherCAT: Requesting master 0...
[  +0.000003] EtherCAT: Successfully requested master 0.
[  +0.001335] EtherCAT 0: Domain0: Logical address 0x00000000, 112 byte, expected working counter 12.
[  +0.000003] EtherCAT 0:   Datagram domain0-0-main: Logical offset 0x00000000, 112 byte, type LRW.
[  +0.000009] EtherCAT 0: Master thread exited.
[  +0.000002] EtherCAT 0: Starting EtherCAT-OP thread.
[  +0.000030] EtherCAT WARNING 0: 1 datagram UNMATCHED!
[  +5.267792] EtherCAT WARNING 0-0: Slave did not sync after 5000 ms.
[  +6.879988] EtherCAT ERROR 0-3: Timeout while setting state OP.
[  +0.007489] EtherCAT 0: Slave states on main device: SAFEOP, OP.
[  +3.132001] EtherCAT 0: Slave states on main device: SAFEOP, OP + ERROR.



 

@grandixximo
According to my understanding, the P-tolerance is intended to prevent the need to slow down to speed 0 in a corner and to reduce the speed before the corner to such an extent that the 'corner' can be driven through without exceeding the jerk and acceleration.

RotarySMP wrote:

I just realised that the main Gmoccapy screen now behaves nicely. However, now I no longer have the tab for Andy's lathe macros.

You still have it - it just moved to the user tab section.

Damn,
I am sick of this device!
Had it working with one register, redid it all properly, wrote a script to set them all and now none of the vfd registers are showing. I'll check dmesg again later....

I'll look at your tip... 

10K wrote:

Problem came back! Something updated, possibly Debian, and now my screen is too wide again. See attached photo. What's different is that I can't set it to Start as Windowed and then back to Start as Fullscreen to fix the problem. Also, the modification to Gmoccapy by alex_sar earlier in this thread does not work anymore, although it did for 6+ months.

My touchscreen monitor is set for native resolution, 1024 x 768

I have the latest version of Debian Bookworm.

I thought that updating everything might help, so I updated LinuxCNC from 2.92 to 2.93 and Gmoccapy from 3.4.6 to 3.4.8. No joy.

Help me newbynobi, you're my only hope!

Can you post your configuration? Its this still an issue with LinuxCNC 2.9.8 ?

I just realised that the main Gmoccapy screen now behaves nicely. However, now I no longer have the tab for Andy's lathe macros.

It's not really an error in your INI file. The lathemacro just needs too much space for some reason and in ntb_user_tabs is just more space available

Hans got me sorted out. I had an error in my INI.
He had me replace:
EMBED_TAB_LOCATION = ntb_preview
with
EMBED_TAB_LOCATION = ntb_user_tabs
Now it has stopped scaling.
Big thanks to Hans for getting that straight.
Cheers,
Happy Mark