Advanced Search

Joco wrote:

Don’t make that assumption re feedhold. And tapping. The planner is taking things direct from the Hal pin. Not from the lcnc motmod module.

it needs to be explicitly tested. Unless someone has done so already and can confirm what the behaviour is. 
 

I can confirm that LCNC 2.10 disables feed-hold while synchronized motion is in progress.  FH has no effect during a G33 or G33.1 command until after the synch move is complete.

cmorley wrote:

I would expect speed changes to change the sync start point.
There is another post about non syncing threads with speed changes.
In linuxcnc, you can't change the spindle speed and track the same thread, the sync point moves.

Linuxcnc moves the axis as fast as possible till it is moving at the approximate correct 'pitch' speed. It then works to get the exact pitch and maintain it. This distance to get up to speed is kept track of for use of calculating position and position error.

IIRC linuxcnc allows 10 rotations past target before deciding there is an error.
 

Thanks for giving this some thought Chris.

I understand that LCNC needs a short period/distance of max accel to go from zero speed to synch speed - the other thread has a code snippet which explicitly states this in a comment.

What I don't get is why LCNC is moving so far to get up to speed.  My lathe has pretty aggressive acceleration - it can reach max speed in a very short distance.

But it's moving over an inch at rapid speed to get to the synch point at only 200rpm with a 0.05" thread pitch.  Feedrate at that spindle rpm and pitch is really slow - like 4ipm or so.

I guess what I'm seeing is identical behavior from the other thread... but much worse start point variation.

The other thread:
forum.linuxcnc.org/38-general-linuxcnc-q...s-with-speed?start=0

If you can use any Linux x86_64 on it, then download the official ISO and try installing it, but might have issues with RT kernel running in a VM.
You can try Debian 12.x or Linux Mint Debian Edition 6, and if they work just install LinuxCNC with
sudo apt install linuxcnc-uspace
This will also install the RT kernel, but it will not be the default one, so LinuxCNC should work OK in SIM mode.

Position mode = 0
Velocity mode = 1
Step type 0 = step/dir
Step type 1 = up/down
Step type 2 = quadrature
I think, been 15 years since i was deep into it, there are 10 step types, so 3 should be 3 phase at 120 degrees, and some more 4 pin types, and 5 pin/phase types.
Yes, LinuxCNC can do 3 and 5 phase stepper control, i still have some of those in the shop, i was very happy when i saw 5 phase motors move using 5 darlington transistors and 5 resistors from a parallel port.

tommylight wrote:

Did you notice i edited the post?
I saw you clicked on thank you after i posted so probably not.

I see it now, i will try your suggestions. Thanks a lot again

tommylight wrote:

No.
Several things will cause that:
-not enough air pressure
-worn nozzle
-to slow cut speed
-
I would also say to much current, but not exactly the same result, so i will just throw a stab in the dark at very, very low accelerations set in the config.

I thought so
I tried with many different air pressures and not much different, that was with 75psi at nozzle.
Consumables was new and speed getting low because i guess is small part??
I will check to set higher accelerations and see.
Accelerations is best to be the same on X-Y axis or not matter?

Thanks a lot

Did you notice i edited the post?
I saw you clicked on thank you after i posted so probably not.

No.
Several things will cause that:
-not enough air pressure
-worn nozzle
-to slow cut speed
-
I would also say to much current, but not exactly the same result, so i will just throw a stab in the dark at very, very low accelerations set in the config.
Edit:
I did not read the settings when i replied, i just browsed over them, so yes it is low acceleration, but probably there is not much more you can do, that is very small for plasma cutting and would in the end require very small nozzles or "fine cut" as Hypertherm calls them.
Now i would say it might be improved a bit, but not "earth shattering" changes.
Try something thicker, at least 2mm or 3mm then some 5 or 6mm.
There is a procedure i do with my machines to tune the cutting and i can not recall if i posted those here, so here it goes in short:
-tune the cut speed and THC roughly with 3mm material
-tune the air pressure till you get the best cut, does not have to be perfect
-tune the THC by raising or lowering the voltage
-tune the feed rate/cut speed
rinse and repeat for other material thicknesses and nozzles.
This is something you have to do for anything that changes, sometimes even moisture in the air will affect the cut quality, and we do get a lot of moisture despite no sea in site. All mild steels are not the same, etc.
Also, after all that you can test with lower or higher cut current and how it effects the cut speed.

How is the drive connected to LinuxCNC?

@Julian,

Thanks for the interesting file.

The first problem is solved :
1. Buffer overflow.
-> cause :  tcqFull() function was not used.
-> solution : tcqFull() function now uses a max buffer size to prevent buffer overflow. Default = 1000.

The second problem is not solved today:
2. It run's up to ca 1000 lines. Then it hang's.
-> cause : tiny segment length's : 0.007mm.
-> solution : review, recode the abc, uvw part of the planner step by step.


 

I fixed the homing.

But I do need to figure out this Z-Axis.

I would expect speed changes to change the sync start point.
There is another post about non syncing threads with speed changes.
In linuxcnc, you can't change the spindle speed and track the same thread, the sync point moves.

Linuxcnc moves the axis as fast as possible till it is moving at the approximate correct 'pitch' speed. It then works to get the exact pitch and maintain it. This distance to get up to speed is kept track of for use of calculating position and position error.

IIRC linuxcnc allows 10 rotations past target before deciding there is an error.

Now it's getting weird.

New discovery - I can get G33.1 to repeat during the same session, but the distance Z rapids is dependent on the spindle speed.

(Starting from G54 Z0.5)
G33.1 Z-0.5 K0.05 $0

• 60rpm: Z-rapids to ~Z0.1 before waiting for the index signal
• 100rpm: Z-rapids to ~Z-0.2 before waiting for index
• 200rpm: Z-rapids to ~Z-0.6 and immediately reverses spindle and moves Z-pos

The above is repeatable as many times as I want within the same LCNC session.

I think my earlier observation about Z-rapiding down and then synchronized movement up was correct... but the 'only works once per session' was incorrect.  I didn't realize that spindle speed was driving the Z-axis 'start' point to different locations.

Why is the Z-axis moving at all - much less a rapid - before the index signal triggers the start of synch'd motion?

Further update

G33.1 sort-of works after discovering an M03 command must preceed the G33.1 line.  Simply turning on the spindle doesn't work.

Starting a Z0.5, when G33.1 is commanded (via MDI) the Z-axis rapids to Z0 and then starts the synchronized move to K-value.  I didn't think that rapid move is correct behavior, but at least it was 'tapping' in and out.

However... G33.1 only works properly once per session.  Subsequent attempts still have the Z-axis diving past Z0 to the K-value at rapid speed, reversing the spindle, and then a synchronized move back to the Z-start location.  Spindle reverses again at that point, ending the sequence.

Restarting LCNC doesn't change the 'works once, then never again' behavior.

I've gone over all the spindle, encoder, c-axis, and PID settings & connections with a fine-toothed comb and can't see a problem.

• encoder velocity: S200 M03 $0 results in 200rpm, verified with a tach
• encoder position: 1 revolution of the spindle increases position by 1
  • encoder position is connected to spindle.0.revs per the manual
• encoder sign: rotating the spindle forward (i.e. M03 direction) increases the encoder counts & position, and rotating it backwards decreases counts and position
• encoder scale: 1 turn of the spindle increases counts by 40000 (10kppr encoder)
• index: manually set signal index-enable HIGH and rotate the spindle.  Signal goes LOW when the index pin is triggered, and it happens at the same place every time.
  • encoder, spindle, and both velocity & axis PID index-enable pins are all connected and react together.
• spindle-at-speed works as expected, and is connected to spindle.0.at-speed

At no point can I see the index-enable signal go high and then low when G33.1 is commanded.  If I set the spindle speed very low (30rpm), I would expect to see the index-enable signal/pins all get set HIGH by LCNC prior to starting the synchronized move.  But no change in state, even in halscope.

I'm at a loss here.