Advanced Search

(LCNC 2.9.8, new APT develop installation of Probe Basic)

Trying to get all the user buttons, ATC buttons, user tabs, etc to show up and I'm struggling.  PB starts up ok.

Edited INI file:
Copied all the template folders to my config folder.

Edited the copied template UI files slightly so I could see if 'my' versions were loading, but left them named 'template_user_nnn..."

Started PB and no user main/sidebar tabs were present.  User button area was empty/blank, as was the ATC user button area.

Is there something else I have to edit?  Any of the .py files need paths or file renaming?

TANDEM Y AXIS CLOSED LOOP STEPPER MOTOR OSCILLATION WHEN IDLE

Hello all, 

threre is a stepper motor ticking issue going in my cnc mill that has been bothering me for a while now. I have devoted a fair amount of effort trying to "trial and error" my way out of it by playing around with tuning parameters but i cant reach a satisfactory solution, so hopefully someone here can provide some much needed assistance or reference to a similar case if the topic has already been covered before.

CONTROL SETUP DESCRIPTION
In my setup i use a rpi4b as the linuxcnc host along with an EC300 bob in which i have flashed the 3.1.3 version of the remora firmware for the RT1052 processor. Rpi and EC300 exchange packets over ethernet. 4 integrated closed loop stepper motors are driven by the the EC300 board using the corresponding STEP/DIR pins. Feedback from the motor's encoder does not reach linuxcnc so it essentially acts an an open loop system.

INITIAL ISSUE DESCRIPTION
In the capture5.txt i have recorded the following signals of J1 (one of the two joints driving the Y axis, the other is J3) in the servo_thread period while the joint reached the idle state after executing a jog command in my initial attempts to get my machine going.

1st column       : j1pos-cmd
2nd column      : j1pos-fb
3rd column       : j1freq-cmd

at the time of the recording J1 was configured this way (scale and the rest of the dynamic parameters remained the same to my current configuration) :

• remora.joint.1.pgain = 5
• remora.joint.1.ff1gain = 1
• remora.joint.1.deadband = 0.02

having all of the above parameters configured in the .hal file, the joint could not reach a steady state even though a single step equals to 0.0016 mm as you can see in the screenshot which is way below the deadband specified. The pos_cmd was constant but apparently a freq_cmd was calculated and sent to the EC300 that lead to the oscillation (how is that possible that the deadband gets defeated??).

CURRENT STATUS

In the attached .ini, .hal , remora-config-file.txt you can see how my machine is currently configured. Apparently by not specifying a p_gain and ff1_gain they both default to 1 according to the remora-eth-3.0.c  and with a deadband of 0.005 the Y axis is able to reach a steady state. The reason i am not satisfied is because the motion isn't smooth when approaching the final position, instead the motors "click" in place after a few steps and the axis makes a squeeking noise. When i try to modify the gain parameters to smooth the axis i roll back to the initial issue of not being able to reach a steady state. 

Probably i did a poor job explaining the chronicle of my situation but hopefully i ll be able to further clarify in follow up questions.
Thank you






 

Hallo,
ich arbeite an meiner "Simulationsdrehbank" immer etwas weiter. Ich habe jetzt noch einen Spindelencoder eingebaut. An der 7I96
habe ich folgende Anschlüsse: +5V, Gnd, EncA+, EncA-, EncB+, EncB-, Index+ und Index-. Mein Encoder hat nur +5V, Gnd, KanalA,
KanalB. Kann ich den trotzdem benutzen? Kann ich den so anschließen:
+5V         >     +5V
Gnd         >     Gnd
EncA+     >      KanalA
EncB+     >      KanalB       ?
MfG
Hans

sorry i didnt get a notification

tool diameter is 10mm

N2 T1 M6
N3 G43 H1
N4 G64 P0.010
N5 S10000 M03
N6 G0 X50.000 Y-30.500 Z20.000
N7 G0 Z0.000
N8 G1 X55.500 F2000 G41
N9 G3 X50.000 Y-25.000 I-5.500 J0.000
N10 G1 X-50.000
N11 G3 X-55.500 Y-30.500 I0.000 J-5.500
N12 G1 X-50.000 G40
N13 G0 Z20.000

also i had read the documentation and also tried to compensate the tool diameter to the right but it will never do the compensation move that i need. i will see if i can use #5410 for my postprocessor so i dont need to use g41/42 ever again.

Hi kzali:

Of course，I post files on Github：https://github.com/TimBBB/DirtyBee-Carrot-BOB

However, I want to remind you that if you plan to make the PCB yourself, manual soldering will be a bit difficult.


Tim
 

Thanks, I think so too, but I had started to doubt it.

Before you start laying out the electrical cabinet for a CNC machine, it looks like there’s more than enough space and everything will fit without a problem.But once you actually start putting everything in place, the picture changes pretty quickly, and you start wondering if you’ll have enough room after all. ?Here’s the progress on the CNC electrical cabinet so far.

 

I think you will be fine, CNC machines rarely use full power and its highly unlikely you'll have all 3 motors at full power at the one time.  For example, the Z axis is only likely to use full power on plunging and then your X & Y are not moving.

Hi Aciera,

you are right, but i see almost any buttons of my keyboard, i can not figure out how tog et the rest of them.
Another problem is, when i press Button "Z", the menu in LinuxCNC shows up, because "Z" sends "F10" ( or was it F11 ) which brings up the menu. Another Axisbutton toggles between Windowmode and fullscreen of LinuxCNC UI, which is not what i would call a reliable system..
So i think i have to remap the keys send by my keyboard, or is there another solution to this problem i don't know about?

Thanks in advance, Greet, Tom

Hi Tim if you can post the Gerbers or Kicad that would be useful since Riocore is a open source project.

Cheers
Kam

I think github.com/LinuxCNC/linuxcnc/pull/4453 alleviates the issue, testing on machine needed for anyone available to do so

had a stab at #3455 as well github.com/LinuxCNC/linuxcnc/pull/4453
also check
github.com/LinuxCNC/linuxcnc/pull/4441

Dear meister:

item.taobao.com/item.htm?abbucket=19&id=...%7D&xxc=taobaoSearch

This is the link.

1.I am not sure if it is appropriate to post my product link here; if not, I will remove it.
2.I’m not a merchant; I’m just a DIY enthusiast. In my spare time, I make small items and sell them in my little Taobao shop. When a customer from China makes a purchase, I get a notification via the Qianniu app; the order includes a Chinese shipping address, and after paying about $1 for shipping, the buyer receives the item in a few days. However, I have absolutely no idea how to handle orders with foreign addresses. I’ve had English-speaking customers place orders before, but the delivery addresses were still within China, so I assume they were likely foreigners living there.
3.I would be very happy if someone could check out my little shop.


Tim
 

besriworld wrote:

What do you guys think? Is 4.7 kVA enough for three 1.5 kW servo motors?

No, depending on efficiency, 4.7KVA is roughly 4-4.5KW, and your motors/drives will use more than 1.5KW at full power.
Still, it will work fine, do not worry about it.
1. chances are you will never have all 3 motors art full tilt at the same time,
2. even then the transformer should survive just fine, although the output voltage might drop a bit, but the drives should handle that,
3. transformers handle overloads quite good, might heat up a bit more, but they do not cut power when overloaded as switching power supplies do.

Quick development update

Over the past few weeks I've mainly been working on two areas that weren't reliable yet: internal machining and automatic thread reliefs.

Fixed internal roughing: The internal roughing cycle was in some cases producing a single, way-too-coarse cut instead of properly staged passes ("not really a cutting strategy"). Found the root cause and fixed it – now it produces even, multi-pass stepping just like external turning does.
Reworked DIN 76 thread reliefs: Previously only a simple rectangular pocket was generated. Now the actual DIN 76 profile is output (30° lead-in, tangential arcs, relief bottom, shoulder radius), using real values from the DIN 76-1 table instead of estimated widths. The relief is now correctly anchored to the end of the thread (not the end of the contour), and it now also works in the middle of a longer contour, not just at the very start or end.
Safer internal approach moves: When feeding into a bore, the tool now always moves axially to the clear retract plane first and only then feeds in radially – no more diagonal rapid moves straight through the bore.
Also fixed a few real LinuxCNC crashes along the way (e.g. tool nose compensation at the inside relief corner, a "Not monotonic" error with G71/G72 when a relief is spliced into the contour).
Important: All of this is currently only on the dev branch, not yet in the stable release. Not everything has been verified with a real dry run on the machine yet, but preview and generated code now look correct for the cases tested so far. As always: feedback, test cases, and criticism are welcome.

Shaft Gauge - a free browser tool to measure STL files (mm)

I built a small, single-file web tool for quickly reading dimensions off an STL mesh, handy when you just need a measurement without opening full CAD (with claude AI).

Drag in an STL and you get:
- Diameter and length (auto shaft-axis detection) plus full bounding-box dimensions
- Point-to-point distance with vertex snapping
- 3-point circle fit to read a bore/round feature diameter (works on tilted planes)
- 2D cross-sections on the XY / XZ / YZ planes with a movable slider
- Volume, triangle count, light/dark theme

No install, no sign-up, and your file never leaves your device - everything runs locally in the browser (binary and ASCII STL). Open source (MIT).

Live demo: emresensoy.github.io/MeasureFromSTL/
Source: github.com/emresensoy/MeasureFromSTL

Feedback and feature ideas welcome.