Search Results (Searched for: )
- NT4Boy
- NT4Boy
09 Apr 2024 17:12
Replied by NT4Boy on topic Unable to set boot order on Raspberry PI5 with pcie NVMe on HAT.
Unable to set boot order on Raspberry PI5 with pcie NVMe on HAT.
Category: Computers and Hardware
- Frogmate
- Frogmate
09 Apr 2024 14:44
Replied by Frogmate on topic Probe Basic OpenGL requirements don't match the documentation
Probe Basic OpenGL requirements don't match the documentation
Category: QtPyVCP
- TangentAudio
- TangentAudio
09 Apr 2024 14:35
Replied by TangentAudio on topic Dual PID loops with motor encoder + scale encoder per axis
Dual PID loops with motor encoder + scale encoder per axis
Category: Advanced Configuration
- meister
- meister
09 Apr 2024 14:25
Replied by meister on topic WebGUI
WebGUI
Category: Other User Interfaces
my biggest problem is solved, the preview with working/moving tool pos:
- TangentAudio
- TangentAudio
09 Apr 2024 14:15 - 09 Apr 2024 14:28
Dual PID loops with motor encoder + scale encoder per axis was created by TangentAudio
Dual PID loops with motor encoder + scale encoder per axis
Category: Advanced Configuration
Greetings,
I'm working on a new conversion project, modernizing an old Bridgeport Series 1 2HP that has a working Prototrak Plus system on it. For "phase 1" of the project I am keeping the original Trak servo motors/drives/encoders and the Trak "wheel" style encoders for table position. The machine has ballscrews, though they are a bit worn and have some backlash. I know this isn't ideal and may present some problems, but I hope to tackle that later. My goal is to get it up and running for close to zero cost initially, then figure out what I want to improve later.
My past conversions have all been open-loop stepper systems, and while I'm quite comfortable with LinuxCNC down to the programming level, closed loop config and tuning with LinuxCNC is new to me. I've been focused on getting one axis working first. I was able to at least get a basic tune on a single loop using just the motor encoder. But I'd like to take the next step and use the axis position encoders in a dual loop setup to improve positional accuracy.
I'm using a Mesa 7i43 parallel port FPGA card that I had left over from my first CNC conversion 12+ years ago. I designed and hand-wired a board to handle the differential signal conversion to/from the servo drives and encoders. I've got that all working just fine on the bench - all of the encoders and servo outputs work through Hostmot2 and HAL as expected.
I did a bunch of searching on the forum but there seems to be fairly sparse info on this topic. I followed the guide that's on the wiki , but it's also a bit light on details. I've got a rough config put together based on the steps in that guide, but I know it still has some problems and I could use a more experienced set of eyes. My controls engineering class was almost 30 years ago at this point (!), and while I do have some basic experience and understanding of PID control, I'm getting a little ahead of my skis on this dual loop implementation.
Happy to share full configs, but I figured I'd start with just the X axis snippets. My guess is my config problems are in the feedback part near the bottom of the HAL config, that's where the wiki guide seemed most vague.
Thanks!
Here's the INI:
and the HAL:
I'm working on a new conversion project, modernizing an old Bridgeport Series 1 2HP that has a working Prototrak Plus system on it. For "phase 1" of the project I am keeping the original Trak servo motors/drives/encoders and the Trak "wheel" style encoders for table position. The machine has ballscrews, though they are a bit worn and have some backlash. I know this isn't ideal and may present some problems, but I hope to tackle that later. My goal is to get it up and running for close to zero cost initially, then figure out what I want to improve later.
My past conversions have all been open-loop stepper systems, and while I'm quite comfortable with LinuxCNC down to the programming level, closed loop config and tuning with LinuxCNC is new to me. I've been focused on getting one axis working first. I was able to at least get a basic tune on a single loop using just the motor encoder. But I'd like to take the next step and use the axis position encoders in a dual loop setup to improve positional accuracy.
I'm using a Mesa 7i43 parallel port FPGA card that I had left over from my first CNC conversion 12+ years ago. I designed and hand-wired a board to handle the differential signal conversion to/from the servo drives and encoders. I've got that all working just fine on the bench - all of the encoders and servo outputs work through Hostmot2 and HAL as expected.
I did a bunch of searching on the forum but there seems to be fairly sparse info on this topic. I followed the guide that's on the wiki , but it's also a bit light on details. I've got a rough config put together based on the steps in that guide, but I know it still has some problems and I could use a more experienced set of eyes. My controls engineering class was almost 30 years ago at this point (!), and while I do have some basic experience and understanding of PID control, I'm getting a little ahead of my skis on this dual loop implementation.
Happy to share full configs, but I figured I'd start with just the X axis snippets. My guess is my config problems are in the feedback part near the bottom of the HAL config, that's where the wiki guide seemed most vague.
Thanks!
Here's the INI:
#******************************************
[AXIS_X]
MAX_VELOCITY = 2.0
MAX_ACCELERATION = 10.0
MIN_LIMIT = -12.0
MAX_LIMIT = 12.0
[JOINT_0]
TYPE = LINEAR
HOME = 0.0
FERROR = 0.025
MIN_FERROR = 0.008
MAX_VELOCITY = 2.0
MAX_ACCELERATION = 10.0
P = 650
I = 0
D = 25
FF0 = 0
FF1 = 0.1
FF2 = 0.01
BIAS = 0
DEADBAND = 2e-06
MAX_OUTPUT = 10
ENCODER_MOTOR_SCALE = 27940
ENCODER_LINEAR_SCALE = 5080
OUTPUT_SCALE = 10
MIN_LIMIT = -12.0
MAX_LIMIT = 12.0
HOME_OFFSET = 0.0
HOME_SEQUENCE = 2
#******************************************and the HAL:
#*******************
# AXIS X JOINT 0
#*******************
# PWM Generator
# The ProtoTrak Plus servo drives use offset mode PWM, 50% duty = off, 0% duty = -100%, 100% duty = +100%
# HostMot2 PWM generator supports this natively by setting offset-mode to true
setp [HMOT](CARD0).pwmgen.00.output-type 1
setp [HMOT](CARD0).pwmgen.00.offset-mode true
setp [HMOT](CARD0).pwmgen.00.scale [JOINT_0]OUTPUT_SCALE
# Velocity Loop (P, D, FF1)
setp pid.vx.Pgain [JOINT_0]P
setp pid.vx.Igain 0
setp pid.vx.Dgain [JOINT_0]D
setp pid.vx.bias [JOINT_0]BIAS
setp pid.vx.FF0 [JOINT_0]FF0
setp pid.vx.FF1 [JOINT_0]FF1
setp pid.vx.FF2 [JOINT_0]FF2
setp pid.vx.deadband [JOINT_0]DEADBAND
setp pid.vx.maxoutput [JOINT_0]MAX_OUTPUT
setp pid.vx.error-previous-target true
# Position Loop (I only)
setp pid.px.Pgain 0
setp pid.px.Igain [JOINT_0]I
setp pid.px.Dgain 0
setp pid.px.bias 0
setp pid.px.FF0 0
setp pid.px.FF1 0
setp pid.px.FF2 0
setp pid.px.deadband [JOINT_0]DEADBAND
setp pid.px.maxoutput [JOINT_0]MAX_OUTPUT
setp pid.px.error-previous-target true
# Sum output of the two PID loops together to produce the PWM output signal
net x-output-motor pid.vx.output <= sum2.x.in0
net x-output-linear pid.px.output <= sum2.x.in1
net x-output <= sum2.x.out => [HMOT](CARD0).pwmgen.00.value
# Motor encoder
setp [HMOT](CARD0).encoder.00.counter-mode 0
setp [HMOT](CARD0).encoder.00.filter 1
setp [HMOT](CARD0).encoder.00.index-invert 0
setp [HMOT](CARD0).encoder.00.index-mask 0
setp [HMOT](CARD0).encoder.00.index-mask-invert 0
setp [HMOT](CARD0).encoder.00.scale [JOINT_0]ENCODER_MOTOR_SCALE
# Axis scale encoder
setp [HMOT](CARD0).encoder.02.counter-mode 0
setp [HMOT](CARD0).encoder.02.filter 1
setp [HMOT](CARD0).encoder.02.index-invert 0
setp [HMOT](CARD0).encoder.02.index-mask 0
setp [HMOT](CARD0).encoder.02.index-mask-invert 0
setp [HMOT](CARD0).encoder.02.scale [JOINT_0]ENCODER_LINEAR_SCALE
# Neither of the ProtoTrak encoders have index signals, should this always be forced off?
net x-index-enable joint.0.index-enable <=> [HMOT](CARD0).encoder.00.index-enable [HMOT](CARD0).encoder.02.index-enable
net x-index-enable => pid.vx.index-enable pid.px.index-enable
# Velocity and position feedback into the PIDs
net x-motor-fb pid.vx.feedback <= [HMOT](CARD0).encoder.00.position => joint.0.motor-pos-fb
net x-linear-fb pid.px.feedback <= [HMOT](CARD0).encoder.02.position
# Home / limit switch signals - none yet.
net x-home-sw => joint.0.home-sw-in
net x-neg-limit => joint.0.neg-lim-sw-in
net x-pos-limit => joint.0.pos-lim-sw-in
- cncsparko
- cncsparko
09 Apr 2024 13:52
- polskleforgeron
- polskleforgeron
09 Apr 2024 13:25 - 09 Apr 2024 13:29
Replied by polskleforgeron on topic The blacksmith's plasma table
The blacksmith's plasma table
Category: Plasma & Laser
- Cant do this anymore bye all
09 Apr 2024 13:14
Replied by Cant do this anymore bye all on topic Unable to set boot order on Raspberry PI5 with pcie NVMe on HAT.
Unable to set boot order on Raspberry PI5 with pcie NVMe on HAT.
Category: Computers and Hardware
- NT4Boy
- NT4Boy
09 Apr 2024 12:29
Replied by NT4Boy on topic Unable to set boot order on Raspberry PI5 with pcie NVMe on HAT.
Unable to set boot order on Raspberry PI5 with pcie NVMe on HAT.
Category: Computers and Hardware
- tommylight
09 Apr 2024 11:48
Replied by tommylight on topic Thoughts on material properties for fiber laser machines?
Thoughts on material properties for fiber laser machines?
Category: Plasmac
Thank you for all the info you share here and on other topics, and for persisting on things working the way you need/like them, utmost respect.
-
Question:
Does the kerf width change so much as to be relevant when laser cutting up to, say 5, 10, 20mm ?
Gas types seem the same as for plasma cutting used on industrial machines, using those does change other cut parameters, like more speed or thicker material, although way out of hobby reach, i have a Hypertherm 3070 with O2/N2 capabilities collecting dust.
Focus height is a must for manual focus laser heads, probably cut_height works for that, it should.
As for material properties, adding steel, alu, stainles would be nice to have, although thin plates work with same settings, thicker stuff will probably need changes, especially alu.
I resorted to naming materials
03 for 3mm mild steel
13 for 3mm alu
And so on, makes it also easier to find quickly, rarely cut alu so those remain down on the drop down menu.
-
Question:
Does the kerf width change so much as to be relevant when laser cutting up to, say 5, 10, 20mm ?
Gas types seem the same as for plasma cutting used on industrial machines, using those does change other cut parameters, like more speed or thicker material, although way out of hobby reach, i have a Hypertherm 3070 with O2/N2 capabilities collecting dust.
Focus height is a must for manual focus laser heads, probably cut_height works for that, it should.
As for material properties, adding steel, alu, stainles would be nice to have, although thin plates work with same settings, thicker stuff will probably need changes, especially alu.
I resorted to naming materials
03 for 3mm mild steel
13 for 3mm alu
And so on, makes it also easier to find quickly, rarely cut alu so those remain down on the drop down menu.
- Cant do this anymore bye all
09 Apr 2024 11:46
Replied by Cant do this anymore bye all on topic Low Cost Raspberry Pi Contoller Board
Low Cost Raspberry Pi Contoller Board
Category: Driver Boards
- Cant do this anymore bye all
09 Apr 2024 11:40
Replied by Cant do this anymore bye all on topic Unable to set boot order on Raspberry PI5 with pcie NVMe on HAT.
Unable to set boot order on Raspberry PI5 with pcie NVMe on HAT.
Category: Computers and Hardware
- tommylight
09 Apr 2024 11:29
- tommylight
09 Apr 2024 11:24
Replied by tommylight on topic How to convert old Acromatic 850 language to Fanuc OiT-f+
How to convert old Acromatic 850 language to Fanuc OiT-f+
Category: General LinuxCNC Questions
- Murphy
- Murphy
09 Apr 2024 11:22 - 09 Apr 2024 15:12
Replied by Murphy on topic Remora - ethernet NVEM / EC300 / EC500 cnc board
Remora - ethernet NVEM / EC300 / EC500 cnc board
Category: Computers and Hardware
Time to create page: 1.174 seconds