Advanced Search

Hi again Mark,

I realized the technical manual was a bit large so I ran it through a later version of my document compositor and replaced it on the Google drive. It's about half the size at 151MB or so.
You asked about my Maho speed capability. There is a video called 13_Gearchange.mp4 at the same link that explains it in full. The video is relevant to both yours and my machine and probably other Mahos. I think your machine goes 2 gears higher and misses the lower 2. It is relevant to LinuxCNC programing on Mahos also, I think
I'm envious of a lot of the high quality kit you have. Even your metal vise looks superb. Your tool setter is wonderful. But the way you do the Z with the round bar is a bit nasty. There is a video of my Z setter in the link which I think works more elegantly. It's a real cheap one but it works really well. I could possibly replace the dial gauge in it with a micron scale one, but it seems repeatable if I put it exactly on the dial line with respect to the micrometer scale on my Heidenhain VDU.
I'm thinking that many of the things we are talking about aren't directly relevant to many LinuxCNC users, mainly Maho stuff, and this to and from is very slow. So if you want you can Skype me on david.jensen11, I can share screen and go through myriad topics and photos. Switzerland is 10 hours behind Sydney at the moment. 3 months either side of new year we are UT+11, Otherwise UT+10

Cheers,
David

I would set the drives steps per turn equal to the encoder resolution
(10K counts per turn with 4x quadrature)

Then the step scaling in LinuxCNC is 100000
 (10000 CPT *5 TPI *2 Belt_division)

Hello, i am looking for some guidance on the interaction of LinuxCNC PID Tuning +  EtherCAT + Servo Drive Tuning.

Would i still need to setup  PID + FF components in the .hal file and .ini file to tune a ethercat system or should all the tuning be in the drive only ?  or a combination of both ?  i would think the servo and the drive are tuned for being dynamic and critically damped to step responces then LinuxCNC  tuning terms are used to reduce following errors.

I have grown my .hal and and .ini file from bare bones and i am unsure on what i need to add to make a PID system work with ethercat.

I am not having any sucess in reducing following error, and my feeling is i need to alter gains in Linuxcnc.
as most people are using steppers and i am using AC servos over ethercat i have not been able to find much direction of how to setup files to enable the calibration feature to work with ethercat. i get an error opening the calibration module saying no tunable gains are in the .ini file

i realize its not as simple as adding the gains to the .ini file
looking at some examples do i need to load the following into a .hal file:-
load at_pid num_chan=2             (for two axis system)
addf pid.0do-pid-cals servo-thread  (for each axis)
set pid.0.Pgain [JOINT_0]P            (for each gain)

then add the gains to the .ini file ?

does anyone have any insight into this type of configuration.








 

Hi all, hope you can help.

Having a issue when trying to use a while loop. I am using lcnc 2.9.3 & gmoccapy 3.4.8
But lcnc just freezes or really slows down the system when the comp turns true, see below 2 examples that work and two that dont.
the only change is that it should go in to the while loop
I am trying to use imputsignals: 
output: Start tool state 0.000000 after (edited)
 
output: Start tool state 0.000000 greater after (edited)
 
output = nothing, not even start

output = nothing, not even start

Hi Guys.
I'm trying to figure out my motor tuning for steps per inch.  I'm guessing I need to use electronic gear ratio in my servo drive to be able to handle the amount of pulses per inch.  Right now if I set an encoder count of (and i'm just playing right now to speed it up)  2500, and a step count of 500000 and I tell it to move .100 it only moves approx. .002 with different variations I can only get approx .011 per .100 request.
So what would be the best electronic ratio
As near as I can figure at this time my ball screws are 5 turns per inch or .200 per rev. (Hardenge SuperSlant, Maybe someone knows for sure the pitch)
it is belt driven from the survo 2 turns of the motor to one turn of the ball screw
The information I find for the motor and encoder says it has 2500 lines per rev.
So I'm thinking  2500x 5(ballscrew)= 12,500 x 2(gearing)= 25000  and this is nothing for a computer to handle but if i work on this figure the axis doesn't move enough to read on a dial indicator.

I am  obviously missing something big because of the amount of pulses I actualy have to put out to get any readable movement per .1 request.
In case you might think it helps.  My survo's and drives are chinese the drives are lichuan A6 drives  Below copied from manual.  I only wanted one line but it would only let me copy from top of page to that line.

A6 Servo Drive User Manual
2
Chapter 2 Electrical Specifications
2.1 Specification
Input power
CONTROL
POWER Single phase 220 VAC
MAIN POWER
SUPPLY Single Phase/Three Phase 220VAC
Working
environment
Temperature 0～45℃
Humidity No condensation ≤90% RH or less
Elevation Altitude ≤1000M
Installation
environment Non-corrosive gases, flammable gases, oil mist or dust, etc.
Installation
method VERTICAL INSTALLATION
Encoder feedback 2500 p/r (resolution: 10000), incremental encode

In the ini file, joint 1 section
HOME_IGNORE_LIMITS = NO
should be
HOME_IGNORE_LIMITS = YES

Use
latency-histogram --nobase --sbinsize 1000 --show
There are a lot of excursions out of the visible area.

You are welcomed, always.

  uploaded the configs.

When you say drives are you meaning the drive boards.. or the servos themselves?
"DIFF/ INPUT Designates differential input speed command terminals."
Thats what the book is calling those terminals if you are meaning on the driver board or cards themselves.

While I was seeing the exact terminology I figured I'd post why I commented on the frequency thing... this is the books description of how the drives work... 
PWM indicates pulse-width modulation which is the operating mode of each transistor in the
controller. In a PWM controller, a power transistor bridge switches a fixed DC bus supply to
the motor load at a constant frequency with varying on-off ratio for output power control. The
SERVOPAK differs from other controllers of its type in that two types of switching techniques
(standard-H and Uni-Switching) are employed.

Anyway, thats a long explanation to say yes I'll find some time tomorrow to see what happens when I do that. The drives are enabled with a ground, and I'm sure they are enabled since it told me how to bypass the enables in the book, and the motors make noise when I enable them. 

These drives also have a lot of adjustments on them, so I assume they are a little different than just a straightforward drive. It appears as though the drives themselves did the tuning more so than the computer did. 

No idea, also didn't see an option to compile for non lts, sometimes things are easier than they seem
Better to set isolcpus=5 or 2-5 because linuxcnc always takes the last core and the quad core a53 is after the 2xa72

Ideally you would set isolcpus=1 but then you need to set the rt tasks to core 1 manually, via a script or change the source of linuxcnc.

royka wrote:

You can try this image, with desktop and gpu hw acceleration but under 1GB:
drive.google.com/file/d/1PCetphqJUFfS3aL...eD7/view?usp=sharing

It's laughable...did not worked tweaking & reweaking to modified it for opi4 ...
your image simply worked as is ...lesson learned here..
i dont understand , if armbian image for opi4 lts work on opi4 why the hell armbian link to non lts ?!?
You tell me , because everything seem to be good .
Trying some isolcpu 1,2 , no latency miracle ,should be ok , cutting air .
Let'see if i can install 2.9.3 now.

Good job Roya ..