Advanced Search

If these are voltage mode (bare PWM) drivers, the tuning is somewhat like
velocity mode servos (with low velocity loop gain) That means that FF1 is critical.

I would first set a relatively  low P term and adjust FF1 so that during constant velocity
the motor neither lags nor leads the desired position (too much FF1 will make the
position lead and too little will make it lag) Note that this is just for tuning during constant
speed motion.

Next I would adjust D as high as possible while remaining stable.

Next do the same with P and back off about 30% from where it becomes unstable

You can now add some FF2 to compensate for the following error during acceleration
and decceleration.

Finally you can add as much I term as is stable (watch out for big oscillations)
and then back off about 50% from the point where it became unstable

Note: If you need to retune FF1 , you need a low P term and zero  I term

Also if you can decrease the Ethernet latency and run a higher frequency servo
thread, the performance will be improved.

Ignore what Langdon suggests, his advice is unreliable at best, mostly uninformed and most formed in an empty head and produced rectally.
My apologies for sourcing the thread and being so crass but many a time I have seen the above mentioned member not answer the topic and speak non sense.

Whilst I myself don't have correct answer Tommy or some other more knowledgeable member may chime in.

As for decreasing the thread period, aka increasing thread frequency I would run a latency test using the new figures the see how this affects latency, just to be sure there are no surprises, the Linuxcnc docs will have this information, but it will be need to run from the command line.

Once again apologies for the crass warning.

The guidelines would be in the raspberry pi config.txt docs. As I'm sure has been mentioned previously.
You will have to read the main thread to find the location of this file as it is not in the same location as the raspberry pi images.
Honestly for your case use I would ditch the Linuxcnc image, install Linuxcnc via apt to a RaspiOS, there is information on the forum regarding this, or as I mentioned earlier build from source, you don't require an RT kernel to build from source.
The zswap fix is to actually have swap a compressed swap in memory. It is not just to "fix an error log message", which you don't seem to understand.
If you aren't going to bother to try I can't be fucked to give you help on what is really a non issue for an image created to run a CNC machine which is something you admitted you are not intending to do. And in that case you have been given suggestions to overcome what is a very minor issue.
Regarding the zswap it seems you didn't see or ignored the issue it caused for a user when the system was left idle. I'll repeat this one more time, it is not a fix for an error log message, the error log message pointed to an actual issue and the fix was to fix the actual issue.
So far I have offered some suggestions, yet you have not bothered to attempt any. Maybe this image is not suitable for your intended use and another solution maybe a better path to take.
Good luck in your endeavours but I have exhausted my patience and will to help

I have a question. How to use Fixed PDO mapping on Lichuan LC Servo?

I used 10 entries in PDO 1600h ( Variable PDO ) in ethercat-conf.xml ( 10 entries are maximum )
So I just moved some <pdoEntry>  to FixedPDO (1704h )
But it didn't work.

Do I have to do something to use FixedPDO ?
Are there any good example for it ?
 

There is the LinuxCNC homing documentation:

linuxcnc.org/docs/html/config/ini-homing.html

There are also some you tube videos around
that will show up if you google "linuxcnc homing tutorial"
 

The machine worked before, but basically your saying due to the amount of power being used aka the P needed to get the motor to move and then the drift is evidence that the motor is probably a little under powered? So maybe I need to redo the tuning overall instead start from P = 300, D = 0, I = 0, BIAS = 0 FF0 = 0 FF1 = 0 and FF1 = 0 and see if im still getting drift? The other option is to see if i can increase the servo-thread update frequency? from 1000000 to 900000? That would get more more cycles per P right giving me more response, but is essentially the same as increasing right?

Just to close the loop on this from a Teknic/ClearPath perspective, they have a Jitter delay that can smooth out motion, that was set at 14ms, when turned off the f-error dropped to 2ms which is in the ball park of a timing loop. 

Long story short, I am turning off everything PID related since I am using CSP and DC as long as my jitter setting is the same on all the motors they will all end up at the same commanded place at the same commanded time even if LinuxCNC is 2ms behind. 

Hello.
I had to recreate LC10 homing code from zero reading manual designated for a driver with a motor with incremental encoder . ( I thought original code was difficult to fix )

If you still have some problems, I consider to open my codes...( I have something to do for some improvement )

The zswap fix seems intended to fix a system log error message "Failed to start zramswap.service" but I checked systemctl and have no failed units, so it doesn't seem relevant to this case, is it?

Is there a guideline on how much memory to allocate to the GPU? This is a 4GB Pi 400. One idea I have is to compare the config.txt with that of other images that work well.

The other thought is regarding codecs or configuration that may be missing as suggested by the attached report from the chrome://gpu

The existing VFD is an Invertek and it can not work with Ethercat - the reason I think about the exchange.
Why that specific one? Mmhh - I had Yaskawa components in another mill and was very satisfied with the quality and the GA500 with the interface card is relatively cheap to get here in Germany.

As Servos I use 4 x A6 OMC/Stepperonline

Have you implemented the zswap fix or tried looking at the memory allocated to the GPU ?

To be quite blunt this is a low priority issue, as the primary function of the image is to run a CNC machine. Which by all accounts it has done quite well, the only show stopper was the zswap issue for one user, but this was resolved.
I may sound like an ass and appear to be unhelpful, but as I said before it's a low priority issue and so far has not, no reports, has affected the operation of a Linuxcnc session.

Sorry, took a while to check this out as I did not have a working
test system with a EPP parallel port and Mesaflash/LinuxCNC.

Unfortunately, I am unable to duplicate the issue.

What I did was take a 7I90HD and corrupt the S EEPROM so the boot block was broken
then booted from the P EEPROM , reselected the S EEPROM and attempted to update
The S EEPROM via mesaflash:

 

So it still may be a host EPP issue of some kind.

Test system: LinuxCNC 2.9.4 mesaflash 3.4.9
Hardware: HP Elite 8200 SFF with on board parallel port
 

7i96S + 7i83 ?
Wen sie mehr als ein encoder brauhen, noch 7i85 oder 7i85S

A very common method was to start on Y configuration and then switch to Delta. The motor starter switched between Y and Delta (contactors).
Ask google or chatgpt for the reason.

Depends on what you are aiming at, but a standard chinese VFD controlled by Modbus/USB is a common way to get speed control of the spindle.
The rpms can be somewhere 50%-150% of nominal rpm. Low torque at low speeds, possibly balancing issues at high speed, but at least some variation up and down is possible.
USB/Modbus is somewhat sensitive to electric noise so it is usual to have USB serial/RS485 with optocouplers and line filter for the mains.

Good point about checking chrome://gpu, it shows a number of problems with graphics features, such as hardware decode being software only.

Graphics Feature Status
Canvas: Hardware accelerated
Direct Rendering Display Compositor: Disabled
Compositing: Hardware accelerated
Multiple Raster Threads: Enabled
OpenGL: Enabled
Rasterization: Hardware accelerated on all pages
Raw Draw: Disabled
Skia Graphite: Disabled
TreesInViz: Disabled
Video Decode: Software only. Hardware acceleration disabled
Video Encode: Software only. Hardware acceleration disabled
Vulkan: Disabled
WebGL: Hardware accelerated
WebGL2: Hardware accelerated
WebGPU: Disabled
WebNN: Disabled

 

Problems Detected
WebGPU has been disabled via blocklist or the command line.
Disabled Features: webgpu
Accelerated video encode has been disabled, either via blocklist, about:flags or the command line.
Disabled Features: video_encode
Accelerated video decode has been disabled, either via blocklist, about:flags or the command line.
Disabled Features: video_decode
Disable partial swaps on Mesa drivers (detected with GL_VERSION): 339493
Applied Workarounds: disable_post_sub_buffers_for_onscreen_surfaces
Disable KHR_blend_equation_advanced until cc shaders are updated: 661715
Applied Workarounds: disable(GL_KHR_blend_equation_advanced), disable(GL_KHR_blend_equation_advanced_coherent)
Expose WebGL's disjoint_timer_query extensions on platforms with site isolation: 808744, 870491
Applied Workarounds: enable_webgl_timer_query_extensions
Some drivers can't recover after OUT_OF_MEM and context lost: 893177
Applied Workarounds: exit_on_context_lost
Avoid waiting on a egl fence before swapping buffers and rely on implicit sync on Broadcom GPUs: 938286
Applied Workarounds: rely_on_implicit_sync_for_swap_buffers
Disable GL_MESA_framebuffer_flip_y for desktop GL: 964010
Applied Workarounds: disable(GL_MESA_framebuffer_flip_y)

Full report attached below.

I don't think the issue is hardware or resolution, since it is just this image that has these problems.