Advanced Search

copied

Thanks Andy. In fact I amusing Open Loop control fo this spindle. its a 2hp three phase device and seems pretty stable as is. Tried the improive it with pid tuning but only made things worse.
Still not got the orient to work. Swiching pids ok, but somewhere the orient count, angle and units screwed up and not reseting on the index.
might take a while longer.

Hello All,

Im a first timer trying to setup linuxcnc and fairly new to linux in general.  Im hoping there is something im overlooking and my purchase wasn’t a waste but not optimistic…

I was originally going to get a Pi5 but price went from $80 to $130 locally just for the board that would have needed cost of case and screen on top of that.  Picked up the laptop below for $200 so it seemed like a sweet upgrade and saw what appeared to be success from other users on what seemed like similar hardware.  I understand that laptops are not ideal due to power saving settings built into them but I will be running a Mesa 7i96s so figured I could get away with it.  I wanted something with a bit more power as in a perfect world I WOULD like to be able to run fusion 360 on it (I know its not directly compatible) if I need to make simple tweaks to a program.  This would just save me carting around two computers.


Computer:
Lenovo T14 Gen 2 AMD
Ryzen 7 Pro 5850u (8c/16t 1.9ghz)
Radeon Integrated Graphics
16gb DDR4 3200
512gb SSD
Realtek RTL8111H-CG Ethernet


Installed LinuxCNC 2.9.8 Debian 13 Trixie PREEMPT-RT without major issue except it not finding network drivers itself.  Everything seemed to work ok but I installed r8168-dkms to hopefully prevent issues.  Get into the system and run the latency test.  Not getting any consistent results but from the latest run I got the following.

Latency Histogram Results with nobase – 2450 microseconds running 10 glxgears and a couple videos
Latency Test Results – Servo thread jitter 3,370,408ns 
Latency from laptop to Mesa card through a 3 ft cable is 0.857/0.891/0.937/0.014ms


Results seem crazy based on what I have read so I started following various peoples advice on tweaks.

Going through the BIOS its completely LOCKED down unfortunately.  There are basically no options other than security and boot options.  I went through and did everything I could but no ability to disable SMT or anything else that everyone recommended as first steps.

Walked through several videos on youtube from MrRodW on system and kernel tweaks and none of the tweaks seem to have a significant impact unfortunately.  I could get a little better but still miles off what people say is acceptable.

Changes made
Isolcpus – tried several combinations of multiple and single cpus
nohz_full – matched above changes
rcu_nocbs – matched above changes
intel_ideal.max.cstate=0
processor.max.cstate=1
idle=poll
cpupower frequency-set --governor performance

Checks:
uname -a = Linuxcnc 6.12.63+deb13-rt-amd64
latency-histogram – Note: Using POSIX realtime


So give it to me straight….  Am I screwed?

Thanks,
Looby

What is the required analog  input range, speed and resolution?

Could you suggest any Mesa board that can connect to my Mesa 7i95 and provide analog input support?This is urgent for my project, and I need to finish it as soon as possible. I cannot wait for the 7i87 to come back in stock.Thanks.

<data value="0"/>

Is register initialization data for the init section of the mbccs file
(since the data is initialization data, not data from hal like the command
section)

What does the last error show? (theres a modbus pin for the last reported error)

This is a really strange error since it seems like CRC checked communication
is happening but no data is being transferred.

thanks tommy for the reply. regarding pushing the contactors, i hade to prove that the machines electronics are working properly and that the push button 3sh1 which powers 6a1 safety relays temporarily until 19k1 is latched was working correctly. all the safety systems of the machine are not be touched. similar to Rotarysmp's mh400 retrofit.

Linear scales are reading correctly on the dro and as you can see from the ini file posted above i have accurately populated the scales values.

regarding the indramat drive signals i bet your right and i know i have something wrong here but im having trouble figuring out what. +10 and -10 wires are connected to the 7i97ts and i have left the original drive enable circuit untouched as this is handle by a contactor called 7k1 whose coil is pulled when 19k1 relay has been latched leaving mahos estop and error indictor logic untouched. can this work?

linuxcnc does not show joint.n.following-error but it did before i added an extra ground from +10-10 cable sheathing cables sheathing.

contactors fire and back to estop errors. i know im doing something wrong but writing this and try to explain whats going on might help. thanks for your help

Hey Guys,
bad luck. The communication seems to run between mesa and Vfd. 
Now i measured (in parallel) the connected Mesaoutput with Vfd Input. (mesa_modbus.png)
Compare to the connected USB-Dongle output and Vdf Input (USB_Modbus.png)
Signal Quality looks pretty good. (pktuart seems to work)

I figured out by using the USB-Dongle, that i have to set some Frequency to get this RUN-LED flashing on my OmronMX2 Vdf.
So I implemented all the other necessary Hal Pins to .hal file to check this with the Mesa <--> Vdf communication.
Know, if i click the CW AXIS GUI Button, Iḿ getting the same Error at my Vdf ( E12.1 like a few days ago).

From my point of view, there must be something wrong with the run, reverse or cmd-freq command.
Or still some wrong configuration somewhere.
Iḿ still getting no Vdf temperature out of the Vdf.

so write or read is not Working, 
my current .mbccs:
Can someone explain these?

Once, i had one PDU Error. But i guess it was only a wrong configuration.
Iḿ loosing my motivation, i wasted 3 weeks now 3h a day

 

Hopefully have a completely stable firmware and some type of watchdog to keep track of the communication. By the way, with firmware I speak about the USB dongle and with driver I mean the computer/LinuxCNC component.

I'm not completely happy with how the LinuxCNC xhc-whb04b-6 component works either. And at least for me, the amount of abstraction layers in that code makes it difficult to understand/change.

djdock99 wrote:

.... i can manually latch relay 19k1 by holding on the contactor with a screw driver.

Never smart to do that to machines.

djdock99 wrote:

.... but the x and z servos are rotating slowly but smoothly.

This means that LinuxCNC has no control over the servo drives, usually incorrect or missing wiring, like are the encoders wired to Mesa and do they work correctly counting up/down on the LinuxCNC DRO ?

djdock99 wrote:

Then when i update the hal file to electronically latch 19k1 with f2 it will not hold and drops relays straight away.

What error does LinuxCNC show when that happens?
I would bet it shows "joint.n.following-error" as noted on my second reply.
-
You are working on a complicated machine, so it is imperative to follow instructions strictly and as listed in priority.
See this for wiring and testing:
forum.linuxcnc.org/10-advanced-configura...to-example-mesa-7i77
And for network and WiFi to work at the same time, see here
forum.linuxcnc.org/27-driver-boards/3559...ethernet-mesa-boards

On the HW side the break (VFD + and 0) nodes would also need to be duplicated or relayed over.

In a terminal, what does
ip a
return?

That typically means a connection issue

Either the link was not up when LinuxCNC started, the IP address was wrong or similar

 

Well, after racking my brain for several days, I've managed to understand and solve the problems. The problem with the reverse gear is that the servo's DI4 signal is 24V, so when connected to pin 16 of the bobbin, it always has voltage, and the signal remains active even when you activate or deactivate reverse. The solution was to use a 5V to 24V optocoupler.

Regarding the RPM problem, I've managed to understand how the spindle scale works, allowing me to calibrate the RPM perfectly.

There are many factors involved, such as the bobbin not delivering an exact 0-10V signal, and you have to play around with the offset in LinuxCNC and the driver itself.

It's a real headache, but when you finally understand it and manage to solve it, it's very satisfying.

Thanks - of course -so simple i should have noticed it myself
Next error after changing name of hal in ini:

Debug file information:
Note: Using POSIX realtime
hm2_eth: rtapi_app_main: Resource temporarily unavailable (-11)
./main.hal:9: waitpid failed /usr/bin/rtapi_app hm2_eth
./main.hal:9: /usr/bin/rtapi_app exited without becoming ready
./main.hal:9: insmod for hm2_eth failed, returned -1
2077
Stopping realtime threads
Unloading hal components
Note: Using POSIX realtime

Why "temporarily" unavailable?