Advanced Search

you can simply use shiftregisters,
it's faster than going via a microcontroller and spi, and cheaper too. i always say that you shouldn't run encoders via the expansions, but it's still faster than using an MCU. and as long as the encoders aren't connected to an hf-spindle, it works too

Thanks for the earlier reply meister, regarding using extra pins on your Tangoboard, I downloaded the datasheet for the nano 9K and that also cleared up a few things.

I was thinking, since you are already using SPI for your expansion IO boards, would it be feasible to use a hardware quadrature decoder chip, or something along the lines of DigitalChicken's OctoQuad code running on a Pi-Pico? Basically it will read 8 high speed quadrature signals (1MHZ) and convert each to a 32 bit position value per encoder that the controller calls for via SPI when the servo loop is required to be updated.

The idea is to have a dedicated expansion board for this, with 8 RJ45 connectors, associated line receivers and the decoding hardware. I know closed loop to LinuxCNC isn't everybody's cup of tea but you could also use it for spindle encoders, MPG's etc.

Cheers.
Russell.

From your log file:You have to fix latency issues first, that will for sure case rough running of the machine.
Check your motherboards BIOS and disable hyperthreading, turbo, TMP, power saving option, C states, speed step, whatever you do not actually need, then try again.
If you do get latency warning after disabling the above, check the temperatures with
sensors
in a terminal while LinuxCNC is running.

blazini36 wrote:

rodw wrote:

Odd, it works for me CATIO-A

 
Do you have the vendor data from the ESI....I don't think most of that AliExpress stuff is legit.No 

No, I have no idea what an ESI is.  It would be pretty hard to clone an Ethercat chip. The license is embedded into the chip so they just need to buy the right chip.

Moved to "EtherCAT" section.

On the TangNano9k, 13&15 are LED's and they work fine

onceloved wrote:

Now I use the rv901 fpga card with six encoder inputs and more than a dozen input and output
 

Pretty decent looking set-up! Lucky you. I'd move the CAT6 lead far away from the magnetic IC though. Currently sitting right on top of it, a big no-no!!  

It worries me now that Andy did not update the Linuxcnc ISO  with the one I built for him.
Here it is in my Google Drive
drive.google.com/file/d/1gEQzgHsj_LjHiGB...k1Z/view?usp=sharing

Re making an ISO, I use the Debian Live installer. Its meant to install the same version of Debian you build the ISO on so I have been meaning to upgrade to Trixie and try it to see if it does work on testing.

But its not really necessary to do that because Linuxcnc 2.9.3 is in Trixie
packages.debian.org/search?suite=trixie&...ords=linuxcnc-uspace
so sudo apt install linuxcnc-uspace does it all including installing the RT kernel!

I also shared a script to install the Ethercat repos here this week
forum.linuxcnc.org/ethercat/45336-etherc...tep?start=250#307276
Then go back to post #1 on the same thread to install Ethercat the easy way.

rodw wrote:

Odd, it works for me CATIO-A

 

Do you have the vendor data from the ESI....I don't think most of that AliExpress stuff is legit.

Quick update:

Changed all the pinouts to the ones that works. So far PIN13 and PIN15 are persistently HIGH. As if they were physically shorted but they aren't (good soldering). Just putting this out there, to be aware.

So far got stepgen, spindle pwm, spindle encoder & index, and home switches. [TESTED]. Everything that makes up for the essentials!! Super happy about that.

Those screwed-up pins are discouraging me though. But I'm not coming down to any cynical conclusion yet. I'll have a look in their datasheet again and see if I can find any reason for those being HIGH on crack...

When the machine runs, and does the homing cycle it feels like it loses steps. It does not sound like it runs smooth, if that make sense. Then it endes with join 1 following error.
 

Might be something here worth looking at

Hi Forum,

I just found an old PC which has better Latency as my current LinuxCNC machine. I am using soft stepping. However, I can not get the machine working with my PCI Express Parallel Port Card.

The weird things starts already in the bios. Only in Slot 4 the Bios sees the Parallel Port Card. The card works for sure. When i put the card in the old computer it works, just the address is different.

Every now and then it works and starts to home the first joint, but stops once this is homed.

I noticed that the Card get's disabled, and I have to enable it manually after the first Linuxcnc start. But this will not make it work.
Only a fresh restart works and that only sometimes. When it ran I had an error joint following or so.

I also notice that there is a delay in the GUI when I disable the machine. It looks like frozen for a second.

 

Hi,

That inch or mm is covered here in the interpreter.cc file i showed before at this line

I see an issue with having two different versions of interpreters.
Yes, the sai didn't cover everything. But that's not a big deal. Just add some code to the sai to retrieve more info.

The guy who developed JCNCScreen
I know Reinhard. He is good coder. Did a lot off research himself for opencascade. Also on opencascade forum he asked questions.

quite behind the internal interpreter.
Dont let it sqare you. Reinhard was doing a quite diffucult job. He had not the cmake overview as i have these day's. So
he was much complaining about spagetti code i remember.

For that, on ethercat we would need to read the encoder slice from beckhoff and then sync the Z axis according to the spindle encoder
I don't fully understand yet.

Hi Rod,

For me it's a long time ago i did ethercat installing. So i had to look again how it was done step by step.
I installed from gitlab.

I have seen the ethercat-master on gitlab is updated last week, the ./configure script run's now without any
additional arguments. It install flawless. This was not alway's the case.

However i was not able to start the ethercat bus as user, after adding rules to udev, reload, reboot etc.
So i thought how is this possible?

Relating to my time shedule, i leave it now as is.

Maybe i will make a new iso soon.
I have installed the trixie with rt kernel, it's really a lot off work to get this installed and running.
A new iso will be very handy.

On the linuxcnc iso, there was a message about the ethercat after installing, i think it was unable to find
the update sources.

Ok, thanks so far!