Advanced Search

Use the sim as a template

Had to change the RT kernel to 6.13 as when disabling BIOS stuff the 6.10 would fail to boot some 70% of the time.
This was on 2 laptops (Lenovo X1 Yoga and Yoga 370) and my main PC (14400 with RX6800).
So far tested on laptops only, will do more installs and if all is good, replace the ISO.
Stay tuned.
FYI:
6.10 works fine for Mesa boards, even with default BIOS settings, but will do 170000-280000 in latency, and that is on the limit of what i consider safe, disabling just Hyperthreading and ATM (or AMT?) would bring that down to safe numbers.
With everything in BIOS of Y370 disabled and CPU in powersave mode on batteries, i could get the latency under 100000 and testing with base period of 100000 would show no warnings for hours. Did not test with power attached.

PCW wrote:

Yeah, that's not a sine wave (more like a rectified sine wave)

Yeah you will probably get a sine wave when you measure it differentially.
This could be done with a differential probe, or you could power the encoder with a galvanic isolated power supply.
But that is not important now, I only wanted to check if there is a signal at all.

I think the low amplitude level is also ok. I have a Heidenhain rotary encoder and also some linear magnetic scales l worked with -- they all have a level much lower peak-peak value than 1V.


 

I thought I had to configure it the same way I did with the windmill. 


The simulator started correctly, but at this point, there was no way to proceed.

I agree with Rod, naming can and will cause issues as soon as more people start using this, so changing it sooner rather than later would be beneficial.
Also agree on calling it Grotius, for us who spend more time here on the forum it would obvious, and new user will have to learn whatever the name may become.
P.S.
I still have nightmares with Probe Basic and Basic Probe !!!
Naming it QtMill and QtLathe would help a lot... Sorry LCVette, i know you asked us for a name and we came up with nothing.
We already have QtDragon and QtPlasmaC, those should fit nicely in the naming scheme.

there is no "probe basic lathe" just probe basic and a lathe ui. if you install probe basic you should be able to open the lathe sim, start there to verify the installation is correct first.

I wonder if the scurve planner could not be renamed to say "stpmod.so" or even "grotius-tp.so" in the build process?
The trajectory planner is loaded in the ini file settings so the name can be anything.
Changing the name would allow it to coexist in a version of linuxcnc that also includes the default tpmod.so. This would be desirable long term.

Also do try just the add on card with

Yes,
yes,
yes,
...
no!

ATX PSU is OK for testing and running the machine, but will limit the speed.
X220 is the "no" up there, had 2 of those, absolutely not usable for LinuxCNC, but since you have it, get the ISO, burn to USB, boot the X220 from it and do some latency tests as follows
latency-histogram --nobase --sbinsize 1000 --show
Also had 2 of X240 (still have one) both were/are fine for running LinuxCNC.
If the latency turns bad, i can upload an image i use on all my production laptops and PC's with Mint 19.3 and LinuxCNC 2.8.4, but this needs to be installed to be able to test it.

Thanks for getting back to me. I feel like I have tried that but I will confirm later on today.

Also I apologise for the photo they were taken when I was starting to play with getting this going. They were not meant specifically for the forum post more just as a record that I could look up on my phone when I was editing the hal file so all I needed was the address.

Happy to provide further information and better photos. If anyone thinks it will help.

OK, so
Mesa 7i96s it is (should be worth it, since I can us that one for the big machine later)
3 of the TB6600
Leaves the PSU question. I think I have an old ATX PSU. Those deliver 2-4 lines of 12V and several 5V lines. Could that do for the small steppers?
And my Thinkpad x220.