Laptop or PC for EthercatMaster LinuxCNC

Hi,

I new here.

I want to use LinuxCNC with soft Ethercat Master and locking for a Laptop or PC.
Can somebody share his/her experience what is the right Laptop/ PC to buy for this purpose.

Are the realtime needs for the EtherCat port critical because IgH offered a list of suitable chip sets.

Am I wrong with the findhng, that the lactency list and hardware liste often show old stuff ?

Thanks for your answer!

Best regards

Klaus

Laptops, in general, seem to do poorly because of various power saving features kicking in.

So best bet seems to be small from factor PC. However I am not aware what is EtherMaster so can't comment what should be the qualifications of the PC to run it well. Most of the hardware in the list is indeed ancient.

I believe it would be beneficial if the latency-test would have an ability to upload it's results somewhere directly with a press of a button as that would probably increase the sample size substantially - at the moment the list is only composed of entries where users have bothered go to that list specifically and update it.

Hi,

I am not familiar with the ethercat requirements.

If you are looking for a modern board with low latency, the current favorites are the J1800/1900 ones, which are on the latency page
wiki.linuxcnc.org/cgi-bin/wiki.pl?Latency-Test

I believe it would be beneficial if the latency-test would have an ability to upload it's results somewhere directly with a press of a button as that would probably increase the sample size substantially - at the moment the list is only composed of entries where users have bothered go to that list specifically and update it.

Quite a few entries are from results posted here that I have put into the database.

Anyone testing a board or computer not covered, only has to specify the exact spec and full results and I will include it.

Interesting idea though

regards

ArcEye wrote:

I believe it would be beneficial if the latency-test would have an ability to upload it's results somewhere directly with a press of a button as that would probably increase the sample size substantially - at the moment the list is only composed of entries where users have bothered go to that list specifically and update it.

Quite a few entries are from results posted here that I have put into the database.

Anyone testing a board or computer not covered, only has to specify the exact spec and full results and I will include it.

Interesting idea though

I believe that if you would not be putting the results from these forums up there the list would be even more sparse so this is a great job you are doing in this regard. However, this way of collecting information still relies on people to be motivated and aware enough to go and type up the results, post them on forums (or edit the wiki directly if they have the rights) and so on.

My assumption is that most people are pretty lazy (or busy with other stuff) to do all that most of the time. I find the linuxcnc testing tool results to be quite a valuable infromation so I would like for that dataset to be far more complete and statistically significant than it is currently. Now, asking linuxcnc devs to deticate devtime for figuring out how to read all the needed information directly might be a bit too much ofc - but perhaps some exiting tool exposes already that kind of information and is available under linux as well (CPU-z, GPU-z, SpeedFan, HWMonitor, etc all seem to be already capable of reading quite a lot of relevant hardware information down to the exact model number of memory and CPU variant) so if it would be possible for the test to hook up into one of these for pulling the hardware information then all that would remain to be done would be some-kind of location in the web where it would be possible to upload the results automatically at the press of a button (or just automatically if internet connection exists). It would further enable one to do statistics with the results, averaging over the whole set of tests run with the same hardware (and showing also the absolute min and max values observed), taking into account the time the test has run and so on and forth.

newegg still carries the GIGABYTE GA-H81M-HD2 that I have. with 4GB ddr3 and a pentium G3220 I was getting approximately 6900 on the max jitter after about an hour oh heavy use and leaving the screensaver settings on. Not a bad deal considering I spent more on Mesa cards than I did the PC.