Questions about timing and Latency Tests HobbyCNC

Switching from Mach 3 to LinuxCNC.
Specs
Driver: HobbyCNC ez driver (uses same driver chip as the PRO version)
PCI parallel port
dual core amd system ~2.4ghz 4gb ram
linuxcnc installed via wubi

First question,
during the stepconf wizard you have to choose timings for your driver board, I had no clue what mine were as these timings aren't listed on the HobbyCNC website that i know of. So I chose the HobbyCNC pro preset (my board is the HobbyCNC ez driver not the pro). Will these presets for the PRO board be okay for my ez driver board? I tested the motors with this setting and it all seems to work fine. Is there any way to test to make sure its fine? Also, what do these timings actually effect?

Second.

I did the latency test and my reading is like 75 percent over the recommended range (~20000-50000). Mine was going up past 100,000. However, again the motors seemed to jog fine and i even ran a program in the air. All seemed fine.

Is there any dangers to my machine/driver board/computer if i run with these settings?

blitz355 wrote:

First question,
during the stepconf wizard you have to choose timings for your driver board, I had no clue what mine were as these timings aren't listed on the HobbyCNC website that i know of. So I chose the HobbyCNC pro preset (my board is the HobbyCNC ez driver not the pro). Will these presets for the PRO board be okay for my ez driver board? I tested the motors with this setting and it all seems to work fine. Is there any way to test to make sure its fine? Also, what do these timings actually effect?

Second.

I did the latency test and my reading is like 75 percent over the recommended range (~20000-50000). Mine was going up past 100,000. However, again the motors seemed to jog fine and i even ran a program in the air. All seemed fine.

Is there any dangers to my machine/driver board/computer if i run with these settings?

1st) The timing should be ok. If not mistaken, the timing will reflect the maximum pulse rate.

2nd) That PC is not good for linuxcnc. What can happen is a "glitch" during a job, something random as for a instant, linuxcnc could not output the desired steps into the parallel port, and this is something that you do not want. My latest PC had a max of 7500.

Hi

100K base thread is not usable as emcPT says

You need to be very clear as to whether these are occasional spikes or a constant high figure

Have a read of these to see the problem and some possible solutions
www.linuxcnc.org/index.php/english/forum...-the-latency-problem
www.linuxcnc.org/index.php/english/forum...me-latency-solutions

I would start by looking hard at the graphics driver / card and disabling the second core if possible

I don't know if AMD uses hyperthreading or if that is just Intel, there is certainly no SMI isssue, that just affects some Intel chipsets.
Any power saving / management in BIOS that can be disabled needs looking at.

regards

ArcEye wrote:

Hi

100K base thread is not usable as emcPT says

You need to be very clear as to whether these are occasional spikes or a constant high figure

Have a read of these to see the problem and some possible solutions
www.linuxcnc.org/index.php/english/forum...-the-latency-problem
www.linuxcnc.org/index.php/english/forum...me-latency-solutions

I would start by looking hard at the graphics driver / card and disabling the second core if possible

I don't know if AMD uses hyperthreading or if that is just Intel, there is certainly no SMI isssue, that just affects some Intel chipsets.
Any power saving / management in BIOS that can be disabled needs looking at.

regards

I will be sure to look into those links and try to remedy the issue.

As I'm typing the test is running for about 5 minutes now and the base thread max jitter (that is the value I'm supposed to look at right?) hasn't gone past 13,441ns (is this the number i put in the stepconf wizard?). Which is good. not sure what made the values go that high yesterday.

As I'm typing the test is running for about 5 minutes now and the base thread max jitter (that is the value I'm supposed to look at right?) hasn't gone past 13,441ns (is this the number i put in the stepconf wizard?). Which is good. not sure what made the values go that high yesterday.

Yes it is base thread max jitter, but you need to run for a lot longer than that.

The figure shown is a cumulative one, which means one high spike and the whole test is high.
The first FAQ shows how to do a rtai latency test from the command line, which does a per second reading and should assist seeing what is going on if you get a high reading again.

regards