On 19/01/17 23:51, Oliver Rew wrote:
/"I've used a couple of J1900 motherboards and they work great.
Together with a 5i25/6i26 + 7i76 mesa card"
/Which J1900 motherboard did you use with the 5i25? The only one I saw
on newegg with a full PCI slow is the quad-core version(as I
understand it, this is undesirable for LinuxCNC). Thanks!
If you are just using a servo thread and hardware step generation (ie
mesa cards ) it really does not matter. The servo thread max jitter
with a rt-preempt kernel is low enough for that.
I have a quad core mini ITX Pentium that I run a 5i25/7i76 combo from
and found that having a
50K dummy base thread (running but nothing attached to it) kept the cnc
code on one core mostly and brought the servo thread max jitter down
from 100K to about 28K
If you are using a parport on a base thread, then you will probably need
to use `isolcpus 1,2,3` kernel parameter to isolate 3 of the 4 cores to
get the best latency, that is if you are using linuxcnc and a rtai
kernel ( as opposed to machinekit, which no longer supports rtai )
On Thursday, January 19, 2017 at 12:31:42 AM UTC-8, Bas de Bruijn wrote:
On 19 Jan 2017, at 08:45, "[email protected] <javascript:>"
<[email protected] <javascript:>> wrote:
On 18/01/2017 22:10, Oliver Rew wrote:
Hi, I recently got very exciting about running my CNC from
MachineKit on a BBB, but after getting a BBB and a cape, my
hopes were quickly deflated by the lethargic performance of the
GUI's on the BBB. I have tried tkMachineKit and it is indeed
faster, but is nowhere near the convenience and speed of
GMOCCAPY or AXIS running on a very old x86.
Retrofitting old CNC's is a hobby of my father and myself, and
my hope with MachineKit on BBB was for a small, simple, drop-in
system that didn't require me to search for a big and old x86
box and deal with all it's ambiguities in order to run LinuxCNC.
However, these limitations on GUI's have dashed my hopes.
Not sure how running on a x86 can be described as full of
ambiguities compared to a BBB!
People have been doing it for years before the BBB was even made.
A BBB feels slower than running on an old x86, because it is.
I would suggest that you have a basic level rethink and look at
the available mini ITX boards available like the J1900, which are
known to work well and have 10x the processing power of a BBB,
with inbuilt graphics that work well.
I've used a couple of J1900 motherboards and they work great.
Together with a 5i25/6i26 + 7i76 mesa card.
Some of the fanless ones take a 12v supply and can quite easily
be built into a controller head for instance.
I have attempted to research the underlying problems with the
GUI, but these discussion quickly go beyond my knowledge. As far
as I can tell, it mostly has to do with the 3D g-code preview
window and the graphics driver behind it. Is the problem that
the driver doesn't run well on the BBB, or that any sort of 3D
g-code simulation is too process intensive for the BBB? People
have said disabling the g-code simulation does increase speed
significantly, but g-code simulation is a standard feature on
most CNC softwares and is very handy for the end user.
Are there any current, more feature-rich GUI's that work well on
the BBB? Are there any in development?
Have a look at the remote GUI capabilities of qtquickvcp and Cetus.
Thanks!
--
website: http://www.machinekit.io blog:
http://blog.machinekit.io github: https://github.com/machinekit
---
You received this message because you are subscribed to the
Google Groups "Machinekit" group.
To unsubscribe from this group and stop receiving emails from
it, send an email to [email protected] <javascript:>.
Visit this group at https://groups.google.com/group/machinekit
<https://groups.google.com/group/machinekit>.
For more options, visit https://groups.google.com/d/optout
<https://groups.google.com/d/optout>.
--
website: http://www.machinekit.io blog: http://blog.machinekit.io
github: https://github.com/machinekit
---
You received this message because you are subscribed to the
Google Groups "Machinekit" group.
To unsubscribe from this group and stop receiving emails from it,
send an email to [email protected] <javascript:>.
Visit this group at https://groups.google.com/group/machinekit
<https://groups.google.com/group/machinekit>.
For more options, visit https://groups.google.com/d/optout
<https://groups.google.com/d/optout>.
--
website: http://www.machinekit.io blog: http://blog.machinekit.io github:
https://github.com/machinekit
---
You received this message because you are subscribed to the Google Groups "Machinekit" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
Visit this group at https://groups.google.com/group/machinekit.
For more options, visit https://groups.google.com/d/optout.
