There are two modes of operations:
- mode operating on buffer, so for ex running a path from G-code
- mode operating on commands, for ex: homing. In this mode controller works 
really independent. Movement is generated by uC in CSMIO. For ex. CSMIO get 
homing cmd on selected axis from PC and then in generates step/dir or +-10V 
independently. Controller have configuration with inputs (hom, limits etc) and 
it checked when homing pin is reached, then it finishes the homing procedure 
and set one of 'homing is done flag'. Other independent commands is for ex 
probing, jog etc.

So about buffers. I know now there are new positions every 1mS. 
I have one idea. CSMIO Linux driver can have a buffer which contains points in 
this 1ms interrupt. When this buffer will be for ex. 50% full I can start 
sending of these points. I think this method will be quite comfortable.

Dave said:
" Most of these buffered devices must maintain a certain buffer level or else 
they go into shutdown mode - so they can bring the motors to a controlled stop.
If they run out of buffer, they have no idea what to do next and they can't 
stop the motor in 1 ms so they get into trouble."
Yes. When controller works in buffered mode then buffer can’t be empty (in 
exception when we want to finish this mode). You can imagine for ex situation 
when axis runs with speed few m/s and then run out of points in buffer;]. But 
if the buffer is quite big (I am using now 500 ms) that's no problem. That's 
impossible even in non RT system to have so long transmission pauses.

I thing any RT Ethernet drivers are not needed. Of course it is possible to do 
that on time slots in etherCAT for ex. But buffered method is really ok. We can 
use yours Ethernet port on mother board with standard drivers. And machine can 
operate in company local network. Here I have one machine with CSMIO, 
additionaly two CSMIO on my desk and about 10 computers in local network and 
there is no problems. Machine can operate all the time :).

Lukasz

-----Original Message-----
From: andy pugh [mailto:[email protected]] 
Sent: Thursday, May 31, 2012 5:17 PM
To: EMC developers
Subject: Re: [Emc-developers] CSMIO Eth motion controller + EMC2

On 30 May 2012 15:45, Łukasz Prymula <[email protected]> wrote:

> For me, most interesting thing is that in what method planner transfers 
> trajectory to HAL layer. As you see I need points buffer.

Can your system run with an empty buffer? That would make the interface with 
LinuxCNC much easier. If you can simply accept a new set of axis positions 
every 1mS then there would be very little integration effort.

I think this would require your device to run in a "slightly dumb" mode.

--
atp
If you can't fix it, you don't own it.
http://www.ifixit.com/Manifesto

------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and threat 
landscape has changed and how IT managers can respond. Discussions will include 
endpoint security, mobile security and the latest in malware threats. 
http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Emc-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-developers


------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
Emc-developers mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-developers

Reply via email to