jeetendra_g10 wrote:
>thanks steve,
>the offer is most welcome.
>my only rant with emc2 is i could not become an active developper of emc2,after
a few months of reading around,and could not get " online" together with the dev
community.
>the description ,documentation and location ,are (last time i checked ,around 1
year ago) not as easy to implement in a matter of weeks for practical
systems,for my situation.
>
>
Yes, there's a lot of code and a lot of documentation. Even in the
areas that have sufficient documentation, it can be hard to find what
you want (due to the large amount of information as well as the
organization)
>i will check again how the documentaion for developpers fare,as right now 1 am
building 2 cnc machines,both are to be tested on emc2,for standard cartesian
moves.one is a small 2 way clamp on a xylotex 425 oz-in,the other a 4m by 4m dc
brushed servo router with 1800 oz-in continuous torque on double x ,and 900
oz-in continuous on y axis.
>
>
The double X setup will need a little bit of thought regarding homing
the dual axis. Other than that (which there should be some examples for
in the archives), driving two motors is trivial.
>but since i simply have to use a controller system that i can modify to almost
step level,to send modbus signals to other devices,tcp-ip from master image
processing computer,rs232 and rs 485,i am not quite sure which system i will be
using in final versions.
>
>
There are two or three examples of modbus communication (to drive a
couple of types of VFD), and it's pretty easy to make components to talk
to other devices. RS232 vs. 485 is a non-issue. There are several
methods of using TCP/IP to control an EMC2 machine, and you can even
have a GUI on the EMC2 machine and a remote "master" running at the same
time.
>which is why i find the mach com api,integrated on a vb.net front end quite
attractive,to serve as a hub to receive and send commands to devices,while
controlling a mach kernel in the background.
>
>
Personally, I would be scared to run a commercial machine on vb.net.
Windows is decidedly non-realtime, and even the Mach kernel, though it
does an amazing job of making Windows seem like it's realtime, just isn't.
>the other reason is with mach,i get 3 choices of step-dir outputs: g100 and
smoothstepper,or parallel port.
>
>
You also have (at least) 3 choices for step/dir with EMC2: Mesa
(parallel port or PCI-connected), Pico Systems (parallel port
connected), and parallel port. There may be others (in fact, you can
use several kinds of PCI-based I/O cards for software step generation),
and you can use different types at the same time if you run out of I/O
on one device.
>finally,the g100 can be used as an encoder to pc transducer for a CMM-arm
project i have.So having mach will mean i get to use a g100 to full potential,if
i skip on the smooth stepper,which does not have an sdk,afaik.
>
>
As can the Mesa cards, or a parallel port (for slow movement) on EMC.
In fact, a CMM is trivial to build with EMC2, since you can (a) use
non-trivial kinematics, (b) use the already-existing feedback facilities
for reading position and (c) do probes which are logged to a text file.
>so,schematically,this is my road map of what i am doing currently.will take
years to reach full potential though.
>
>
If you've got years in your roadmap, you should absolutely look at
EMC2. Even if it's harder now (it will be, since you're already
familiar with VB and not with the various languages you can use with
EMC2), it will end up being much easier in the long run. Do you know
what changes will occur on Windows 7? How about Windows 7"+1"?
Although it will only be "officially" supported for 3 to 5 years, you
can use and install a distribution of Linux forever, and can ask
questions or fix problems for as long as you like. Kernel 2.4 is still
supported, and it was originally released in January of 2001. You won't
have to concern yourself with any changes Microsoft might make to the
authentication procedure ("Windows Genuine Advantage", anyone?).
I seem to have an opinion on this topic, don't I? :)
- Steve