Re: [Emc-users] 6I24 PCI or 7I80 Ethernet?

[Chris Albertson]

On Fri, Jan 13, 2017 at 3:27 PM, sam sokolik  wrote:

> I have used both.  If you need higher servo thread speeds (>2khz
> depending on hardware)  Stick with pci/pcie.
>

I thought the point of using the Mesa hardware was that it off loaded the
high speed loops from the main computer.  So the servo loop is NOT running
in there FPGA?
--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today. http://sdm.link/xeonphi
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] The noise has about beat me dowm.

[Gene Heskett]

On Friday 13 January 2017 22:52:45 Peter C. Wallace wrote:

> > So something is lieing to me, just like a glued down rug.
> >
> > Now I am trying to cobble up a reset but I've run into something I
> > gave up on once before. Namely that hal apparently has no clue what
> > the hell to do with a pin that is supposed to be a bi-directional
> > pin according to the message spit out when the watchdog bites. That
> > says I should drive that bit false to re-enable the 7i90. But I have
> > only seen that led on once, months ago.  And show config, when I try
> > to setp the pin, setp complains the pin is held high by another
> > signal and it cannot access it to set it.  Why?
>
> Do you have a non-tristate hal pin driving the has_bit pin?
> That would prevent you from setting it by hand

in this case I was using a oneshot.out-not

> resetting works for me:
>
> pcw@pcw-G41M-Combo:~/linuxcnc/configs$ halcmd setp
> hm2_7i76e.0.watchdog.timeout_ns 100
>
> hm2/hm2_7i76e.0: Watchdog timeout (100 ns) is dangerously short
> compared to hm2_write() period (100 ns)
> hm2/hm2_7i76e.0: Watchdog has bit! (set the .has-bit pin to False
> to resume) hm2/hm2_7i76e.0: trying to recover from IO error or
> Watchdog bite
>
> pcw@pcw-G41M-Combo:~/linuxcnc/configs$ halcmd setp
> hm2_7i76e.0.watchdog.timeout_ns 1000
> pcw@pcw-G41M-Combo:~/linuxcnc/configs$ halcmd setp
> hm2_7i76e.0.watchdog.has_bit false
>
> hm2/hm2_7i76e.0: error recover successful!
>
I'll print this as its shut down for the night, and adapt it to the 7i90 
tomorrow. This looks totally different compared to what was in the 
example 3 axis file I started with. That logic simply didn't work, only 
a complete pwrdown restart worked. When it was a real, lit the leds, 
watchdog bite.  You have an advantage on me however, you show a 
setp watchdog.timeout_ns parameter, which is not documented anyplace in 
the very hard to find 7i90 documentation.

The pdf of it that I got from your site, seems not to even contain the 
word "watchdog" anyplace in it, so where is this sort of info actually 
squirreled away?

Thank you Peter.
>
> Peter Wallace
> Mesa Electronics
>
>
> --
> Developer Access Program for Intel Xeon Phi Processors
> Access to Intel Xeon Phi processor-based developer platforms.
> With one year of Intel Parallel Studio XE.
> Training and support from Colfax.
> Order your platform today. http://sdm.link/xeonphi
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users


Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today. http://sdm.link/xeonphi
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] The noise has about beat me dowm.

[Peter C. Wallace]

> So something is lieing to me, just like a glued down rug.
>
> Now I am trying to cobble up a reset but I've run into something I gave
> up on once before. Namely that hal apparently has no clue what the hell
> to do with a pin that is supposed to be a bi-directional pin according
> to the message spit out when the watchdog bites. That says I should
> drive that bit false to re-enable the 7i90. But I have only seen that
> led on once, months ago.  And show config, when I try to setp the pin,
> setp complains the pin is held high by another signal and it cannot
> access it to set it.  Why?

Do you have a non-tristate hal pin driving the has_bit pin?
That would prevent you from setting it by hand

resetting works for me:

pcw@pcw-G41M-Combo:~/linuxcnc/configs$ halcmd setp 
hm2_7i76e.0.watchdog.timeout_ns 100

hm2/hm2_7i76e.0: Watchdog timeout (100 ns) is dangerously short compared to
hm2_write() period (100 ns)
hm2/hm2_7i76e.0: Watchdog has bit! (set the .has-bit pin to False to resume)
hm2/hm2_7i76e.0: trying to recover from IO error or Watchdog bite

pcw@pcw-G41M-Combo:~/linuxcnc/configs$ halcmd setp 
hm2_7i76e.0.watchdog.timeout_ns 1000
pcw@pcw-G41M-Combo:~/linuxcnc/configs$ halcmd setp hm2_7i76e.0.watchdog.has_bit 
false

hm2/hm2_7i76e.0: error recover successful!


Peter Wallace
Mesa Electronics


--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today. http://sdm.link/xeonphi
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] The noise has about beat me dowm.

[Gene Heskett]

Greetings all;

I did find one or 2 more places I could reduce the noise with. I stuck a 
17" piece of the wide tinned braid to the single point ground and found 
that connecting it to one of the pi's usb sockets with some suture 
clamps helped a bit, as did several of those tear off a wire and plug it 
onto a pin, in this case any of the p1-p2-p3 ground pins, all clamp to 
this same hunk of braid.  But nothing could completely nullify it.

So next I duplicated 2 more minmax modules so I could watch the xz axis's 
which also suffered even when the spindle vfd wasn't powered up. And 
rigged them all to the flood button for a reset.  Then the watchdog bit 
but because I'd broken the path, it did not disable the machine, and the 
pwr led on the 7i90 stubbornly remained the only one lit. And it 
continued to run the motors I did have hooked up.

So something is lieing to me, just like a glued down rug.

Now I am trying to cobble up a reset but I've run into something I gave 
up on once before. Namely that hal apparently has no clue what the hell 
to do with a pin that is supposed to be a bi-directional pin according 
to the message spit out when the watchdog bites. That says I should 
drive that bit false to re-enable the 7i90. But I have only seen that 
led on once, months ago.  And show config, when I try to setp the pin, 
setp complains the pin is held high by another signal and it cannot 
access it to set it.  Why?

My thanks to anybody who can help me.

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today. http://sdm.link/xeonphi
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] EDM machining, user interface, increments

[TJoseph Powderly]

Nicklas, Sarah and all interested in EDM
i found this resource
unlike most tech sites , the papers are free :-)
http://www.sciencedirect.com/science/journal/22128271/42/supp/C
regards, tomp tjtr33

btw: lots more than edm here!

--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today. http://sdm.link/xeonphi
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] 6I24 PCI or 7I80 Ethernet?

[sam sokolik]

I have used both.  If you need higher servo thread speeds (>2khz 
depending on hardware)  Stick with pci/pcie.

The matsuura was converted using mesa ethernet.  (7i80+lots of other 
daughter boards)  while configuring the machine - the computer / mesa 
hardware was up for weeks on end - with linuxcnc running.  Not one 
glitch.  Very happy with all of peters hardware.

sam

On 01/13/2017 05:09 PM, Les Newell wrote:
> I have a couple of rebuild projects coming up (a router and a milling
> machine) and I am trying to decide between PCI or Ethernet. I like the
> idea of Ethernet because it will probably be around long after PCIe has
> disappeared but the idea of it running in user space instead of realtime
> makes me a little nervous. Both machines will be using their original
> servo drives connected to 7I33 analog servo interfaces.
>
> Does anyone here have any recommendations?
>
> Les
>
>
> --
> Developer Access Program for Intel Xeon Phi Processors
> Access to Intel Xeon Phi processor-based developer platforms.
> With one year of Intel Parallel Studio XE.
> Training and support from Colfax.
> Order your platform today. http://sdm.link/xeonphi
> ___
> Emc-users mailing list
> Emc-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/emc-users
>


--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today. http://sdm.link/xeonphi
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

[Emc-users] 6I24 PCI or 7I80 Ethernet?

[Les Newell]

I have a couple of rebuild projects coming up (a router and a milling 
machine) and I am trying to decide between PCI or Ethernet. I like the 
idea of Ethernet because it will probably be around long after PCIe has 
disappeared but the idea of it running in user space instead of realtime 
makes me a little nervous. Both machines will be using their original 
servo drives connected to 7I33 analog servo interfaces.

Does anyone here have any recommendations?

Les


--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today. http://sdm.link/xeonphi
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users

Re: [Emc-users] Two questions.

[Gene Heskett]

On Thursday 12 January 2017 22:06:23 Gene Heskett wrote:

[snip pages of history]

> I have it running fairly noise free today, and had no such problems.
> However it appears that I may have blown the #0 stepgen, as data is
> going in, its enabled, but nothing is coming out of other either pin
> 1-1 (step) or pin 1-3 (dir). So I added another one in the config, and
> will re-configure the hal file to use it for the X axis. Where I'm
> working on the back side of this is just a hair crowded, and more than
> 10 times my clothing has rubbed on the 2" of blue styro glued to the
> inside of the garage door, giving be a weak, just enough to make me
> aware of it shock when I touched the machine for the first time after
> walking around to the back of it.

And in the middle of this, with the X driver not yet connected to 
stepgen#2, the noise is back, with a vengeance.

I have done everything I can to reduce the noise in terms of a central 
single point ground. So I just broke down and bought an assortment of 
clamp-on ferrite chokes. I have now rigged up the minmax module, 
watching the encoder velocity times 60 so I get rpms, and rigged the 
flood button to serve as a thru or freeze the mins and maxes. Even with 
the spindle stopped, it takes less than 10 seconds to record rpms in 
the - and + territory ranging from -25 to + readings that force the 
halmeter into scientific notation, by +13 for the exponent.
When a burst of noise occurs, the ringing frequency exceeds 100 MHz.

And my best guess is that it hits when all the various switching supplies 
hit a switch point at a common time and it adds up but at such a slow 
repeat rate my analog scope is too dim to see it. So the gigahertz 
digital comes out of its case when I go back out.

When they arrive, a clamp on to help kill longitudinal noise goes on each 
motor cable, as close to the drivers output as I can get it, and as many 
as there is room for in the run from the ground point to the 7i90 on the 
cable from the encoder. The maximum pulse rate into the encoder would be 
700 KHz if I try to explode the chuck so in terms of beads, I ought to 
be able to cover that from the point where its drain wire is grounded as 
it goes past that bolt on its way to the 7i90.. When the door is open, 
thats about 14". At least 7 beads will fit on that cable...

Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page 

--
Developer Access Program for Intel Xeon Phi Processors
Access to Intel Xeon Phi processor-based developer platforms.
With one year of Intel Parallel Studio XE.
Training and support from Colfax.
Order your platform today. http://sdm.link/xeonphi
___
Emc-users mailing list
Emc-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/emc-users