Re: missed keystrokes problem, back with a vengeance.
On Monday 30 January 2017 04:51:59 Jonathan Wilson wrote:
> On Sun, 2017-01-29 at 13:10 -0500, Gene Heskett wrote:
> > On Sunday 29 January 2017 12:16:52 Jonathan Wilson wrote:
> > > On Sun, 2017-01-29 at 08:47 -0500, Gene Heskett wrote:
> > > > Greetings everybody;
> > > >
> > > > I am in the process of bringing a nearly 70 yo Sheldon lathe
> > > > back to life, useing an raspberry-pi 3b for the machine
> > > > controller.
> > > >
> > > > However, I have now had 5 different keyboards plugged into it,
> > > > some wired, some wireless, and none of them can give me a
> > > > dependable response to a key, and the error seems much worse for
> > > > the key-up event. When driving the machine by hand as we often
> > > > do for one-offs, missing a keyup event can be disastrous for the
> > > > part being made because it keeps on cutting until you've given
> > > > that, or another key, a quick tap to stop the unwanted motion.
> > >
> > > I'm wondering if the problem is noise on the power, or RFI noise
> > > being picked up/interfering with the usb cable and/or PI in
> > > general. I'm guessing that the lathe is probably kicking out a
> > > shed load of RF noise into the air or on to the power line and the
> > > pi is not overly hardened for such a harsh environment.
> >
> > Understatement of the month... :)
> >
> :-) Its a good old British tradition to always understate ;-)
> :
> > This application, linuxcnc, is largely written in python, but I'm
> > not fam enough with python to go mucking about in its internals. And
> > the fact that 3 other installations of the same code on x86 boxes do
> > not have this problem tends to point a pretty stiff finger at the
> > pi.
>
> Yeah, the PI is made down to a price point (Although I'm fairly sure
> it meets some CE requirements:
> https://www.element14.com/community/docs/DOC-43262/l/updated-29-march-
>frequently-asked-questions-about-raspberry-pi but it may be that any
> standards regarding "input noise", as opposed to it generating RFI,
> are possibly not met, likewise input shielding, de-coupling, etc.) ...
> X86 boards not only have to meet Intel standards but also a host of
> International and US standards and I'm guessing that some of the
> standards, either international or Intel, exceed the minimum
> legal/ISO/CE requirements (for being able to sell into markets) when
> taken as a whole.
>
> > > > There is something funkity in the usb keyboard handling that can
> > > > get much worse with a reboot, or get almost perfect with a
> > > > reboot.
> > > >
> > > > Its all uptodate a/o yesterday. The psu is a 4 amp box, making
> > > > 5.07 volts solidly. Verified with a 100 Mhz scope.
> > > >
> > > > What can I do to fix this?
> > > >
> > > > Cheers, Gene Heskett
> >
> > Thank you. There might be the seeds of a fix in this.
>
> You're welcome.
>
> > Cheers, Gene Heskett
There was that kernel of truth in it. I put a couple clamp-on ferrite
chokes on the usb cable in the first 2" beyond the end molding of the
cable where it plugs into the pi, problem totally solved.
For about an hour.
So I next fished out the UVW wires leaving that fake ebay VFD, and put a
bigger copy of one of them on that group of wires, about 4 inches from
its output terminals. When I came around in front of the machine, and
re-enabled the VFD's power, the pyvcp tachometer went to full scale and
nothing would move. It appears I have totalled the 7i90 i/o card. So I
have to call that vendor and get a couple more. The power of this noise
had already blown two of the 4 stepgen modules in that one. Its not
rated for more than 5 volts, and is internally a 3.3 volt card. And
I've seen it at well above 15 volts p-p when I had switching power
supplies for the motors.
I've also been given a link to a different way to construct the spi
cable, out of a piece of cat5e or cat6 and two std IDC connectors. Much
more compact than 2 pieces of ribbon with a junction pcb between them.
The only potential problem I can see is that the conductors in the cat
cable are 23 or 24 gauge whereas the ribbon cable is smaller, 28 gauge
IIRC. But it works 3 feet long w/o any sources terms, whereas I only
need 7". What I could use is a bigger box so I could spread this out
another 2 or 3". But bigger boxes, new, start at 250 USD so this old
rusty salvage will have to do. And my local home store to carry a BX
style conduit so I could replace two pieces of plastic with loose wire
in them. They are excellent broadcast antennas for this switchmode crap
and the spindle motors wiring, which currently shares some conduit with
the rest of the power wiring is going to get its own conduit at least
above the machines chip pan. I also bought a pair of 20 amp Corcom brick
wall filters for the VFD's inputs but have not installed them yet, I
need to make room for them in that junction box by putting a 4square
extension ring on it, making it a true 4x4x4. Or better yet, find a 4x6
box. Sigh.... It never ends.
Thank you very much, it did solve one problem.
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 <http://geneslinuxbox.net:6309/gene>
Reply to: