On Friday 09 December 2016 12:23:49 Peter C. Wallace wrote:
> On Fri, 9 Dec 2016, Gene Heskett wrote:
> > Date: Fri, 9 Dec 2016 12:05:42 -0500
> > From: Gene Heskett <[email protected]>
> > Reply-To: "Enhanced Machine Controller (EMC)"
> > <[email protected]>
> > To: [email protected]
> > Subject: [Emc-users] Need advice on joint errors
> >
> > Greetings all;
> >
> > I could use some advice since this raspi-7i90 uses a com protocol,
> > spi, that is relatively new, and I believe that is where a noise
> > problem might lie.
> >
> > I got it back to running the right rt kernel yesterday, and I am
> > waiting on another psu & driver for the x axis. But I did get z
> > moving. Nowhere near optimized though. That's being moved by the
> > motor and psu I took off the mill because it was too slow.
> >
> > But I am being plagued by joint following errors, both axis's, when
> > its not moving. There is not any connections yet to the x motor pins
> > on the 7i90.
> >
> > This armhf setup has no PID's, so I've yet to locate where this
> > error is generated.
> >
> > My hookup for the spi is currently some of those rip off the next
> > wire and plug it in, and each is a piss-poor connector, very poor
> > spring pressure unless I first take a tiny punch and bend the
> > contact tab inward so it will even stay on the header pin. I've seen
> > some mention of 80 ohm series damping resistors, but this initial
> > hookup is straight thru, and a bit shy of 200mm long. At the speed
> > of this interface, it really should be treated as a transmission
> > line, but its not a very good one when its all done with loose
> > wires.
> >
> > I got some of the 10 pair female connector blocks from digi-key
> > yesterday, and the gold flashed female pins, and the crimping tool
> > came in Tuesday, so I think the first thing I'll do today, is cut
> > those junk connectors of these jumpers, and make it back up using
> > these good connectors. That will shorten each wire something under
> > an inch. It also gives me a chance to put some damping resistors in
> > the signal lines, possibly cutting down on the ringing and the data
> > errors that I'm convinced is the source of the joint errors being
> > reported when there is not any motion.
> >
> > So my question is, are 100 ohm R's, which I have in bulk, too much?
> >
> > The one mention was of 80 ohm chip R's on the OHPark adaptor board I
> > assume, which puts them right on the r-pi's header, but when the
> > door of this box is closed, I don't have room for any additional
> > stickup above the r-pi, but the interface on the 7i90 faces straight
> > out, so putting them there, on the other, 7i90 end, is quite easily
> > done.
> >
> > I'm also aware this is new ground, so what have the explorers here
> > found is the best solution for an spi driven setup?
> >
> > Did you use loose jumpers, or did you use a ribbon cable, arranged
> > to put a grounded wire between the signal wires? I have around 80
> > feet of 26 wire ribbon cable, so thats not a problem.
> >
> > The pin pattern on the r-pi is easily covered by one, or a piece of
> > one of these header blocks I have now.
> >
> > Using ribbon cable makes it a lot easier to think of it as a
> > transmission line, but how best to terminate it for least vswr?
> >
> > So lets talk.
> >
> > Thanks everybody.
> >
> > Cheers, Gene Heskett
>
> The SPI pinout is designed to use flat cable which has interleaved
> grounds. any flying wire interface is liable to be troublesome
>
> There are 2 issues here, 1. Is controlled impedance (the flat cable
> has ~!30 ohm single ended impedance) and 2. Is ground bounce (you want
> as many grounds between the 7I90 and the SPI master to avoid ground
> bounce, 8 are available on the flat cable)
I've had my pain from that, in a circuit running at 1% of the speed of
this one. That thought had also come to mind, and at least 3 are in my
existing scheme. However, available docs that I have has reduced me to
searching for grounds on the r-pi header with an ohmmeter.
> Also regards grounding, you need to watch out for ground loops that
> let high frequency noise currents flow between the 7I90 and the CPU.
Humm, would one of those clamp-on ferrite chokes help? Packrat that I am,
I'm pretty sure I could find one.
> The most critical signal is the clock from the SPI master. Depending
> on the drive strength it may require a series termination resistor
> (the 7I90 SPI output are series terminated so need no termination )
Thank you Peter, this points me in the right direction. And I'll term the
signals _from_ the r-pi at the r-pi. Would it be too much loss if I used
120 ohm r's? I am pretty sure I can find enough of those. Obviously I
haven't a clue what the source impedance of a pi gpio pin is or what its
swing is. The r-pi of course is running on 5 volts but I'd expect thats
cut way down for power savings.
I have a 2 gigahertz sampling scope, 1 ghz per trace but just looking at
this stuff with a 10pf scope probe reminds me of Schroedingers famous
cat.
> Peter Wallace
> Mesa Electronics
>
> (\__/)
> (='.'=) This is Bunny. Copy and paste bunny into your
> (")_(") signature to help him gain world domination.
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>
------------------------------------------------------------------------------
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
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users
