Date: Wed, 22 Jun 2016 01:45:38 +0100 From: "James C Elstone" <james.c.elstone@ntlworld.com> To: <freebsd-arm@freebsd.org> Subject: RE: pl2303 lockups on rpi2 Message-ID: <000001d1cc1f$656511b0$302f3510$@c.elstone@ntlworld.com> In-Reply-To: <1466551193.34556.76.camel@freebsd.org> References: <fbsd@www.zefox.net> <20160619020311.GC38492@www.zefox.net> <20160619034248.1D097406057@ip-64-139-1-69.sjc.megapath.net> <20160621041439.GA2449@www.zefox.net> <CANocirkCFXr8zDQZQv_Cm6j4UYDFHEuvQ2Fwn7wmwPuqkvTWaw@mail.gmail.com> <20160621223531.GC2449@www.zefox.net> <1466551193.34556.76.camel@freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Hi Bob, Just a few more thoughts to ponder: It may be an idea to check the status / strength (voltage) of the TX>RX / RX<TX lines to see if they are being driven at MARK or SPACE, or floating before "lifting them", as if the failure status is consistent in a specific pattern you could look at an external circuit that handles that eventuality correctly. Also have you tried passing the RS232 through a line driver - something like the MAX238 in back to back mode, although this could be compensating for the "UART" on either side being a bit "off"? (no pun intended)... Is the RS232 cable length between the devices more than a few metres, and is the cable shielded? There may be some mileage in using a less quality network cable for a shorter period, which doesn't have a continuous earth connection for the shield, thus taking out potential ground loops, floating earths and any other interesting and relevant hoo-ha's? Kr, James (P.S. Excuse me if this is irrelevant, I have joined this thread half way through, so feel free to shoot me down in kind flames)! -----Original Message----- From: owner-freebsd-arm@freebsd.org [mailto:owner-freebsd-arm@freebsd.org] On Behalf Of Ian Lepore Sent: 22 June 2016 00:20 To: bob prohaska; James Elstone Cc: freebsd-arm@freebsd.org Subject: Re: pl2303 lockups on rpi2 On Tue, 2016-06-21 at 15:35 -0700, bob prohaska wrote: > On Tue, Jun 21, 2016 at 10:34:41PM +0100, James Elstone wrote: > > Hi Bob, > > > > What sort of flow control (rts/cts, xon/xoff, or none)? > > > > I suspect you cleared the control signals rather than unfreeze?? > > There are none to clear: The serial end is on an RPI2, with only TX, > RX and ground. The USB end is on a second RPI2. > > In that particular case, I tried unplugging and replugging the USB > end, to no avail. Then I dragged over an iMac with Prolific's drivers > installed and plugged the USB end into the iMac. The > PL2303 still wasn't recognized. While it was still plugged into the > iMac, I lifted the TX, RX and ground connections on the PL2303 cable. > The PL2303 recognition message immediately showed up on the iMac's > console. > > As it happens, the two RPI2s in question share a ground through the > wired Ethernet. They're connected to two different switches, which are > in turn connected by a length of cable. > > The ground cable on the PL2303 forms a second, parallel ground, > amounting to a loop. This is universally considered bad practice but > is often gotten away with. Perhaps I'm not so lucky. > > The Prolific driver installation instructions specify installing the > software, connecting the USB end next and connecting the serial end > last. I thought they were just pacifying the pedantic among us, but > maybe there's a physical reason behind the advice. > > Next time a PL2303 locks up I'll begin by lifting the serial cable > connections, ground first, to see what the uplcom driver does. If it > subsequently recognizes the PL2303 I _think_ that supports the notion > of a wiring problem. > > Thanks for reading and special thanks to Hal Murray for introducing > the idea of serial devices being "held" by signaling voltages, > > bob prohaska > If the serial side has active signals and the usb side is unconnected then you are essentially applying power to some of the IO pins of an otherwise unpowered chip. Sometimes you end up accidentally powering the chip that way and it kinda-sorta works. Other times it just fries the chip. And sometimes it's harmless (usually only because the active signal line(s) can't source enough current to cause any harm). -- Ian _______________________________________________ freebsd-arm@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/freebsd-arm To unsubscribe, send any mail to "freebsd-arm-unsubscribe@freebsd.org"
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?000001d1cc1f$656511b0$302f3510$>