Linux + serial port
in [UK Linux]
|
Prev: World domination foiled again...
Next: Missing?
From: Daniel Smith on 16 Dec 2008 12:27 Martin Gregorie wrote: > - The Linux system has only the two serial ports and one spare PCI > socket: the latter is longer that the other two which contain a USB II > adapter and the system's only ethernet adapter. this might be a PCI-X slot but have you considered it might be an ISA slot if, so im sure ISA serial cards still exist out there!
From: Unruh on 16 Dec 2008 14:05 Martin Gregorie <martin(a)see.sig.for.address.invalid> writes: >On Tue, 16 Dec 2008 01:30:50 +0000, Unruh wrote: >> >> How about attaching the time receiver to the usb serial port? Since the >> time is a linux only thing (and since the linux kernel etc keeps running >> under Wine), and since linux sees that usb serial port, it would seem >> that should work. >> >I doubt that this would work over USB since the MSF receiver's port >handling is very non-standard. Its Jonathon Buzzard's design, built round >the Galeon module. The problem is that it doesn't use the Rx and Tx lines >at all. It only uses GND, DTR and DCD. DTR powers the receiver. The DCD >level follows the Galeon's output, which the driver decodes and passes to >ntpd via shared memory. Well, since the usb port has a 5 V power supply, using that to power the receiver is trivial. Then the question is whether or not one can drive the signal lines of the usb appropriately is the only question. I am not sure how or how well the usb-serial ports mimic the control lines. But that would seem to me to be the best bet for you. >-- >martin@ | Martin Gregorie >gregorie. | Essex, UK >org |
From: Unruh on 16 Dec 2008 14:08 Ian Rawlings <news06(a)tarcus.org.uk> writes: >On 2008-12-16, Martin Gregorie <martin(a)see.sig.for.address.invalid> wrote: >> I doubt that this would work over USB since the MSF receiver's port >> handling is very non-standard. Its Jonathon Buzzard's design, built round >> the Galeon module. The problem is that it doesn't use the Rx and Tx lines >> at all. It only uses GND, DTR and DCD. DTR powers the receiver. The DCD >> level follows the Galeon's output, which the driver decodes and passes to >> ntpd via shared memory. >Even if it did work, the USB<->serial interface isn't likely to help >accuracy at all, kludges on top of kludges on top of kludges aren't >good things to have in a fairly precise timing chain. It depends on the errors those kludges introduce. The accuracy of the radio receiver is anyway in the many millisec range ( eg even the distance from the receiver introduces a few msec errors) If the usb port introduces only 100s of usec errors it is irrelevant. If it introduces 100s of msec, it is a problem. All engineering is a matter of controlled kludges. >-- >Blast off and strike the evil Bydo empire! >http://youtube.com/user/tarcus69 >http://www.flickr.com/photos/tarcus/sets/
From: Martin Gregorie on 16 Dec 2008 14:37 On Tue, 16 Dec 2008 17:27:01 +0000, Daniel Smith wrote: > Martin Gregorie wrote: > >> - The Linux system has only the two serial ports and one spare PCI >> socket: the latter is longer that the other two which contain a USB >> II adapter and the system's only ethernet adapter. > > > this might be a PCI-X slot but have you considered it might be an ISA > slot if, so im sure ISA serial cards still exist out there! No ISA, only PCI. All three connectors look the same (same xsection and colour) except that one is longer than the other two. I've just been onto the IBM site and double checked. It says 3 x PCI 2.2. -- martin@ | Martin Gregorie gregorie. | Essex, UK org |
From: Martin Gregorie on 16 Dec 2008 14:54
On Tue, 16 Dec 2008 14:13:09 +0000, Martin Gregorie wrote: > On Mon, 15 Dec 2008 23:56:21 +0000, Theo Markettos wrote: > >> Can you see the USB adaptor in Linux - ie does /dev/ttyUSBx appear? If >> not, fix that first. >> > Yes - tail -f /var/log/messages shows it being connected and > disconnected. > >> In my ~/.wine/wine.conf I have: >> >> Com2=/dev/ttyUSB0, 115200 >> >> which must have done the trick at some point (but I can't remember >> precisely what I put it there for). >> > That sounds like what I need. I'll try it shortly (as com3) and report > back. Well, I tried it and it didn't work, unfortunately. FYI, this part of the WINE config has changed entirely: .wine/wine.conf no longer exists. Instead, you configure non-default serial ports by putting COMx symlinks in .wine/dosdevices alongside the drive letter definitions. Anyway, I tried linking COM2, COM3, COM2: and COM3: to /dev/ttyUSB0 but without any success for any of them. I know that's the right name for the USB adapter because I could see it in /var/log/messages when I plugged the adapter in. None of these allowed the downloader to see the EW logger. At this point I stopped the radioclock server and unplugged the MSF clock, replacing it with the lead I pulled out of the USB-serial adapter. Now booting the logger got a reaction from the download program, but something is still wrong, as the download program wasn't able to complete its login to the EW logger: it looks as though some sort of handshake gets lost between the uploader and the serial port. I'll see if I can get any help from EW and r.a.s, since the EW uses non- standard D-9 connections at its end and the interface operation is not documented. -- martin@ | Martin Gregorie gregorie. | Essex, UK org | |