AirCard 750 on Linux Platform :NO CARRIER - PPP

This is a discussion on AirCard 750 on Linux Platform :NO CARRIER - PPP ; Has anyone connected to the Internet via the AirCard750 wireless modem card on the Linux 2.4.18-14 # platform or any version of Red Hat above 7x?? I found a couple of docs including the guide presented at Sierra's website and ...

+ Reply to Thread
Results 1 to 13 of 13

Thread: AirCard 750 on Linux Platform :NO CARRIER

  1. AirCard 750 on Linux Platform :NO CARRIER

    Has anyone connected to the Internet via the AirCard750 wireless modem
    card on the Linux 2.4.18-14 # platform or any version of Red Hat
    above 7x?? I found a couple of docs including the guide presented at
    Sierra's website and configured with no success. All configuration
    adjustments result in the following:

    ***********************************************
    [root@localhost peers]# pppd call ac750
    send (ATZ^M^M)
    expect (OK)
    ATZ^M^M
    OK
    -- got it

    send (AT+cgdcont=1,"IP","internet2.voicestream.com"^M^M)
    expect (OK)
    ^M
    AT+cgdcont=1,"IP","internet2.voicestream.com"^M^M
    OK
    -- got it

    send (AT+csq^M^M)
    expect (OK)
    ^M
    AT+csq^M^M
    +CSQ: 15,7^M
    ^M
    OK
    -- got it

    send (ATD*99#^M^M)
    expect (CONNECT)
    ^M
    ATD*99#^M~~~^?}#@!}!p} }<}!}$}&@}#}$@#}%}&_8^G}"}&} } } } ^M
    NO CARRIER^M
    alarm
    Failed
    ************************************************** ***
    It seems to bug out after the ATD*99# command as if that number does
    work. I've also tried the command manually in 'minicom'. I've
    increased my timeout value in ac750chat for the 'chat' command:

    ac750: (File)
    *************
    -detach
    /dev/ttyS2
    57600
    debug
    defaultroute
    usepeerdns
    user ''
    show-password
    crtscts
    lock
    connect "/usr/sbin/chat -v -s -t480 -f /etc/ppp/peers/ac750chat"

    ac750chat (File)
    ****************
    '' ATZ
    OK AT+cgdcont=1,"IP","internet2.voicestream.com"
    OK AT+csq
    OK ATD*99#
    CONNECT ''



    This was in order to give the modem time to come back with a response
    where as before I would get a 'Failed' on script sooner. the AT+csq
    was to let me know my signal strength and according to Sierra's notes
    I have good reception. Is there any other options I can try or is
    there some kind of configuration I'm missing?? I have T-Mobile as my
    carrier and they stated to me I do not need to use a userid/password.

    Thanks in advance.

  2. Re: AirCard 750 on Linux Platform :NO CARRIER

    juggernaut7_@hotmail.com (-juggernaut7) writes:
    > send (ATD*99#^M^M)
    > expect (CONNECT)
    > ^M
    > ATD*99#^M~~~^?}#@!}!p} }<}!}$}&@}#}$@#}%}&_8^G}"}&} } } } ^M
    > NO CARRIER^M


    Looks like a bad chat script. This device just doesn't bother sending
    back the usual CONNECT string before connecting through to the peer.

    > ac750chat (File)
    > ****************
    > '' ATZ
    > OK AT+cgdcont=1,"IP","internet2.voicestream.com"
    > OK AT+csq
    > OK ATD*99#
    > CONNECT ''


    Delete that last line, and you should be good to go.

    --
    James Carlson, IP Systems Group
    Sun Microsystems / 1 Network Drive 71.234W Vox +1 781 442 2084
    MS UBUR02-212 / Burlington MA 01803-2757 42.497N Fax +1 781 442 1677

  3. Re: AirCard 750 on Linux Platform :NO CARRIER

    James Carlson wrote in message news:...
    > juggernaut7_@hotmail.com (-juggernaut7) writes:
    > > send (ATD*99#^M^M)
    > > expect (CONNECT)
    > > ^M
    > > ATD*99#^M~~~^?}#@!}!p} }<}!}$}&@}#}$@#}%}&_8^G}"}&} } } } ^M
    > > NO CARRIER^M

    >
    > Looks like a bad chat script. This device just doesn't bother sending
    > back the usual CONNECT string before connecting through to the peer.
    >
    > > ac750chat (File)
    > > ****************
    > > '' ATZ
    > > OK AT+cgdcont=1,"IP","internet2.voicestream.com"
    > > OK AT+csq
    > > OK ATD*99#
    > > CONNECT ''

    >
    > Delete that last line, and you should be good to go.



    Ok great! That allowed me to establish a good connection, however it
    seems the modem needs more time to confirm good configuration
    information packets it sends:

    [root@localhost peers]# pppd call ac750
    send (ATZ^M^M)
    expect (OK)
    ^M
    OK
    -- got it

    send (AT+cgdcont=1,"IP","internet2.voicestream.com"^M^M)
    expect (OK)
    ^M
    ATZ^M^M
    OK
    -- got it

    send (AT+csq^M^M)
    expect (OK)
    ^M
    AT+cgdcont=1,"IP","internet2.voicestream.com"^M^M
    OK
    -- got it

    send (ATD*99#^M^M)
    Serial connection established.
    using channel 5
    Using interface ppp0
    Connect: ppp0 <--> /dev/ttyS2
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    Modem hangup
    Connection terminated.
    [root@localhost peers]#


    I searched around a bit and found that I may need to increase the
    'lcp-max-configure' value. Is that correct? I'm not sure which file
    it suppose to go in but default I think its /etc/ppp/options but I
    still get the same results or it doesn't seem to try to send 'ConfReq'
    after 10 or 12 times (default =10). Any additional ideas or is there
    something missing?? Here's what I have:

    /etc/ppp/options
    *****************
    lock
    passive
    lcp-max-configure 60
    ipcp-max-configure 30

    ....and I've added these to my 'ac750' file in /etc/ppp/peers/ac750.

    Your help is much appreciated!
    -juggernaut7_@hotmail.com

  4. Re: AirCard 750 on Linux Platform :NO CARRIER

    juggernaut7_@hotmail.com (-juggernaut7) writes:
    > send (ATD*99#^M^M)
    > Serial connection established.


    It's possible that you need more in your chat script here, but since I
    don't have your communications device to play with, it's very hard to
    tell.

    I would start by using 'miniterm', 'kermit', or 'cu' to dial the
    device directly and figure out what belongs in the chat script.

    > I searched around a bit and found that I may need to increase the
    > 'lcp-max-configure' value. Is that correct?


    No. Timing out here means that the peer isn't speaking PPP at all,
    and that's usually due to a chat script failure, not due to a problem
    with PPP itself.

    --
    James Carlson, IP Systems Group
    Sun Microsystems / 1 Network Drive 71.234W Vox +1 781 442 2084
    MS UBUR02-212 / Burlington MA 01803-2757 42.497N Fax +1 781 442 1677

  5. Re: AirCard 750 on Linux Platform :NO CARRIER

    James,

    I've tweaked my ac750chat script to reflect the following that gave me
    a different output (sorry its taken me so long..)

    '' ATZ
    AT&F0 OK
    AT&K00 OK
    AT+cgdcont=1,"IP","internet2.voicestream.com"
    ATDT*99***3# TIMEOUT 884

    This allows me now to get a 'recv' back but I guess it doesn't mean
    anything..(or that's what I'm trying to find out). After several sent
    requests I get:

    sent [LCP ConfReq id=0x1
    ]
    rcvd [LCP ConfReq id=0x1
    ]
    sent [LCP ConfNak id=0x1 ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]
    sent [LCP ConfReq id=0x1
    ]

    eventually it'll come back with:

    LCP: timeout sending Config-Requests

    Does that mean the peer I'm trying to connect to (T-Mobile) will not
    talk to me yet because I haven't sent the right type requests? I've
    still to work this in minicom (temporarily forgot how to use it)...

    Thanks man!
    juggernaut7

    James Carlson wrote:
    > juggernaut7_@hotmail.com (-juggernaut7) writes:
    > > send (ATD*99#^M^M)
    > > Serial connection established.

    >
    > It's possible that you need more in your chat script here, but since

    I
    > don't have your communications device to play with, it's very hard to
    > tell.
    >
    > I would start by using 'miniterm', 'kermit', or 'cu' to dial the
    > device directly and figure out what belongs in the chat script.
    >
    > > I searched around a bit and found that I may need to increase the
    > > 'lcp-max-configure' value. Is that correct?

    >
    > No. Timing out here means that the peer isn't speaking PPP at all,
    > and that's usually due to a chat script failure, not due to a problem
    > with PPP itself.
    >
    > --
    > James Carlson, IP Systems Group


    > Sun Microsystems / 1 Network Drive 71.234W Vox +1 781 442

    2084
    > MS UBUR02-212 / Burlington MA 01803-2757 42.497N Fax +1 781 442

    1677


  6. Re: AirCard 750 on Linux Platform :NO CARRIER

    juggernaut7_@hotmail.com writes:
    > I've tweaked my ac750chat script to reflect the following that gave me
    > a different output (sorry its taken me so long..)
    >
    > '' ATZ
    > AT&F0 OK
    > AT&K00 OK
    > AT+cgdcont=1,"IP","internet2.voicestream.com"
    > ATDT*99***3# TIMEOUT 884
    >
    > This allows me now to get a 'recv' back but I guess it doesn't mean
    > anything..(or that's what I'm trying to find out). After several sent
    > requests I get:
    >
    > sent [LCP ConfReq id=0x1
    > ]
    > rcvd [LCP ConfReq id=0x1
    > ]
    > sent [LCP ConfNak id=0x1 ]


    You're talking to yourself. Either the modem didn't connect (and is
    still in command mode) or the peer is broken.

    Either way, the serial link is looped-back on itself. The bits you
    transmit are the ones you're receiving.

    > Does that mean the peer I'm trying to connect to (T-Mobile) will not
    > talk to me yet because I haven't sent the right type requests? I've
    > still to work this in minicom (temporarily forgot how to use it)...


    No, it means that the serial link itself isn't working. Most likely,
    it's something wrong in the chat script, but it's hard to tell here.

    --
    James Carlson, IP Systems Group
    Sun Microsystems / 1 Network Drive 71.234W Vox +1 781 442 2084
    MS UBUR02-212 / Burlington MA 01803-2757 42.497N Fax +1 781 442 1677

  7. Re: AirCard 750 on Linux Platform :NO CARRIER

    When you say "No, it means that the serial link itself isn't working".
    (I know its a silly question) Can you elaborate? I can plug this
    aircard into my XP laptop and it works just great of course with
    Tmobile's watcher software but as soon as its up on my linux box the
    above happens. I was finally able to get minicom working but when I
    ran 'Alt-a K' to run the chatscript I would get scribble accross the
    screen and in the logs it would re-attempt what I did on the command
    line which I still recieved the:

    sent [LCP ConfReq id=0x3b ]
    LCP: timeout sending Config-Requests

    Like you said somehow the link is dead. Could there be something with
    the irq settings? I ran accross a posted forum where the person had
    the wrong irq for his modem. Once he changed it to a non-used one it
    worked. Here are my settings:

    [root@localhost peers]# cat /proc/interrupts
    CPU0
    0: 139748198 XT-PIC timer
    1: 3481 XT-PIC keyboard
    2: 0 XT-PIC cascade
    3: 43 XT-PIC serial
    5: 245177 XT-PIC eth0
    8: 1 XT-PIC rtc
    9: 2909 XT-PIC i810@PCI:0:1:0
    10: 0 XT-PIC PCI device 1524:1410 (ENE Technology
    Inc), Intel ICH 82801AA
    11: 0 XT-PIC usb-uhci
    12: 4960 XT-PIC PS/2 Mouse
    14: 865267 XT-PIC ide0
    15: 5561605 XT-PIC ide1
    NMI: 0
    ERR: 0

    Where at first I had both aircard(modem) and the PCI card (where the
    aircard plugs in) on IRQ 10. I've since then changed the
    /dev/modem(/dev/ttyS2) to IRQ3 as above shows. Here is an output of my
    setserial command:

    [root@localhost peers]# setserial -G /dev/ttyS2
    /dev/ttyS2 uart 16550A port 0x03e8 irq 3 baud_base 115200 spd_normal
    skip_test

    I had in the past changed the uart setting to 16550 but I think it
    stopped working all together so I put it back to 16550A.


    Also I wanted to let you know I went back to Sierra's unofficial doc
    for linux and moved the O/S to the supposed confirmed working RH7
    whereas initally I had RH8. I've read also on slash.org

    http://slashdot.org/article.pl?sid=0...37221&from=rss

    how most were able to get this working using merely a phone instead of
    a PCMCIA card. But I really need to have it work PCMCIA style.

    I'm a bit confused on what files are the minimum for it to work as
    well. The only files I'm using/modifying are:

    /etc/ppp/peers/ac750 (Vendor specific from Sierra's website)
    /etc/ppp/peers/ac750chat (Vendor specific chatfile Sierra's website)
    /etc/ppp/options (empty right now because it uses /etc/ppp/peers/ac750
    right?? It seemed to not make any difference either way)
    /etc/ppp/resolv.conf (I copied the nameservers to /etc/resolv.conf to
    be safe)
    /etc/ppp/pap-secrets (Not sure the exact settings because Tmobile
    doesn't require userid/password but I copied this file to
    'chap-secrets' also)

    In [pap-secrets] I literally have:
    # Secrets for authentication using PAP
    # client server secret IP addresses
    (none) * password '' *

    where '(none)' is typed in as I show here. Inside ac750 I put: user '
    ' but now I took it out of the file

    I've also tried:
    "guest" * "guest" *

    Using 'guest' I entered in the ac750 file: user 'guest'. None of
    these seem to make a difference.

    In my ac750chat:
    '' AT
    OK ATZ
    OK AT+cgdcont=1,"IP","internet2.voicestream.com"
    OK AT+CGDCONT?
    OK AT+CGDCONT=?
    OK AT+csq
    OK ATD*99***1# TIMEOUT 480

    What could possibly go after the 'ATD*99***1# ? it seems when I
    typed this in 'minicom' I wasn't able to type anything else to see what
    modifications my chatscript needed. The CGDCONT? / CGDONT=? helped me
    see what CID range/settings there was I guess that's why you would use
    ATD*99***(1)# as suppose to (2 or 3 or a different number) when
    AT+CGDONT=1 (where CID is =1 not 2 or 3 or a different number) Could
    there be another setting change needed here or additional commands
    missing?

    Since Tmobile doesn't require userid/password will there still be a
    need to insert (ogin: , ssword: ) commands? if so at which point?
    because after the number is dialed garble takes over the screen (that
    is in minicom) . I ran across a couple of postings where I vaguely
    understood I think they inserted something like: ppp ogin: , or just
    ppp ?? (is that like starting up pppd on my end??) not sure of what the
    postings were saying so I didn't put anything.


    Could there be a setting fix in my profile of the Aircard? Here is
    what the profile looks like currently:

    AT&V^M^M
    ACTIVE PROFILE :^M
    E1 L0 M0 Q0 V1 X4 &C1 &D2 ^M
    S0:0 S2:43 S3:13 S4:10 S5:8 S6:2 S7:60 S8:2 S10:15 S12:10 S13:60 S1:0
    ^M
    +CBST: 7,0,1^M
    +CSMS: 0^M
    +CRLP: 61,61,128,6,1,3^M
    +CRC: 0^M
    +CR: 0^M
    +FCLASS: 0^M
    +IFC: 2,2^M
    +IMODE: 0^M
    +ICF: 3,3^M
    +DR: 0^M
    +CMGF: 0^M
    +CSDH: 0^M
    +CNMI: 2,1,0,0,0^M
    +ILRR: 0^M
    +IPR: 115200^M
    +DS: 3,0,512,20^M
    +CMEE: 0^M
    +CREG: 0^M
    +CCUG: 0,0,0^M
    +CLIP: 0^M
    +COLP: 0^M
    +CCWA: 0^M
    +CAOC: 1^M
    +CLIR: 0^M
    +CUSD: 0^M
    +CFUN: 1^M
    +CSCA: "+12063130004",145^M
    +CSMP: 17,167,0,241^M
    ^M
    OK
    -- got it

    Some postings stated to use AT&F0 or AT&F1 for the fix. My card
    registers this as an ERROR each time I inserted that into my
    chatscript.

    I'm running:
    -Linux RH 7 2.4.18-3 #1 Thu Apr 18 07:37:53 EDT 2002 i686
    -ppp-2.4.1-3
    -chat (version not able to determine yet but it came with the install)


    What about the settings in 'ac750' versus 'options'? Is there a link
    that explains all of the various parameters? I stumbled across another
    posting that had the option 'persist' which kind of helped me in
    troubleshooting. Also another posting said to insert 'debug' 3 times
    like:

    debug debug debug

    to get more info but didn't really make a difference. Also there were
    these others I saw listed in other postings so I added them just to see
    what would happen. Here is what I have currently and I'm not sure if I
    need any of these or if they are in the correct order in the file:

    -detach
    debug debug debug
    /dev/modem -> has been changed to /dev/ttyS2 (same as modem) before
    57600 -> has been changed to 115200 before
    noipdefault
    defaultroute
    usepeerdns
    passive
    local
    -crtscts
    persist
    connect "/usr/sbin/chat -v -s -t480 -f /etc/ppp/peers/ac750chat"
    ipcp-accept-local
    ipcp-accept-remote
    mru 1500
    mtu 1500
    novj
    nobsdcomp
    novjccomp
    nopcomp
    noaccomp
    [root@localhost peers]#


    Its taking me a while to get back to you (if you're still around)
    because I've been trying all these different scenarios and researching
    different forums. I'm almost willing to ship out my aircard to you if
    you think that would help you better understand a fix if need be but I
    really appreciate any additional help/insight you can provide. In the
    mean time do you know of any books I can purchase on PPP or GPRS that
    would point me in the right direction?

    Thanks again,

    juggernaut7


  8. Re: AirCard 750 on Linux Platform :NO CARRIER

    juggernaut7_@hotmail.com writes:
    > When you say "No, it means that the serial link itself isn't working".
    > (I know its a silly question) Can you elaborate?


    The bytes that are transmitted are not being received by the peer, or
    the bytes sent by the peer aren't being received locally.

    > I can plug this
    > aircard into my XP laptop and it works just great of course with
    > Tmobile's watcher software but as soon as its up on my linux box the
    > above happens. I was finally able to get minicom working but when I
    > ran 'Alt-a K' to run the chatscript I would get scribble accross the
    > screen and in the logs it would re-attempt what I did on the command
    > line which I still recieved the:
    >
    > sent [LCP ConfReq id=0x3b ]
    > LCP: timeout sending Config-Requests


    That still means the low-level link is non-functional.

    > Like you said somehow the link is dead. Could there be something with
    > the irq settings?


    Sure. Could be almost anything, unfortunately. In general, the most
    common reason this happens is that the chat script itself is
    incorrect. Since I know nothing about the device you're using, I
    really don't know what might be wrong.

    I suggest starting with something simpler (like miniterm or kermit)
    and trying to get a basic serial link working *first* and then try to
    translate that into a working chat script.

    > I'm a bit confused on what files are the minimum for it to work as
    > well. The only files I'm using/modifying are:
    >
    > /etc/ppp/peers/ac750 (Vendor specific from Sierra's website)
    > /etc/ppp/peers/ac750chat (Vendor specific chatfile Sierra's website)


    That's a strange place to put a chat file. Probably doesn't matter
    though.

    > /etc/ppp/options (empty right now because it uses /etc/ppp/peers/ac750
    > right?? It seemed to not make any difference either way)


    Pppd options can be placed in any file you like. The only real
    difference is that options on the command line and $HOME/.ppprc are
    unprivileged and the ones under /etc/ppp/ are privileged. The files
    in the /etc/ppp/peers location can be loaded up using the "call"
    option.

    Other than that, it's not really an issue. It has nothing to do with
    whether you have a working serial link or not. It won't matter until
    you get PPP messages on the link, which you don't have now.

    > /etc/ppp/resolv.conf (I copied the nameservers to /etc/resolv.conf to
    > be safe)


    This file gets written dynamically. I suggest using a link instead.

    > /etc/ppp/pap-secrets (Not sure the exact settings because Tmobile
    > doesn't require userid/password but I copied this file to
    > 'chap-secrets' also)


    This won't matter until you get a working serial link.

    > Using 'guest' I entered in the ac750 file: user 'guest'. None of
    > these seem to make a difference.


    It won't make a difference until you get to the point of trying to
    authenticate a PPP link. You don't yet have a serial link.

    > In my ac750chat:
    > '' AT
    > OK ATZ
    > OK AT+cgdcont=1,"IP","internet2.voicestream.com"
    > OK AT+CGDCONT?
    > OK AT+CGDCONT=?
    > OK AT+csq
    > OK ATD*99***1# TIMEOUT 480
    >
    > What could possibly go after the 'ATD*99***1# ?


    This is the $100,000 question. The answer depends on what the heck
    those AT commands mean for the device you have. I don't have a
    programmer's reference manual for that device. I suggest that you
    find one, or find someone else who has the device and understands how
    to use it properly.

    > it seems when I
    > typed this in 'minicom' I wasn't able to type anything else to see what
    > modifications my chatscript needed. The CGDCONT? / CGDONT=? helped me
    > see what CID range/settings there was I guess that's why you would use
    > ATD*99***(1)# as suppose to (2 or 3 or a different number) when
    > AT+CGDONT=1 (where CID is =1 not 2 or 3 or a different number) Could
    > there be another setting change needed here or additional commands
    > missing?


    Sure. I have no way of knowing. You'll need to find a technical
    reference for that device.

    > Since Tmobile doesn't require userid/password will there still be a
    > need to insert (ogin: , ssword: ) commands? if so at which point?


    No. Don't do that. Really. Don't do that.

    > Some postings stated to use AT&F0 or AT&F1 for the fix. My card
    > registers this as an ERROR each time I inserted that into my
    > chatscript.


    It depends on the device. AT&F1 works great with USR modems. What
    you have isn't a USR modem.

    > What about the settings in 'ac750' versus 'options'? Is there a link
    > that explains all of the various parameters?


    Try the pppd man page. It explains all the options.

    > I stumbled across another
    > posting that had the option 'persist' which kind of helped me in
    > troubleshooting. Also another posting said to insert 'debug' 3 times
    > like:
    >
    > debug debug debug


    No, that doesn't work. Whoever told you that was just grossly
    mistaken. pppd doesn't do that.

    > to get more info but didn't really make a difference. Also there were
    > these others I saw listed in other postings so I added them just to see
    > what would happen. Here is what I have currently and I'm not sure if I
    > need any of these or if they are in the correct order in the file:


    Order doesn't matter.

    In general, fewer options are better, but it's almost certainly that
    *none* of these options has anything to do with the problem you're
    seeing.

    > -detach
    > debug debug debug


    Just "debug"

    > /dev/modem -> has been changed to /dev/ttyS2 (same as modem) before


    I prefer to use the real device (ttyS2), but it probably doesn't
    matter.

    > passive
    > local
    > -crtscts


    You don't want those unless the hardware is horribly broken.
    "Passive" means that PPP doesn't start up on its own, and is wrong for
    a dial-up client. "Local" means that pppd ignores the modem control
    leads (DCD) and thus won't detect hangup. "-crtscts" means that flow
    control is disabled, which is unlikely to be good.

    > ipcp-accept-local
    > ipcp-accept-remote


    Those are redundant with "noipdefault".

    > mru 1500
    > mtu 1500


    Those are the default for PPP; no need to specify.

    > nobsdcomp


    No ... use "noccp" instead.

    > novjccomp


    If you say "novj" then this option does nothing.

    > nopcomp
    > noaccomp


    Those shouldn't be needed unless the peer is really broken.

    > Its taking me a while to get back to you (if you're still around)
    > because I've been trying all these different scenarios and researching
    > different forums. I'm almost willing to ship out my aircard to you if
    > you think that would help you better understand a fix if need be but I
    > really appreciate any additional help/insight you can provide. In the
    > mean time do you know of any books I can purchase on PPP or GPRS that
    > would point me in the right direction?


    I don't have GPRS service and I'm not about to get it. There's no way
    I could debug it for you.

    I suggest working *first* on getting a functional serial link. You
    don't have that today. I don't know what it takes to make this card
    work, and the "AT" commands it uses are all proprietary.

    Good luck.

    --
    James Carlson, IP Systems Group
    Sun Microsystems / 1 Network Drive 71.234W Vox +1 781 442 2084
    MS UBUR02-212 / Burlington MA 01803-2757 42.497N Fax +1 781 442 1677

  9. Re: AirCard 750 on Linux Platform :NO CARRIER

    juggernaut7_@hotmail.com wrote:
    > Thank you both for your responses. I was able to find me a Command
    > Reference Guide for the AirCard 555 which works with my AirCard750. I
    > initially started out with the few commands in the chatscript provided
    > by Sierra for Linux but that only produced hanging on the part of the
    > modem. When I deleted the last line (CONNECT '') like mentioned above
    > said I was able to make a GPRS packet data connection to the Public
    > Data Network.


    I would be interested in seeing the chat log messages. I'd like to
    know whether or not a CONNECT is issued by the PCMCIA card when you
    dial into it.

    My problem lies in establishing a link with the APN
    > which is my access point to the Internet(if I understood last postings
    > correctly). I need this before I can start to communicate via PPP.
    > Inside this GSM/GPRS AT Command Reference it said:


    > Rev 1.1A Sep 03 p.71 Make GPRS connection:
    > -note: After issuing ATD in GPRS mode, no further AT commands may
    > follow on the command line until the GPRS connection is terminated.


    This is likely to be almost always true.

    > My questions are related to the AT GPRS Commands I found....I noted
    > some options for setting the card's PDP context (cid=value). This is
    > first defined by the command
    > (+CGDCONT=,,,,,):


    > AT+CGDCONT=1,"IP","internet2.voicestream.com",,0,0


    Is this command issued before all the other commands shown below?

    > This PDP context (cid=value) however is either by default suppose to be
    > already active (AT+CGACT) and the card/modem setting value (AT+CGATT)
    > should be attached to the GPRS network or this may be something you
    > have to set yourself (What the 'Watcher' software might normally do
    > automatically?). Ordinarily with a GSM/GPRS enabled phone these
    > settings would already be taken care of and even using this pcmcia
    > AirCard with the Watcher software the settings would be I think
    > pre-programmed in.


    Wasn't Watcher something specific to a MS OS? Not many here use such
    a system - including me.

    > The Sierra AirCard 750 only allows (cid=1) to work. Here are my cards
    > default settings:


    > (inside minicom) echo is on...


    > OK
    > ath
    > ath
    > OK


    You might try ATZ at this point and see then see what AT+CGACT? shows.
    It's hard to guess what will be triggered by what with a GPRS device.

    > at+CGATT? (GPRS attach or detach; returns current GPRS-attach state of
    > at+CGATT? ...modem)
    > +CGATT: 1 <- (0 detach, 1 attached)


    > OK
    > at+CGACT=? (GPRS PDP context activation/deactivation; request info on
    > at+CGACT=? ...supported PDP context activation states)
    > +CGACT: 0, - CID NOT ACTIVE
    > +CGACT: 1, - CID ACTIVE


    > OK
    > at+CGACT? (GPRS PDP context activation/deactivation; returns current
    > at+CGACT? ...state syntax to set: +CGACT=,)
    > +CGACT: 1,0 <- (??? does this mean cid=0 is active?? if so then
    > ...that's my problem!)


    I don't think it means CID 0 is active. What little info I've found
    suggests it means that the state of PDP context activation for CID
    1 is "deactivated." The AT_CMD_GPRS.pdf from www.rfsolutions.co.uk
    seems to agree with this interpretation.

    > Since Sierra only uses a PDP context cid value of 1 after viewing my
    > settings by default on the card (above) it appears I have active cid=0.
    > Would this would mean each time after connecting via ATD*99***1# (or
    > ATD*99# -> this by default uses cid=1) the card doesn't see the PDP
    > context cid=1 as active at all so it doesn't know what to do?


    It looks like CID 1 is not active (deactivated). So if there is a
    default setting of CID 0 then it should be active. But CID 0 seems
    to be an invalid CID - see below.

    > If I'm reading this correctly then I've tried to deactivate cid=0 and
    > activate cid=1 by doing the following..but with the following results:


    > at+CGACT=0,0
    > at+CGACT=0,0
    > +CME ERROR: invalid cid value


    It appears that 0 is not a valid CID, assuming the error message means
    what it says.

    > I can seem to deactivate cid=1 ...


    > at+CGACT=0,1
    > at+CGACT=0,1
    > OK


    Apparently deactivating a deactivated CID is okay.

    > ...but interestingly here lies my problem (I think). After
    > deactivating cid=1 I check the status and it doesn't seem to affect the
    > settings:


    > at+CGACT?
    > at+CGACT?
    > +CGACT: 1,0


    > OK


    Perhaps because CID 1 is still deactivated.

    > And...when I try to activate cid=1 I get:


    > at+CGACT=1,1
    > at+CGACT=1,1
    > NO CARRIER


    What does "AT+CGACT?" show after you do this? If CID 1 is activated
    then the NO CARRIER could be moot (assuming you have yet to dial out
    and connect).

    > What in the world does that mean?? If I can't show the cid=1 in an
    > active state and I can't activate it without this 'NO CARRIER'
    > response, I'm wondering if there is some sort of permissions issue
    > preventing me to change settings in the card? or Could it be I'm
    > missing some other command(s) to allow me to change the activation
    > state of cid=1?


    Lots of questions, but none for which I have firm answers. I've not
    found any example configuration that uses AT+CGACT .

    > Is there somewhere on the net I can get a copy of the source code for
    > the Sierra Wireless Watcher program? I could then find out the actual
    > sequence of these AT GSM/GPRS commands that make the AirCard establish
    > a link.


    I rather doubt you'll find such source code, but good luck. I am at
    least as confused as you. ) Maybe more, you decide.

    --
    Clifford Kite Email: "echo xvgr_yvahk-ccc@ri1.arg|rot13"
    PPP-Q&A links, downloads: http://ckite.no-ip.net/
    /* The generation of random numbers is too important to be left
    to chance. */

  10. Re: AirCard 750 on Linux Platform :NO CARRIER

    ....It took me a bit to figure out what I found here next before
    responding..sorry it took so long.

    FYI - I found a chatscript with these commands in them..that's what
    kind of started me on this path:

    (http://www.cisco.com/univercd/cc/td/...f/m517modm.htm)

    chat-script tmobile "" "AT" TIMEOUT 5 OK "ATE0V1" TIMEOUT 5 OK "ATS0=0"
    TIMEOUT 5 OK
    "AT+CGATT=1" TIMEOUT 5 OK "AT+CGACT=1,1" TIMEOUT 5 OK "ATD*99***1#"
    TIMEOUT 10 CONNECT


    I found Rev 3.x of the AT Command Reference GSM/GPRS for the 700 series
    pcmcia AirCard but It appears you are correct when you stated the
    command AT+CGACT=1,0 is actually referring to instead of the
    reverse like the ACT+CGACT? makes it out to appear. Looking into the
    referenced link you posted I noticed the syntax as the following:

    AT+CGACT=,[,,,,]

    I stumbled across the following (initially a typo) which I found quite
    interesting:


    ath
    ath
    OK
    atz <- This command didn't seem to make a difference on affecting
    atz ...the output of AT+CGACT
    OK
    at+CGACT?
    at+CGACT?
    +CGACT: 1,0 (apparently is NOT ACTIVE)

    OK
    AT+cgdcont=1,"IP","internet3.voicestream.com",,0,0
    AT+cgdcont=1,"IP","internet3.voicestream.com",,0,0
    OK

    (AT+cgdcont defines so it has to be before below comands)

    at+CGACT=1,1,
    at+CGACT=1,1, <- notice the added ',' at the end.
    OK <- Amazing!! it worked!
    at+cgdcont?
    at+cgdcont?
    +CGDCONT: 1,"IP","internet3.voicestream.com","208.054.156.201",0,0

    OK
    at+CGACT?
    at+CGACT?
    +CGACT: 1,1 (NOW...the is in ACTIVE state)

    OK


    Somehow does this prove I'm connected to my ISP because I can pull an
    IP address?? That IP address pulled ended up being:

    (name = mc99c36d0.tmodns.net.) <- nslookup output

    At that point I was estatic. however now in minicom I started to dial
    the GPRS number to connect to my APN (even though I'm not sure if I
    need to at all) and this is what I got:

    atd*99#
    atd*99#
    +CME ERROR: phone-adaptor link reserved

    I'm thinking this means I have the IP already that my APN wants to
    assign me and now won't let me connect because of it or I pulled it too
    soon??

    I backed out this command by issuing AT+CGACT=0,1 to deactivate
    . I also had to redefine the by running the AT+CGDCONT
    command again. I've streamlined my chat script to the following to
    get right to trying to connect (minicom is extremely slow!):

    '' AT
    OK ATZ
    OK AT+cgdcont=1,"IP","internet3.voicestream.com",,0,0
    OK "ATD*99#" TIMEOUT 480
    \~ ''

    Looking at the PPP HOWTO I changed the last string to:

    \~ ''

    because I could never get the expected CONNECT with the (CONNECT '')at
    the end of the chatscript. Deleting this line would respond similar to
    this (\~''). My logs now look like this:


    send (ATD*99#^M)
    timeout set to 480 seconds
    expect (~) <- '~' added in the chatscript
    ^M
    ATD*99#^M~
    -- got it
    send (^M) <- The '' added after the (~) in chatscript
    expect (^M)
    ~~^?}#@!}!}?} }<}!}$}&@}#}$@#}%}&_8^G}"}&} } } }^M
    -- got it
    Serial connection established.
    Using interface ppp0
    Connect: ppp0 <--> /dev/pts/2
    net.agent: assuming ppp0 is already up
    LCP: timeout sending Config-Requests
    Terminating on signal 2.

    Now whats interesting is when I changed the 'chat' command's options
    (to get a bit more output) I came closer to what I think is happening.
    Prior to change the chat command was:

    connect "/usr/sbin/chat -v -s -t480 -f /etc/ppp/peers/ac750chat"

    now changed to:

    connect "/usr/sbin/chat -e -V -s -t480 -f /etc/ppp/peers/ac750chat"

    The (-e, -V) options disabled logging to my syslog and outputed the
    echo from the modem as if I'm in minicom with E1 set. Here's what I
    got on the command line:

    (This below has an older chatscript with added commands and no CONNECT
    '' or \~ '')

    [root@localhost peers]# pppd call tmobile
    AATT

    OOKK

    AATTZZ

    OOKK

    AATT++CCMMEEEE==22

    OOKK

    AATTSS00==00

    OOKK

    AATT++CCGGAATTTT==11

    OOKK

    AATT++CCFFUUNN==11

    OOKK

    AATT++ccggddccoonntt==11,,""IIPP"",,""iinntteerrnneett33..vvooiicceessttrreeaamm..ccoomm"",,,,00
    ,,00

    OOKK

    AATTDD**9999##~~~~~~^?^?}}##@@!!}}!!}}++}}
    }}<<}}!!}}$$}}&&@@}}##}}$$@@##}}%%}}&&__88^^GG}}""}}
    &&}} }} }} }}

    NNOO CCAARRRRIIEERR <- This is before an actual connect from pppd!

    Failed <- a Ctrl-C by me caused this.
    Connect script failed
    Waiting for 1 child processes...
    script pppd (charshunt), pid 28764
    Script pppd (charshunt) finished (pid 28764), status = 0x0

    Without breaking out of this the pppd would at this point on the
    command line establish what it thinks is a connection and begin sending
    LCP requests until it times out with LCP: timeout sending
    Config-Requests.

    (output with original 'chat' command options in place same as above):

    Feb 3 18:26:54 localhost chat[28803]: send (ATD*99#^M)
    Feb 3 18:26:54 localhost chat[28803]: timeout set to 480 seconds
    Feb 3 18:26:54 localhost chat[28803]: expect (~)
    Feb 3 18:26:54 localhost chat[28803]: ^M
    Feb 3 18:27:24 localhost chat[28803]: ATD*99#^M~
    Feb 3 18:27:24 localhost chat[28803]: -- got it
    Feb 3 18:27:24 localhost chat[28803]: send (^M)
    Feb 3 18:27:24 localhost chat[28803]: expect (^M)
    Feb 3 18:27:54 localhost chat[28803]: ~~^?}#@!}!}?}
    }<}!}$}&@}#}$@#}%}&_8^G}"}&} } } }^M
    Feb 3 18:27:54 localhost chat[28803]: -- got it
    Feb 3 18:27:54 localhost pppd[28801]: Serial connection established.
    Feb 3 18:27:54 localhost pppd[28801]: Using interface ppp0
    Feb 3 18:27:54 localhost pppd[28801]: Connect: ppp0 <--> /dev/pts/2
    Feb 3 18:27:54 localhost /etc/hotplug/net.agent: assuming ppp0 is
    already up
    Feb 3 18:28:25 localhost pppd[28801]: LCP: timeout sending
    Config-Requests
    Feb 3 18:29:03 localhost pppd[28801]: Terminating on signal 2.

    Again no show of the NO CARRIER but it happens at the time of the
    garbled above.

    I compared this with an output from pppdump without the 'chat' options
    in place. Here's my output at a differnt time earlier:

    rcvd "\0d\0aOK\0d\0a" <- OK from ATDGCONT command prior
    sent "A"
    time 0.1s
    sent "T"
    sent "D"
    sent "*"
    sent "9"
    sent "9"
    time 0.1s
    sent "#"
    sent "\0d"
    time 19.8s
    rcvd "ATD*99#\0d"
    time 10.0s
    rcvd "~~~\ff}#\c0!}!}!} }<"
    sent "C" <- This is when I had CONNECT '' in the chatscript
    time 0.1s
    sent "O"
    sent "N"
    sent "N"
    sent "E"
    sent "C"
    time 0.1s
    sent "T"
    sent "\0d"
    time 9.8s
    rcvd "}!}$}&@}#}$\c0#}%}"
    time 10.0s
    rcvd "&_\b8\de\c7}\"}&} } } }"
    time 10.0s
    rcvd "\0d\0aNO CARRIER\0d\0a" <- Just after this pppd would
    time 1.1s ...establish a connection and send LCP Requests
    sent "~\ff}#\c0!}!}!} }4}\"}&} } } } }%}&\be6g}(}'}\"}(}\"\8fH~"
    time 3.0s
    sent "~\ff}#\c0!}!}!} }4}\"}&} } } } }%}&\be6g}(}'}\"}(}\"\8fH~"
    time 3.0s
    sent "~\ff}#\c0!}!}!} }4}\"}&} } } } }%}&\be6g}(}'}\"}(}\"\8fH~"
    time 3.0s
    sent "~\ff}#\c0!}!}!} }4}\"}&} } } } }%}&\be6g}(}'}\"}(}\"\8fH~"
    time 3.0s
    sent "~\ff}#\c0!}!}!} }4}\"}&} } } } }%}&\be6g}(}'}\"}(}\"\8fH~"
    time 3.0s
    sent "~\ff}#\c0!}!}!} }4}\"}&} } } } }%}&\be6g}(}'}\"}(}\"\8fH~"
    time 3.0s
    sent "~\ff}#\c0!}!}!} }4}\"}&} } } } }%}&\be6g}(}'}\"}(}\"\8fH~"
    time 0.9s
    rcvd "~\ff}#\c0!}!}!} }4}\"" <-this would scroll infinitely
    ....until you killed the session

    Again in the logs it would continue to try and send LCPs until timeout
    like above shown but never show the NO CARRIER.


    I'm thinking doesn't this tend to still lean more on my side perhaps
    with configuration or something? Or Could I not actually be talking to
    a PPP server? What other kind of servers are out there for this kind
    of connection? My reason for asking this is in the PPP HOWTO it poses
    questioning the ISP for what kind of server you're connecting to. But
    after issuing an AT*99# (AT*99***1)I get garbage/garble and doesn't
    that mean its always a PPP server responding?

    How do I go about enabling the kernel debug? Is it more than 'kdebug
    31'(??) in my 'options' file?

    At anyrate, I did call my T-Mobile provider and spoke with Tier3
    (highest level). The guy on shift said he didn't know much about the
    commands (GSM/GPRS) what I was asking about and just referred me to
    already posted stuff on the net. I plan to call back at a different
    time to speak with possibly someone a bit more knowledgable.


    I tried the exact same commands on a Windows XP laptop in
    Hyperterminal. Big difference -speed only...minicom is sooo slow!
    But ...exact same responses after pulling an IP. Once I pulled an IP
    on the DOS window prompt I was not able to see the IP established
    (typing 'ipconfig') but instead it showed my card as disconnected.
    Obviously... because I'm thinking there was no way for me to manually
    start up PPP. I did run 'Connection Wizard' at this point and only had
    to insert the connection number '*99#' and it connected immediately..go
    figure! (This was after I left in place manual configured AT+CGDCONT,
    etc. and just disconnected from hyperterminal). So this too leans me
    toward thinking there still is a problem somewhere with my pppd
    configuration.

    I orignially have the default pppd from the O/S install (Red Hat 7.3)
    and now I'm thinking of downloading the latest ppp-2.4.3 package but
    I'm not sure if the default pppd compiled in the kernel is sufficient.
    In fact how do you verify that it is in the kernel or not? I saw
    something in the PPP HOWTO doc but wasn't clear on this (not a kernel
    guru). If I upgrade I'm thinking I just may need the user-level
    compiled and not the kernel. Actually from pppd package do I need to
    use such scripts as 'ppp-on' 'ppp-on-dialer' at all? if so then I will
    start from scratch.

    One last command I noticed in both minicom and hyperterminal:

    ATO0

    this command changes the modem Aircard into 'data mode'. No commands
    can be issued afterwards like when dialing the AT*99# string. However
    its the only one that gives me the CONNECT:

    OK

    ATO0
    ATO 0

    CONNECT

    I can't imagine a way to get this inserted before the dial string and
    it complete the chatscript or afterwards and get the CONNECT I'm
    suppose to get. To get back to command mode you enter (+++).


    I also looked at 'wvdial'. But this won't recognize my pcmcia card as
    a modem I think. Here is an output while I'm connected in minicom:

    [root@localhost peers]# wvdial
    --> WvDial: Internet dialer version 1.41
    --> Cannot open /dev/modem: Device or resource busy

    And finally when I'm free and clear of /dev/modem (/dev/ttyS2):

    [root@localhost peers]# wvdial
    --> WvDial: Internet dialer version 1.41
    --> Initializing modem.
    --> Sending: ATZ
    --> Modem not responding.

    or

    [root@localhost peers]# wvdialconf now
    Scanning your serial ports for a modem.

    ttyS2<*1>: ATQ0 V1 E1 -- ATQ0 V1 E1 -- ATQ0 V1 E1 -- nothing.
    ttyS0<*1>: ATQ0 V1 E1 -- ATQ0 V1 E1 -- ATQ0 V1 E1 -- nothing.
    ttyS1<*1>: ATQ0 V1 E1 -- ATQ0 V1 E1 -- ATQ0 V1 E1 -- nothing.
    Port Scan<*1>: S3 S4 S5 S6 S7 S8 S9 S10
    Port Scan<*1>: S11 S12 S13 S14 S15 S16 S17 S18
    Port Scan<*1>: S19 S20 S21 S22 S23 S24 S25 S26
    Port Scan<*1>: S27 S28 S29 S30 S31 SA0 SA1 SA2
    .....
    .....


    It doesn't see it.


    I didn't think it would be this hard but I'm going to keep poking
    around 'till I find something.

    Thanks,

    juggernaut7


  11. Re: AirCard 750 on Linux Platform :NO CARRIER

    Your problems seem very strange, but have you tried the my chat scripts
    and pppd configuration:

    http://www.iki.fi/mikko.rapeli/linux_gprs.html

    To me your serial link layer has is problem, so perhaps adding this one to
    /etc/modules.conf would help:

    options ppp_async flag_time=0

    Usually with GPRS the PPP server is in the GSM phone, so there are all
    kinds of strange features, but if the device works as standard modem in
    Windows, then it will work with Linux too.

    -Mikko

  12. Re: AirCard 750 on Linux Platform :NO CARRIER

    Thanks Mikko for the link. This was one I saw and used a little bit
    before. After some time and effort I think I'm closer to what my
    problem is.


    One thing I neglected to mention in my posting was that this was a
    desktop I am developing on and I purchased a Desktop PCMCIA PCI slot
    card (P111)

    http://www.elandigitalsystems.com/adapter/p111.php

    where my Aircard 750 is inserted to. This I believe is where my
    problem is. Just the other day/night, I took my card and inserted it
    into my laptop where I installed Linux finally onto and it worked like
    a charm! The configuration setup as various postings has it works.
    The PCMCIA PCI slot card seems to slow down the reception and
    transmission just enough to cause or experience packet dropping of some
    sort resulting in a dead link on my end. I would send out but not
    recieve. Going thru 'minicom' (same thing as 'hyperterminal' on a
    windows machine) typing in AT commands were extremely slow in each
    response whereas on a laptop with Linux it was fast responding like
    hypertiminal. One MAJOR difference was typing in the dial string
    'ATD*99#' where it would instantly give me back a CONNECT on the linux
    laptop which is what I couldn't get on my linux desktop and I had to
    remove or change up the CONNECT '' in the chatscript.

    I'm still researching this issue and I think there is some resource
    conflict going on like irq or possibly port addresses. I've looked a
    bit into the PCMCIA HOWTO:

    http://pcmcia-cs.sourceforge.net/ftp...A-HOWTO-2.html

    Under section 2.7 -Interrupt Delivery Problems; Serial and modem cards
    may respond very sluggishly.

    however I've decided to look into purchasing a Single Board Computer
    (motherboard) with an embedded PCMCIA slot already on it. So far I've
    seen the Intel based LV-671

    http://www.bwi.com/scripts/site/site...tion/configure

    ...which appears quite appealing but I have to ensure whatever I get
    that Linux can run on it.

    Any Ideas?? as to what hardware would be good?

    Clifford Kite, Thanks for your help man..really appreciate it the
    details on the HDLC stuff this led me to get this book by James.

    James Carlson, Thank you too for help..and your book is cool man!
    (Just bought it!..PPP Design Implementation and Debugging)


    juggernaut7_@hotmail.com




    Mikko Rapeli wrote:
    > Your problems seem very strange, but have you tried the my chat

    scripts
    > and pppd configuration:
    >
    > http://www.iki.fi/mikko.rapeli/linux_gprs.html
    >
    > To me your serial link layer has is problem, so perhaps adding this

    one to
    > /etc/modules.conf would help:
    >
    > options ppp_async flag_time=0
    >
    > Usually with GPRS the PPP server is in the GSM phone, so there are

    all
    > kinds of strange features, but if the device works as standard modem

    in
    > Windows, then it will work with Linux too.
    >
    > -Mikko



  13. Re: AirCard 750 on Linux Platform :NO CARRIER

    "juggernaut7_@hotmail.com" writes:
    > where my Aircard 750 is inserted to. This I believe is where my
    > problem is. Just the other day/night, I took my card and inserted it
    > into my laptop where I installed Linux finally onto and it worked like
    > a charm! The configuration setup as various postings has it works.
    > The PCMCIA PCI slot card seems to slow down the reception and
    > transmission just enough to cause or experience packet dropping of some
    > sort resulting in a dead link on my end. I would send out but not


    I'm no expert in this area, but that sounds remarkably like the
    descriptions I've heard of IRQ conflicts.

    > Under section 2.7 -Interrupt Delivery Problems; Serial and modem cards
    > may respond very sluggishly.


    Yep.

    > Any Ideas?? as to what hardware would be good?


    Sorry ... not at all a Linux expert. I'd recommend posting to a
    newsgroup related to Linux instead.

    > James Carlson, Thank you too for help..and your book is cool man!
    > (Just bought it!..PPP Design Implementation and Debugging)


    Thanks!

    --
    James Carlson, IP Systems Group
    Sun Microsystems / 1 Network Drive 71.234W Vox +1 781 442 2084
    MS UBUR02-212 / Burlington MA 01803-2757 42.497N Fax +1 781 442 1677

+ Reply to Thread