Aironet AP4800 Firmware

Discussion in 'Cisco' started by DJ Allen, Mar 1, 2005.

  1. DJ Allen

    DJ Allen Guest

    I know this is a shot in the dark, but would anyone happen to have the 3.xx
    and 4.xx firmwares for pre-Cisco Aironet AP4800 AP? Thanx for any help.
    DJ Allen, Mar 1, 2005
    #1
    1. Advertising

  2. DJ Allen

    DJ Allen Guest

    My apologies...plz respond to texas<at>vfemail.net

    "DJ Allen gmail.com>" <texas<at> wrote in message
    news:...
    >I know this is a shot in the dark, but would anyone happen to have the 3.xx
    >and 4.xx firmwares for pre-Cisco Aironet AP4800 AP? Thanx for any help.
    >
    DJ Allen, Mar 1, 2005
    #2
    1. Advertising

  3. DJ Allen

    Dan Lanciani Guest

    In article <>, at (DJ Allen) writes:
    | My apologies...plz respond to texas<at>vfemail.net

    Doesn't work...

    | "DJ Allen gmail.com>" <texas<at> wrote in message
    | news:...
    | >I know this is a shot in the dark, but would anyone happen to have the 3.xx
    | >and 4.xx firmwares for pre-Cisco Aironet AP4800 AP? Thanx for any help.

    Did 4.x even exist for the AP4800? I bought one when they first came out
    and it used with 6.x (which I still have). The 3000 series had 4.x, though.

    Dan Lanciani
    ddl@danlan.*com
    Dan Lanciani, Mar 1, 2005
    #3
  4. DJ Allen

    DJ Allen Guest

    I'm going to have to claim ignorance...a friend has a couple that he wants
    to upgrade to latest Cisco 340 8.xx but says he needs 3.xx and 4.xx to get
    there (?) I should have done a little more research, thanks for the info.
    Can you send me the 6.x you have? If so, plz send to texas<at>vfemail.net.
    Rgds, Dennis


    "Dan Lanciani" <ddl@danlan.*com> wrote in message
    news:...
    > In article <>, at (DJ Allen) writes:
    > | My apologies...plz respond to texas<at>vfemail.net
    >
    > Doesn't work...
    >
    > | "DJ Allen gmail.com>" <texas<at> wrote in message
    > | news:...
    > | >I know this is a shot in the dark, but would anyone happen to have the
    > 3.xx
    > | >and 4.xx firmwares for pre-Cisco Aironet AP4800 AP? Thanx for any help.
    >
    > Did 4.x even exist for the AP4800? I bought one when they first came out
    > and it used with 6.x (which I still have). The 3000 series had 4.x,
    > though.
    >
    > Dan Lanciani
    > ddl@danlan.*com
    DJ Allen, Mar 1, 2005
    #4
  5. DJ Allen

    Uli Link Guest

    DJ Allen <texas schrieb:

    > I'm going to have to claim ignorance...a friend has a couple that he wants
    > to upgrade to latest Cisco 340 8.xx but says he needs 3.xx and 4.xx to get
    > there (?) I should have done a little more research, thanks for the info.


    When Cisco aquired Aironet the 4800B was brand new.
    The Cisco 340 is nearly equal the 4800B. A 50MHz PowerPC860 CPU with
    16MB SDRAM and 2 MByte flash. They run VxWorks.

    The older 4800 (same chassis than 3500 and 4500) has some weird
    restriction, it used a pre standard 11MBit mode. The radio was only able
    to WEP encrypt at 1 or 2 MBit rates. It lacks the ability of a short
    preamble. The case is large (same case than a 340 Bridge!)
    The CPU is a Motorola 68020 at 25Mhz. The AP boots in few seconds. The
    browser GUI is blue, red and white.

    The latest fw is 8.80 (which support an LMC352) as radio and I don't
    know any reason why to use a version older 8.24_2

    --
    Uli

    These opinions are mine. All found typos are yours.
    Uli Link, Mar 1, 2005
    #5
  6. DJ Allen

    Dan Lanciani Guest

    In article <>, at (DJ Allen) writes:
    | I'm going to have to claim ignorance...a friend has a couple that he wants
    | to upgrade to latest Cisco 340 8.xx

    This doesn't make a lot of sense unless he is talking about the 340 *bridge*
    series which in fact is the same hardware platform as the AP4800E. The 340
    access point family didn't have version numbers in those ranges that I recall.

    | but says he needs 3.xx and 4.xx to get
    | there (?)

    Regardless of the platform I can't think of a reason for that. Is he
    trying to convert an access point to a bridge? There was a story about
    the 630 series that claimed you could do such by running through a set
    of about six old firmware versions. It's possible that it worked (I never
    checked because I found much easier ways to convert an access point to a
    bridge) since the firmware does fudge the device type in some cases if
    it doesn't have a matching table entry. I'm pretty sure that it will not
    work for the AP4800E.

    | I should have done a little more research, thanks for the info.
    | Can you send me the 6.x you have? If so, plz send to texas<at>vfemail.net.

    That address does not work.

    Dan Lanciani
    ddl@danlan.*com
    Dan Lanciani, Mar 2, 2005
    #6
  7. DJ Allen

    Dan Lanciani Guest

    In article <4224bda2$0$26554$-online.net>, (Uli Link) writes:
    | DJ Allen <texas schrieb:
    |
    | > I'm going to have to claim ignorance...a friend has a couple that he wants
    | > to upgrade to latest Cisco 340 8.xx but says he needs 3.xx and 4.xx to get
    | > there (?) I should have done a little more research, thanks for the info.
    |
    | When Cisco aquired Aironet the 4800B was brand new.
    | The Cisco 340 is nearly equal the 4800B. A 50MHz PowerPC860 CPU with
    | 16MB SDRAM and 2 MByte flash. They run VxWorks.
    |
    | The older 4800 (same chassis than 3500 and 4500) has some weird
    | restriction, it used a pre standard 11MBit mode.

    That wasn't a problem--even the oldest 4800 series radio cards supported
    standard CCK modulation for 11 & 5.5 Mb/s rates with the right firmware.
    (The non-standard mode was MOK.)

    | The radio was only able
    | to WEP encrypt at 1 or 2 MBit rates.

    That was a problem, but it was solely the fault of the radio card. Sadly,
    Aironet seemed to have some sort of marketing policy to keep this a secret.
    When you called to ask why your nearly-new AP4800E couldn't do WEP at 11 Mb/s
    they told you that it was a hardware problem with the chassis and that you
    needed to "upgrade" (where "upgrade" meant to toss the AP4800E and buy new)
    to an AP4800B series with its low-power 35mW radio and all sorts of silly
    license key management. All you really needed to do was replace the radio
    card with an LM4800A (a 100mW card with an A503 cpu just like the B series)...
    if you could find one. In fact, Aironet (and then Cisco) shipped AP4800Es
    (and BR342s which are the identical hardware) with LM4800As for years to
    support customers who demanded 100mW solutions. Cisco even shipped LM4800As
    by themselves after the 350 series was introduced; I'm not sure why.

    | It lacks the ability of a short
    | preamble.

    That is unfortunately true of even the LM4800A. Of course, there are some
    advantages to long preambles (diversity selection).

    | The case is large (same case than a 340 Bridge!)

    Same case, same hardware, same everything. :) (Ok, the bridge doesn't
    being out both antenna connections.)

    | The CPU is a Motorola 68020 at 25Mhz. The AP boots in few seconds. The
    | browser GUI is blue, red and white.
    |
    | The latest fw is 8.80 (which support an LMC352) as radio and I don't

    8.81 seems to be floating around.

    | know any reason why to use a version older 8.24_2

    Another good version to hang onto is 8.65_2. This was the last version
    that included card firmware that was compatible with the LM4500 and the
    original (non-A) LM4800 (anything newer breaks WEP support on the card).
    But unlike 8.24, the card firmware in 8.65 does not require the masterky
    upgrade bits to be set to support 128-bit WEP.

    Dan Lanciani
    ddl@danlan.*com
    Dan Lanciani, Mar 2, 2005
    #7
  8. DJ Allen

    DJ Allen Guest

    >>That address does not work.>>>>

    Hi Dan,

    It's my main email account (been working al day!)... replace "<at>" with @
    .... texas@vfemail.*net as you have yours listed for spam evasion. I've done
    a bit more research and found that to get an Aironet AP4800 to load Cisco
    340 8.xx fw, I must upgrade incrementally from 3.xx...been told this is true
    for any fw below 7.xx. Question: will flashing up to 8.xx fix the bandwidth
    limitation when using WEP encryption? Thanks for the help!

    Dennis Allen




    "Dan Lanciani" <ddl@danlan.*com> wrote in message
    news:...
    > In article <>, at (DJ Allen) writes:
    > | I'm going to have to claim ignorance...a friend has a couple that he
    > wants
    > | to upgrade to latest Cisco 340 8.xx
    >
    > This doesn't make a lot of sense unless he is talking about the 340
    > *bridge*
    > series which in fact is the same hardware platform as the AP4800E. The
    > 340
    > access point family didn't have version numbers in those ranges that I
    > recall.
    >
    > | but says he needs 3.xx and 4.xx to get
    > | there (?)
    >
    > Regardless of the platform I can't think of a reason for that. Is he
    > trying to convert an access point to a bridge? There was a story about
    > the 630 series that claimed you could do such by running through a set
    > of about six old firmware versions. It's possible that it worked (I never
    > checked because I found much easier ways to convert an access point to a
    > bridge) since the firmware does fudge the device type in some cases if
    > it doesn't have a matching table entry. I'm pretty sure that it will not
    > work for the AP4800E.
    >
    > | I should have done a little more research, thanks for the info.
    > | Can you send me the 6.x you have? If so, plz send to
    > texas<at>vfemail.net.
    >
    > That address does not work.
    >
    > Dan Lanciani
    > ddl@danlan.*com
    DJ Allen, Mar 2, 2005
    #8
  9. DJ Allen

    Dan Lanciani Guest

    In article <>, at (DJ Allen) writes:
    | >>That address does not work.>>>>
    |
    | Hi Dan,
    |
    | It's my main email account (been working al day!)... replace "<at>" with @
    | ... texas@vfemail.*net as you have yours listed for spam evasion.

    I understand the format; the address didn't work. More specifically the
    mail server for vfemail.net said that there was no user texas. Send me
    some email directly.

    | I've done
    | a bit more research and found that to get an Aironet AP4800 to load Cisco
    | 340 8.xx fw, I must upgrade incrementally from 3.xx...

    This still doesn't make sense on several levels. 8.x is a recent version
    for the 340 bridge series and the original 4800 access point, but not for
    the 340 access point series. There have been a few obscure cases that
    (sort of) require incremental upgrade, but nothing like what you are
    saying. What exactly do you have and what firmware is it running?

    | been told this is true
    | for any fw below 7.xx. Question: will flashing up to 8.xx fix the bandwidth
    | limitation when using WEP encryption?

    No. If you have an original (non A series) LM4800 radio card you must replace
    it with a 4800A or better to run WEP at > 2 Mb/s.

    Dan Lanciani
    ddl@danlan.*com
    Dan Lanciani, Mar 2, 2005
    #9
  10. DJ Allen

    Uli Link Guest

    Dan Lanciani schrieb:

    > When you called to ask why your nearly-new AP4800E couldn't do WEP at 11 Mb/s
    > they told you that it was a hardware problem with the chassis and that you
    > needed to "upgrade" (where "upgrade" meant to toss the AP4800E and buy new)
    > to an AP4800B series with its low-power 35mW radio and all sorts of silly
    > license key management. All you really needed to do was replace the radio
    > card with an LM4800A (a 100mW card with an A503 cpu just like the B series)...
    > if you could find one. In fact, Aironet (and then Cisco) shipped AP4800Es
    > (and BR342s which are the identical hardware) with LM4800As for years to
    > support customers who demanded 100mW solutions. Cisco even shipped LM4800As
    > by themselves after the 350 series was introduced; I'm not sure why.


    Drop in a LMC352, in Europe limited to 50mW, but much better sensitivity
    than the LMC4800 or LMC34x.
    I only kept one LMC4800 for the collection's completeness.

    > | It lacks the ability of a short
    > | preamble.
    > That is unfortunately true of even the LM4800A. Of course, there are some
    > advantages to long preambles (diversity selection).


    On my 342 AP diversity works (with both a LMC342 or a LMC352 radio)
    together with short preamble.

    I once got a bunch of 3500E AP from a migration project, and "converted"
    them with LMC352 radios and FW 8.80 :)

    --
    Uli

    These opinions are mine. All found typos are yours.
    Uli Link, Mar 2, 2005
    #10
  11. DJ Allen

    Dan Lanciani Guest

    In article <422599e8$0$26545$-online.net>, (Uli Link) writes:
    | Dan Lanciani schrieb:
    |
    | > When you called to ask why your nearly-new AP4800E couldn't do WEP at 11 Mb/s
    | > they told you that it was a hardware problem with the chassis and that you
    | > needed to "upgrade" (where "upgrade" meant to toss the AP4800E and buy new)
    | > to an AP4800B series with its low-power 35mW radio and all sorts of silly
    | > license key management. All you really needed to do was replace the radio
    | > card with an LM4800A (a 100mW card with an A503 cpu just like the B series)...
    | > if you could find one. In fact, Aironet (and then Cisco) shipped AP4800Es
    | > (and BR342s which are the identical hardware) with LM4800As for years to
    | > support customers who demanded 100mW solutions. Cisco even shipped LM4800As
    | > by themselves after the 350 series was introduced; I'm not sure why.
    |
    | Drop in a LMC352,

    At the time the 350 series did not exist, so that was not an option. :(

    | in Europe limited to 50mW,

    Well, that's easy to fix. :)

    | but much better sensitivity
    | than the LMC4800 or LMC34x.

    I'm surprised it's much better than the 340 series (or for that matter
    significantly better than the 4800 series). What sensitivity figures
    have you measured? In my highly unscientific range tests I found that
    unit-to-unit differences in performance seemed to dwarf any differences
    between the 4800 and 350. (I didn't test much with the 340 series because
    of the lower power.) I also found that a 4500 series radio running at
    1 Mb/s often gave better range than a 4800 or 350 series also running at
    1 Mb/s.

    | > | It lacks the ability of a short
    | > | preamble.
    | > That is unfortunately true of even the LM4800A. Of course, there are some
    | > advantages to long preambles (diversity selection).
    |
    | On my 342 AP diversity works (with both a LMC342 or a LMC352 radio)
    | together with short preamble.

    Can you tell me exactly how this works? I've gotten several different and
    confusing stories about per-MAC (rather than per-packet) diversity, but
    none of them make much sense. Also, can you explain the transmit diversity
    algorithm for beacons? I've observed some confusing behavior where the
    access point will at first alternate ports but after some days of non-use
    switch to only one.

    Dan Lanciani
    ddl@danlan.*com
    Dan Lanciani, Mar 2, 2005
    #11
  12. DJ Allen

    Uli Link Guest

    Hi,

    > | in Europe limited to 50mW,
    >
    > Well, that's easy to fix. :)


    How with keeping the channels 12 and 13 available?
    It isn't nice when an FCC only card roams to a cell on channel 13 :-(

    I have two ETSI AIR-342. Drop in a FCC LMC34x or LMC3x.
    After one reboot I have a ETSI LMC3x2 :)
    Because when my customer orders 20 ETSI LMC's he always got 2 or 3 FCC
    cards among them. Too late for return. :-(

    I expect the same trick does vice versa with an FCC AP.
    The Rev4 PCM352 now have 100mW available. The LMC only 50mW.
    We use shorter pigtails or better antennas instead ;-)

    > | but much better sensitivity
    > | than the LMC4800 or LMC34x.
    >
    > I'm surprised it's much better than the 340 series (or for that matter
    > significantly better than the 4800 series). What sensitivity figures
    > have you measured? In my highly unscientific range tests I found that
    > unit-to-unit differences in performance seemed to dwarf any differences
    > between the 4800 and 350.


    I haven't tested much with 4800 radios. Really rare here in Germany.
    The 4500 were quite common, although I still have at least 30 LMC3500
    around from migrations and only a few 4500 radios.

    > (I didn't test much with the 340 series because
    > of the lower power.) I also found that a 4500 series radio running at
    > 1 Mb/s often gave better range than a 4800 or 350 series also running at
    > 1 Mb/s.


    The throughput of the 340 and 350 is equal, the 4800 is still slightly
    faster than a Hermes based ORiNOCO card, but the difference to the
    340/350 is what the short preamble makes. The ETSI 4800 cards were
    limited to 50mW in later fw versions.

    The PCM350 compared the PCM340 has a much better coverage, but the
    PCM350 has inferior throughput in IBM ThinkPad T2x series. No problem
    with other cards. Now I use a CB21AG in my ThinkPad T23.

    When I compare LMC340 (a radio from an AP) with a LMC352 with the old
    snap-on antenna, the coverage of the 352 is better, but the difference
    isn't as large as for PCM cards.
    When reducing the tx power to 30 mW the 350 cards are still ahead, more
    than the 2dB difference.
    Over the thumb I would say about 3 to 4 dB better sensitivity. You still
    have a 11MBit connection, where the site survey with a 340 shows deep
    red with the first dropouts.

    I don't have long range expierence, mainly factory plants, production
    area with fully redundant at least 5.5MBit coverage specified. A strong
    focus on seamless fast roaming for handheld or vehicle mounted computers.
    Therefore we often reduce the tx power to save battery life time.
    Most cards in terminal emulators are configured for CAM, when I use
    FAST-PSP every echo from the host is delayed, to slow for scanned
    barcodes sometimes. Not an issue with the Lucent cards.


    > Can you tell me exactly how this works?


    No.

    > I've gotten several different and
    > confusing stories about per-MAC (rather than per-packet) diversity, but
    > none of them make much sense.


    What should a per-MAC-what should be???

    > Also, can you explain the transmit diversity
    > algorithm for beacons?


    The preamble is transmitted at 1MBit or 2MBit for using short preamble.
    The selector decides on the better antenna from the received preamble
    for the following frame and the eventually sent frame.
    This is for data.
    For the beacons, I haven't found anything.
    I believe they use the primary antenna for sending or simply alternating
    every beacon. Wild guess only.

    > I've observed some confusing behavior where the
    > access point will at first alternate ports but after some days of non-use
    > switch to only one.


    All I tried for verifying diversity is detaching/reattaching one of the
    two antennas, and then same play with the other with constant traffic
    and watching dropped packets. My observations gave the expected results,
    no need for further investigation.
    Most APs in the field have only one 7dBi omnidirectional antenna
    connected. In my home lab, I never noticed any problems. But 2 AIR-1230
    plus one or two AIR-342 for temporary customer test environments are
    fully redundant at every corner.


    --
    Uli

    These opinions are mine. All found typos are yours.
    Uli Link, Mar 3, 2005
    #12
  13. DJ Allen

    Dan Lanciani Guest

    In article <4226591f$0$26539$-online.net>, (Uli Link) writes:

    | > | in Europe limited to 50mW,
    | >
    | > Well, that's easy to fix. :)
    |
    | How with keeping the channels 12 and 13 available?

    As far as the card is concerned the power limit and the channel setting
    (country code) are independent. (Well, that's not completely true since
    setting the country to "Test" also removes any limit from the power.
    Maybe they've added more special cases for ETSI, so you might actually
    have to use "Test.") On an AP4800E or similar you can set either/both
    from the eeprom option of the zz menu.

    | I have two ETSI AIR-342. Drop in a FCC LMC34x or LMC3x.
    | After one reboot I have a ETSI LMC3x2 :)

    One reboot of what? An AP4800E doesn't reset the card's eeprom unless
    you explicitly request it to. I don't think any PC drivers touch the
    eeprom at all. You might try the little eeprom dumper I posted to the
    Aironet mailing list to see exactly what is going on. The country code
    is tag 0x1001 and the power level is tag 0x1002.

    | > | but much better sensitivity
    | > | than the LMC4800 or LMC34x.
    | >
    | > I'm surprised it's much better than the 340 series (or for that matter
    | > significantly better than the 4800 series). What sensitivity figures
    | > have you measured? In my highly unscientific range tests I found that
    | > unit-to-unit differences in performance seemed to dwarf any differences
    | > between the 4800 and 350.
    |
    | I haven't tested much with 4800 radios. Really rare here in Germany.

    Maybe you had some bad samples then.

    | > (I didn't test much with the 340 series because
    | > of the lower power.) I also found that a 4500 series radio running at
    | > 1 Mb/s often gave better range than a 4800 or 350 series also running at
    | > 1 Mb/s.
    |
    | The throughput of the 340 and 350 is equal, the 4800 is still slightly
    | faster than a Hermes based ORiNOCO card, but the difference to the
    | 340/350 is what the short preamble makes.

    I think you are sort of mixing apples and oranges with sensitivity and
    throughput. When I tested I forced the cards that were capable of short
    preambles to use long preambles. (On an AP4800E this option is available
    in the radio configuration menus if you have enabled debug menus from the
    zz menu.)

    | > I've gotten several different and
    | > confusing stories about per-MAC (rather than per-packet) diversity, but
    | > none of them make much sense.
    |
    | What should a per-MAC-what should be???

    One person from Cisco claimed that the access point kept diversity information
    on a long-term per-destination (per-MAC) basis. I can see where this would be
    useful for subsequent transmissions when you haven't recently received a
    packet from that MAC, but it seems you could get into trouble depending on
    how you update it.

    | > Also, can you explain the transmit diversity
    | > algorithm for beacons?
    |
    | The preamble is transmitted at 1MBit or 2MBit for using short preamble.
    | The selector decides on the better antenna from the received preamble
    | for the following frame and the eventually sent frame.

    I was told that the short preamble does not allow enough time to measure
    the signal level at both ports, select, and still synchronize reliably.
    If this is true it implies that the decision has to be made over several
    packets (which is a problem since you don't know that the next packet is
    coming from the same source).

    | This is for data.
    | For the beacons, I haven't found anything.
    | I believe they use the primary antenna for sending or simply alternating
    | every beacon. Wild guess only.

    They certainly start out alternating.

    | > I've observed some confusing behavior where the
    | > access point will at first alternate ports but after some days of non-use
    | > switch to only one.
    |
    | All I tried for verifying diversity is detaching/reattaching one of the
    | two antennas, and then same play with the other with constant traffic
    | and watching dropped packets. My observations gave the expected results,
    | no need for further investigation.

    I connected an amplifier with a visible transmit indicator so I could see
    when a port was actually sending. The behavior changed over time. I
    really don't understand what was going on. I also tried some tests with
    diversity off and a single antenna connected to the correct or incorrect
    port. Often there was little difference, so there must be a lot of leakage
    between ports. Either that or diversity wasn't really off.

    Have you noticed that the single antenna on an ISA carrier is connected
    to the opposite port of that on a PCI carrier? Does the driver know to
    enable the correct port based on card type?

    Dan Lanciani
    ddl@danlan.*com
    Dan Lanciani, Mar 3, 2005
    #13
  14. DJ Allen

    Uli Link Guest

    Dan Lanciani schrieb:


    > | I have two ETSI AIR-342. Drop in a FCC LMC34x or LMC3x.
    > | After one reboot I have a ETSI LMC3x2 :)
    >
    > One reboot of what?


    Cisco AIR-342E2R, running 12.xx fw.

    > An AP4800E doesn't reset the card's eeprom unless
    > you explicitly request it to.


    The MC68020 based APs take the regulatory domain from the radio.
    Flashing the fw by the AP doesn't alter the cards PDA in eeprom.
    But a pre-set wep key survives until you touch it through the AP admin
    interface.

    > | The throughput of the 340 and 350 is equal, the 4800 is still slightly
    > | faster than a Hermes based ORiNOCO card, but the difference to the
    > | 340/350 is what the short preamble makes.
    >
    > I think you are sort of mixing apples and oranges with sensitivity and
    > throughput.


    Not really, if s/n falls below a given level, the data rate is reduced.
    So with better sensitivity you'll have 11MBit where another card falls
    back to 5.5 or 2 MBit. The throughput doesn't differ much between 5.5
    and 11MBit, but noticeably for 2. When s/n is good or better, there is
    no difference in throughput.

    When using the link test from the AP to a station, I noticed, that a 352
    radio usually gives better values for quality. Whatever they define for
    quality. Or the values from /proc/driver/aironet/eth0/Status on Linux.

    > | What should a per-MAC-what should be???
    >
    > One person from Cisco claimed that the access point kept diversity information
    > on a long-term per-destination (per-MAC) basis. I can see where this would be
    > useful for subsequent transmissions when you haven't recently received a
    > packet from that MAC, but it seems you could get into trouble depending on
    > how you update it.


    Even the wildest guess: Something like mobility/non-mobility for WGBs?
    Don't know.

    > I was told that the short preamble does not allow enough time to measure
    > the signal level at both ports, select, and still synchronize reliably.


    AFAIK with 4800b and 350 radios this isn't a problem.
    The Lucent cards did diversity only rx, tx primary only. And they aren't
    able doing short preamble at all.
    No matter for Atheros cards.
    The Symbol Spectrum24HR radios are good hw, but the drivers are a whole
    mess, so a try to avoid them whenever possible.
    Usually my job is to make the operating system work and build the images
    for the client devices. There aren't many out there still remembering
    how easy MS-DOS was ;-)


    > | This is for data.
    > | For the beacons, I haven't found anything.
    > | I believe they use the primary antenna for sending or simply alternating
    > | every beacon. Wild guess only.
    >
    > They certainly start out alternating.
    >


    I called out for a few people, waiting for responses.
    First response covering Atheros based APs: Alternating if both antennas
    are in use by connections, else the select antenna only, for tx the beacons.
    For Hermes based APs. no tx diversity. The simple solution.
    Still waiting for some opinions and internals.

    > I connected an amplifier with a visible transmit indicator so I could see
    > when a port was actually sending. The behavior changed over time. I
    > really don't understand what was going on. I also tried some tests with
    > diversity off and a single antenna connected to the correct or incorrect
    > port. Often there was little difference, so there must be a lot of leakage
    > between ports. Either that or diversity wasn't really off.


    There is a lot of leakage. Or the diversity configuration is simply ignored.

    In a vehicle mount computer running RedHat 7.3, the manufacturer started
    to use the PCM352. There were enormous problems, because the integrated
    antenna was halfways covered by the computers die cast case. Weird
    signal quality due to the reflections from the case. When we switched to
    the LMC (which cost a lot of money more together with an external
    antenna), I switched of diversity. But it worked quite well with the
    external antenna connected to the wrong port of the LMC.
    I didn't noticed a significant difference between selecting the correct
    port or diversity.

    The whole diversity seems a little bit esoteric. If you believe in
    diversity it will help, else it will work too. I use it, because from my
    personal expierience it never hurts.

    As I wrote I never used 4800/4800A radios in a real world project.
    My only 4800 card is a PC4800 A9.08, s/n JW4203****. Can't remember
    where I pulled this card from the dumpster.

    The Lucent point of view of avoiding diversity on both sides works very
    good and simple.


    --
    Uli

    These opinions are mine. All found typos are yours.
    Uli Link, Mar 3, 2005
    #14
  15. DJ Allen

    Dan Lanciani Guest

    In article <4226e7cf$0$26549$-online.net>, (Uli Link) writes:
    | Dan Lanciani schrieb:
    |
    |
    | > | I have two ETSI AIR-342. Drop in a FCC LMC34x or LMC3x.
    | > | After one reboot I have a ETSI LMC3x2 :)
    | >
    | > One reboot of what?
    |
    | Cisco AIR-342E2R, running 12.xx fw.

    There was a reason I put "upgrade" in quotes...

    | > An AP4800E doesn't reset the card's eeprom unless
    | > you explicitly request it to.
    |
    | The MC68020 based APs take the regulatory domain from the radio.
    | Flashing the fw by the AP doesn't alter the cards PDA in eeprom.

    Correct, though it will ultimately load new firmware on the card if
    the version included with the AP firmware is different.

    | But a pre-set wep key survives until you touch it through the AP admin
    | interface.

    Correct. Everything in the eeprom (and the flash parameter block) survives.

    | > I think you are sort of mixing apples and oranges with sensitivity and
    | > throughput.
    |
    | Not really, if s/n falls below a given level, the data rate is reduced.

    But my point was that if you are trying to infer sensitivity from throughput
    it is inappropriate to configure the preamble length differently on the
    subject cards. Look at it this way. If you took two pairs 350 series cards,
    one pair configured for short preamble and one for long, then under otherwise
    identical conditions the throughputs would be different. This would not mean
    that the sensitivities were different.

    | > I was told that the short preamble does not allow enough time to measure
    | > the signal level at both ports, select, and still synchronize reliably.
    |
    | AFAIK with 4800b and 350 radios this isn't a problem.

    Maybe someone from Cisco can chime in with a definitive answer on this.

    | When we switched to
    | the LMC (which cost a lot of money more together with an external
    | antenna), I switched of diversity. But it worked quite well with the
    | external antenna connected to the wrong port of the LMC.
    | I didn't noticed a significant difference between selecting the correct
    | port or diversity.

    Good, this effect isn't just my imagination. :)

    | As I wrote I never used 4800/4800A radios in a real world project.
    | My only 4800 card is a PC4800 A9.08, s/n JW4203****. Can't remember
    | where I pulled this card from the dumpster.

    That's pretty old (presumably not even an A version) though I still
    have the first one I bought (new) which is JW12...!

    Dan Lanciani
    ddl@danlan.*com
    Dan Lanciani, Mar 4, 2005
    #15
  16. DJ Allen

    Uli Link Guest

    Dan Lanciani schrieb:

    > | > I think you are sort of mixing apples and oranges with sensitivity and
    > | > throughput.
    > |
    > | Not really, if s/n falls below a given level, the data rate is reduced.
    >
    > But my point was that if you are trying to infer sensitivity from throughput
    > it is inappropriate to configure the preamble length differently on the
    > subject cards.


    ACK.
    The effect of long vs. short preamble must be cleared when comparing
    throughput results. On the other side if the throughput difference is a
    concern, you have to use short preamble capable equipment.

    A great plus of the Cisco APs is, you can set them to short, and they
    fallback to long automagically. And the best: only for the unicasts to
    the "long" stations. Capable station are still served with short preamble.

    For most other vendors better select "long" or a ORiNOCO card won't
    connect.

    > | As I wrote I never used 4800/4800A radios in a real world project.
    > | My only 4800 card is a PC4800 A9.08, s/n JW4203****. Can't remember
    > | where I pulled this card from the dumpster.
    >
    > That's pretty old (presumably not even an A version) though I still
    > have the first one I bought (new) which is JW12...!


    Contest: who has the oldest 802.11 card

    What about my LM3500 03-A0 JL21001303 June 98? Mint condition.
    There was a time when the LM3500 and PC3500 were different hardware.
    If you try to flash fw. 3.68 into the card, the PC3500 stops working
    except it's s/n is JL4* or greater. The LM3500 works all.
    The PSP modes worked much better and they were capable of WEP! Starting
    with fw 2.xx WEP was available as buyable option :-(
    The only frequenzy hopper with *working* WEP and hidden SSID I know.
    Even with MS-DOS paket driver! Works reliable with Linux, *much* better
    than the official Symbol Spectrum24 driver. But all installations were
    made long before WEP appeared and those days changing the default SSID
    from "zephyr" was considered a great *security* feature. The only
    security was, that those days a 802.11 radio was really! expensive and
    hard to find.

    --
    Uli

    These opinions are mine. All found typos are yours.
    Uli Link, Mar 4, 2005
    #16
  17. DJ Allen

    Dan Lanciani Guest

    In article <42282bef$0$24925$-online.net>, (Uli Link) writes:

    | A great plus of the Cisco APs is, you can set them to short, and they
    | fallback to long automagically. And the best: only for the unicasts to
    | the "long" stations.

    Yes, this kind of scared me the first time I saw it. Do you know of a way
    to display which nodes have been determined to need a long preamble? And
    what happens with broadcasts? They seem to get through as well.

    | Contest: who has the oldest 802.11 card

    I have some 3000 series cards. I forget, did they ever make 802.11
    firmware for them? I remember talk about it, but I sort of lost
    interest before anything came of it.

    I bought several Symbol FH access points and cards, but the range was
    terrible even though they were the 500mW versions. Maybe it's time to
    look at FH again for security through obscurity. Probably nobody's
    snooping on it. :)

    Dan Lanciani
    ddl@danlan.*com
    Dan Lanciani, Mar 5, 2005
    #17
  18. DJ Allen

    Uli Link Guest

    Dan Lanciani schrieb:

    > | A great plus of the Cisco APs is, you can set them to short, and they
    > | fallback to long automagically. And the best: only for the unicasts to
    > | the "long" stations.
    >
    > Yes, this kind of scared me the first time I saw it. Do you know of a way
    > to display which nodes have been determined to need a long preamble? And
    > what happens with broadcasts? They seem to get through as well.


    With IOS APs "show dot11 association [MAC address]"
    There is a field "capabilities" with "ShortHdr" or not.
    The VxWorks APs show equal info on the stations statistics. Never used
    the cmd line with the VxWorks Aironets.

    > I have some 3000 series cards. I forget, did they ever make 802.11
    > firmware for them? I remember talk about it, but I sort of lost
    > interest before anything came of it.


    I once had a single 3100 card, single antenna port. 1MBit only.
    Worked w/o problems in a large 3500 series infrastructure.
    Most of the APs in this infrastructure were labeled with 3000 series and
    not 3500 series, but had LM3500 radios installed.

    > I bought several Symbol FH access points and cards, but the range was
    > terrible even though they were the 500mW versions. Maybe it's time to
    > look at FH again for security through obscurity. Probably nobody's
    > snooping on it. :)


    The exorbitant power consumption disqualified this card for battery
    powered devices.
    All Symbol FH installations were the most insecure I ever saw.
    When asked for security the Symbol guys always talked of Kerberos, but I
    never saw any real world installation with working Kerberos outside a
    lab. They told teir customers it is completly secure.
    One reason I still carry an Aironet LM3500 card with my XP Notebook, to
    prove them wrong :)

    --
    Uli

    These opinions are mine. All found typos are yours.
    Uli Link, Mar 5, 2005
    #18
  19. DJ Allen

    Dan Lanciani Guest

    In article <4229c5de$0$26550$-online.net>, (Uli Link) writes:
    | Dan Lanciani schrieb:
    |
    | > | A great plus of the Cisco APs is, you can set them to short, and they
    | > | fallback to long automagically. And the best: only for the unicasts to
    | > | the "long" stations.
    | >
    | > Yes, this kind of scared me the first time I saw it. Do you know of a way
    | > to display which nodes have been determined to need a long preamble? And
    | > what happens with broadcasts? They seem to get through as well.
    |
    | With IOS APs "show dot11 association [MAC address]"
    | There is a field "capabilities" with "ShortHdr" or not.
    | The VxWorks APs show equal info on the stations statistics. Never used
    | the cmd line with the VxWorks Aironets.

    What about the 68k-based boxes with Aironet's proprietary OS? (That's
    all I have. :)

    | I once had a single 3100 card, single antenna port.

    You've got me there; I've never even seen a 3100 series card.

    | > I bought several Symbol FH access points and cards, but the range was
    | > terrible even though they were the 500mW versions. Maybe it's time to
    | > look at FH again for security through obscurity. Probably nobody's
    | > snooping on it. :)
    |
    | The exorbitant power consumption disqualified this card for battery
    | powered devices.

    Yet Symbol loved to put them in their bar code scanners. (I have several
    on the shelf. Actually kind of nice little DOS computers.)

    | All Symbol FH installations were the most insecure I ever saw.
    | When asked for security the Symbol guys always talked of Kerberos, but I
    | never saw any real world installation with working Kerberos outside a
    | lab.

    I still use Kerberos a lot. Partly it's being lazy about installing ssh
    everywhere and partly it's because people put more effort into attacking
    ssh these days...

    Dan Lanciani
    ddl@danlan.*com
    Dan Lanciani, Mar 5, 2005
    #19
  20. DJ Allen

    macgd016

    Joined:
    Sep 6, 2010
    Messages:
    2
    QUOTE=Dan Lanciani]

    | in Europe limited to 50mW,

    Well, that's easy to fix. :)

    Hi I realise that this is a really old thread but I have three Cisco Aironet 350 Series AP 12.05 bridges. They appear to be the European spec and only give 50mw output. We are operating them at the very far limit of their range and if we could change them to the USA spec and have 100mw we might just get away with replacing the whole network. So if here is anyone out there still who could give me a noddy's guide to doing this or tell me where I can get the information I would be very much obliged.
    macgd016, Sep 6, 2010
    #20
    1. Advertising

Want to reply to this thread or ask your own question?

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. Dan Lanciani

    Re: AP4800 Config HELP !!!

    Dan Lanciani, Jul 12, 2003, in forum: Cisco
    Replies:
    0
    Views:
    1,096
    Dan Lanciani
    Jul 12, 2003
  2. Dave
    Replies:
    2
    Views:
    977
    Dan Lanciani
    Jul 15, 2003
  3. km
    Replies:
    0
    Views:
    843
  4. Andre Paetzold

    Aironet 1200 and Aironet 1300 compatible?

    Andre Paetzold, Dec 8, 2004, in forum: Cisco
    Replies:
    2
    Views:
    855
    Andre Paetzold
    Dec 9, 2004
  5. Sven Juergensen

    aironet <-> aironet bridging?

    Sven Juergensen, Dec 15, 2004, in forum: Cisco
    Replies:
    1
    Views:
    447
    Walter Roberson
    Dec 16, 2004
Loading...

Share This Page