Thanks both for replies

I have to say I put in the "????" for just this reason.  This whole RSSI% vs
RSSI dBm mapping leaves me very bewildered !!  Although Steve I think your
results are more like what I would expect to see.

The default AB mapping (internal formula) is
2% -93dBm
5% -91dBm
7% -89dBm
10% -87dBm
12% -85dBm
15% -82dBm

20% -78dBm
30% -70dBm
40% -61dBm
50% -53dBm
60% -44dBm
70% -36dBm
80% -27dBm

which looks like a very sensible way to calibrate: lowest workable signal at
close to 0%, raging hot signal at fsd.  As far as agc dynamic range will
take it.

So this says we need
 2% for 1Mbps (@~ -94dBm see product spec)
10% for 5.5Mbps (@~ -87dBm)
13% for 11Mbps (@~ -84dBm)

All very reasonable.

Lets add ~15dB safety margin (thats a very well engineered link) to
give -70dBm - a roaring signal. This corresponds to just 20%.  So, thats it
then, anything above 20% is a knock out rock solid signal !!!!! Everyone
agreed ?

Why doesn't that stack up with reality ?  Not leaving a customer 'til you've
got 80% !!  Well, the receiver should have steam coming out of its ears
with -27dBm at the antenna socket !

With +17dBm Tx, 13dBi antenna at each end, direct LOS in free space, no
feeder loss etc  you will get -46dBm at 300m (~58%).  Thats a ~40dB fade
margin for 11Mbps.    But who expects to put up 13 dbi at both ends for 300m
in perfect conditions ? Even with an ABI (~2dBi antenna indoors) the result
is usually much better than 58% at this distance.  It just doesn't add up

I only wish I had the correct test gear to calibrate a receiver.  sB support
team, are you there ?  What am I doing wrong ?  Is the internal formula
sooooo far out ?  If so could you give us a sample table for typical user
defined values.

bw

----- Original Message ----- 
From: "Steve Good" <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>
Sent: Tuesday, November 25, 2003 3:48 PM
Subject: RE: [smartBridges] Minimum working config suggestion (was Tranzeo
CPE / APPO Compatibility)


In response to your RSSI part. I use some WUSB (linksys) units because the
customer is close to the tower and they work like a champ getting T-1 speeds
from them. But in SIMPLE NMS or SIMPLEMONITOR they show they might only have
50-55% RSSI. Does anyone know if non SB products will show a true RSSI
and/or LQ in Simple NMS or SimpleMonitor?



THanks,
Steve

________________________________

From: [EMAIL PROTECTED] on behalf of [EMAIL PROTECTED]
Sent: Tue 11/25/2003 7:15 AM
To: [EMAIL PROTECTED]
Subject: Re: [smartBridges] Minimum working config suggestion (was Tranzeo
CPE / APPO Compatibility)


The only thing I disagree with is the use of a spectrum analyzer.

By choosing from an available access point the only way to see the access
point is to either know the SSID and statically enter it in the the CPE or
the device that is potential interference must be set to broadcast the SSID
or it will not show up in the available access points list.

Microwaves, cordless phones, and other wireless ethernet devices will not be
detected without one.


Another thing, the recommendation for signal strength I believe is no less
than 75%.  That should be no less than 80% average based on my experience.
I had one client that I left running around 77% average, it was a mistake
because I rolled the truck out to replace the ABO with 13db to an outdoor
with a Maxrad 18db flat panel.  That bumped the 77% average up to around 87%
and I have not heard from them since.  Marginal LQ and RSSI = marginal
service.  I now do not leave a customer site with less than an 80% average,
no acceptions.  If I cannot get that I deem the service unavailable in that
location.

I have also had very good luck by eliminating the Auto Fallback ability.  I
set all my CPE's to 2Mps and my APPO's are set with 1Mps and 2Mps enabled.
Both devices have Auto Fall back disabled.  This has provided stability.
Marginal signal devices continualy trying to change the rate is additional
overhead and results in latency as the rate change takes place, busies the
AP during that time which affects all users, and if you have multiple
devices doing this on a regular basis times the problem time the number of
users and you end up with a busy system.  I have mentioned this before on
this forum, but a great way to test if this is being an issue is to watch
the link status in Simple NMS and while performing a performance test on
Toast.Net to see if the LQ ever drops to zero during the test.  If so, back
the data rate off (with auto fallback disabled on both ends) and continue to
back it off until the link remains stable during the test.  Basic internet
browsing will not behave this way, it is only during periods of heavy
traffic.  If you get down to an unacceptable speed then it is time to go
back to square one for that particular client.  I get just under full T1
speed set to 2Mps.  Some folks say that in doing this it negatively affects
the AP because it keeps the communcation throttled back and busies the AP,
but not as busy if it is continually changing rates and/or re-associating
and the end result is retransmitting packets.  I would rather see the
packets get through reliably the first time.   So the theory of let it get
through as fast as it can I would say is great under perfect conditions, we
are all not under perfect conditions when we go in the field.  Auto fallback
in my opinion is meant to be used in a WLAN environment where the signal
levels can be kept high in a more controlled environment to eliminate the
data rate changes.  I am not sure it is set this way by default because it
is the recommended setting, I think it is more to do with the 802.11b
standard, very much like by default the SSID broadcasting is enabled and is
one of the first things that gets enabled, but to be 802.11b compliant it
must be enabled to meet the standard.

My 2 cents.

Very good post Brian.
----- Original Message ----- 
From: Brian Winter <mailto:[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Sent: Tuesday, November 25, 2003 4:53 AM
Subject: [smartBridges] Minimum working config suggestion (was Tranzeo CPE /
APPO Compatibility)

I'm one who currently believes that many "issues" seen on this forum
(lockups etc) are down to the way we, all of us, expect to utilise this
fragile airtime protocol in the real world - sB have little control over
this if their kit is to remain standards compliant (however bad the
standards are!).  There will be some problems which are due to faulty kit.
I wonder if it would be useful for list members to jointly devise a minimum
working configuration which does not push the protocol to its working
boundaries, and a check list of observations to best get through the most
common areas of concern.  Once basic operation is achieved, each parameter
can be backed off towards its expected operational state or 'til it stops
working.  Time consuming, but scientifically satisfying  ;)    Does this
already exist on sB site ?

I have also had very good luck by eliminating the Auto Fallback ability.  I
set all my CPE's to 2Mps and my APPO's are set with 1Mps and 2Mps enabled.
Both devices have Auto Fall back disabled.  This has provided stability.
Marginal signal devices continualy trying to change the rate is additional
overhead and results in latency as the rate change takes place, busies the
AP during that time which affects all users, and if you have multiple
devices doing this on a regular basis times the problem time the number of
users and you end up with a busy system.  I have mentioned this before on
this forum, but a great way to test if this is being an issue is to watch
the link status in Simple NMS and while performing a performance test on
Toast.Net to see if the LQ ever drops to zero during the test.  If so, back
the data rate off (with auto fallback disabled on both ends) and continue to
back it off until the link remains stable during the test.  Basic internet
browsing will not behave this way, it is only during periods of heavy
traffic.  If you get down to an unacceptable speed then it is time to go
back to square one for that particular client.  I get just under full T1
speed set to 2Mps.  Some folks say that in doing this it negatively affects
the AP because it keeps the communcation throttled back and busies the AP,
but not as busy if it is continually changing rates and/or re-associating
and the end result is retransmitting packets.  I would rather see the
packets get through the first time.   So the theory of let it get through as
fast as it can I would say is great in theory under perfect conditions, we
are all not under perfect conditions when we go in the field.  Auto fallback
is much better and in my opinion meant to be used in a WLAN environment
where the signal levels can be kept high in a more controlled environment to
eliminate the data rate changes..

The main technical concerns seem to be (list start - please add)
1. Poor receive signal strength    (long distance, non LOS, leaves and other
obstructions, poor aerial/feeder etc, multipathing, dead spots)
2. Interference - Poor link quality (co-channel traffic, adjacent channel
traffic, multipathing, microwaves, heavy current switches, electric
railways, high power broadcast transmitters, pulsed radar etc)
3. Inappropriate data rate for low sigs or interference
4. Use of software which needs broadcast messages (eg dhcp)
5. Non-reciprocal radio paths


My first suggestions for config are:
1. Fix to one channel for results gathering - different chans may show
different results - confusing.  When one channel is eliminated (noisy etc)
fix to another one and start again.

2. Disable WEP - WEP protocol overhead means 50% of the traffic across the
air is unnecessary for testing.  On top of this many (most) probs are in
fringe coverage, so much of the real and overhead traffic will be retried,
reducing throughput further.  If you have congestion this might relieve some
of it.


3. Select "Basic Rate" = 1Mb "AutoFallBackRate" = off (should not matter as
basic rate is set to minimum, but will prevent any rate negotiation).

4. Set fixed IP for client side radio, and the first device on the client
side.  DHCP involves broadcast messages which are unreliable in the
protocol - VERY IMPORTANT (esp thro AB)

5. Set preamble = ???.  For best performance in poor signal conditions.

6. Set allow "Non IP Traffic" off - less extraneous traffic to deal with.

7. Set Fragmentation = ???.  For best performance in poor signal conditions.

8. Set RTS/CTS = ???.  For best performance in congested conditions.
Messages of length < this value are sent without regard for others on the
channel, which can give contention/retry problems on a busy channel.
Messages of length > this value tell other radios to "shut up" so it can get
this message out and hear the reply without someone else butting in.  This
adds a significant overhead of its own - much of the available airtime is
spent "busy waiting" - Although this parameter is most beneficial in high
capacity systems, thats where you need least overhead - theres the rub !!
Also note that this is only worthwhile where the other radios can hear you
properly (see discussion on reciprocity later).  Also read up on the "hidden
node" problem - google "hidden node".

9. Use a cheap router at the client end to get rid of all the default
Windoze network management traffic (eg broadcasts from Computer Browser
service)

10. Set the radio transmit power (not eirp) to be the same at both ends !!!
With tx/rx on the same aerial, there is nothing to be gained by whacking up
the output power at the client end, nor by changing it for other kit with
higher power output.  The link is only as good as the worst leg, and can be
much worse than that.  By increasing only one end you are forcing the far
end to send (and resend) replies that you just can't hear - this is just
more "noise" for everyone else on the same channel, and no benefit for you.



Observations to report should include:
1. RSSI (should be > ??? for 1Mb data rate) - Link Status tab

2. LQ (should be > ???) - Link Status tab

3. RSSI/LQ observed on other channels ? SM for AP(CB mode), "Site Survey"
tab, "Select from Available Acess Points"  (I could never figure why anyone
should need a spectrum analyser when we have a perfectly good sensitive
receiver in exactly the right location, with exactly the antenna
arrangement, on exactly the freq of interest etc etc that we want to
measure - toys for the boys)

4. Compare stats for transmitted "Unicast Packets" and received "Received
ACK" - They should be very similar.

5. Received RTS and CTS will give some idea of how the system is managing
contention

6. Failed Packets and Aged Packets are the results of contention, poor
reciprocity, one end interference etc.  Measure over fixed periods at
different time of day


Its a real shame that SimpleNMS/Monitor can't let us change parameters (at
both ends) and watch statistics at the same time

Comments welcome.






Discussion on Reciprocity:

Consider a pretty normal path between A (say APPO, AP mode) and B (say ABI,
CB).  For this discussion ignore feeder losses, assume LOS, perfect path
etc - I'm not suggesting that a link of 10Km is viable in this way !!


>From A > B

Tx A       +17 dBm
Antenna A  +13 dBi   (eirp = +30dBm Illegal in UK)
PathLoss  -120 dB    (eg 10Km @ 2437Mhz)
Antenna B  +02 dBi
==================
Rx sig str -88 dBm   (at B)
Best Rate    2 Mbps  (assuming no fading etc !!!)



>From B > A

Tx B       +17 dBm
Antenna B  +02 dBi   (eirp = +19dBm)
PathLoss  -120 dB    (eg 10Km @ 2437Mhz)
Antenna A  +13 dBi
==================
Rx sig str -88 dBm   (at A)
Best Rate    2 Mbps  (assuming no fading etc !!!)

You have a good reciprocal path.  No matter how you change the aerials you
will always have a reciprocal path - Good




Now, on a whim, you decide to increase the client end transmit with an amp
(yuk) or a 200mW pcmcia card


>From A > B (no change)

Tx A       +17 dBm
Antenna A  +13 dBi   (eirp = +30dBm Illegal in UK)
PathLoss  -120 dB    (eg 10Km @ 2437Mhz)
Antenna B  +02 dBi
==================
Rx sig str -88 dBm   (at B)
Best Rate    2 Mbps  (assuming no fading etc !!!)



>From B > A

Tx B       +23 dBm   (=200mW)
Antenna B  +02 dBi   (eirp = +25dBm)
PathLoss  -120 dB    (eg 10Km @ 2437Mhz)
Antenna A  +13 dBi
==================
Rx sig str -82 dBm   (at A)
Best Rate   11 Mbps  (assuming no fading etc !!!)


This has not improved the system even slightly.  In fact all we have is a
client install spraying out louder messages that it can't hear the reply
(ACK) to - further cluttering up the AP with useless noise.  It might even
be worse than this as the dialog during association will be full of retries
etc.  DHCP might be OK for client AP but pointless even trying for AB.  This
could lead to all sorts of strange issues.



----- Original Message ----- 
From: Nimesh D. Parikh <mailto:[EMAIL PROTECTED]>
To: [EMAIL PROTECTED]
Sent: Tuesday, November 25, 2003 3:04 AM
Subject: RE: [smartBridges] Tranzeo CPE / APPO Compatibility


Hi Ian,

I am sorry to hear of the problems you are having with the aBOs. Since you
are running 1.08 and still having to reset, I think the problem should be
some thing else. This firmware had solved the issue of occasional
dis-association of the aB. This dis-association used to happen on a
relatively small number of installations, depending on some combination of
conditions. After many months of investigation we could never figure out
what were these conditions. During these investigations we learned that this
is a common and unpredictable problem with almost all manufacturers'
devices. Well anyway, now we are happy that 1.08 solves the dis-association
issue.

As others have pointed out, this is a tech support forum. So most times you
get to hear only the "I got problem" cry for help. Once the issues are
resolved you would seldom find a post that clarifies the issue. We think
that there are perhaps less than 1% or 2% of our users posting regularly on
this forum. Typically when somebody is having problem, jumps on the forum,
gets help from the experienced users and then becomes a passive reader. This
is a great forum with so many friendly and willing users contributing their
expertise. The forum helps us stay in close contact with our customer base.
A little balance should be kept while reading the forum to get the best out
of it.

With best regards,

Nimesh
sB



________________________________

From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
On Behalf Of Ian Ellison
Sent: Tuesday, November 25, 2003 10:39 AM
To: [EMAIL PROTECTED]
Subject: Re: [smartBridges] Tranzeo CPE / APPO Compatibility

Yes, all of my ABO's are running 1.08.  At least 25% of them still need to
reset frequently.  It isnt as bad as it was, and I am tyring to setup a
"ping" keepalive system by recommendation of this list.  It shouldnt be
necessary, but it it works for now, I'm game.

Ian


Pascal Losier wrote:


Are those corporate account running the 1.08 firmware and still have to
power cyle their unit ????

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]
On Behalf Of Ian Ellison
Sent: Monday, November 24, 2003 8:55 PM
To: [EMAIL PROTECTED]
Subject: Re: [smartBridges] Tranzeo CPE / APPO Compatibility
It is great that there are those of you who are experiencing no problems
withe sB equipment, sincerely.  However, that does not dismiss the fact that
there are SEVERAL people using these products, who are all experiencing the
same difficulties, that do not seem to exist (at least as much) in so many
other manufacturers products.  I made the mistake of buying several units at
an early stage (all of my units came with the old TRULY transparent
firmware), with the (and Im so glad this is no more) 50' of cable attached
to the radio.  Perhaps since then they have better quality control, or other
factors.  In the meanwhile, those of us who are so unfortunate to be
experiencing these problems are hardly seeing any "price performance" when
we are constantly doing tech support and truck rolls to alleviate the
issues.  the first unit i bought was an Alvarion DS.11, which has (is) run
flawlessly (reaching WELL beyond 200 feet i might add ;o).  For "price
performance" reasons, I made the change to smartBridges.  Noise and
interference could definitely be an issue, but we are not finding any noise
of significance at our location.  The suggested steps for resolving my 100%
RSSI,  0% Quality problem included checking for interference, etc.  When all
of htis turned up nothing, i simply replaced the ABO, and it was fine (as
well as could be expected anyway).  This unit had only been in (highly
problematic) operatoin for about 3 months.  We  havent even  made enought
money from the subscriber to cover the cost of the unit, and we have to
replace it....

I just  hope you can see why this is so frustrating to those of us who ARE
experiencing these issues.  I havent tried RMA yet, but im going to have to
try soon.  Hopefully smartBridges will make good on their products, and
perhaps the newer units will be more robust and less problematic.  It goes
without saying that one has to appreciate their online customer interaction
and responsiveness to suggestions.  they get an A+ in that category.   I
truly hope that the issues get resolved.  I have almost all smartBridges
equipment in use, and didnt really want to mix manufacturers for CPE.

I still have about 10 ABO units brand new in the box (the ones with the
attached 50' cable).   It would make my day to "trade" them back to
smartBridges for the newer units so i can see if there is in fact hope to
continue deploying smartBridges.

I apologize for the rant, but my largest corporate account has been on the
phone ranting to me about how ridiculous it is that they have to reset their
internet connection once a day....  I'm about to deploy a PTP 5.8 Ghz link
as a temporary solution to make them a happy customer.  Another $1000 spent.
thats one expensive band aid :o\

Looking forward to a brigher future!

Ian


[EMAIL PROTECTED] wrote:


Just a reminder, there are many of us not having to reset the AB's or any
problems for that matter.  If you have a problem with 2.4 in general due to
noise and interference a different manufacturer is not going to help you.

I had a potential competitior who attempted to deploy the "higher priced"
Alvarion gear.  The furthest they could reach out from the tower was 200',
although I think it was stable at 200'.  : )

They threw in the towel and are no longer pursuing the wireless internet
industry.

Be careful in what you spend.  A $300 AP and a $1500.00 AP will both become
obsolete within 5 years.

Again, not everyone is experiencing problems.

----- Original Message ----- 
From: "Ian Ellison" <[EMAIL PROTECTED]> <mailto:[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]> <mailto:[EMAIL PROTECTED]>
Sent: Monday, November 24, 2003 4:07 PM
Subject: [smartBridges] Tranzeo CPE / APPO Compatibility



Has anyone tried using the Tranzeo CPE units with an APPO?  Due to the
way too many headaches I'm having with ABO's needing reset (constantly),
I am venturing out and trying some new CPE until (and then, maybe) sB
can get some issues resolved.  I'm not even using APPO at my next AP,
but am using high-priced Alvarion equipment to test the stability.
However my other AP's are still running APPO's, so I'm wondering if
anyone else has tried this combination, or has any opinions on the
Tranzeo line of products.

Thanks,

Ian

The PART-15.ORG smartBridges Discussion List
To Join: mailto:[EMAIL PROTECTED] (in the body type subscribe

smartBridges <yournickname>

To Remove: mailto:[EMAIL PROTECTED] (in the body type unsubscribe

smartBridges)

Archives: http://archives.part-15.org




The PART-15.ORG smartBridges Discussion List
To Join: mailto:[EMAIL PROTECTED] (in the body type subscribe
smartBridges <yournickname>
To Remove: mailto:[EMAIL PROTECTED] (in the body type unsubscribe
smartBridges)
Archives: http://archives.part-15.org


The PART-15.ORG smartBridges Discussion List
To Join: mailto:[EMAIL PROTECTED] (in the body type subscribe smartBridges 
<yournickname>
To Remove: mailto:[EMAIL PROTECTED] (in the body type unsubscribe smartBridges)
Archives: http://archives.part-15.org

Reply via email to