stds-802-16-tg2: Re: Letter Ballot Comment 39 Resolution

["Barry Lewis" <radcom25@dircon.co.uk>]

Jack,
I meant to reply to the message you sent regarding my comments on "guard
bands" etc.. Although to some extent this is water under the bridge as you
have now sent your resolution comments.
The problem is that it seems what is good for one scenario is not
necessarily good (but it might be!) for another which makes it extremely
difficult to come with an all encompassing recommendation regarding guard
frequency without lots of caveats. Therefore I would support your proposal
to somehow retain the existing text and supplement it with my comment. The
difficulty is finding a balance between a firm recommendation and something
that recommends nothing except a list of options with no way out. (To some
extent this was the rationale behind my proposed deletion of the word
"usually" in the second line of Recommendation 8).

But, now I've looked at your resolution of Comment 39 and your sixth
paragraph dealing with p22, line 7 doesn't make sense as you've mixed "same
channel bandwidth" and "wider system" in the same sentence.

Can I make another proposal to resolve the comment;
p22, line 7: Delete the existing first sentence and replace with:
"In most co-polarised cases, where transmissions in each block are employing
the same channel bandwidth, the guard frequency should be equal to one
equivalent channel. Where the transmissions in neighbouring blocks employ
significantly different channel bandwidths, then, it is likely that a guard
frequency equal to one equivalent channel of the widest system will be
adequate. However analysis suggests that under certain deployment
circumstances this may not offer sufficient protection and that guard
frequency equal to one channel at the edge of each operators block may be
required."

I think this combines both concepts in an acceptable manner and the last two
sentences can also used to resolve comments 45 and 46.

The only other issue I have on this is what is the equivalent of guard
frequency on p21 line 44? I don't understand this. Surely some guard
frequency is required unless specific mitigation techniques are employed
which are detailed elsewhere. These may have the effect of reducing
(possibly to zero) the guard frequency requirement.

Lastly, thanks for your help on comments 108, 110 and 112.

Cheers

Barry Lewis
RA London



-----Original Message-----
From: Jack Garrison <gjg@telus.net>
To: Philip Whitehead <pw@radiantnetworks.co.uk>; Demos Kostas
<dkostas@adaptivebroadband.com>; Adrian Florea <adrianf@newbridge.com>; Vito
Scaringi <vscaring@newbridge.com>; Bob Whiting <rwhiting@gabrielnet.com>;
Remi Chayer <rchayer@harris.com>; Paul Thompson <paulcom@paulcom.com>; Barry
Lewis <lewisb@ra..gtnet.gov.uk>; Reza Arefi <reza.arefi@wfinet.com>; Andy
McGregor <mcgregor@nortelnetworks.com>
Cc: Bob Foster <bfoster@wavtrace.com>
Date: Friday, December 29, 2000 12:57 AM
Subject: Letter Ballot Comment Resolution


>References:
>
>Ballot Comments 9, 15 and 16
>Contribution 802.16.2c-00/22
>
>Hi All,
>
>I decided to look at Barry's comments in further detail as I had placed
>a "not sure" action response against them. Here, Barry wants added or
>replacement wording so that the same area guard band "may" require a
>spacing equal to the carrier bandwidth of both operators. This would
>apply when the 2 operators are employing significantly different carrier
>bandwidths.
>
>In his prior contribution, Barry modified the ETSI Type B emissions mask
>to account for reality, i.e. the emissions will probably continue to
>roll off beyond a 2'nd adjacent spacing. But across the spacing we are
>dealing with, this does not come into play and Barry's mask is the same
>as that of ETSI.
>
>In his contribution, Barry examined the impact of a spacing equal to the
>sum of a 1/2 carrier spacing of the wide bandwidth carrier (BW1) and the
>narrow bandwidth carrier (BW2). Hence, the spacing would be BW1/2 +
>BW2/2. The contribution also looked at a spacing equal to BW1 + BW2. But
>what was not looked at was a spacing equal to just BW1. Would it be much
>different than BW1 + BW2?
>
>I decided to run it through my archaic computational computer system.
>
>Using Barry's example of a bandwidth ratio BW1/BW2 = 4/1, I got the
>following results:
>
>Cxr BW                    Spacing                    ACI
>
>BW2=BW1                BW1                       47 dB
>
>BW2=1/4BW1           BW1                       50.5 dB
>
>BW2=1/4BW1           BW1+BW2            53.8 dB
>
>So, by increasing the spacing by an extra BW2, we pick up about 3 dB.
>Assuming that the narrow band BW2 system runs at a 6 dB lower threshold,
>we are now back to about the same effective ACI values as the equal
>bandwidth systems. I therefore conclude that the existing wording should
>stand and be supplemented by Barry's proposed wording or the equivalent.
>
>But there is one thing that our document does not address; who gives up
>the BW1 + BW2 bandwidth? Should a BW1 operator be the victim because his
>neighbor decides to run narrow band or vice-versa? I know we can't
>answer this one, nor should we.
>
>Happy New Year,
>
>Jack
>
>PS. It is now the 28'th and I haven't seen a voting update. I assume
>Roger is off skiing. What happens if  there is less than a 50% voter
>turnout?
>
>
>
>
>