Re: [802.3_EPOC] RF Spectrum Ad Hoc Minutes
Steve, et. al.
I understand. Let me re-state my point in much simpler terms. It doesn't
matter if the devices create some transmissions in the exclusion bands
because the purpose of the exclusion bands is to "avoid"
ingressors/interference.
The only reason such transmissions would be a problem would be if one wanted
to use exclusion to wrap around some hfc use, which - as I described below,
is not a good idea. So the clearly stated purpose of the exclusion
capability should be to avoid ingresses and other interference. If we agree
on this, then it would avoid the need to address the problem that Tom raises
which would ONLY be a problem if one tried to use it (as I think we all
agree should not be done) to 'exclude' same range of frequencies for other
HFC T/R.
I think Tom is right and someone MIGHT try to do this, but protocols cannot
prevent poor choices. All we can do is state the intended use. At this time,
in the ad-hoc, if we can agree on the use of exclusion, then we don't need
to address how to fix a problem which is NOT the intended use.
Victor
From: Shellhammer, Steve [mailto:sshellha@xxxxxxxxxxxxxxxx]
Sent: Wednesday, December 12, 2012 7:11 PM
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_EPOC] RF Spectrum Ad Hoc Minutes
Victor and Tom,
The OFDM waveform consists of many subcarriers. Each
subcarrier would be infinitely narrow if it went on for every. However, the
subcarrier values change every symbol period. So one can think of these
subcarriers becoming sinc functions in the frequency domain, since the time
domain window is a rectangle. So even though there is no energy from
subcarrier M, there is the impact of the adjacent subcarriers being sync
functions. These sync functions are zero at the subcarrier M frequency, but
not exactly zero close to that frequency. So all the other subcarrier have
some impact in the exclusion sub-band. This is like out-of-band emissions
in any communication system, the spectrum drops off but there is still some
residual out-of-band emissions. This morning Avi mentioned the idea of
windowing. If that is done you no longer get a sync function in the
frequency domain, and the out-of-band emissions drop off faster.
It would be really good to have an out-of-band emission
requirement. It is listed on the Open Issues list, but we have not
addressed it yet. If we had that we could see if we need windowing, and if
so what type.
Steve
From: Victor Blake [mailto:victorblake@xxxxxxxxxxxxxxx]
Sent: Wednesday, December 12, 2012 12:44 PM
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_EPOC] RF Spectrum Ad Hoc Minutes
Tom, Steve, et. al.
And I think this gets, perhaps better, to my point. If the purpose of
exclusion is, for example, to avoid ingressors, than not using the
frequencies is fine because the receivers (in either direction) won't expect
anything there. So RF there will just be ignored. This, I think is the
intended use and for it, 50 db down is certainly sufficient.
That is very different from using the exclusion to drop a band on-top of or
around other HFC uses, in which case, as Tom points out, even 50db down
signals could cause problems. I'm not sure of the solution, but I'd at least
ask, why do it this way ? (wrench as hammer analogy).
-Victor
From: Tom Staniec [mailto:tom.staniec@xxxxxxxxxx]
Sent: Wednesday, December 12, 2012 2:57 PM
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_EPOC] RF Spectrum Ad Hoc Minutes
Steve
Not to be a pain but I can still have a carrier with a power spectrum with
no modulation in an RF world but it doesn't mean the RF power is "0" or off.
So in the OFDM sub-carrier world does "modulation set to 0" mean there is no
RF carrier present in the sub-carriers set to 0? Stated another way, if I
connected a spectrum analyzer to a point in the network where I can a set a
band of sub-carriers set to "0" would I see a "hole" where the carriers were
and the noise floor? If I see anything other than that then I have to
consider the carrier as being "on" and not "off."
In short, my expectation is turning the sub-carriers off yields NO
measurable RF power within the defined band. While 50 dB down is low, it is
not zero power. Going back to the analogy of the light switch, throwing the
switch on fills the room with light. Off makes the room dark or no presence
of light. So if I set the sub-carrier to "0" modulation is the room dark (no
light) or do I still have remaining light but very dim 9think 5w night
light)?
Tom
From: Shellhammer, Steve [mailto:sshellha@xxxxxxxxxxxxxxxx]
Sent: Wednesday, December 12, 2012 2:04 PM
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_EPOC] RF Spectrum Ad Hoc Minutes
Jim, Tom, Victor, et. al.,
In terms of turning off the subcarrier versus any filtering,
the exclusion sub-bands is a set of subcarriers whose modulation is set at a
"zero" meaning that the subcarrier carries no signal. So the subcarrier is
effectively "off."
Depending on if we do windowing of the OFDM symbols or not,
the suppression of the PHY signal in the exclusion sub-band may be
suppressed, as was suggested up to 50 dB or more.
We should see if we can get some volunteers to do some
simulations.
In terms of the questions about "proposals" versus
"recommendations" the proposal column is to capture ideas suggested by
anyone on the call, while the recommendation column is an agreement of the
Ad Hoc to recommend to the task force. Then the task force will vote on our
recommendation to see if there is a consensus in the task force.
So any recommendations we agree to before the January meeting
will be brought to the task force at the January meeting.
Jim and Tom, thanks for the information about the pilots.
On the upstream band, which is larger than 192 MHz, that
would allow the operator to shift the center frequency, though not by much
in the upstream given the limited frequency band.
Steve
From: Victor Blake [mailto:victorblake@xxxxxxxxxxxxxxx]
Sent: Wednesday, December 12, 2012 10:20 AM
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_EPOC] RF Spectrum Ad Hoc Minutes
Tom,
I agree about having the choice to shut-off (out), but limiting it to all-on
or all-off 192 leaves little flexibility even with a sliding (any)
start/stop f (that is, no grid). To that end, I think we still need the
ability, within a 192MHz band to exclude specific (narrow) ranges of
frequencies as is currently being discussed.
I don't think that means that one would place the 192 with, for example, a
block of AM smack in the middle (very bad idea), but it more likely means
excluding frequencies that are otherwise problematic, not due necessarily to
other HFC use, but from external sources.
-Victor
From: Thomas Staniec [mailto:staniecjt@xxxxxxxxx]
Sent: Wednesday, December 12, 2012 12:14 PM
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: Re: [802.3_EPOC] RF Spectrum Ad Hoc Minutes
Steve and Jim
I agree with Jim points in the second paragraph below. Generally in view the
best filtering is no filtering meaning the complete ability to turn the
carriers off much like turning of the lights in a room. The 50 dB number
might be acceptable in the "white space" between channels but I don't think
it is good in channel and specifically analog channels that may still exist.
Residual RF energy particularly in and around the color subcarrier will
cause noticable beat issues in the picture.
Unfortunately I did not get back on the call before it ended yesterday but
while the call was in progress I called a close friend who is respondsible
for the manufacter of amplifiers and nodes. I asked what I might expect to
see in the field related to pilot carriers. He told me I would definitely
see a wide cross section of deployments with dual analog slope and gain
pilots, gain only pilots, QAM sensing pilots and cw carrier pilots in
currently deployed networks. He commented that the time frame of the
original deployment will be the best determinent of what we will see in an
operators network. So in my view, I would plan on a survey before an EPoC
deplyment is done to make sure I could tell the operator what the best
options are for his plant.
I guess that leads me back to not wanting to describe exclusion bands at all
but provide a general operational model that allows the operator to decide
how he wants to handle a conflict between an 802.3 bn carrier and a carrier
inside the coax which can't be moved or ingressing into the coax causing an
issue. In the latter case, I would think the operator would want to fix the
issue. In the former case, I would consider it "implementation planning."
;-)
I'm not sure I understand the difference in the third paragraph either. I
think, in the current standard development, the channel is the channel: 192
MHz. The only option - read that as "choice" - the operator has is to
operate at cumulative 192 width or make the channel smaller in size with a
corresponding reduction in actual transmission capacity. The best example is
where we started with 1GE in 120 MHz. I'm not sure of bandwidth variances
beyond 192 MHz per se but
I suspect in the upstream the bandwidth is at issue because we haven't
really determined where the diplex filter edges need to be.
My thought at any rate....
Tom
On Wed, Dec 12, 2012 at 10:03 AM, Jim Farmer <jfarmer@xxxxxxxxxx> wrote:
Thanks Steve. Since we are not meeting again for a while, and since my
feeble mind cannot retain information very long, maybe I should register a
few comments now.
Regarding the minutes, there seems to be an implied concept, which I
intended to bring up on the phone but forgot (I was triple-tasking during
the call, unfortunately), that we would generate signals in exclusion bands,
then filter them out. I see us simply turning off generation of these
signals, so that there will be no noise contribution at all from the
exclusion band, except maybe to the extent that there is limited filtering
of adjacent carriers that ARE turned on. Someone seemed to suggest that we
could expect 50 dB of attenuation of these signals, which is a pretty good
number, though I'd like to understand better how the spectrum would look.
Also, there is a note that in dual pilot systems, both pilots may be within
the 192 MHz passband. I doubt that the two pilots would be this close in
frequency, but assuming several 192 MHz wide bands dedicated to EPoC, then
it is certainly possible that each pilot will fall into one of our bands.
It is true that most all amplifiers out there use an analog carrier as a
pilot, though some recently-deployed amplifiers may use a digital carrier.
I would expect these to need either a 4 or 6 MHz exclusion band in order to
function normally. Fortunately we don't have to have much attenuation in
order to let the pilot work normally, but we do have to have on the order of
50 dB attenuation I order to provide a good carrier-to-noise ratio for
analog signals. I'm not sure how many analog signals will exist by the time
this system gets deployed, but some probably will still be deployed.
Regarding the open issues, I'm still not clear of the distinction (and need)
between mandatory and optional FDD downstream channels, but maybe it'll get
through my thick skull one of these days. I suppose it has to do with
frequency. The upstream FDD bands are curiously wider than 192 MHz, and not
by a consistent amount. Ditto for the TDD bands. I'm not sure how much we
need to be specifying band edges, since the industry is going to do what it
sees as most beneficial. We may want to include text that says that while
certain bands are suggested, use of other frequency bands is not a violation
of the specification. Finally, there seems to be an inconsistency between
"Possible Rules" and "Recommendations," in that "rules" has struck the note
about exclusion bands being to control ingress, while that language remains
in the "Recommendations."
Thanks, and Merry Christmas to all.
jim
Jim Farmer, K4BSE
Chief System Architect,
FTTP Solutions
Aurora Networks
1220 Old Alpharetta Rd.
Ste. 370
Alpharetta, GA 30005 USA
678-339-1045 (office)
678-640-0860 (mobile)
jfarmer@xxxxxxxxxx
From: Shellhammer, Steve [mailto:sshellha@xxxxxxxxxxxxxxxx]
Sent: Tuesday, December 11, 2012 4:46 PM
To: STDS-802-3-EPOC@xxxxxxxxxxxxxxxxx
Subject: [802.3_EPOC] RF Spectrum Ad Hoc Minutes
All,
Attached are the minutes from today's call. Also attached is
R4 of the RF Spectrum Open Issues List. The a schedule conflict next week
and the upcoming holidays, our next call will be in January.
Steve
_____
<="" p="">
_____
_____
_____
<="" p="">
_____
<="" p="">
_____
_____
<="" p="">
_____
________________________________________________________________________
To unsubscribe from the STDS-802-3-EPOC list, click the following link:
https://listserv.ieee.org/cgi-bin/wa?SUBED1=STDS-802-3-EPOC&A=1