|
How do I best characterize the cost difference of
plant migration (e.g. to 860 MHz or 1 GHz) now, versus five years ago? I
keep hearing it’s cheaper now, but I need to understand why.
The reason it’s cheaper now than five years ago is that
today’s RF amplifiers have the performance and gain characteristics
needed to do a drop-in upgrade.
If we go back to the last big go around – the 750 MHz upgrades – all of
our amplifier hybrids had about the same gain specifications. The only
thing that changed in performance was the upper spectral boundary. We
went from 400 MHz to 450 MHz, then to 550 MHz, then to 750 MHz -- but
all of the amplifiers had the same gain characteristics for a given
channel loading. As a result, to increase bandwidth, you ended up moving
amplifiers around, changing tap values, moving cables. It got pretty
pricey.
Now, we use discrete components instead of hybrids. Specifically, we use
Gallium Arsenide (GaAs), which gets you better distortion than the old
hybrids. The means higher output and better noise characteristics.
It used to be that upgrading meant you had to change the taps, because
the values were wrong. You were re-splicing down the street. Now, with
the drop-in upgrades, you’re not messing with splitters, couplers, or
taps.
Today’s bandwidth migrations are less than half as costly, and much
faster, than the last upgrade cycle.
If my plant is mostly 750 MHz, and my amplifiers were
spaced for 550 MHz, can I go spectrally higher without having to
re-space?
It depends. There are some 550 MHz amplifiers that can get to
1 GHz, and some that can’t. Here’s how you know.
If you go back to the 550 MHz days, in the days of classic
trunk-and-feeder topologies, we had the old trunk and bridger systems.
Back then, it was long cascades, with some very elaborate methods, like
feed-forward trunks, and power-doubling bridgers, some with dual
outputs. That’ a long way of saying that there was a phase of 550
amplifiers that were all souped up, and they had their levels cranked up
real high, too.
When you want to increase bandwidth today, you have to take your
existing levels up, up, up, to 870 MHz or 1 GHz. And, if you had your
550s tweaked to the max, then you’re probably going to have problems
getting to a Gig.
It comes down to signal level. If the old, 550 MHz system had a feeder
level greater than 46 dBmV, then a drop-in upgrade to 870 MHz is iffy.
It depends on what you have in place. But I’d say that the bigger nodes
– 500 to 1,000 home passed – most of those we can upgrade.
If I have a choice between going to switched digital
video, or migrating plant to higher bandwidth, what factors should I
consider in making my decision?
I’m not convinced it’s an “either/or” answer. I think things
have moved to the point where the two go hand-in-glove.
When people first started looking at switched digital video, the idea
was to put lesser-viewed material on the switched tier. That created
unused bandwidth, which could be used for more HD channels. The goal was
to save 6 to 8 QAM (Quadrature Amplitude Modulation) channels.
Well, between the time that concept came out, and the actual
deployments, the “long tail” phenomenon was born – an increase in
content that, by its very definition, isn’t widely viewed.
Some operators look at switching and bandwidth upgrades as a two-step:
Switch first, as a way to preview new, native language channels or HD
channels. As that’s happening, get the bandwidth upgrade underway. You
can add a switch that serves all your hubs, real fast. It’s just racking
and stacking and feeding narrowcast tiers into the optical transmitters.
Assuming no re-spacing of amplifiers, how quickly can
I shift to a higher bandwidth network, start to finish?
You can do it real fast. It’s just a matter of people. Most
sites with, say, several thousand miles, will stage a bandwidth upgrade
over two to three years. You can go faster, but you need that many more
people.
Part of the reason that it goes quickly is that you don’t have to change
any splitters, passives or taps. You will add in-line equalizers, to
equalize out to the plant. But you do it in such a way that deliberately
doesn’t require you to re-splice your way down the street.
In terms of maintenance windows, when you’re working on main stations,
or those that affect a lot of people, or those that carry voice – those
should be done at night. Then all the other smaller areas, served by
line extenders, can be done in daylight hours. The goal is to make sure
people aren’t off but for 20, 30 minutes at a time – less than an hour,
anyway.
How do you pick levels to do a drop-in upgrade?
Two things matter when you’re planning for bandwidth upgrades: The level
and tilt of the existing system. Let’s pick a set of numbers -- say, 46
dBmV and 750 MHz. If, on that, you had a 10 tilt to channel 2, you’d
have 36 dBmV at channel 2. So there’s your 10 tilt. When you upgrade,
you first extend that tilt upward. Going from 750 to 870 MHz is 2 dB
higher, and going from 750 to a Gig is 4 dB higher. So you hold the
existing levels and extend the tilt.
If the operator has line equalizers out there, you can stay with these
levels. If not, we elevate it 1 to 2 dB. So you take 750 and the channel
2 carrier up some, to let him add those line equalizers. The line
extenders have some insertion loss, but they put you right back to where
you are today.
We’re making the assumption that the existing 750 system works, and that
you’re happy with it. You’ve improved it by adding line equalizers.
You’ve extended the tilt, and it works. That’s the secret recipe of how
you pick levels to do a drop-in upgrade.
|
