Single chipset supports all major cellular standards

> However, a bare minimum level of air-interface stability, equivalent
> to at least one permanently deployed channel per air-interface at each
> cell/sector, must be maintained. Pilots & control channels simply
> cannot be ephemerally flitting in & out as the system dynamically
> optimizes its interface equilibrium. For the sake of idle mobiles &
> system acquisition, at least one CDMA pilot & sync pair plus one GSM
> BCCH plus one IS-136 TDMA DCCH must be statically deployed at a known
> invariable frequency at each cell/sector to serve as a beacon for
> mobiles.

I understand that the beacon/pilot channels will be hard on available
spectrum.
But the DCCH doesn't have to be static for TDMA.
A TDMA phone scans all the channels for a DCCH.
You could put the DCCH at any channel in the network and it will find it.
If the DCCH drops and moves, the phones will have to find it again.
Once the phone loses the signal, it will assume it has went out of rang
and will search for it's neighbors first then go into scan mode again.
It wouldn't require any change in tdma phone setting or firm ware.
That is normal operation

The cdma one channel is assumed.
I am not very familiar with GSM so I won't comment on it.
one thing I wonder is if there is any command in the TDMA, GSM
or CDMA architecture to forcefully tell the phone on the pilot channel to go
to another pilot channel.
That way you could start a new DCCH or BCCH and send the phones
to it from the old pilot channel and then close the old pilot channel.
That would eliminated the need for the phone to look for it again
and eliminate any dropped calls while they are doing so.

> For example, in a hypothetical 10 MHz block of spectrum, a coincident
> 1.25 MHz would have to be dedicated at every cell/sector for a single
> consistent CDMA carrier.
Yeap.

>For the FDMA interfaces, a frequency re-use
> pattern of N=4*3 would be assumed for the beacon "place-holder"
> channels. Thus, w/in each given re-use cluster, 2.4 MHz (200 KHz *
> 12) & 360 KHz (30 KHz * 12) must be reserved at every sector/cell for
> GSM & IS-136 TDMA control channels, respectively, for a total (CDMA +
> GSM + TDMA) aggregation of 4.01 MHz.

That is the main problem, a way to reduce the base line spectrum required.
Of course, that is one that any cellular/pcs provider wished they could
answer.

The TDMA control channels are relatively small at 360 or so Khz.
The CDMA channel is fine because it carries traffic.
But the GSM channels is a problem.
They would not be a problem during high usage, but during
low GSM usage, it would be nice if there was a way to compact the
pilot channels to a bare minimum.
One thing that comes to mind is to reduce the aggregation of
the cells for the channel spacing to reduce the number of channels needed.
(ie)
Transmit the same GSM pilot on all three sectors of a tower.
aggregate the channels on the towers to provide a total channel requirement
of (guestimates....)4 then you would reduce the GSM requirement down
to 800Khz, with each pilot channel covering three sectors.
You can't transmit the same pilot channel on two adjacent towers
because of the mixing effect.

Hmm.... something to think about.....:-)

And if a person had the A or B cell band and one or so PCS licenses,
then you would have plenty of workable space. :-)

---

› See More: Single chipset supports all major cellular standards

> Transmit the same GSM pilot on all three sectors of a tower.
> aggregate the channels on the towers to provide a total channel
requirement
> of (guestimates....)4 then you would reduce the GSM requirement down
> to 800Khz, with each pilot channel covering three sectors.
> You can't transmit the same pilot channel on two adjacent towers
> because of the mixing effect.

Duuu..... that is what the tower in Nashville already does.
It transmits the same TDMA DCCH and analog pilot on all three sectors.

So, if you did that at low tdma usage, you could get the
tdma usage down to 4 channels or 120 khz
plus 1.2meg for cdma and 800k for gsm would leave you
with 2.12Mhz of minimum usage.

On Sat, 27 Sep 2003 16:21:17 GMT, "N9WOS"
<[email protected]> wrote:

>> So, how does the system get the message down to the consumer's cell
>> phone that the PRL needs to be upgraded dynamically, in real time,
>> just before you make the call?
>
>It wouldn't have anything to do with the PRL.
>The system would just be listed in the PRL as a digital system
>with the proper system ID.
>The current flux of the system wouldn't affect the PRL.
>
>> Now, if the base station power amps are over-engineered to handle any
>> RF peak power (expensive)
>
>Cheaper that a bunch of small amps for each carrier.
>
>and passband bandwidths (expensive), and all
>
>Standard old broad band, no tune, monolithic amp technology.
>The monolithic amps in the old 3W analog cell phones can
>operate across the entire A and B bands with no tuning.
>And since they are class A, they can amplify multiple carriers
>at the same time in the pass band with no distortion.
>As long as the resultant additive peak power is less than the peak
>capability of the amp.

So, we're talking a massive retro-fit, and a stringent amp design so
that there's no intermod!

>> the antennae are tunable (expensive) so that they radiate the pattern
>> you want without a nasty SWR (expensive),
>
>The last thing you would want is an electrically steered antenna,
>or point tuned antenna.


Ah, but you were talking about using frequencies from multiple bands,
hence the need for complex antennae systems.

>You want the plain old panel antenna that you aim to
>get the radiation pattern you want.
>They already sell them that will operate across the entire
>transmit or receive portions of the A and B band with no
>tuning required by the installers.
>They have a useable SWR across the entire bandwidth.

I'm not talking about just A & B, I'm talking about non-800 MHz bands.

>> AND the mobiles are
>> frequency agile
>There is no more agility needed by that system than
>what is used today.
>as long as the phone can operate on any cellular channel
>or pcs channel, it is perfectly suited for the application.
>
>>and have the protocol to understand what the base
>> station is trying to tell them, then you've got a system!
>
>The base station will talk to them with their normal protocol.
>There is no need for you to have any new multimode protocol.
>Most phones don't support any cross cdma/gsm "or the like".
>so there is no need for a new protocol.
>the phone just thinks it's talking to a native gsm or cdma system.


So you're really talking about a "hobbled" system that only supports
specific services at specific frequency slots in the 800 MHz bands.
This will be terrible for serving subscribers using different
services, especially for a spectrum deficient carrier.

On Sat, 27 Sep 2003 16:41:22 GMT, Al Klein <[email protected]> wrote:

>On Sat, 27 Sep 2003 10:16:03 -0500, [email protected] posted in
>alt.cellular.verizon:
>
>>Now, if the base station power amps are over-engineered to handle any
>>RF peak power (expensive)
>
>No, just the peak power needed by the system that needs the highest
>peak power. Controlling it down from there is cheap.
>
>> and passband bandwidths (expensive)
>
>Again, only the bandwidth needed by the system needing the highest
>bandwidth. A wideband amplifier can amplify a narrow signal.
>
>>and all the antennae are tunable (expensive)
>
>We're either talking about an 800 MHz system or a 1900 MHz system,
>right? Antennas for either system cover the entire band.
>
>>so that they radiate the pattern you want without a nasty SWR (expensive)
>
>SWR and pattern aren't really related.

They can affect each other!

>>AND the mobiles are frequency agile
>
>Any cell phone can hop frequency.
The mobile should hop 800, 1900, 700, 450 Mhz, 1500, and 1700 MHz.
(I'm talking world-phone, here)

>>and have the protocol to understand what the base
>>station is trying to tell them
>
>Bingo! And that, supposedly, is the new chip set.

The new chipset is for the base.

I think for what I really want for the mobile, the eventual best
solution is software defined radio.

> Ah, but you were talking about using frequencies from multiple bands,
> hence the need for complex antennae systems.

Uuuu......no
I was talking about multiple frequencies from one large band.
Up to 70Mhz wide.

> >You want the plain old panel antenna that you aim to
> >get the radiation pattern you want.
> >They already sell them that will operate across the entire
> >transmit or receive portions of the A and B band with no
> >tuning required by the installers.
> >They have a useable SWR across the entire bandwidth.
>
> I'm not talking about just A & B, I'm talking about non-800 MHz bands.

I am talking about the A and B system AND the PCS bands.
They also sell PCS panel antennas that cover the entire PCS
spectrum with no tuning or setup required by the installer.

> So you're really talking about a "hobbled" system that only supports
> specific services at specific frequency slots in the 800 MHz bands.
> This will be terrible for serving subscribers using different
> services, especially for a spectrum deficient carrier.

Hobbling stuff together makes a great learning experience.
I have learned a lot from home brewing electronic stuff. :-)

> >Bingo! And that, supposedly, is the new chip set.
>
> The new chipset is for the base.
>
> I think for what I really want for the mobile, the eventual best
> solution is software defined radio.
>

Now that is pushing it.
The chipset they have for the base is a software defined radio.
But....
They can only implement it with a super conducting chip being
driven by gigabit Ethernet with multiple PC's producing carriers
to mix into the final broad band output to the antenna.
It takes about 1Ghz worth of computing power to produce
one GSM carrier.

When the day comes that you can have a super conducting chip
on you cell phone, running 12Ghz plus, then you are talking.

> So, if you did that at low tdma usage, you could get the
> tdma usage down to 4 channels or 120 khz
> plus 1.2meg for cdma and 800k for gsm would leave you
> with 2.12Mhz of minimum usage.

A big problem I see is when you throw WCDMA into the mix,
with it's 5Mhz bandwidth, everything goes out the window.

> So how much guard band are we talking about?
> 1.2500 - 1.2288 = 0.0212 MHz or 21.2 KHz
>
> That's the gap between two IS-95 CSMA carriers edges. Not enough for
> one AMPS channel (30 KHz, three ID-136 timeslots) or one GSM channel
> (200 KHz, eight timeslots). Between a band edge and a carrier edge
> would be half that, 10.6 KHz.
>
> There isn't enough room, unless we're willing to accept some
> interference.

The spreading sequence and data correction should tolerate
some fringe signals without loss of data.
And it will just show up as white noise to the TDMA carrier.
And it will be so much lower than the signal level
of the TDMA carrier that the TDMA carrier won't notice.

Technically, a CDMA signal should be able to tolerate a
narrow band signal or two within it's pasband without any problem.
Remember the whole ruckus in the ham community about
sharing ham bands with spread spectrum modes.

Having two tdma signals right in the middle of the 1.2meg pas band
should only raise the noise floor for the CDMA channel a few DB
and the tdma signals won't even notice a DB raise in noise floor.
CDMA gets that off it's military heritage.
It is it's one true distinct capability.
But that capability is never utilized with the cellular application.

My opinion is direct sequence SS TDMA would be a better interface tech
than the chipping method that cdma uses.
The code spreading sequence doesn't gain anything that you
can't get with other interlacing methods besides the
secure communications junk which is no longer secure when
the spreading sequences are public knowledge.
Go figure?????

"N9WOS" <[email protected]> wrote in message
news:[email protected]...
> > (ie)
> > 7Mhz aggregated into tdma carriers for one sector and aggregate
> > The rest into tdma channels for the rest of the system in the area.
>
> And there is nothing that says that the channel system has to be
contiguous.
>
> You know those big CDMA channels, and that guard band between them.
> That is many fine Khz of spectrum going to waste.
> I am sure you could plant one or two TDMA carriers in the CDMA
> guard band without causing any cross mode interference.
>
> Stick the narrow band channels in where ever you have guard bands
> or where you have any left over space.
> (ie) The last few Khz that is left over above the last CDMA channel
> that is too small to hold another CDMA channel.

So how much guard band are we talking about?
1.2500 - 1.2288 = 0.0212 MHz or 21.2 KHz

That's the gap between two IS-95 CSMA carriers edges. Not enough for
one AMPS channel (30 KHz, three ID-136 timeslots) or one GSM channel
(200 KHz, eight timeslots). Between a band edge and a carrier edge
would be half that, 10.6 KHz.

There isn't enough room, unless we're willing to accept some
interference.

If I have made a error, let me know gently.

John C. N4BKN

> My opinion is direct sequence SS TDMA would be a better interface tech
> than the chipping method that cdma uses.

Nick that..............
Just go with a non spread high bandwidth QPSK TDMA.
Who needs any spreading sequence. :-)

"N9WOS" <[email protected]> wrote in message news:<[email protected]>...
> > Transmit the same GSM pilot on all three sectors of a tower.
> > aggregate the channels on the towers to provide a total channel
> requirement
> > of (guestimates....)4 then you would reduce the GSM requirement down
> > to 800Khz, with each pilot channel covering three sectors.
> > You can't transmit the same pilot channel on two adjacent towers
> > because of the mixing effect.
>
> Duuu..... that is what the tower in Nashville already does.
> It transmits the same TDMA DCCH and analog pilot on all three sectors.
>
> So, if you did that at low tdma usage, you could get the
> tdma usage down to 4 channels or 120 khz
> plus 1.2meg for cdma and 800k for gsm would leave you
> with 2.12Mhz of minimum usage.

Yes, your train of thought seems promising.

By deploying the GSM BCCHs at a re-use of N=4 omnidirectional, the
dedicated GSM spectrum requirement is reduced from 2.4 MHz to only 800
KHz w/in the re-use cluster. Minimum capacity would be seven
full-rate timeslots per cell, as timeslot zero would be the static
BCCH. And the robust GMSK modulation should be able to manage any
decrease in link margin due to the tighter N=4 re-use, which provides
a consistent one cell diameter offset between like cells in adjacent
clusters.

As for IS-136 TDMA, retain the N=4*3 sectorized re-use, which can
create an additional half cell diameter separation between like cells
in adjacent clusters. Primarily, pi/4-OQPSK, requiring a greater link
margin than GSM, is not so immune to lowered C/I from tighter re-use.
And minimum TDMA capacity would only be two full-rate channels per
cell, as the timeslot 1+3 pair would be reserved for the DCCH. On the
other hand, if N=4*3 re-use were employed for IS-136, the capacity per
cell would be six full-rate timeslots, essentially equivalent to the
GSM capacity, while the spectrum requirement would still only be 360
KHz, again almost exactly comparable to the GSM spectrum outlay.

Andrew
--
Andrew Shepherd
[email protected]
[email protected]
http://www.ku.edu/home/cinema/

On Sun, 28 Sep 2003 17:31:26 GMT, "N9WOS"
<[email protected]> wrote:

>> >Bingo! And that, supposedly, is the new chip set.
>>
>> The new chipset is for the base.
>>
>> I think for what I really want for the mobile, the eventual best
>> solution is software defined radio.
>>
>
>Now that is pushing it.
>The chipset they have for the base is a software defined radio.
>But....
>They can only implement it with a super conducting chip being
>driven by gigabit Ethernet with multiple PC's producing carriers
>to mix into the final broad band output to the antenna.
>It takes about 1Ghz worth of computing power to produce
>one GSM carrier.
>
>When the day comes that you can have a super conducting chip
>on you cell phone, running 12Ghz plus, then you are talking.

I don't think it'll need to be a superconductor. It will, with current
technology, chew up alot of power, indeed.

Give it 5 years. Meanwhile, the GSM1x handsets are starting to come
out, and Qualcomm has some GSM/cdma2000 1x/cdma20001x EV-DO chipsets.
So, aside from the tuning, we're starting to "get there".

Stripline antenna, anyone?

On Sat, 27 Sep 2003 22:31:31 GMT, "N9WOS"
<[email protected]> wrote:

>> The computers can handle that.
>> That is the point of a dynamic system.
>> The computers can allocate space and
>> channels on a real time basses as the loading occurs.
>> There is no need for planning. :-)
>
>Hmmmmmmmmm.........
>That gives me an idea for a network forecast.
>Something that you could get from the 611 menu.
>
>......................................
>Welcome to the N9 dynamic cellular network.
>Current network weather for the day of September 27.
>
>12.5 megahertz currently in use.
>2.5 megahertz currently available.
>5 megahertz unallocated.
>
>Operating spectrum is as follows.
>3 CDMA carriers.
>152 TDMA carriers
>33 GSM carriers
>8 analog carriers.
>
>There has been 3 unsuccessful calls and
>21 dropped calls within the last hour.
>
>To go back to the main menu, hit 1,
>Or hit end to finish the call at any time.

Good thought!

>.....................................................
>Kinda makes me think of the solar forecast on WWV. :-)

The Boulder K Index at 03 UT was 2....

On Mon, 29 Sep 2003 00:57:21 GMT, "N9WOS"
<[email protected]> wrote:

>
>> So how much guard band are we talking about?
>> 1.2500 - 1.2288 = 0.0212 MHz or 21.2 KHz
>>
>> That's the gap between two IS-95 CSMA carriers edges. Not enough for
>> one AMPS channel (30 KHz, three ID-136 timeslots) or one GSM channel
>> (200 KHz, eight timeslots). Between a band edge and a carrier edge
>> would be half that, 10.6 KHz.
>>
>> There isn't enough room, unless we're willing to accept some
>> interference.
>
>The spreading sequence and data correction should tolerate
>some fringe signals without loss of data.
>And it will just show up as white noise to the TDMA carrier.
>And it will be so much lower than the signal level
>of the TDMA carrier that the TDMA carrier won't notice.
>
>Technically, a CDMA signal should be able to tolerate a
>narrow band signal or two within it's pasband without any problem.
>Remember the whole ruckus in the ham community about
>sharing ham bands with spread spectrum modes.
>
>Having two tdma signals right in the middle of the 1.2meg pas band
>should only raise the noise floor for the CDMA channel a few DB
>and the tdma signals won't even notice a DB raise in noise floor.
>CDMA gets that off it's military heritage.
>It is it's one true distinct capability.
>But that capability is never utilized with the cellular application.

The answer is that isn't quite that simple. Todays CDMA activities are
very toleratant of sharing space and rising noise floor , because the
chipping to data rate provides a processing gain on the order of 1000.
In other words a narrow interefering signal can only damage a tiny
portion of the data, because the data is spread so widely.

However when you go broad band applications that people are proposing
for use with
WCDMA, and instead of 10,000 bit per second channel, you want a 2
million bit per second channel, the gain from the chipping to data
rate is suddenly more like 2-3, and at that ratio, the system is a
whole lot less tolerant of raising the noise floor.

> The answer is that isn't quite that simple. Todays CDMA activities are
> very toleratant of sharing space and rising noise floor , because the
> chipping to data rate provides a processing gain on the order of 1000.
> In other words a narrow interefering signal can only damage a tiny
> portion of the data, because the data is spread so widely.
>
> However when you go broad band applications that people are proposing
> for use with
> WCDMA, and instead of 10,000 bit per second channel, you want a 2
> million bit per second channel, the gain from the chipping to data
> rate is suddenly more like 2-3, and at that ratio, the system is a
> whole lot less tolerant of raising the noise floor.

The bandwidth verses power is the primary factor.

A 10kbps signal is still 20Khz of actual bandwidth
no matter how much the CDMA spreading process spreads it.

A 100mw 10Kbps CDMA signal has the same range as
A 100mw 10kbps 20Khz wide signal.

If you take a 1Mbps (2Mhz wide) signal and spread it across 5Mhz of
spectrum.
You will need a lot higher power to spread across the larger 2Mhz
bandwidth.
If you use TDMA, you can cut the TX time down to around 1/1000 of a
second at a rate of around 10 times a second.
For a total TX time of 1/100 of a second.
Of course, the peak TX power during that time will be around 10W. :-)
But you will still have 100mw average.
Granted, it wouldn't be hand held friendly.

During low usage, one phone could pretty much
use the entire 1Mbps channel.
Who needs 3G? :-)

You could stack a couple carriers on the band
with the chipping sequence rotated a few notches.

And since the signal is still Spread spectrum like CDMA,
you can use rake receivers to add reflected signals
and signals from multiple towers.