Dear Infinity,
based on my search over internet and other documents; for TDMA factor GSM cell cannot exceed 35 km. Because it cannot exceed 35 KM, thats why TA value is fixed for 63. please visit   "http://www.mobileworld.org/gsm_faq_04.html" .
here u can see little explain.

So, the MAIN REASON is "TDMA".  if we use 4 time slots ( Australia operaor used this as trial ) the cell radius can be extended to 70 ( notice that  ---double of 35 km)
so, depending on Time Slot numbers in TDMA frame cell radius can be changing.

what u said      * the power of the cell,  * the conditions (terrain around the cell, weather, etc.),* the height of the cell (as well as its directionality)----------------  these factoes are used for "practical cell planning"-----------even by using all these resources in maximum limit u cannot exceed ur cell over  35 KM which is a GSM theorotical standard.

>> Because it cannot exceed 35 KM, thats why TA value is fixed for 63.
The TA value was fixed to 63 as a design decision. They could have decided to make it larger or smaller, but they decided on 63.
TDMA is indeed the system used in GSM to synchronize between the cells and the mobile phones. But, it's only a system - it doesn't define what TA should be used. That's decided by the GSM specs.

The reason it was decided to be 63, is because they had to find a balance between the number of cell phones that can be served, and the distance they can be from the cell. You can serve more cell phones by decreasing the radius (which is why in cities, there are usually more small cells, and in the countryside more larger cells).
63 is a nice round value (6 bits). The people that wrote the specs probably decided that 17km was a bit too small (if the TA would be 5 bits or max. 31), and that having a TA of 7 bits (max. 127) would limit the number of phones that can be served too much.

>> even by using all these resources in maximum limit u cannot exceed ur cell over  35 KM which is a GSM theorotical standard.
The radius of a cell CAN exceed 35km, but it doesn't because the cell has to follow the GSM standard - otherwise it would be unusable. The theoretical limit in the GSM standard is just that : a THEORETICAL limit. If the power of a cell is increased it can easily reach a lot further than 35km, but you'd have to change the timing/synchronization too. So, you'd have to change the standard for that.
A standard is there to standardize things - in this case, the TA is standardized to max. 63, which imposes a limit on the cell radius of 35km.

my present  understanding, u can change "35 km"  radius only if u change Time Slot numbers in a TDMA frame.
from ur base, by changing what particular parameter u can change this "35 km" ?

>>The radius of a cell CAN exceed 35km, but it doesn't because the cell has to follow the GSM standard - otherwise it would be unusable.
----------------------I am partially agree with u on this. u r right that, only if u break GSM standard then u can go beyond 35 KM..................i 'm not agree on second part because australia opearots already broke the GSM standard by implementing 4 time slots..and as far i heard the network working normal with all kinds of mobile sets.


>>If the power of a cell is increased it can easily reach a lot further than 35km, but you'd have to change the timing/synchronization too.
----------------I am totally agrre with u on this.  .......but my adding question would be " by what u gonna implement timing/synchronization " ?  isn't it by channels? ...........and what is physical entity of a channel?  isn't it Time Slot ?

>> the TA is standardized to max. 63, which imposes a limit on the cell radius of 35km.
----------------i 'm 100% disagree on it. TA is not so imporatant paramter. If i stay in a buildiung ( and BSS also in this building) I can  even disable this TA function from BSS, and i'll still be able to talking normally.

>> from ur base, by changing what particular parameter u can change this "35 km" ?

By changing the max. TA. The thing is that if you change the TA, you also change the number of timeslots. A larger TA for example, would make the timeslots longer, and thus you can have less timeslots for the same amount of time.


>> i 'm not agree on second part because australia opearots already broke the GSM standard by implementing 4 time slots..and as far i heard the network working normal with all kinds of mobile sets.

They most likely made use of the extended radius I mentioned earlier. That's also specified in the GSM standard, and is thus also supported by a lot (if not most) of the phones.


>> by what u gonna implement timing/synchronization " ?  isn't it by channels? ...........and what is physical entity of a channel?  isn't it Time Slot ?

Correct ... you'd also use TDMA for the synchronization, just with different parameters. For example, you'd have to increase the TA.


>> i 'm 100% disagree on it. TA is not so imporatant paramter. If i stay in a buildiung ( and BSS also in this building) I can  even disable this TA function from BSS, and i'll still be able to talking normally.

You're talking about indoor cells now ... I doubt that your building will be 35km long, so they're usually very limited in coverage. I would have to look that up, but I imagine that the synchronization happens slightly differently with indoor cells.

I think one of the big reasons not yet mentioned for the 35km limit is that the speed of light becomes an important factor in the precision timing required by the channel-sharing.

35000m / 2e8 m/s = 175 microseconds

>> 35000m / 2e8 m/s = 175 microseconds
35000m / 3e8 m/s = 117 microseconds

And with the extended radius :

70000m / 3e8 m/s = 234 microseconds

That's still do-able.

As I said earlier, increasing the radius of an antenna, will decrease the number of calls it can treat. That's a choice that has to be made (and the choice they did make in the GSM specs by choosing a max. TA of 63).

But you made a very valid point, NovaDenizen : increasing the radius of a cell will of course increase the delays too (simply because the timeslots will be larger, and the distance to bridge is bigger).

2E8, 3E8, what's the difference? :)

Not much - especially not since the exact speed of the signal will be somewhere in between. But even when using 2e8, 350 microseconds for the extended radius is admittedly a bit on the high side, but it should still be ok for the enduser.

Dear administrator ,  as no comment is added to show the "CALCULATION" i'd like to close the question ...

Euhm ... I've shown you the calculation ... what's not clear about it ?

dear inifinity,
i dont accpet your method that cell radius is based on TA.
even if u r true,
you failed to show any CALCULATION of why TA=63 .

>> you failed to show any CALCULATION of why TA=63 .

There IS no calculation. It is an arbitrary choice balancing the number of users that can be served by a cell, and the radius of the cell.
The reason it's 63, and not 50 for example, is that 63 is the highest 6bit value (so TA can have a value between 0 and 63).

If a 5bit value would have been used, then 31 would be the max. TA, meaning a max. radius of :

    31 x 0.55 = 17.05 km

which most likely was decided to be too small by the people that wrote the specifications.

If a 7bit value would have been used, then 127 would be the max. TA, meaning a max. radius of :

    127 x 0.55 = 69.85 km

which limits the number of users that can be served at the same time too much.



It's a CHOICE they made, not a CALCULATION.

Please re-read my posts - I've already said that.


>> i dont accpet your method that cell radius is based on TA.

Then prove me wrong, or find someone who can prove me wrong ...

i'm trying to find it, when i get.. i'll post it here . till then, many thanks .