To: Asterisk who wrote (12403 ) 7/15/1998 5:57:00 PM From: Clarksterh Read Replies (1) | Respond to of 152472
Michael - Although you brought up some points I hadn't considered, you are making at least one big mistake, and, I suspect, some smaller ones: 1) Lets take a specific example. If we hold both powers at max AMPS transmit (28dbm) then the TDMA pulsed transmit would take less power. If your PA is only on for 1 out of 10 time slots then you would take max 1/8th the power right? No. The reason is that you are making the assumption that they both require the same raw power out, when they do not. They require the same processed Eb/N. Thus, for instance, GSM packs all of their info into one time slot with the result that their bit rate is (assuming a duty cycle of 1/8) 8 times higher. Therefore, for that time, they need 8 times the power. Broadly speaking, you don't save power by going digital (no matter what format) except insofar as you can decrease the transmission of useless data. CDMAOne happens to do this better than GSM. 2) If you have a slotted system then you can afford 2 things, 1) more noise, and 2)less isolation. In a CDMA system you have to have some type of filtering to seperate the transmission of Recieve and Transmit signals. This filtering costs tons in the way of transmit power and power in general. In GSM you have no need for this filtering, your time slot is yours, either recieve or transmit whatever you want. Well, I agree this is a good point, but I suspect it isn't too significant. The send and receive are separated by 45 MHz, which is a lot for for a channel only 1.25 MHz wide. The filter should easily be down well in excess of 100dB. However, I'll admit I'm not a filter designer. 3)The corollary to this is that when you have your own slot you need not worry about the noise you are transmitting inside your band. Noone else is affected so who cares? With CDMA (a noise limited system) ANY noise that you add to the system that is not directly associated to your communication will lower the capacity of the entire system. This never occurred to me, but again I suspect the effect isn't that large because in GSM the transmit noise floor is a much bigger factor in the received S/N (nobody else is transmitting), whereas in CDMA the noise floor created by the intentional transmissions of other users is 30dB higher than the noise transmitted by the broadcaster (assuming a transmit S/N or 40 and 10 handsets in the cell). Thus, even with 10 users in a band of a cell, the noise floor transmissions are only 1/1000 of the 'noise' caused by the other users intentional signals. Clark