- This topic has 11 replies, 1 voice, and was last updated 11 years ago by SHAMEER.
28th January 2008 at 12:19 #50655MKTGuest
this is just for discussion sake.
In a EDGE n/w we use 8-PSK for modulation. This gives us 3 times higher data rates as compared to GSM.
If we try 8-PSK for speech also then
we can have a capacity gain of 3 times also.
But we are not doing so. Any reasons.28th January 2008 at 14:32 #50656pixGuest
it can be done, it is defined in the 3GPP standards. However it is not implemented by vendors YET, because it is not really useful.
In the future, you will see the AMR WIDEBAND introduction. With this technology, you can expect to see half rate channels carrying 8PSK voice, with more than 12 kb/s codec.29th January 2008 at 04:40 #50657MKTGuest
“However it is not implemented by vendors YET, because it is not really useful”.
I can understand that it may not be useful for vendors.But I think it operators will love to have such feature.Why?…..
Let us say that if vendor implements any feature which is going to benefit operator in such a way that operator can reduce the
purchase of hardware then……this picture will be good for operator and bad for vendor.
A BTS site with GMSK = 10 TCH
A BTS site with 8-PSK= 30 TCH
Thus We need three BTS sites with GMSK and one BTS site with 8-PSK to offer same number of traffic.
I think vendors do have some influence over the 3GPP or like.
MKT.29th January 2008 at 09:23 #50658TNSGuest
“…I can understand that it may not be useful for vendors.But I think it operators will love to have such feature.Why?…..”
Is this true? My opinion:
If we want to transmit, lets say 40kbit/s ( 3x13kbit/s voice ), we will need 6 Abis TS per TRX and at least MCS7.
To transmit 40 kbit/s using MCS7 we need at least C/I=20dB; MCS8->22dB; MCS9->23dB.
Do you still think it’s good idea from operator point of view?
I think the vendors of transmission and BSS equipment will love it…
I will mention only that retransmission of bad frames is not possible so we will need more then 40kbit/s for 3 frame of coded speach.
So… what do you think?
TNS29th January 2008 at 09:38 #50659pixGuest
one timeslot could carry 2 calls max, in HR. Not 3
So 10 TS in FR or 20 TS in HR. On the half-channel, one call could be modulated in 8PSK.
The feature is interesting and is going to be available from vendors soon. It’s just that it’s a new feature and quite complex to implement (new codecs, new transcoder boards, need new phones on the market, etc)
I think it is R6 feature, isn’t it?
We’re going as fast as we can you know…
pix30th January 2008 at 09:30 #50660MKTGuest
Pix and TNS…….thanks for reply.
It is a coincident or what that a thought comes into my mind. And i put it for a discussion and then i came to know it is there already. I mean 3GPP or like already thought the same way as i did. But ironically they did it first.
This time it is 8-PSK, earlier it was diversity at MS and may be some more cases which i may be going to discuss soon.
From the day one (in GSM) i was of the view that sooner or later they have to give up with TDMA. I have seen that TDMA and Population growth can never ever match their pace. And see they are now moving towards WCDMA. Anyway that is a different story.
Back to 8-PSK:
A Burst in GSM contains 156.25 bits. These bits are coming at a speed of 22.8 Kbps from behind. Now with GMSK i will need exactly the same time of RF carrier as that of burst to send 157 bits. but with 8-psk i will need 1/3 time. Hence 4.615 msec can now have 24 FR TCH.
regards.30th January 2008 at 09:39 #50661PixGuest
MKT, too bad you’re not in 3GPP, you’d make them think faster 🙂
“Back to 8-PSK:
A Burst in GSM contains 156.25 bits. These bits are coming at a speed of 22.8 Kbps from behind. Now with GMSK i will need exactly the same time of RF carrier as that of burst to send 157 bits. but with 8-psk i will need 1/3 time. Hence 4.615 msec can now have 24 FR TCH.”
As far as i know, it will not be done this way.2nd February 2008 at 22:14 #50662TNSGuest
Please answer me:
Do you realy prefer to have 24 FR TCH at C/I = 20-23 dB (8-PSK) instead of 8 FR TCH at 9 dB (GMSK)?
1G was substituted by 2G, which is now substituted slowly by 3G, which will be substituted by 4G and etc…
At 2G we have mainly two resources:TCH and SDCCH.
At 3G we have: spreading factor, channel element and Tx power.
The picture is so complex that no one can compute the capacity in fact…
With 2G we have 8 trx/sector = 60 TCH FR.
With 3G 65-70 voice connections (theoreticly) and only up to 70% of them “on air”.
So WCDMA is not panacea for the ” …I have seen that TDMA and Population growth can never ever match their pace. And see they are now moving towards WCDMA …” problem.
TNS2nd February 2010 at 09:41 #50663bjushkaGuest
we are implementing EDGE in our sistem, but if this feature is activated in Half Rate TRXs, EDGE doesn’t work. How can this happen? Aren’t calls in HR timeslots modulated in 8-PSK??
Bjushka3rd February 2010 at 04:32 #50664PixGuest
No, the HR calls are modulated in GMSK @ 6kb/s approx.
a HR call uses only 1 out of 2 bursts. In other words, it is transmitting only half the time.
HR calls or FR calls in 8PSK is a FUTURE improvement, that is not yet implemented in the GSM networks. The goal would be to keep a high throughput (12kb/s) but be able to place 3 “FR” calls in the timeslot. But let’s wait & see how it is really done.
pix22nd February 2010 at 16:04 #50665bjushkaGuest
Thanks pix, you were very helpful. 🙂
bjushka26th February 2010 at 16:22 #50666SHAMEERGuest
plz help me with the details of RBLT in between BTS and BSC