« Opera Mini - Still My Preferred Browser | Main | Cheap 2.6 GHz Licenses in European Nordic Countries »


Topper Harley

Thank you very much for those calculations. I will use them many times as examples at work!
Always a pleasure reading you, keep it up! :-)

A radio tech

You folks have it wrong DMA techs have more bandwidth than the GSM standards. In fact GSM's 3G is a CDMA variant. Public safety will be moving to P25 phase 2 which is a TDMA system....and or something close to AMPS...for you elder cell folks. GSM, is an aging form of bandwidth seperation, and or narrowbanding.


Any mention about LTE's terminal power consumption? I guess MIMO and similar tricks are taken in account, not easy and cheap to be implemented.

Given current business situation, perhaps we can forget about AMR and the like: how many Skype calls per LTE MHz? :-)


Hi Martin,

When they evaluate GSM capacity for voice, what kind of features they are considering? Is OSC taken into account?

Best regards,


Hi David,

I think you are probably referring to the MUROS studies of 3GPP. That's indeed interesting and here are some more details: http://tinyurl.com/yauexdo

As far as I can tell, this was not looked at in the book as it wasnt in the standards when they wrote the book. For GSM, they've looked at other potential improvements such as SAIC, DARP and DFCA which brings again some improvement over GSM AMR but nothing compared to the HSPA and LTE values.

I wonder where MUROS went in the meantime!? Anyone?


David Boettger

After 8 or so years in the field, cdma2000 1x capacity is agreed to be around 36 calls/MHz/cell. Comparison of cell capacity numbers is notoriously difficult because the results are critically dependent on assumptions. However, even with a significant margin allowed for differing assumptions, the numbers given by Homa and Toskala indicate that WCDMA and even HSPA are well off the pace in terms of the state-of-the-art in voice capacity. A couple of reasons for this are that WCDMA, in particular, is rather inefficient at the physical layer, and because it is code-limited with AMR12.2. (12 calls/MHz/cell is actually a bit optimistic, in my opinion.) Even if LTE does indeed yield up to 80 calls/MHz/cell, that's "only" an improvement of 2x or so from what can be done today.

Capacity calculations for N=1 systems aren't really "simpler" than those for systems that utilize frequency reuse; they're just differently complex.

The comments to this entry are closed.

My Photo

The Books to this Blog

Secure Hotel Wi-Fi Sharing

My Pictures on Flickr

  • www.flickr.com
    martin.sauter's photos More of martin.sauter's photos


  • Clicky
    Clicky Web Analytics