Hi all!

I'm looking for 'good' and detailed reading material on : "multi-mode, multi-band base station transmitter design for GSM, 3G and LTE" - books, tutorials, app. notes, design guides, theses etc.

There are a few companies out there that work in this area, but hardly any that share info (incl. ones like Agilent!).

I wasn't sure where to post this question, so I put it here thinking it might be most relevant to this subject-topic.

B.t.w., IEEE Xplore and Google searches didn't help significantly, so any solid info and/or suggestions would be much appreciated!

Thanks in advance!

Cheers!

Hi pancho_hideboo,

I'm trying understand the basics of transmit RF-front end architectures (incl. the modulation aspect of it), for mobile comm. base stations, multi-mode and multi-band ... ones that support several mobile communications in parallel - this, because I'm studying them in comparison...

So,.. system architecture of the RF - TX  but at the same time, a focus on hardware power consumption as well!

If you have info on how to implement these requirements, that would also be very welcome!  :)

DaveB wrote on Mar 20^th, 2009, 5:04pm:

You should also look at Papers on Polar Modulator.

Based on the analysis I've done, I've come to the conclusion that polar modulators are difficult, if not impossible, to implements for wide bandwidth signals (signal bandwidths greater than a MHz or so). two things come into play here, One is the difficulty of generating phase/frequency modulation of the required bandwidth, the other is the difficulty of achieving the required Amplitude/phase matching. You need to be able to align the signals to within a small fraction (think 1/20 of a symbol) to meet mask performance. If you are looking at multimode designs, you might be better off looking at a DDS source, as current consumption might not be such an issue. Galton at UCSD has been doing some interesting work on Delta Sigma RF transmitters, which might be interesting to look at if you can do IC design.


I'm not familiar with any published books that go into an detail on Polar transmitters. There are a few PhD dissertations I've come across that discuss Polar transmitters, but nothing I currently have access to.

Dave

Polar modulation works OK for narrower bandwidths, like GSM and EDGE. The advantage there is that the out of band emissions of a polar modulator are somewhat lower than an I/Q modulator, allowing you to meet the 20 MHz offset noise requirements without a filter in the transmit lineup. Take a look at the JSSCC for 2004 and 2005, there were several polar transmitter papers published that go into the details.

Dave

ACPR and EVM are both system level performance requirements. You'll need to meet those regardless of whether you are using Quadrature/Polar/DDS transmitter designs.

Dave

My opinion is that pure polar modulation is not good for multi-standard multi-mode. However, envelope tracking (UCSD) offers some possibilities that pure polar cannot.

It is more difficult to meet ACLR specifications for WCDMA and EVM for OFDM using polar.
Why? Because in pure polar even static errors in the system yield nonlinearity in the output signal.
Why? Because polar only works perfectly when two nonlinear functions are multiplied perfectly (amplitude and phase). Any error in the alignment of amplitude and phase generates nonlinear distortion.

Envelope tracking offers more flexibility since the amplitude modulator does not need to track the amplitude modulation perfectly, but there is a slight reduction in efficiency for this offset.

A predistortion system is needed to overcome the fine alignment errors even in envelope tracking systems.

The next problem is bandwidth. The amplitude and phase signals are nonlinear functions of the I/Q modulation, so each takes significantly more bandwidth than the original signal, so the dataconverters must be higher speed and the modulator bandwidths must be 2-3x what would be necesary for I/Q modulation.
This is an obvious problem for wideband OFDM signals.

But the benefit is potentially significantly better overall power efficiency which is beneficial for basestations even if there is some additional cost and complexity.

DaveB wrote on Mar 25^th, 2009, 7:35pm:
I don't think so.
In I/Q modulator, we have to have PA after I/Q mod. This efficiency of this PA can not so good even if we use predistortion.
Maybe you assume post PA after Polar Modulation System. Right ?

My opinions are :
- Polar Modulation System is not suited to Multi-Band
- Polar Modulation System is not suited to achieve high ACPR(ACLR)
- There could be a possibility to achieve good power added efficiency in Polar Modulation System
　　if compared to I/Q Modulator + Post-PA with Predistortion which is currently widely adopted.