I am new to the field of MIMO, but we are working on a new technique to decouple multiple antennas that are in close proximity. As I understand it, ECC=0 means the antennas are completely decoupled, while ECC=1 means they are essentially short circuited.
What values of ECC are typical for antennas in cell phones today? What value of ECC would be considered poor, good, or great? I can find some numbers from the literature, but I am not confident those are representative of actual cell phones in use today. I am just trying to get a feel for the numbers and for the state-of-the-art.
Thank you!
Hi Raymond:
One thing to remember about ECC is that it's not JUST an antenna related quantity. There's an assumption of the field distribution of the local channel which will have an impact. Some antennas may have a high correlation in one environment and a low correlation in another. For example, a line-of-sight, plane wave illumination of a device would show correlated antennas regardless of what you do, since the two antennas are receiving the exact same plane wave, altered only by the magnitude and phase difference of the patterns in the line-of-sight direction. There's nothing to decorrelate the signals. Most often, however, ECC is just evaluated for a uniform isotropic field; something that never happens in the real world!
That said, a correlation below about 0.3-0.4 is usually considered "good enough" for MIMO. A truly uncorrelated channel is highly unlikely for any real propagation condition. That's essentially equivalent to a cabled connection to the radio.
If you're interested, here's a link to an article I published in IEEE that talks about a set of "good", "nominal", and "bad" antennas designed specifically for evaluating MIMO test methodologies. Since then we've learned that real devices can potentially have antennas better than the good and much, much worse than the bad!
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=6397087&url=http%3A%2F%2Fieeexplore.ieee.org%2Fiel5%2F5962381%2F6397038%2F06397087.pdf%3Farnumber%3D6397087
Hi Raymond:
One thing to remember about ECC is that it's not JUST an antenna related quantity. There's an assumption of the field distribution of the local channel which will have an impact. Some antennas may have a high correlation in one environment and a low correlation in another. For example, a line-of-sight, plane wave illumination of a device would show correlated antennas regardless of what you do, since the two antennas are receiving the exact same plane wave, altered only by the magnitude and phase difference of the patterns in the line-of-sight direction. There's nothing to decorrelate the signals. Most often, however, ECC is just evaluated for a uniform isotropic field; something that never happens in the real world!
That said, a correlation below about 0.3-0.4 is usually considered "good enough" for MIMO. A truly uncorrelated channel is highly unlikely for any real propagation condition. That's essentially equivalent to a cabled connection to the radio.
If you're interested, here's a link to an article I published in IEEE that talks about a set of "good", "nominal", and "bad" antennas designed specifically for evaluating MIMO test methodologies. Since then we've learned that real devices can potentially have antennas better than the good and much, much worse than the bad!
http://ieeexplore.ieee.org/xpl/login.jsp?tp=&arnumber=6397087&url=http%3A%2F%2Fieeexplore.ieee.org%2Fiel5%2F5962381%2F6397038%2F06397087.pdf%3Farnumber%3D6397087
This is great and thank you for sharing the paper! It will be very helpful.
We made a mockup mobile phone with two antennas about 3 mm apart operating at around 750 MHz. We were able to change the ECC from 0.7 to below 0.02. Our HFSS simulations and lab measurements are in very good agreement so I think we are on to something. Our biggest problem is that I am a beginner in the MIMO world that that is slowing us down.
I really appreciate your input!
@Raymond how did you calculate the ECC in lab. If you have used S parameter approximation (as proposed by Smith et al.) it is not very acceptable in real environment as explained by Michael. Since the approximation from S parameters is based on isotropic assumption in a rich scattering environment. The best way to calculate ECC is by using far field parameters.
- Badges
- Science topic