# Communication and DSP

Make it to the Right and Larger Audience

### Coherence interval = coherence bandwidth * coherence time?

Resolved Coherence interval = coherence bandwidth * coherence time?
2 Voices |4 Posts |300+ | Discussion Rooom: Communication

This topic contains 3 replies, has 2 voices, and was last updated by  ARao 4 years, 4 months ago.

• Author
Posts
• The massive MIMO tutorial says:

“assume that the coherence interval is T = 200 symbols (corresponding to a coherence bandwidth of 200 kHz and a coherence time of 1 ms)”

Not sure why so? I understand coherence bandwidth comes from delay spread in frequency selective fading and coherence time comes from Doppler shift/freq in time selective fading.

But what is coherence interval?

• Nobody?

Wiki says coherence bandwidth is 1/D and D is channel delay spread. Coherence time is 0.423/fm and fm is Doppler frequency shift and equals to vf/c where v is moving speed, f is carrier freq, and C is the light speed constant. So how come coherent bandwidth * coherence time becomes # of symbols during which channel is static? Symbol rate is not in the picture at all?

• jmong
Engineer

ARao, good question and here is my understanding. As you said, coherence time comes from Doppler shift and it tells during this time the channel is static. Coherence bandwidth comes from multiple path effect (delay spread) and it says if the symbol rate is not higher than this the channel fading is a flat fading rather than a frequency selective fading.

Back to your question of coherence interval =coherence time * coherence bandwidth, I think what it means is during the coherence time of 1ms and if the symbol rate is just 200KHz (same as coherence bandwidth), the 200 symbols (1ms * 200KHz) don’t suffer time selective fading (due to Doppler shift) and frequency selective fading (due to multipath delay spread)

• Thanks jmong. I think you are right.

You must be logged in to reply to this topic.