The next portion of the prior art MU-MIMO timeline shown in FIG. 1 is the downlink portion including the downlink transmission and receipt acknowledgement phases “C” and “D” respectively of an MU-MIMO downlink over a representative interval, e.g. 90 milliseconds. The precoding matrix “Q” 118 for each tone or subchannel, a.k.a. the beamsteering matrices, are starting at time t2, used for spatially mapping the transmitted MU-MIMO packet to the stations 102A-B. The MU-MIMO packet 124 includes a header 124A and a payload portion. The spatial mapping of the payload portion of the MU-MIMO packet 124 allows the distinct block of data for station 102A and the distinct block of data for station 102B to arrive at their respective target stations without crosstalk or interference from the other block of data. The precoding applied by the spatial mapper to the transmitted payload produces the distinctly anisotropic RF signal strength multi-lobe pattern 120A-120B. At time t3 each station receiving the MU-MIMO packet 124 acknowledges receipt of the packet in an “ACK” packet. In this example where the MU-MIMO group comprises two stations 102A and 102B, each of which separately acknowledge receipt of the packet in ACK packets 132, 130 respectively. The WAP will resend any packet for which it has not received a receipt acknowledgement. Since slight changes in the respective communication channels are expected the soundings are frequently repeated, e.g. 100 millisecond intervals in other words ten times per second, with the precode matrix recalculated each time. Thus the overhead related to maintaining the MU-MIMO downlink is 10% or 10 milliseconds out of every 100 millisecond interval.
