Hello, loginx, you wrote: L> the Time delay is and when at me on the channel of 30 networks of different capacity (19 hours) and when I one (4 o'clock in the morning) At 4 o'clock in the morning they are ungeared physically? They generally are not present on air or you think that time people sleeps, networks do not occupy frequency? And different independent devices do not swing, do not wake up, updates do not swing? L> clearing up of the reasons was led by testing different vaj-faj routers in the same conditions. L> outputs - all NEW i.e. approximately since 2014 vaj-faj routers import a time delay, and even purposely L> lose packets if to hollow them 50 times in 1 second That such to "hollow"? ? and any icmp the traffic has the least priority, the processor for answers to it is given on residual to a principle. It generally can not answer on , and it is normal. L> for games without . Simply switch the new router in a mode "only 802.11g". L> OUTPUTS: Experiment is delivered not correctly, the theoretical basis is not brought, outputs are erratic. New routers by default work in 802.11n. In this mode, for capacity growth (those "300 ") it is used MIMO aggregation of channels, simultaneously goes transmission and reception on two frequency channels (still there is MIMO to magnification of an amount of antennas, but it for another). Accordingly and noises from "neighbors" are caught not on one channel in a band of 20 Mgts, and in two, in a band 40 (but, see a picture more low). Each badly accepted packet . As, in new wifi chips sensitivity is refined. If earlier, to old chips the signal-65 actually was not audible, the new hear and to hold more or less a stable relation even at-80. From this follows that they are forced to process and noises with level-85, not only on the but also on adjacent channels. Working on channels 1+6, to process noises (and the useful signal) on 1,2,5, 6,7 channels. Taking into account more broad band and the best sensitivity, in the conditions of a typical city (when each coffee maker creates ), noises it is caught essentially more is more often (than in 802.11g), arrive beaten packets, , the general ACTUAL throughput is more often is necessary more low. Switched in 802.11g you you narrow down working frequency and you lift a threshold of cutting off of noises. That not to make including 802.11n on operation in one frequency channel, there there is such possibility for struggle against an amount of noises, but all the same it is necessary to process low-level noises of adjacent channels (5, 6,7 for example).