What if slew rate and clock speed could be controlled and only run at max when the speed is really needed and no interference issues are present? Only run as fast and steep as needed for device service needded. Also can a better USB cable help this?
Careful of USB3.0 RF noise.
Intel released a white paper in April this year titled USB 3.0* Radio Frequency
Interference Impact on 2.4 GHz
Wireless Devices. This outlines the problems that can be created for WiFi through the use of USB3.0. These warnings also apply to other wireless (Cellular) frequencies. If we are not careful we will kill the radio pipe in the name of widening the wired one
In terms of bandwidth, USB3.0 is limited when compared to the PCIe and its upcoming mobile version(s). In terms of power consumption, USB3.0 is not as capable as MIPI-HSI. In terms of protocol overhead, USB3.0 is heavier than PCIe3.0 and beyond. For any SoC or platform to include the various flavors of 802.11, going through USB3.0 simply adds another layer of latency and dies size for NO gain at all. In particular, the argument of LTE is wrong because, as a network function, IP packetization is intrinsic, what a waste if protocol translation like IP-USB--connection--USB-IP is added! USB1.x met its objective, for all low cost peripherals. USB2.0 serves certain applications well but add the unnecessary cost. USB3.0 adds yet another cost burden and now also makes it difficult for further integration. The latter is because nobody dares to take the route to implement the pure LVDS electrical and drop the backward compatibility. Wake up, don't keep on putting bandage on a tired interface because this aligns with the business line of the corporates as well as personal performance within the company. There are other serial interfaces which have headrooms, protocol benefit than USB3.0 for the intended applications!