Electronic – Crossover design

The LM386 amplifier's inputs are an internally biased differential pair, with common mode input range extending to the indicated ground, V_ee, or actually 0.4 Volts below ground.

In order to not disturb this internal biasing, the input signal must be capacitively coupled (or perfectly ground-referenced) - a capacitor in series on the input line is needed.

Next, for a voltage gain of 200, the input signal peak to peak needs to be under 1/200 of the output voltage range of the amp. While I could not find this specified in a quick scan of the datasheet, if we assume 7 volts output voltage range with a 9 Volt supply, the input signal needs to be under 35 mV peak to peak, to avoid clipping the signal. Clipping would result in mild to severe distortion of the output - though this does not explain the complete lack of signal you find on output.

If the incoming signal is higher than this 35 mV P-P, a potentiometer as an attenuator on the input is suggested.

Once these fixes are done, please revert with results, so the answer can be added to if needed.

1. You don't want to send a single packet for every sample. That's because the packet has headers as well, and that means inefficiency in the transfer. Putting several samples into a single packet will both increate the efficiency and get you over your minimum packet size. The upper limit on the number of samples is the maximum latency that you can tolerate. At 48kHz, adding one more sample into your packets would increase the latency by ~21\$\mu\$s. You should also calculate the latency introduced by the whole system end-to-end. Having 8-16 samples per packet would increase your efficiency above 70-80%.

2. I'd go for a microcontroller with a built-in Ethernet MAC. Having a built-in PHY as well would be a nice plus, however, that extra constraint would decrease your choice by large. There many-many microcontrollers with a built-in 100Mbps MAC, but only a few with a built-in PHY as well.