Linear Power Supply for Network Findings
Interestingly enough, removing the LPS supporting a modem and switch have greatly improved my network performance.
I’m running 1000 Mbps service, and use an Arris 8200SB modem that feeds the Eeros brand mesh network that includes a wireless router. The router connects the switch with RJ45 and wirelessly supports 4 mesh beacon antennas for greater home WiFi coverage.
Despite the service speed at the modem, the wireless network would bog down enraging a teenage daughter, and because of this it forced me to look into something I felt I did not need to.
I then ran tests using the Trifield EMI detector into the AudioQuest distribution used for network gear, an integrated receiver and disc player. Tests reported TriField readings in the 800/900’s. Then I pulled the LPS, and now the readings are between 42 - 46, that’s a huge drop! Note: two LPS were used and the results were essentially the same.
I speed tested using the application Speedtest, the change was drastic, wireless speed went from about 80 Mbps north to over + 400 Mbps. Then comes the sound quality, it significantly improved! As good as the Taiko Extreme is, the EMI was distorting the signal prior to the USB/Fiber output.
Whatever the reason, sonically I’m ahead. In my hypothesis the LPS was rippling current causing fluctuations in the modem power supply resulting in swings in speed, and the Transformer was creating large amounts of EMI that was riding on the RJ45 copper into the switch and subsequently into the server.
The setup looks like this: Modem w/ Audioquest Diamond > Router w/ Blue Jeans RJ45 > Intona w/ AQ Diamond > Taiko Extreme w/Intona Ultimate USB > MSB ProUSB w/fiber > MSB DAC. Power for the modem and switch are now OEM SMPS.
Note: The Blue Jeans is an unshielded CAT6 RJ45 cable, the idea is to break the noise carrying ground brought on via the modem and router to the switch. I also circuit ground the modem and switch to the panel.