USB ground loop isolator

[phofman]

Not only didn't help with my mic's 1kHz noise but increased it from -75 to -66 dB

IMHO the 1kHz noise is caused by the 1ms USB fullspeed packets interval, not by a ground loop which the isolator is intended to break. The 1kHz noise may be improved by either a better power supply filter in the USB device, or perhaps a better generation of the ADC master clock from the USB clock which would have a heavy 1kHz USB packet component.

 

[taisho]

Sorry for the 1000th post on ground loops, but after looking at how expensive regular USB isolators are, would something like this cheap aliexpress isolator work?

Today I tested it on UMIK-1, and although it didn't raise 1kHz like it did with my standard mic, it did not affect the -80dB peak at 50 Hz at all.
Edit, explanation in my next post

 

[phofman]

The isolator eliminates ground loops. There is no opportunity for a ground loop with UMIK-1, no need to use the isolator. Those 50Hz get most likely into the weak analog signal from the microphone, before AD conversion. The isolator has no effect on the digital signal passed.

 

[taisho]

Today I tested it on UMIK-1, and although it didn't raise 1kHz like it did with my standard mic, it did not affect the -80dB peak at 50 Hz at all.

I have an update, this -80 dB was the best-case scenario for my UMIK-1, and for this relatively good value, the isolator made no difference. After testing multiple USB ports, with different devices plugged in configurations, the default 50 Hz noise was sometimes as bad as -60 dB. For -60 dB, depending on the config, the device managed to lower the noise to -65 dB in the worst case and -74 dB in the best case. Thus I can't say it doesn't work at all.

 

[Peterinvan]

USB isolators can only prevent noise coming from the source to the DAC.
When there are groundloops between (active) monitors or connected amplifiers or other devices the isolator will do nothing.
You need to determine where the issue comes from. When removing the USB cable from the DAC results in a silent system then chances are an isolator will be the solution.
Which one depends on the used bitrates and amount of channels.

My system does not have any noise issues.

However I have an old iFi iUSBPower unit in my chain. It outputs the data and power on separate USB outlets, and also has a "ground" switch.

iFi iUSBPower review | Darko.Audio

What must the digital audiophile bat away at source? Two long-time adversaries: 1) jitter and 2) electrical noise both emanate from your computer transport or streaming device. 2) feeds back into 1). These insidious twin menaces must be tamed for optimal sound reproduction. Case in point: the...

darko.audio

You may be able to find a used unit.

 

[HaveMeterWillTravel]

The isolator eliminates ground loops. There is no opportunity for a ground loop with UMIK-1, no need to use the isolator. Those 50Hz get most likely into the weak analog signal from the microphone, before AD conversion. The isolator has no effect on the digital signal passed.

"ground loop" is often a misnomer. It's primary purpose/effect is to greatly reduce or eliminate common mode current flowing between the two devices. The reason why it's a misnomer is this current may be present even if both devices have no ground, such as with Type II power supplies.

The reason why it really wont have any effect on an end device like the UMIK-1 is that it has no electrical connection other than the USB. Only in rather narrow circumstances such as acting like an antenna near a large source of EMI.

Even if it's not enough to disrupt communication, "noise" on a digital link may still be able to affect the audio at the ADC, DAC and/or the analog circuitry.

 

[phofman]

"ground loop" is often a misnomer. It's primary purpose/effect is to greatly reduce or eliminate common mode current flowing between the two devices. The reason why it's a misnomer is this current may be present even if both devices have no ground, such as with Type II power supplies.

Yes, the name is a simplification. But still it is a current loop (DC or only AC) formed by a secondary path, which not necessarily being the ground wire.

I am not sure the ADUM isolator can also suppress noise on the digital link. Also the DC/DC switcher for the secondary Vbus implemented on the cheap boards may be quite noisy too.