USB 5V noise filtering

[Maquis]

Recently the internal audio device in my laptop went dead. I replaced it
with a USB sound adapter tapping its power from the USB +5V. The adapter
works (almost) fine but...
the audio, especially the microphone input, is noisy. Some kind of a digital
noise of switching noise. After inquiry, the noise is coming from the laptop
5V supply coming down the USB cable, about 50mVp-p on a scope, may be more
as my scope is low bandwidth. The adapter is surely lacking effective
filtering.

What would be the most simple and effective way to filter this out without
to much voltage drop?
The adapter can work down to 3.5Vdc and draws about 20mA. I look for
suggestions or ideas form people who got the same kind of noise problem from
USB supply.

Thanks!
John

 

[Ian Stirling]

Recently the internal audio device in my laptop went dead. I replaced it
with a USB sound adapter tapping its power from the USB +5V. The adapter
works (almost) fine but...
the audio, especially the microphone input, is noisy. Some kind of a digital
noise of switching noise. After inquiry, the noise is coming from the laptop
5V supply coming down the USB cable, about 50mVp-p on a scope, may be more
as my scope is low bandwidth. The adapter is surely lacking effective
filtering.

Maybe - the noise may be coming from somewhere else.

What would be the most simple and effective way to filter this out without
to much voltage drop?
The adapter can work down to 3.5Vdc and draws about 20mA. I look for
suggestions or ideas form people who got the same kind of noise problem from
USB supply.

First, I'd try a clean 5V supply.
Maybe even 4.5V from batteries.

 

[artie]

Recently the internal audio device in my laptop went dead. I replaced it
with a USB sound adapter tapping its power from the USB +5V. The adapter
works (almost) fine but...
the audio, especially the microphone input, is noisy. Some kind of a digital
noise of switching noise. After inquiry, the noise is coming from the laptop
5V supply coming down the USB cable, about 50mVp-p on a scope, may be more
as my scope is low bandwidth. The adapter is surely lacking effective
filtering.

What would be the most simple and effective way to filter this out without
to much voltage drop?
The adapter can work down to 3.5Vdc and draws about 20mA. I look for
suggestions or ideas form people who got the same kind of noise problem from
USB supply.

Thanks!
John

USB uses half-duplex differential signaling. One of the tricks you can
try is to get a short USB cable, open up the +5 line (RED), put a 10
ohm resistor in series, and bypass caps from the load side of the
resistor to the ground (BLK) line. I usually dig through the junk box
and put in a surface-mount slug tantalum (Hobson's choice, usually 4 to
10 uF) paralleled with a .01 uF. That will tell you if you've got
noise on the +5 line messing things up.

But how about ground loops? Open up BOTH the +5 and ground lines, and
put 1 ohm resistors in series (1 to 10 ohms, again, hobson's choice
depending on what you have and how much current the thing on the end of
the USB cable wants), and bypass that.

A pair of these cables makes it easier to find/diagnose noisy (cheap)
designs.

 

[Maquis]

USB uses half-duplex differential signaling. One of the tricks you can
try is to get a short USB cable, open up the +5 line (RED), put a 10
ohm resistor in series, and bypass caps from the load side of the
resistor to the ground (BLK) line. I usually dig through the junk box
and put in a surface-mount slug tantalum (Hobson's choice, usually 4 to
10 uF) paralleled with a .01 uF. That will tell you if you've got
noise on the +5 line messing things up.

But how about ground loops? Open up BOTH the +5 and ground lines, and
put 1 ohm resistors in series (1 to 10 ohms, again, hobson's choice
depending on what you have and how much current the thing on the end of
the USB cable wants), and bypass that.

A pair of these cables makes it easier to find/diagnose noisy (cheap)
designs.

I build a small filter with 22 ohms with bypass caps 15uF/0.01uF on the
supply line. I still have 4.3V at the output. This got rid of most noise by
about 80% (subjectively...).
But I still have some kind of steady high frequency wistle noise when
recording with the mic. I know for sure the mic is good.
Just to be sure, I tried 4a clean 4.5V supply from three "AA" batteries as
suggested by Ian and the result is the same. At least the filter is good for
filtering noise from the computer.
I guess the remaining noise is "internal" to the adapter just before the
microphone signal is sampled and converted into digital form.

 

[Maquis]

I build a small filter with 22 ohms with bypass caps 15uF/0.01uF on the
supply line. I still have 4.3V at the output. This got rid of most noise
by about 80% (subjectively...).
But I still have some kind of steady high frequency wistle noise when
recording with the mic. I know for sure the mic is good.
Just to be sure, I tried 4a clean 4.5V supply from three "AA" batteries as
suggested by Ian and the result is the same. At least the filter is good
for filtering noise from the computer.
I guess the remaining noise is "internal" to the adapter just before the
microphone signal is sampled and converted into digital form.

Finally I reduced (drastically) the mic remaining noise by adding a bigger
cap right at the supply input of the adapter. At least 220uF does the trick.
I kept the resistors in the circuit.

Thanks for your suggestions guys!
John