RC filter from power bank, to get rid of high frequencies

martinsson93 · Oct 14, 2021

Hi!

I have a question regarding a RC-filter I'm trying to create.

I have an amplifier with a microphone connected to power supply in a power bank. The power from the power bank is switching power which causes a lot of noise for my analouge circuit. So I want to filter the high frequencies with a passive RC filter. How do I know which values to use for the resistor and capacitor? And do I connect them right after the power bank but before the amplifyer?

It's a 5 V power bank and my circuit needs at least 2 V.

Lucien Nunes · Oct 14, 2021

An RC filter is probably not the right choice.

Before deciding a strategy to control the noise, you need to know whether it is ripple on the voltage itself, or common-mode i.e. propagating from the power supply along both conductors. If it is common-mode, especially if anything else sharing the supply is connected to the computer that is also connected to the amplifier, then you cannot remove the noise by filtering the DC because it will continue to circulate around the ground loop. Sometimes the only solution is galvanic isolation using an isolated DC-DC converter although that can introduce additional noise of its own.

The way to test for this is to try the battery again as the power source for the amplifier, but connect the power bank negative wire to the amplifier negative supply (=ground) terminal at the same time. If the noise returns, it's ground-borne. If not, it's voltage ripple that can be solved using a low-dropout adjustable linear regulator with its output set slightly below the the 5V input. This will get rid of it much better than an RC filter and it is a common method used in analogue equipment: Get the voltage close using switched-mode conversion for efficiency, then skim off the noise with a couple of volts of headroom using a linear post-regulator.

If this all sounds a bit involved and complicated for what looks like a simple task of powering a small circuit, I would say that the correct treatment of audio noise sources in mixed analogue / digital environments is one of the factors that distinguishes good circuit designers from bad ones. I spend a lot of time and effort on it, to make my equipment quieter than the competitors' products.

martinsson93 · Oct 18, 2021

Thank you so much!

I measured like you told me, and there was no noise when connecting the power supply to ground and at the same time have a connection to a battery. So I measured the voltage ripple from the power supply, and got the following scope (attached image), with a peak to peak voltage of 60 mV.

Now I'm not sure what to do, should I try to implement an LC filter or a linear regulator? And what values would then be suitable to use?

Furthermore, my amplifier specifies that it has a supply voltage of 3 - 12 VDC and a power consumption of 2.2 mA at 12 V. What should be my cut of frequency?

Lucien Nunes · Oct 19, 2021

Use a low-dropout adjustable linear regulator. I would pick something like an LP2951 and set it to 4.25V or thereabouts. The regulator needs to drop about 0.35V minimum to operate. That will give some headroom for the amplifier, it's best not to run at the very lowest possible voltage.

martinsson93 · Oct 19, 2021

Ok, I realised that I was measuring the wrong thing on the oscilloscope. So I'm not sure anymore if it's ripple that is causing the noise or if it's just broad spectrum noise, i'll attach an image.

Either way, I tried with an LC-filter with capacitor 10 uF and an inductor of 100 uH. It helped a bit with the noise but it's still there and also the signal is not as much amplified (at least it sounds like that, it sounds like I'm in a box when speaking close to the microphone?)

You can't tell from the image, but the oscilloscope measured frequency around 2 kHz from the signal. Do you still think a low-drop out adjustable linear regulator would be a good fit?

Thank you, once again!