Skip to content
USB Noise suppresso...
 
Notifications
Clear all

USB Noise suppressor

2 Posts
1 Users
0 Reactions
105 Views
marcdraco
(@marcdraco)
Posts: 476
Member
Topic starter
 

This is something of a long shot (for reasons I won't go into at the moment) since the design requires part values that I can't know for sure ahead of time.

 The problem a lot of people have been suffering with the USB microphone project (and USB audio in general) is the noise breakthrough on the USB power line. Traditional methods to remove noise like this might use large smooth capacitors but they tend to perform poorly at high frequencies, but still well inside the audio band. These spikes are due to electrical bounce on ground lines, poor decoupling and badly designed grounding. 

Earth loops? They're a cakewalk compared to this - those are the ones that cause that deep "humm" shown as a spike at 50/60Hz and the harmonics 100/120 and so on...

So I figured I'd try some "proper" ( 😝 😜 🧐 ) design and see what pops out. 

As of today I'm sat on my hands waiting for parts, but if anyone has the correct parts (or similar) lying around, fill yer boots!

The circuit has two major sections - a primary mass filter with a soft-start and a bypass followed by one or two LC filters giving a 2nd or 4th order LC filter.

image

Those USB receptacles at each end depend on how you're going to fit this, I've specified mine with USB 2 "B" types because they won't be easily confused, although it might be worthwhile putting a B on the input and A on the out put.

The soft start works by limiting the inrush current through R1. The theoretical limit for USB 2 is 10uF but you can use more by picking a suitable current limit to stop >500 mA surging through.

If the power line dips below 4.8V (give or take) the diode conducts and allows a short burst of current to hold up the device demanding some extra oomph.

The LC filters are a bit trickier but each one is a potential divider formed from two reactive elements that take (almost) no power from the circuit. Real components have losses that theoretical one don't.

So the idea is to take a combination of two of these set to act as two low-pass filters with a 3 dB point somewhere below the frequency interference. The further below the better the effect, but then the inductors become unreasonably large.

Note also that the second filter loads the first one so you can't put two identical sections on the same board (it will work but not as well as it can). Here's a tool to do that fiddling maths for you:

LC Filter Design Tool (markimicrowave.com) 

I assume the input impedance of the USB ports to be around 10 ohms (less for USB 3/SS) but the loading circuit is more difficult and that can have a significant effect on the performance (the greater load tends to drag the energy out of the LC filter faster than it can respond) meaning larger inductors are needed.

I'll report back when I know more myself. I'll need to test this with the various version of the USB C microphone before it'll be possible to get a decent comparison.


Take everything I say with a pinch of salt, I might be wrong!

 
Posted : 21/05/2024 3:25 pm
marcdraco
(@marcdraco)
Posts: 476
Member
Topic starter
 

This is just 100 uH and 680 nF. A second (matched) pair to not load the first ones unduly would have an even greater effect.

image

And that's without any other protection or filtering direct from a USB wall wart.


Take everything I say with a pinch of salt, I might be wrong!

 
Posted : 25/05/2024 3:19 pm