noise/hum in my switcher

[Dimitree]

hello guys
I finished my home made switcher (true bypass relays), but I got some hum in the audio path.
I recorded the problem:
https://soundcloud.com/dimitree-1/noise-hum

First, the switcher is off (no power), and there's no hum. Then I plug the 9V AC power supply (Line6 one, noiseless with my Line6 M13), and the hum comes!
I measured the hum frequency and it's 50hZ (I'm in Italy).
Then I engage the onboard buffer (this one uses a MAX1044 charge pump) and looks like an hiss adds to the hum.
Where should I start to fix that??? 

[GibsonGM]

Can you post a schematic of your switcher?  Why are you powering it with AC?  That creates much more of a problem than DC.    Your thoughts on the noise are correct regarding the frequency and the change when you turn on the buffer.

You probably need to add shielding, and more filtering for your power supply.   Inside the Line 6 there may be a network (R-C) to further filter the AC supply (in fact, I guarantee it, if it takes AC power).     

Just trying to get a handle on what you have built, and why you chose to use AC.   It would be much much easier to use DC.    You are rectifying inside your switch box?

[Dimitree]

thanks for the help,
I'm using AC since I need many mA to power 17 relays rated at 12V, so I'm using a 9V AC and I rectify it.
Unfortunately I don't have a schematic since I built it using bits/schematics taken from multiple places.
The 9V AC is rectified with a W06M bridge and a 1000uF cap. Then the DC signal is regulated to 9V with a 7809 (+ 100uF and 0,1uF caps) and passed to the MAX1044. The relays are powered from the unregulated DC coming out of the rectified (more or less 15V DC)

[GibsonGM]

Is the rectifier or filters near any signal lines?   Can you replace your signal lines with shielded cable, grounded at one end only?   Does playing with the routing of power wires vs. signal wires change anything?

Sounds like interference being radiated from power lines into signal runs, and maybe a faulty grounding scheme.....wire routing can easily cause these problems.
You want to keep the regulator(s) away from signal stuff, but then add some capacitance near where the power is being used....see the application notes for the LM7809.

[PRR]

Perhaps 0.8Amps (I do not know your relay consumption). 1,000uFd will filter that enough to "be DC", but not to be "clean! DC" that you can run next to sensitive audio signals.

I figure about 3/4V of ripple. Around signals that run 0.02V-0.5V, that is a LOT of 50Hz and higher harmonics.

Find a 3 Ohm (2 Ohm-4 Ohm) 5 Watt resistor. Build a C-R-C filter with 6000uFd, 3 Ohms, 6000uFd. That gets ripple down to 0.05V, still high, but a lot less.

Also route your relay coil-leads AWAY from your relay signal leads.

Of course your transformer and rectifier are FAR away from all signal stuff?

A well-regulated and filtered external DC supply may be a better bet.

[GibsonGM]

Perhaps 0.8Amps (I do not know your relay consumption). 1,000uFd will filter that enough to "be DC", but not to be "clean! DC" that you can run next to sensitive audio signals.

I figure about 3/4V of ripple. Around signals that run 0.02V-0.5V, that is a LOT of 50Hz and higher harmonics.

Find a 3 Ohm (2 Ohm-4 Ohm) 5 Watt resistor. Build a C-R-C filter with 6000uFd, 3 Ohms, 6000uFd. That gets ripple down to 0.05V, still high, but a lot less.

Also route your relay coil-leads AWAY from your relay signal leads.

Of course your transformer and rectifier are FAR away from all signal stuff?

A well-regulated and filtered external DC supply may be a better bet.

+1   I wasn't sure just how ineffective the 1,000 uF would be with that kind of current draw. SEEMS sufficient, but you nailed it....relays are stealing a bit more than just a couple of opamps!   Crazy filter, Paul, 6000uF x2!   That will do it fo' SHO!

He might need an "all of the above" approach, I think!    I'm wondering if he gets *POPs* when it switches?

[Dimitree]

I was checking connections, and I noticed that I forgot to connect digital ground and analog ground togheter at one point!! Now I connected them togheter, and the unit is totally silent!
thanks for the suggestions

[GibsonGM]

Great!  Sometimes things like that happen - yes, I would expect a LOT of noise from that, ha ha!   )