Crosstalk in Passive Summing Mixer

[cpsmusic]

Well, anyway, being non-insulated jacks, they are all interconnected via the front panel. That is a dead certain way of creating ground loops. You should use insulated jack sockets.

Just checked the jacks and they're insulated.

 

[cpsmusic]

I checked the jack socket with four connectors (they're now out of the mixer) and the fourth connector is a shunt for the Tip. Don't think it would have affected anything.

 

[abbey road d enfer]

Just checked the jacks and they're insulated.

So you replaced the ones on the picture?

 

[cpsmusic]

So you replaced the ones on the picture?

No, I checked them with a meter. The ground connector doesn't connect to the front panel.

 

[cpsmusic]

Let me check them again.

 

[cpsmusic]

I checked again and there's no connection between the socket ground and the front panel. I think the panel has some sort of non-conductive coating (?) as it seems to be a very good insulator.

 

[abbey road d enfer]

Relying on paint or anodization for insulation is not good practice.
So, how many lugs on these sockets? Can you draw a sketch or take a closer pic?

 

[cpsmusic]

Here are a couple of pictures - one is the inputs and the other is the outputs.

 

[cpsmusic]

All the sockets that are currently in the unit have three connectors.

 

[Matt Syson]

You need to step back and reconsider the whole setup.
If all the 'sources' are balanced, and the mix is into a balanced amplifier input then the grounding of the sockets is rather less important HOWEVER almost whatever you do that panel of jacks will represent a 'central' ground point. If most of the 'sources' are from a convertor (say 8 or more channels then they too will be a 'common' ground. If all sources are well balanced then the location of the 'central pint is less important. If you are building a studio using a large mixing desk 48 or more channels perhaps, there is a certain logic to calling that the central ground point. You are hitting the area where idealistic 'star ground' theories hit 3 dimensional practicalities.
Your panel appears to be anodised and even if it wasn't, you have correctly linked all the sleeve tags of the jacks so that in itself isn't an issue. Having the summing bus wires 'hanging in space' is not really good either but as you have a 300 Ohm? termination that is probably not an issue. Busbar sections of 'old' mixers often had a large aluminium unit which ran the length of the desk, with multiple 'channels milled (extruded profile actually) with the bus wires suspended in the slots. If you wanted to go tortally overboard you could use aluminium extrusion like the 2 'channel' strips used for sliding doors to 'encase' your bus wires. Not sure a day of drilling holes and fitting PTFE feed through connections is really warranted though.
Matt S

 

[abbey road d enfer]

Here are a couple of pictures - one is the inputs and the other is the outputs.

Well, it seems correct. So you have to do some detective work.
Start with all inputs disconnected and the panel floating. Does it hum?
If yes, check the preamp you use for make up.
If not, ground the panel to the rack. Does it hum? If yes, you have a ground loop, that you probably can't solve.
If not, connect one input. And so on...

 

[joeblack007]

A good reading for you that should help : Console Modification: Soundcraft 200b grounding

 

[abbey road d enfer]

A good reading for you that should help : Console Modification: Soundcraft 200b grounding

The reason this mod works is not because it uses star ground. It works because the flimsy ground connection provided by the ribbon cable is replaced with a much larger gauge wire.
Ribbon cable has a resistance per meter of 0.2 ohm, when loudspeaker cable is 0.02. The loom in a Soundcraft mixer is about 2 meters long, resulting in about 0.13 ohm (there are 3 wires in parallels). With the ground wires attached, the resistance from one end to the other is way smaller.
For a 32 ch, the resistance of the ground bus is about 0.95 ohm, with the ground wires added, its 0.025 ohm.

 

[dyamakuchi]

In this particular case, no. X-talk is caused by the topology, that allows mixed signals to be reinjected into the sources. Depending on the actual impedance of said sources, the effect is more or less pronounced. Zero-ohm sources suffer no x-talk, when open circuit is tehemost prone.

OK, whatever you say.

 

[dyamakuchi]

A quick question about grounding my summing mixer - what I did was connect all the earths of the sockets together, including the L/R outputs. Is that the best way to do it? Doesn't this mean that there are two separate paths to ground (via the Left and Right outputs)?

Single point ground is what you want. All ground connections lead directly to a single point, which is connected to the earth electrically. Return currents should not travel via these ground paths.

On another note, usually low frequency coupling is magnetic field (basically, always via a ground loop,) high frequency (crosstalk) electric field i.e. capacitive coupling dominates. I hope this helps.

Good luck!

 

[abbey road d enfer]

Single point ground is what you want. All ground connections lead directly to a single point, which is connected to the earth electrically. Return currents should not travel via these ground paths.

OK, whatever you say...

 

[radardoug]

I'm so tired of seeing people try to make passive mixers work, when it is so easy to make an active mixer.
Here you go, took me 20 minutes to draft up. Easy to build, no crosstalk problems, balanced. Will drive down to about 1.2K impedance.
Less than unity gain to allow for lots of inputs.