Crosstalk in Passive Summing Mixer

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

cpsmusic

Well-known member
Joined
Dec 3, 2013
Messages
292
Location
Melbourne, Australia
Hi All,

A while back I put together a simple passive summing mixer. It uses balanced paths and has four mono inputs as well as nine stereo inputs (with separate Left and Right). I was trying the mixer out today and noticed that if I put a signal into one side of one of the stereo inputs it comes out both sides of the output but around 12dB lower on the opposite side. This happens on both the Left and Right channels.

For example:

Signal -> Left Input -> Left Output + 12dB lower signal Right Output

Have I done something wrong here or is this "crosstalk" a "feature"? :D

The design of the mixer is pretty simple - the mono inputs go to the four busses (R+, R-, L+, L-) via 15k resistors while the L/R inputs feed their respective busses via 10k resistors. The ground is common to all sockets.

The only way I can think of for the L/R signal to cross to the other buss is via the mono inputs - would that be right?

Any thoughts on what's happening here and how to fix it?

Cheers,

Chris
 
Hi All,

A while back I put together a simple passive summing mixer. It uses balanced paths and has four mono inputs as well as nine stereo inputs (with separate Left and Right). I was trying the mixer out today and noticed that if I put a signal into one side of one of the stereo inputs it comes out both sides of the output but around 12dB lower on the opposite side.

Could be a grounding problem. Could be a wiring layout problem. Post a pic and a schematic.
 
Sorry there's no schematic, but I based the build on a combination of the Forsell, DIYRE and Ian Thompson-Bell summing mixers.

I've attached an image of the unit - the four mono inputs are on the left and the outputs are on the right.Summing Mixer.jpg
 
The mono inputs need to be connected, either a low impedance source or grounded if unused.

Are these feeding a mic pre for make up gain? Did you install the -200 ohm resistor? That will improve crosstalk even more.
 
The mono inputs need to be connected, either a low impedance source or grounded if unused.

Are these feeding a mic pre for make up gain? Did you install the -200 ohm resistor? That will improve crosstalk even more.

It's feeding either a Focusrite Scarlett interface or a Warm Audio WA273-EQ (Neve-ish Mic Pre). I'll give it a test with the mono inputs grounded and see what happens.
 
The mono inputs need to be connected, either a low impedance source or grounded if unused.

Are these feeding a mic pre for make up gain? Did you install the -200 ohm resistor? That will improve crosstalk even more.

With plugs in the four mono inputs, the crosstalk is still there but reduced quite a bit. The channel with the crosstalk is now about 22dB down.

What I'm not clear about is how to ground the unused inputs?
 
You do not need to ground the unused inputs, indeed you should not. You just short hot and cold together. If you short hot and cold of the mono inputs together the crosstalk should disappear.

Edit: Are those jack sockets stereo types (TRS)??

Cheers

Ian
 
You do not need to ground the unused inputs, indeed you should not. You just short hot and cold together. If you short hot and cold of the mono inputs together the crosstalk should disappear.

Edit: Are those jack sockets stereo types (TRS)??

Cheers

Ian

What's the best way to short hot and cold? A shorted TS plug? A switch?

Also, yes, all the jack sockets are TRS.
 
If you are using a TRS then the best way to make sure an unused input has its hot and cold shorted together is to use a TRS socket with additional contacts that you wire so that it does this automagically when you unplug.

If you don't have these type of TRS sockets then the best way to short an unused input its to plug in a TRS plug with tip and ring shorted. You only need to do this for the mono inputs because these are the only ones that can cause crosstalk.

Cheers

Ian
 
This issue is just one of the ways a 'passive summing' unit may not be 'optimum'.
I am presuming you are ignoring the fact that the summed level depends on the number of 'inputs' connected and whether they are fed from a low impedance or not. it may not be important but the summed level will change if you add or remove inputs to the mix.
Audix, and I think Neve 'standardised' at a summing bus level of -35dBu so the bus terminating resistor was fitted to suit. The channels contributing to the bus 9through the summing resistors) were either terminated to ground or from the output of an amplifier so that there were no 'sneak' paths to other bus mixes and the level would remain constant irrespective of the number of channels switched to the bus. One 'downside' to this is that as the makeup gain is always requiring 35dB, it fixes the noise floor.
When you use a 'virtual earth' summing system, if channels are NOT routed to the bus, then the 'makeup gain' automatically reduces so for a low number of channels ROUTED to the bus, the noise floor falls. Of course this philosophy starts to 'fail' if you leave most or all channels permanently routed to the bus, where the summing amplifier 'acquires' more gain and sums any ground noise and there is then a number of channels which then gives the same noise floor whether virtual earth or 'passive'. i forget what that number is, I think around 35 or so channels.
A 'benefit' of having a [passive summing system is that the frequency response should remain constant irrespective of the contributions to the sum whereas the altered gain of a virtual earth sum will affect the HF response (and distortion) depending on the number of channels routed. As ever you are juggling with several variables at the same time.
Note a balanced summing bus can introduce crosstalk if some of the sources are actually unbalanced as the network of resistors becomes a bunch of interlinked attenuators. You would need to sketch out exactly what you are doing to work it out.
Matt S
 
Yes, my guess is there is no terminating resistor, which results in enough level to operate with line inputs or mic inputs with little gain, but has this problem of x-talk.
Passive mixers are intended to work in conjunction with a mic preamp.
 
If you are using a TRS then the best way to make sure an unused input has its hot and cold shorted together is to use a TRS socket with additional contacts that you wire so that it does this automagically when you unplug.

If you don't have these type of TRS sockets then the best way to short an unused input its to plug in a TRS plug with tip and ring shorted. You only need to do this for the mono inputs because these are the only ones that can cause crosstalk.

Cheers

Ian

I'm going to make 4 shorted TRS plugs, however I was wondering about the TRS sockets with additional contacts that you mention - do you have a link to one of these. Also, are they designed for this or does it depend on how they're wired up?
 
You do not need to ground the unused inputs, indeed you should not. You just short hot and cold together. If you short hot and cold of the mono inputs together the crosstalk should disappear.

Edit: Are those jack sockets stereo types (TRS)??

Cheers

Ian

I've made four shorted TRS plugs (TR is shorted) and tested the mixer with them in the four mono inputs - no difference.

I'm thinking I've wired something wrong here! I'm tempted to disconnect the mono inputs and just test the unit with the L/R inputs. Anything else I should try first?
 
Back
Top