Help understanding a simple Differential Amp circuit

Hello, everyone,

I've attached the main block of the schematic for an eight channel balanced to unbalanced PCB that I use for various applications, such as summing mixers. For a particular build I am doing for a client, the signal leaves the INA134  then goes to a on/off switch, then to the summing network (47k resistors) then to the summing amp. What I am confused about is when a particular channel is muted (signal is broken to the summing network) I still get a faint amount of signal from that same channel. In the grand scheme of real world use, it will never be noticed. But what I can't understand is how the signal is getting to the summing amp in the first place when the signal is being disconnected before the network. The best I can figure is that signal is somehow bleeding down the ground connection (all of the connectors have pin 1 tied together and going to the star ground) and is faintly being heard when referenced to the audio path. When I disconnect signal to the summing amp the faint signal bleed goes way, so I've at least narrowed it down to the INA137 circuit. I currently have the chassis and ground connections going to the star ground on separate connections. I had seen and read about separating audio and chassis ground with a 10 ohm resistor and a 100nF film capacitor connected parallel, but  don't understand why.

What I have currently will be more than fine and I've had no complaints about crosstalk/bleed from other clients, but I'm always in pursuit of better understanding even though it causes me great mental pain.

Thanks!

Paul

How are you muting the signal? That would be the first thing Ild check.

john12ax7 said:

How are you muting the signal? That would be the first thing Ild check.

Click to expand...

It's in the description and not on the schematic. INA134 output to switch to summing network to summing amp. The signal is being broken at the switch. I can even pull the connections at the switch and I still get the faint signal. Only when I pull the connection between summing network and summing amp is there no signal.

Thanks!

Paul

It is probably capacative coupling across the internals of the switch. There are a couple of things you can try. 

First you can change the way the mute switch works. Use a SPDT switch and use it to disconnect the signal and also to ground the 47K bus resistor.

Secondly you could leave the switch as is and change all the bus feed resistors to 10K and also change the feedback resistor in the summing amp pro rata. This should reduce the effect of any capacative leakage by more than 12dB.

Cheers

Ian

ruffrecords said:

It is probably capacative coupling across the internals of the switch. There are a couple of things you can try. 

First you can change the way the mute switch works. Use a SPDT switch and use it to disconnect the signal and also to ground the 47K bus resistor.

Secondly you could leave the switch as is and change all the bus feed resistors to 10K and also change the feedback resistor in the summing amp pro rata. This should reduce the effect of any capacative leakage by more than 12dB.

Cheers

Ian

Click to expand...

Ian,

Thanks for the info. I tried to make the switch connect the signal before the bus resistor to ground but it caused more noise, something that was akin to a feedback loop. I will give your bus resistor value change suggestion a go and see where that lands me. Only three signal are being summed at any one time, and I had used 47k resistors since I am using an API amp circuit, but I didn't do the calculation as seem to have buried that info in the piles of other bits of info I've accumulated here, but just by knowing the number of inputs being summed I should know that 47k is a higher value than needed.

Thanks!

Paul

Potato Cakes said:

... what I can't understand is how the signal is getting to the summing amp in the first place when the signal is being disconnected

Click to expand...

Star ground is the culprit IMO. Each stage draws current that passes through a wire that goes to the star point. Your local ground actually carries a minute but significant parasitic voltage to which the output voltage is referenced to, so it comes in addition to the output voltage, even if it's muted..
All mixer designers know that the grounding of the mixing buss and summing amp must be as short as possible, and in particular not take a detour.

abbey road d enfer said:

Star ground is the culprit IMO. Each stage draws current that passes through a wire that goes to the star point. Your local ground actually carries a minute but significant parasitic voltage to which the output voltage is referenced to, so it comes in addition to the output voltage, even if it's muted..
All mixer designers know that the grounding of the mixing buss and summing amp must be as short as possible, and in particular not take a detour.

Click to expand...

The grounding scheme being involved is something I thought about as well, but not nearly in as much detail as you explain. All of the boards have separate connections to the star ground point and are not making other connections beforehand. Being this is in a rack case, I have the PSU on one side (closest to the IEC inlet and the star ground point) and the audio circuit boards are on the other side to minimize EMI radiating from the PSU transformer. The question I have is should I have another chassis ground point on the other side of the case closest to the audio circuits?

Thanks!

Paul

Potato Cakes said:

The grounding scheme being involved is something I thought about as well, but not nearly in as much detail as you explain. All of the boards have separate connections to the star ground point and are not making other connections beforehand. Being this is in a rack case, I have the PSU on one side (closest to the IEC inlet and the star ground point) and the audio circuit boards are on the other side to minimize EMI radiating from the PSU transformer. The question I have is should I have another chassis ground point on the other side of the case closest to the audio circuits?

Thanks!

Paul

Click to expand...

No. You should rewire all the input module grounds to the ground of the summing amp. The PSU ground should be connected also to the summing amp ground. You need a stiff connection.