Any technical problem with this pre summing?

[k brown]

Often need to mix two preamps together in a very simple way (equal gains, or one at -3 to 9dB).

Would like to experiment with doing this without adding a seperate mixer to the setup. Thinking of adding 'bus jacks' to two M-Audio Duos (orig, not M-Track Duo), to use a linking cable to connect the two at the 5532 inverters that follow the INA163 pres. Essentially using the inverters as virtual ground summing amps (never more than two pre outs would be linked to this point). Would the 3.2k between summing point and each INA provide sufficient isolation between their outputs?

Actual values from the Duo are shown.

See any issues with doing this?

 

[ruffrecords]

Your method connects two negative feedback loops together. Could possibly be a recipe for instability.

Cheers

Ian

 

[abbey road d enfer]

Much simpler this way.

 

[k brown]

Much simpler this way.
View attachment 115715

Hmm - INA and 5532 don't need DC block? Certainly alleviates Ruff's concern.

Was hoping to avoid that much board surgery, though.

 

[k brown]

Your method connects two negative feedback loops together. Could possibly be a recipe for instability.

Cheers

Ian

Just the sort of thing I was afraid I'd overlooked. Thanks.

I suppose I could also use a switched jack to disconnect the unused 5532 stage; both wouldn't be used in this application, just the one the two pres are mixed to. But I do want to retain the unit's original function as well.

 

[abbey road d enfer]

Hmm - INA and 5532 don't need DC block? Certainly alleviates Ruff's concern.

Was hoping to avoid that much board surgery, though.

Hard to say withoutcircuit's details, but the opamp needs correct biasing, assuming the output of the preamp is at 0V, that's what the doctor ordered for the opamp.

 

[Hubbub]

Referring back to the original post schematic simply move the interchannel connection to the left (-) side of the 22uF capacitors. Since the INA163s are followed by 1K resistors you are not shorting outputs, so that's legal. Any slight difference in INA163 DC output level won't be an issue unless the interchannel linkage is done with a variable resistor that gets changed in operation. Since you are not connecting after the 22uF caps you are not so significantly affecting the 22uF-2K2 highpass corner frequency or introducing 5532 DC or stability issues.

 

[k brown]

Referring back to the original post schematic simply move the interchannel connection to the left (-) side of the 22uF capacitors. Since the INA163s are followed by 1K resistors you are not shorting outputs, so that's legal. Any slight difference in INA163 DC output level won't be an issue unless the interchannel linkage is done with a variable resistor that gets changed in operation. Since you are not connecting after the 22uF caps you are not so significantly affecting the 22uF-2K2 highpass corner frequency or introducing 5532 DC or stability issues.

Thanks - I guess I assumed the INA outputs wouldn't be happy only being isolated with 2k between them . . .

But, no - no pot between them just a short link cable.

 

[FIX]

When you join them, each will drop 6dB, which might be OK if they are headed to the same input, but you might want to do this:

Add caps if you want. The input is hooked inverted, as the summing is also inverted. The pink is a DPDT switch, Green is ground. This will give you unity gain.

 

[Hubbub]

That (#8) should work. The actual highpass corner will move slightly as the series resistance seen by the 22uF changes from 2k2 +1K to 2k2 + 500R, effectively 3k2 to 2k7.

 

[k brown]

When you join them, each will drop 6dB, which might be OK if they are headed to the same input, but you might want to do this:

Add caps if you want. The input is hooked inverted, as the summing is also inverted. The pink is a DPDT switch, Green is ground. This will give you unity gain.

No, the 6dB loss is not at all an issue in this application; actually results in a more usefull gain range!

 

[abbey road d enfer]

The 6dB drop would happen when only one signal, L or R is present. in most cases teh signals on L & R are highly correlated, which results in summing close to unity.

 

[k brown]

The 6dB drop would happen when only one signal, L or R is present. in most cases teh signals on L & R are highly correlated, which results in summing close to unity.

Especially in this use case, as it's for summing two stereo pairs on a single bar; L+L, R+R.

 

[k brown]

Thanks for everyone's input - I went with @Hubbub's suggesttion, as it was the simplest, and required zero board surgery; only needed to solder wires from the neg side pins of the 22uF caps, run to TRS jacks on the back- works perfectly!

Four pres to a single SPDIF out, without mixer/computer/interface/DAW, etc. - just two Duos and a bit bucket; love it!