passive switcher/splitter for monitor controller

[redmaster]

Hi,
Friend of mine is building for me monitor controller.
design is simply :
- 3 input and 3 sources selector - A,B,C
- 2 output selector - 1,2,3 (main out, subwoofer, tape out - non variable)
output 3 is always ON - it's copy from input
- passive attenuator controlling volume (I've chosen khozmo, as it supposed to be 0.1 dB channel difference)
it's not balanced, rca and TRS inputs/outputs
3 outputs work not only as output selector, but splitter/mixer.
output 1 - main out - signal on main out plus copy on output 3 - for tape monitoring/recording back to DAW
output 2 - main out plus sub out and copy on output 3
I wanted to have it completely passive.
When I checked the model I noticed when engaged sub out, I have a drop on main output (around 0.6-1.2dB)
I noticed that drop using my ears only, but couldn't give a detailed info about drop, so checked it with RME digicheck software plus looped waves from 40Hz up to 20kHz.
Gentleman who's building it for me told me it's because of passive design, and it has to be active opamp to balance this 1dB audio drop.
I am not in position (lack of knowledge in area) to argue about it, but I am trying to establish if it's possible at all to have it completely passive, without any audio drop or degradation in this design.
Please advise.

many thanks in advance

PS. I did some internet research and found coleman audio being passive design (don't know if it has any drops ) :
Coleman Audio M3PH MkII
or others like this :
Pro Co Sound, Inc. Pro Co Switch Witch Stereo In 4 Stereo Out Monitor Switcher Audio Switchers at Markertek.com
although I don't need even the cheapest switcher LS3,
my controller is almost ready, all sockets, ins/outs, cabling done, also this controller design is custom for my desk (size is crucial) - I'd like to use something similar to have it completely passive

[redmaster]

after a bit of research I got to conclusion that it is not possible to have it all passive.
the only way is active buffer,
advised on THAT chips, the 1200 or 1240
any more opinions very welcome

[AndrewT]

the drop in output when adding a second line out is due to the resistive ladder effect.
eg.
Rs = 1k, Rload = 10k
emf = 1Vac,
output voltage = emf * 10k / [10k+1k]

Now add a second output feeding Rload = 20k.

The effective load impedance is 10k//20k ~= 6k667
The output voltage is emf * [6k667+1k]

The only way to ensure that the output voltage does not change with loading (number of outputs connected) is to set Rs=0r0

This cannot be done with an attenuator, nor can it be done with most normal Sources which generally have Rs>=50r

You can reduce the multiple loads effect by making Rs very low and making all the Rloads very high. But be wary of increasing noise and interference.