Skip to main content

how to sum 2 line level stereo signals ?


To mix several signals in stereo, there is the mixing console, of course ! But i need no eq, no pan, no level control etc... no active circuitry... What i want is just to mix 2 stereo signal that goes out from 2 preamps (BG1), into a 2 tracks recorder (BB+)... (to 'copy' the Nagra V with its 4 inputs !)

Is there a simple passive circuitry, a small passive box that does the job? the contrary of the multing... There is the Folcrom, but the Folcrom is in a 1U rack and it's for 8 stereo channels... I read somewhere it's quite easy to sum 2 signals, just soldering a resistor on the linked cables or something like that, but i'm not sure that's the best way to really preserve the quality of the audio signal... Is there an audio electronician here who could enlighten me on this ?...

Thanks in advance,



ghellquist Mon, 02/13/2006 - 03:53
It sounds to me that all you need is a few resistors, short pieces of cable and connectors. A bit of soldering, maybe 10 minutes, and you are fit for fight. I would expect no loss of sound quality.

The purpose of the resistors is to prevent the two outpus to "shout" at each other as they both are low resistance.

I would use 1k resistors 1/4 Watt. I would use metal film resistors (not carbon body) for this, but they are common and not very expensive. I cannot see much need for close tolerances, so standard resistors is OK.

Maybe someone with more experience can advice on other values but this has worked for me in other places.

For summing two balanced signals you need 4 resistors. They can be soldered directly into the contacts if you want to making into a nice little Y - calble. (Remember to lable it).

Simply put the resistors in-line with the positive and negative signals (pin 2 and 3 in an XLR) on the "inputs", in the female connector. After the resistors the two positive signals simply meet in the male connector(as does the two negative). The two grounds also meet (pin 1). I tend to not connect the shell to anything as this can lead to ground loops. In some circumstances that can be necessary, but not here.

Keep the Y-cables short as it is more susceptible to loss of high frequencys due to the higher output resistance after the resistors. If you have to go a long cable (tens of meters or more), put the summing resistors close to the input, if you are good at soldering the can be made to fit in the male connector (takes a bit of practice though)..


anonymous Tue, 02/21/2006 - 16:25
Zilla wrote: As ghellquist already suggested, you just need some passive mixer cables. I have a design for such a thing, and would be happy to post it if someone would tell me how. I have it as a pdf. Your preamp will need to be able to drive 600ohms, though.


thanks, i would be glad to have a look on your schematics! :)
i think you need your pdf file be put on an url and link it here or somethink like that so that everyone can download it... i don't know if you can do that. If you can't, would you mind send it to me here ?

but i don't know if the BB+'s preamp can drive 600ohm ! yes, probably... but actually i don't even know what that means ! (i know, shame on me!:oops: i should better go back to study my electronic, lesson one... and i will (try)!! :wink:)

But.... if it's so simple and so cheap to make... why the folcrom costs so much ?!

anonymous Wed, 02/22/2006 - 01:10
Zilla wrote: [quote=taro]i don't know if the BB+'s preamp can drive 600ohm !
The ouput drive capability should be specified in the manual.

I sent the pdf via email. Good luck.
ok, right, t'was the *output* capability ! :D... (the output drive capability of the bg1 is not specified in the specs but i'll send an email to dav.)

Thank you, i received the pdf... the schematic is very clear. i thank you very much !!

now, on the theory side... what exactly is the formula applied to specify the resistor's capacity according to the output drive capability, or... ? well, ok, what audio electronic's litterature should i read to understand better all this ?! ;)

Simmosonic Wed, 02/22/2006 - 01:27
A simple resitor network will do the job, but to do it optimally you'll need to consider all the input and output impedances.

My advice? Assuming this is for your ARES, you ought to send an email to the guys at Nagra in Switzerland, telling them what equipment you're using and what you're hoping to do. Their technical people are good at offering such advice (myself and other Nagra V users have received numerous similar useful tips from their design engineers), and they know the input circuitry you're using so they'll be able to give you the best advice.

Also, I agree with Gunnar, avoid the cheapo carbon composition resistors - go for the metal film.

If you want mixing capability (i.e. pots) you'll get better results from the more expensive plastic film components like Alps or, if you can afford it, Penny & Giles (P&G). In fact, P&G make low capacitance pots for passive applications, and the difference is quite significant in terms of HF response. But they're very expensive...

Zilla Wed, 02/22/2006 - 16:49
taro wrote: now, on the theory side... what exactly is the formula applied to specify the resistor's capacity according to the output drive capability, or... ?

With any passive circuit there will be an insertion loss (there is no active stage to make up gain). In the case of our simple two channel summing network we need enough resistance to isolate the two micpre outputs from each other while minimizing the signal level loss. Since all my micpre's are capable of driving 600ohms, I was able to designed that "mixer cable" for minimum loss. With low 174 ohm resistor values (presenting ~680ohm balanced load to the pre) the loss is minimized to about 6dB. If you where to use 1kohm (presenting ~3.6kohms load, balanced), you would gain more isolation but lose another 0.5dB to insersion loss.

Simmosonic Wed, 02/22/2006 - 19:37
ghellquist wrote: I cannot see much need for close tolerances, so standard resistors is OK.

Agreed. However, considering how cheap metal film resistors are, it may be worthwhile buying a dozen or two of each required value, measuring them and matching them in pairs. I would be using closely matched pairs on each balanced line (one on each leg), to maintain the impedance balance and therefore the CMRR.

It may not make much difference, if any, particularly at line level. But, for what it's worth, I see no harm in doing it. It costs very little and takes very little time (assuming one has a sufficiently accurate resistance meter), and it's one less potential problem to worry about in future. Especially if you are recording on location - you never know when you might find yourself in a 'hostile' audio environment and you really need that CMRR...


Your recently read content