|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
In Reply to: How do you mix 6 channels into 2, without a mixer? [Mapleshade does it] posted by High-end Dreamer on March 10, 2002 at 07:30:29:
I work a lot with an engineer friend of mine Mike Skeet, you might have read some of his reviews of equipement in Audio Media. Mike uses very small discreet mixers that he builds himself, and they inlude a single stage mixer. His are designed around a "virtual earth amp" using a single chip, and in real terms very effective, due to there "minimalist design". However whereas in classical music production this is desirable, the trade off is lack of facilities for eq, sends, ect (although these could be added via inserts). I'm not an electronics expert myself so I can't detail the design of the virtual earth amp, but its a pretty proprietry circuit that I'm sure any local electronics engineer would be able to build for you.Regards
Roland
Follow Ups:
Hi,
All that a mixer is, is a "summer". It adds whatever signals are fed to its inputs. You can take a high quality op-amp integrated circuit, and set it up with a positive and negative dc supply voltage on its vsupply pins, usually pin 4=- and pin 8 = +. Opearate it at unity gain by tying its output pin, usually pin 7, to its - input pin,(usually pin 5), through a 100k resistor. Connect (6) 100k resistors at one point to the - input pin, and keep the other ends of the resistors seperated as inputs.
If you feed each of your 6 input signals to one of the 6 100k input resistors, the output at pin 7 will be the sum of all 6 inputs.
Op amps used in this configuration, are incorporated into commercial mixers .
If you want to control how much of each of the 6 inputs level are available at the output of the op-amp, replace the 6 100k input resistors with 100k potentiometers, (sliders if you want). You can then adjust the levels of each signal fed to the mix.
Commercial mixers add additional op-amp circuitry configured for fx, eq, panning etc., after the main mix output.
mg16
...Mostly correct. The circuit you describe is a virtual earth circuit, and is right up to replacing the resistors with potentiometrs, also it is important that the non-inverting (+) input is connected to ground. The output will in fact be the *inverse* sum of the inputs, since this is an inverting configuration.It's called a 'virtual earth' ("earth" is the British term for 'ground') because the junction of the seven resistors will always be ground, since the Op-amp will always trim it's output to make the two inputs -inverting (-) and non inverting (+) the same. Hence, +1V at any input gives -1V at the output, and the mid-point of the resistors will be the mid-point between the input & output voltages... (i.e. zero volts) thus keeping the potential at the (-) input the same as the (+) input, which is tied to ground.
Here's the problem with *replacing* the resistors with potentiometers: -if you take any one (or more) of the wipers to ground, then the (-) input is held at ground, and the op-amp goes into open-loop gain. However, since this also shorts the combined summing point to ground, then you have no signal, but lots of noise!
The correct way to do it is to leave the summing resistors in place, and put the potential dividers *before* the summing resistors.
An additional benefit of the virtual ground circuit is that the input impedance of each input is precisely controlled (It's the same as the summing resistor at all times, since the othe end of the resistor is connected to a virtual ground) and also that there can be no 'back-bleed' between outputs, since the only way that they are coupled is through a virtual ground, which kills any crosstalk.
For microphone use, this is no good as a front end, since the virtual earth circuit is unbalanced, so a good balanced preamp must be used. This will also allow the signal level to be raised away from the circuit's internal noise.
A virtual Earth circuit's noise level is linked to it's gain, and it's gain is proportional to the number of inputs connected. -For this reason, it is advantageous to add switches to one end of the summing resistors. (Hence all Rupert Neve-designed consoles have buss de-assign switches to the main busses.)
Keith A.
I'm not sure if I'm correct on this, Keith will know the answer, but by using a 33k resistors instead of 100k I believe you get 3db of gain. In Mikes circuits I believe that he uses a second virtual earth amp to restore positive phase, a derives a signal from the first in order to be able to produce a M/S matrix. (I must pay more attenttion!!!!) :o) lolRegards
Roland
Not bad, grasshopper... ;-)The actual gain with 33k would be three (a fraction under 10dB) rather than 3dB, since the gain is always R1/R2, where R1 is each individual input summing resistor, and R2 is the negatave feedback resistor (between the '-' input and the output pins)
In this instance, +1 Volt input would give -3 Volts output. Since the -3V would be across the 100K R2, and the +1V would be across the 33k R1, the mid-point would still equal zero volts, hence still virtual earth. -You can happily mix & match values of R1 input resistors with any given R2. Making R2 also 33k would return you to unity gain, since r1/r2 would equal 1.
For 3dB gain, you'd need R1=70.7k with a R2=100k feedback resistor, which gives a gain of 1.41 (3dB).
A second, unity gain inverting stage would indeed correct the polarity inversion.
Actually, I've built a couple of these things for local recording engineers who often like to sum exotic mic preamps into one channel (for example, several mics -close, ambient, behind, etc.) on a guitar cabinet to be recorded to a single channel on tape. This allows then to avoid too much electronics in the console, especially if they're working in a budget facility! -my own boxes are comprised of a power supply, a few unity-gain balanced front-ends, each feeding the circuit described above, then a final balanced output driver. -They seem to be very popular!
Mike has drawn these for me several times, but I always leave the napkin in the restaurant! lol. I do remember him mentioning that he doesn't like to build a circuit where there isn't some small amount of gain (think he uses 47k), and I do remeber that he uses two circuits to get positive phase, and meantime alowing him to make an MS decoder to boot! Perhaps you could draw a simple diagram of the circuit? I'm sure many here would find that useful.Regards
Roland
This post is made possible by the generous support of people like you and our sponsors: