DIY summing mixer schematic help

[The Scum]

So I just picked up a crap-ton of resistors, and it seemed to really have put a hurt on the guy at Fistel's electronic cavern... so I really don't want to return them if I don't need to but here's the thing. They are all 1/4W. All of the other designs I have seen have what look like at least 1/2W resistors and I do not know how to figure the current from the outputs of my interface (especially since I just sold it ).

I seem to remember that a full-scale +4 signal is up around 25V but without knowing the current that number is kinda meaningless. I would hate to wire up 64 of these things (16channel) and find out I have to return them. At least they are 1% so I figured I could get away with 5K all around.

So calculate the current. Then calculate the power. Wikipedia and Google will be happy to give you background on Ohm's Law and the Power Law.

Start with Mr Ohm: V = IR.

So you'd have 24 V (not an unreasonable number, maybe a little high, but that just adds some margin), into a 5K load (assuming either an inverting summing amp, or an infinite number of other channels in parallel...reality will be somewhere above 5 K, but again, round numbers give us margin).

24 = I * 5000

24/5000 = 4.8 mA

Then we use Mr Watt: P = IV

24V * .0048 = .1152 W.

Well below .25 Watts, And in real systems, not all channels are driven to max capacity all the time...the signals are transient in nature, lowering the dissipation even further.

[gabe real]

i walked around the electronics store reluctant to ask the guys there about
the difference between the wattage. all the guys looked like
they could build a robot from scratch, i was a little intimidated.

but, i just said the hell with it and asked. they said it doesn't
really matter on the wattage for audio due to the load being pretty low.
you can even mix and match different ones and it will still work.
some of my resistors are 1/4 watt and some 1/2 watt
thats what they told me at least.

the bigger the wattage the bigger the resistor.
it makes a difference in the percentages though.
the lower the better. so if they are 10%, they could be 10% up
or down from the stated value. correct me if im wrong please.
so a 100 ohm could actually be a 110 ohm or a 90 ohm.

[thesimulacre]

Thanks Scummy & Gabereal,

B> My only lead on those equations was PIE or something, and the results were lacking in wiki, so I appreciate the actual calcs. PS how is life treating you?

G> I wanted to go with .1% (or better?) because according to my research thus far, the closer matched they are in a balanced design, the lower one can make the actual resistance without risking too much crosstalk. Or something. What they had at Fistel's was 5K, 1%... and for 16 inputs I figure the make-up gain will end up the same as an 8-input with 10Ks.

-now to figure out the preamp insert, post-pre inserts 1&2, and a selector switch for that mess. sigh.... it could all be so simple.

[Z-Plane]

So...between this thread, this one, and however many more were spawned, was there any consensus reached on adding a simple internal makeup stage ?

[thesimulacre]

Z- I was going to use a pair of SC1-MK2 pre's because they are clean, cheap, upgradeable, and fit vertically in my 2U chassis. And since they are fully functioning preamps I can use them as such with the flick of a switch. The project is on hold for now until I get the panel layout finalized and sent out.

[Z-Plane]

Thanks - that's a great idea. Just out of curiosity, how simple can the make up stage get before a sonic trade-off is made ? Let's say the output from the summing stage was -30db, what's the minimum design required to recover that loss while remaining transparent, and without introducing anything undesirable ?

[The Scum]

was there any consensus reached

Consensus?

The main point of this design is that it can be used with whatever mic preamps you have handy, and switch them as you see fit.

If you want something with some makeup amps as part of the design, check out the stuff Fred Forsell has on his site.

what's the minimum design required to recover that loss while remaining transparent, and without introducing anything undesirable ?

Plug the output back to the input of your DAW. Apply the 30 dB in software. With 24-bit ADC and floating point math, it should be workable.

Define undesirable. Some of us like a bit of distortion, and don't mind if 20K dips a bit.

The simplest actual hardware amp might be a single transistor per bus...consult your Forrest Mims books for an example. Scott Hampton's JFET module could probably do it, too. But discrete transistor circuits become very dependent on the load they're driving, and power supply stability.

For low oveall parts count, reasonable performance, and corresponding power supply independence, opamp circuits become a consideration. Again, check Fred F's website for a pretty good whitepaper. He makes some pretty good arguments for configuring the opamps in an inverting summing amp config.