Few questions about Linkwitz' mic amp circuit

[454Casull]

http://www.linkwitzlab.com/images/photos/mic-amp.jpg

First - can two channels share the same virtual ground (I assume that's what it is at the bottom) and +/- rails?

Second - doesn't the virtual ground circuit need capacitors or something? http://tangentsoft.net/elec/vgrounds.html

[454Casull]

Also, does anybody know what the current draw might be?

[TheSeekerr]

No, the vground taken from the centre of two batteries doesn't need a capacitor to stabilise it in this design, because, unlike a headphone amp, it's not driving a particularly reactive load, so from a DC perspective, the potential at vground should never move far enough from the centre to be a problem.

That said, a better virtual ground couldn't hurt anything.

[454Casull]

Fantastic. Thank you.

Is there anybody who might be able to answer the other questions?

[infinia]

Quiesent current (Iq) is approx 4mA typical per channel each rail at 15V. Using 9V batteries I wouldn't expect too much downward change.
So using a OPA2134 would be ~8mA and a OPA134 would lower that by 4mA.
Supply current due to Vo(RMS) and Voffset shouln't be a problem due too low MIC levels and reasonable loads.

edit>Virtual ground is for replacing dual supplies (usually batteries) with a single one.

[infinia]

Originally posted by 454Casull

Second - doesn't the virtual ground circuit need capacitors or something?

No.. but the opamp needs them not the batteries. Most people leave them off, or show them in the corner of the schematic for simplicity sake. Use 2x0.1 uF decoupling caps close as you can at the DIP pins.
The common batteries connection is the actual ground (reference)
not really called a virtual ground.

[454Casull]

Thanks for the reply.

I'd like to use one of these circuits to provide the voltage so that I don't have to worry about matching the voltage between the two batteries (wire them up in series). Do you foresee any problems?

http://tangentsoft.net/elec/bitmaps/.../tle2426cp.png

On that note: do I need to put the 0.1 uF caps between the op-amp supply pins and ground no matter what I do with the supply/ground? I have the caps (non-polar MKTs) and I certainly don't mind putting them in, but if I use the TLE2426 rail splitter and for some reason it doesn't need the caps, then I don't want to do anything pointless.

EDIT: Is this how it's done? (attached image)

[TheSeekerr]

When using TLE2426 in a low current situation, it is acceptable to use it as shown, with a largish (220uF or bigger, I usually use 470uF) capacitor before the rail splitter.

However, in situations where more current will be drawn, it is necessary to add 2 more capacitors, one from the positive rail to the vground, and one from the vground to the negative rail, and because these capacitors are essentially in series, they must be fairly large, say 330uF or larger.

In this situation, the additional capacitors should not be necessary, but the decoupling capacitors mentioned above will be. The diagram you have shown is correct - the distance from each power leg of the opamp to its respective capacitor should be as short as possible, and ideally soldered directly to the leg, although having it in an adjacent protoboard hole should be sufficient. The leg of the capacitor going to vground can be as long as necessary.

[454Casull]

Thanks for the input. Pretty much everything I was unsure of has now been cleared up.

Except for one thing. What effect does the size of the cap before the TLE chip have?

[TheSeekerr]

It has two purposes - to lower the impedance of the battery supply, and to supply current quickly - essentially, these are the same thing. Since a mic preamp will, typically, be delivering into a high impedance (the input of an amplifier, or the line in of a PC/recorder), current demands are very, very low. Consider that, with a typical opamp, maximum output swing would be about 2V below the rail to rail voltage - so, running from 18V, delivering into a 50 kOhm input, this requires less than 0.4mA of current.

In this case, size is relatively unimportant, but you should select a low ESR cap. This circuit would likely perform fine with no cap there at all, but a good, low ESR cap of about 220uF should have the circuit performing optimally. Combine this with good 0.1uF decoupling caps (NP0 ceramic is ideal, but you're not likely to find one in this high a value, so use what you've got), and it should sound great and be perfectly stable.