ic power consumption?

[arjespen]

Hey.
I'm working on a project of making a cabinet sim from the LXH marshall schematic:
http://www.hexeguitar.com/images/schm/LXHmarshall_schem.gif

I need to put together a bipolar power supply for it, and I'm trying to figure out the ratings I need for the components - specifically for the transformer.
I know I need a 2 x 18v, trafo, but what power rating do I need?
I've done a pcb layout in eagle, and I'm using five tl074's and three tl072's.
However, reading their datasheets leaves me scratching my already baldening head....
Could one of you guys help me out with the ratings??

Regards
Anders

[Groovenut]

Unless I am missing something, the datasheet says the max supply current for each amplifier is 2.5mA. Looks like you have 26 amplifiers. My math makes that as 65mA max. I would add another 20% just because for safety. That's 78mA. So shooting for 100mA rated transformer seems like a legit target. IMO

[R.G.]

1. Learn to look up and read datasheets. A few ma per amplifier in an opamp is typical, but by no means universal. Some chips are power hogs, some chips use micro- if not nano-watts.
2. Go read "Power Supplies Basics" at geofex.com
3. It is entirely possible to get yourself converted to permanently dead or start fires and burn buildings down either immediately or years later by doing AC mains  wiring wrong. Really, really think about whether you already know how to do this safely. If you do, you already know that you know. If you have any question about whether you know how to do it safely, you don't.

[amptramp]

The power supply current for an amplifier in the spec sheet is given for a quiescent condition with no load.  If the op amp is driving a load, the extra current has to be accounted for otherwise adding output loads like a level control will put the load above the quiescent condition.  If you have an output swing of 9 volts and a 12 K load, there will be an additional 0.75 mA for the amplifier.

There are plenty of power supplies you can buy as packaged units.  It is quite possible to make a bipolar power supply out of two wall warts that go into isolated jacks.  There is almost no danger doing it that way and if you increase the voltage, you can regulate it down to the value you need and eliminate power supply ripple.

[PRR]

> If you have an output swing of 9 volts and a 12 K load, there will be an additional 0.75 mA for the amplifier.

Not that simple. Or that bad. It is over-conservative. Simple, which has merit.

Essentially all our audio chips are totem-pole output. When working hard, the top conducts half the time, the bottom the other half. A unipolar supply only "feels" the suckage of the top half. A bipolar feels load half the time on each side, so the same total.

For Sine-wave and 8 Ohm load, my rule o thumb is to assume the power supply "feels" a 48 Ohm load in addition to the idle. For 12K loading, this would be 12K/8= 72K load on power, and at 9V only 0.125mA added dynamic current.

However in guitar effects we might be wise to assume Square wave. Then the dynamic current of 12K is 4.5V/12K but half the time, 0.187,5mA.

Sometimes the load is not known. Most audio chips can drive 2K, maybe 1K at low voltage and loose THD demands. Then we could suck 4.5mA half the time, 2.25mA added suck.

So a "2.5mA" chip, which might be 4mA if made on a bad day, with heavy load, might be 6mA or so. A dual could run 15mA. (A quad could run more, but we don't usually drive all four amps on one chip at high level in small loads.)

In another world of +/-22V supply and true 600 Ohm loads, a NE5532 is specced 8mA typ 16mA max, but each side may drive 600 Ohms to 20V peak. As a 20V peak Square is a frightening thing, let's assume Sine. Each side may draw another 12mA of dynamic current. If both sides are working hard, we could have 16mA+12mA+12mA or 40mA in one chip.

That's audio. In to-ground DC drive, such as an opto-limiter's LED, most opamps can deliver 30mA and all of this comes from the one power pin. And some optos work this hard, at least on test-bench. (In real speech/music they don't.) So that could be 32mA-36mA just in one amplifier.

> cabinet sim from the LXH marshall

FWIW, U1 is pointless and may be omitted. Saved a few mA right there.

The bulk of the amps have, as Amptramp noticed, 12K loading. Power voltage is not specified. If as high as 12V each side, it probably will not be clipped square. So I'd allow 0.125mA dynamic current. I'd wrap this with the 1.4mA-2.5mA idle (for EACH amplifier, not chip!) and call it 3mA. 21 or 22 of these is 66mA. The final output load is not known, but 2.25mA covers a heavy load at 9V, so another 4mA seems possible, making 70mA total.

[arjespen]

Thanks for the thorough answers!
One question: You mention U1 is pointless. Can you elaborate on that?
(I believe it's a buffer - can you tell me why its pointless?)

[PRR]

The input to "U2" is infinite impedance already. No buffer needed.