power/battery question

[famous beagle]

I'm an electronics newbie, so forgive me is this is 101 stuff. I'm reading everything I can find and building projects, but I'm still having trouble understanding how it all comes together. I mean, I know Ohm's law, and all the permutations of that and all, but just because I know that 1 + 1 = 2, I'm still not entirely sure where the 1 came from and why it's being added to the other 1, etc.

Anyway, I was reading through Craig Anderton's classic Electronic Projects for Musicians book (anyone else notice that book has a period at the end of the title for some reason? ), and I have a question.

I was wondering why he uses 2 9V batteries to power most of his designs. I think I may have figured out why, but I'm not sure. I read in the earlier part of the book that most of the projects are designed to work for line-level instruments (like a keyboard), instead of high-z instruments (like a guitar). He mentions that you can adjust a resistor value if you'd like to use high-z instruments with them.

So is that the reason for the increased power supply? Because they're supposed to operate at line-level?

And if so, why exactly is that? (This is where my understanding falls short.)

Thanks to any who can help me out with this.

 

[mshilarious]

Because he's using the 2 batteries as a split-rail power supply. When you ground the connection between the two batteries, that gives you a +/-9V supply. Opamps like to use a split supply. Strictly speaking, that's not true; opamps don't have a ground reference at all. Look at an opamp sometime, there is no ground pin. All they care about is the difference between their inverting and noninverting inputs.

But the incoming signal has to have a reference to the power rails of the opamp, and ideally that is halfway between the power rails. You can run an opamp off a single-sided supply (one 9V battery); I do that in practically all my designs, because I am using phantom power or a 9V battery (without room for a second). That requires a lot more parts, because you have to create a suitably low-impedance half-rail bias supply for the inputs, and make sure that you eliminate DC offset between stages, etc.

To avoid that, Anderton uses two 9V batteries and there you go. Rackmount gear will use a transformer with a center-tap secondary, usually for +/-15V, sometimes higher. Higher supply voltages get you more headroom, but you have to mind the maximum rating of the opamp (usually 36V to 40V).

Lots more here:

http://focus.ti.com/lit/an/sloa058/sloa058.pdf


It doesn't have much to do with the resistor that sets input impedance, other than the fact that resistor is probably the bias supply in a single-ended design.

A common misconception among newbs is that line-level uses a lot more current than mic level, or instrument level, or whatever. It uses a little more, but it's still trivial. Let's say you have a whopping +20dBV signal--that's 10VRMS. Sounds big, right? OK, Ohm's law - into a nominal 10K line-level load, that's 1mA. Even into 600 ohms, it's still only 17mA. 17mA is a wasteful amount of power; that's why you don't see 600 ohm loads much. But still, no sweat for a 9V battery (except that it doesn't have enough volts!)

I do a headphone amp off a 9V battery that can pump 50mA into 24 ohms . . . still only 60mW (that's plenty loud though), not too much of a challenge for the battery. A 9V is 625mAh, or about 5Wh, so it could drive that headphone for more than three days . . .

 

[dementedchord]

been a long time since i looked at that book... but as i recall it's pretty much all op-amp based projects... most op-amps arent haqppy with real low voltages so 2 9V allows for a good balanced supply for a chip....

 

[famous beagle]

Thanks for the lengthy reply. I guess I'm still more of a newb than I thought, because most of this is still over my head at this point. But thanks for the information. I can't understand it yet, but I'm sure I will eventually!

 

[mshilarious]

Thanks for the lengthy reply. I guess I'm still more of a newb than I thought, because most of this is still over my head at this point. But thanks for the information. I can't understand it yet, but I'm sure I will eventually!

Sure you can! Just consider that all voltage is relative. One 9V battery gives you the ability to have a signal between 0V and 9V (roughly speaking). But what is 0V and 9V relative to? The battery, of course! For a portable device, that's all you care about.

Now let's say you want to interface with the outside world. Rack gear is going to eventually tie signal ground to earth ground. OK, now your battery's 0V is the same 0V as earth potential (again, roughly speaking).

Now let's consider an audio signal. It's AC, right? That means it's going to vary positive and negative around some point. Tie your guitar to earth ground, and that 0V point is earth potential, and your guitar will swing positive and negative around that.

Except you want to introduce a battery-powered stompbox between your guitar and your amp. Hmm. The battery can't go negative with respect to earth if the negative terminal is tied to ground. So what do you do?

Answer: you introduce a DC bias to your AC guitar signal. Every guitar pedal that only uses one 9V must do this (which maybe means Anderton is doing a disservice to his readers, if they want to build 9V stompboxes).

So you create a half-rail bias. How? The easiest way is with a voltage divider--two resistors of equal value, in series, from +9V to 0V (plus a couple of caps). That gives you a 4.5V reference level. You couple the incoming AC signal across a capacitor, and use a resistor to carry that DC bias to the opamp input pin. Et voila, your guitar signal is now (let's say) no longer +/-0.5V, it now ranges from +5V to +4V. Same AC voltage, riding on a +4.5V DC bias.

Now your opamp is happy, because the signal is within its power rails.

Your amp would think that signal was a bit strange, because it isn't expecting a DC bias at its input. That's OK, because both your stompbox and the amp will have DC blocking caps at their inputs and outputs, because it's very bad form to let DC leak out of a device, while at the same time you have to assume horrible things are coming at your input from the outside world, so the amp has input caps too.

The amp will have a split-rail supply (unless it's a tube amp, then it will be biased too), so it is happy with a pure AC signal.

 

[sweetbeats]

HAHAHA! (that's real laughter...the joyful kind)

MSH!

Your post actually makes a lot of sense to me! I've been dialoging with Rick Chinn of Uneeda Audio and he has pointed me to a number of articles that go over opamps and between past posts of yours, demented's and evm1024's (and others...sorry if I'm leaving you out) it is actually starting to make sense!!

I LOVE POSSESSING GEEKY KNOWLEDGE!! Fun and useful too when you have a bunch of repair projects sitting about some with no schematics...

Sorry for the OT post...

 

[dgatwood]

That's OK, because both your stompbox and the amp will have DC blocking caps at their inputs and outputs, because it's very bad form to let DC leak out of a device, while at the same time you have to assume horrible things are coming at your input from the outside world, so the amp has input caps too.

I love how you worded that. Kind of reminds me of how we describe good security practices in computer programming.... First, make sure any data your code outputs is reasonable to minimize the risk of causing a crash in some other code. Second, assume that any data you are passed is designed specifically to cause your code to crash and check it accordingly.