Biasing battery-powered opamps

I'm designing a portable low-level audio circuit to be powered from a 9V battery. I'm aiming for quality that's considerably better than cheap public address gear but not quite audiophool perfection. It uses eight low-noise opamps (4 duals) and involves balanced-to-unbalanced conversion and vice versa, some gain, volume controls, etc.

My question here is about the technique for setting the 0.5Vcc opamp input bias. I could use an opamp voltage divider but that would mean adding an extra IC. Cost is not an issue, but I'm rather partial to simple elegance. Would a parallel R-C divider be good enough or could it raise some quality or stability issues?

Reply to
Pimpom
Loading thread data ...

Don't you mean R-R divider? It is of course cutomary to decouple the centre point to ground with a suitable capacitor.

Many guitar effects I've seen have a pair of 100k resistors divider for a couple of op-amp stages, you could reduce that to 56k if you have a lot of stages.

If you want to use an op-amp stage as a voltage follower 1/2 Vcc buffer, you need a resistor between its output and the 1/2 Vcc rail decoupling capacitor as well as decoupling on the divider that provides the 1/2 Vcc input.

Reply to
Ian Field

lic

t
.
y

I would guess it only depends on how much current is drawn from the reference, and if it changes a lot.

George H.

Reply to
George Herold

Yes, of course. I used the R-C term, perhaps incorrectly, to mean an R-R divider in which each resistor is bypassed by a cap.

I was thinking of using a pair of 10k resistors to swamp the combined input currents of the eight bipolar opamps which could run to several microamps worst-case. My application is rather more demanding than a guitar effects unit in terms of noise, sensitivity and audio quality. (I _am_ also designing a guitar combo with some built-in effects for my son, but that's a separate project).

That's the alternative I mentioned. I was wondering if I could get away with a simple resistive divider without reducing stability or audio quality.

Reply to
Pimpom

If you're using one of those little block 9V batteries (especially the layer cell construction) the Ah capacity is surpisingly low, so you will want to make the divider resistors as high as you can get away with to conserve on current draw.

The quality of the decoupling capacitor is of paramount importance - some people swear by tantalum bead, whatever you use it never hurts to also add a mylar, say 0.1 to 0.22uF.

Reply to
Ian Field

Current drawn will be the input bias currents of the opamps which could be up to several microamps, worst-case. Variations will be a fraction of that at audio frequencies in the presence of a signal, and much of that will be smoothed out by capacitors in parallel with the divider resistors. I'm thinking of using two

10k resistors, each paralleled by about 470uF. I'm not concerned about small changes in the DC level. What I was wondering was if there's any possibility of oscillations or unwanted feedback because of the shared bias point which will have a finite impedance at audio frequencies.
Reply to
Pimpom

Sure. If the V/2 bias voltage just drives non-inverting opamp inputs, there will be almost no signal current. But if you return feedback networks to the rail, you could potentially pump signal current into the bypass cap, and that might cause problems if there's a lot of gain lying around.

Use a big cap!

I sometimes use a c-load opamp, like LM8261, to make such rails.

John

Reply to
John Larkin

What 'ground'? There's no better ground reference than that centre point, is there? That's what I'd strap to the case and shield components.

Reply to
whit3rd

Yes, I mean just that kind of 9V battery and the low Ah capacity caused me to keep revising my design until I could eliminate two opamps (one IC) from the ten I originally laid out - without sacrificing quality. The ICs are NE5532, the only type I know of (with good AF characteristics) that can operate from a partially drained 9V battery. It's closely similar to the LM833, but whereas the LM833 is specced for a minimum supply of +/-4V, the NE5532 is specced down to +/-3V. The main drawback is that it draws about 7mA at 9V against 4.5mA for the LM833. Four ICs will draw about 28mA, so I'm not too concerned about an additional

0.5mA through the voltage divider.

I was thinking of 470uF Al electrolytics paralleled by 0.1uF ceramic discs. I read somewhere a long time ago that ceramics are less prone to oscillations in some filter applications because of their higher loss.

Reply to
Pimpom

You might run into leakage problems with an Al electro that large.

AFAIK the losses in a ceramic will be insignificant until you get to RF, presumably that won't be a problem as you'll have designed in roll-off to keep RF out of your audio stages.

Reply to
Ian Field

Your post caused me to check my design and I have only the non-inverting inputs tied to the V/2 point, no feedback network. Gain is x10 at the input stage and x1.5 at a later stage, with a volume control in between. Other stages are unity gain.

Reply to
Pimpom

Personally I'd go for local decoupling for filter ground points, say 0.22uF metalised film or dipped leaded ceramic chip.

IIRC the bigger the electrolytic; the higher the leakage and more chance of significant parasitic inductance.

Reply to
Ian Field

OK, maybe I should start out with smaller caps and see what happens.

Yes, I have RF filters at the inputs and roll-off in some of the feedback networks.

Reply to
Pimpom

--
Can you post a schematic?
Reply to
John Fields

OK, sounds safe to me. A couple of 50K sort of resistors and a cap to ground, roughly 1 uF or so, should be fine. Bypass the battery, too.

John

Reply to
John Larkin

At audio frequencies, it shouldn't matter. Turns out his overall gain is low, too.

John

Reply to
John Larkin

There are a lot better op-amps available - similar performance and lower current draw.

Reply to
David Eather

"Pimpom"

** Use the two resistor voltage splitter to provide +/- 4.5 volts - and put a 1N4004 diode in reverse across the battery terminals. You will need it.

Then a couple of 6V, 100uF electros from the rails to ground will be all you need, as the op-amps have a very good PSRR.

.... Phil

Reply to
Phil Allison

Can you give some examples? Constraints imposed by my location usually force me to use only devices that have been around for years and are widely available. But one of your suggestions may fit the bill.

Reply to
Pimpom

Good point, thanks. And perhaps a 0.1A fuse to blow in case of reversed battery polarity.

OK. So the general consensus is that there's little danger of instability with that arrangement.

Reply to
Pimpom

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.