# Driving an audio speaker design questions

I have been dealing with some project in which I have some requirements to accomplish. For now, I am dealing with the first part and a quick story is,

• I have to generate a sine wave at a certain frequency and then convert that signal into sound via using an 8 ohm speaker.

• At least 1 W of power on the speaker.

• Adjustable amplitude of the sine wave input of the speaker.

So what did I do:

I am using a wien-bridge oscillator circuit which looks like this to obtain the sinusoidal waveform. And what I get is a nice 16 Vpp (I assume 2 Vpp loss can be considered as normal) sinusoidal-ish waveform while no load is connected to the output of the opAmp. However, when the switch is being closed, i.e 8 ohm load(my speaker in this case) is connected to the opAmp's output, then the voltage on the speaker dramatically drops to 0(nearly 0) volts.

Now, If I did not do the math wrong, I need at least 4 Vpp on the speaker to reach required power rating. After all these, my need reduces to being able to have an adjustable 8 Vpp to 4 Vpp load voltage.

So, how should I proceed? I have thought of using a voltge divider network at the output with a potentiometer and a fixed resistor to tune the output voltage but I don't know that works, or if that can work then how should I arrange the values of the resistor and potentiometer so that I can safely drive an 8 ohm speaker.

p.s: Using job-specific IC's which will bypass design steps and give directly the desired outcomes are not allowed.

Allowed components:

Mosfets, Bjts, Capacitors, Resistors, Potentiometers, DC batteries, opAmps, 8 ohms speaker, Switch, button, Wire(needed to write after some brutal comments..)

• What is "etc", and what is a "design-specific IC"? Because this sounds easy with very general ICs. Nov 27, 2019 at 18:28
• LOL, sounds like a homework project, as in design with transistors. What have you tried so far?
– user105652
Nov 27, 2019 at 18:36
• add all other assumptions such as available power sources , caps, DC blocking to speakers, availability of discretes, like Darlington power transistors or Power FETs, pots or just simulation of Sin Gen, voltage control and and current gain to drive 8 Ohms, Dual supply , batteries only with ESR? or single supply 5V? Nov 27, 2019 at 19:19
• specs are critical.. Power source, voltage , ESR? e.g. lab supply near 0. 9V battery ESR ~ 1 Ohms min Allowed Sine distortion 1%? 10%? Square Wave? triangle wave? Nov 27, 2019 at 19:33
• Here's a discrete speaker driver from Pot onward. that might work for you. usefulcomponents.com/main_contents/projects/… Nov 27, 2019 at 20:35

Two things.

First, calculate the peak current needed to produce 1 W in 8 ohms, and compare this to the maximum continuous current rating for the opamp. I think you will see a problem there. While it is possible to make an audio power amplifier stage be both the signal oscillator and the speaker driver, a more conventional approach is a low-power signal source followed by a volume control and a small audio power amplifier to drive the speaker.

Second, there is a problem with the oscillator circuit. Search for Wein bridge sinewave schematic and you will see many examples. Most of them have a light bulb or a FET in the negative feedback loop to stabilize the output amplitude. The problem with the circuit you have is that in order to maintain oscillation at a constant output amplitude, it needs to have a very precise gain that does not drift with temperature. The standard solution is to add a method of using the output amplitude to control the circuit gain such that when the output amplitude increases, the gain is turned down (and vice versa). Both TI and National Semiconductor (before they were acquired by TI) have excellent app notes on sinewave generation techniques.

• @thanks, for 2nd paragraph, I am aware of that situation. However, I was assuming that not having that stabilization mechanism will not affect so much other than clipping the sine wave a little bit, which was I could tolerate. I saw some examples using tungsten bulbs but I don't know how can I use them in such an electronics projectc. About the 1st paragraph, I have to design the sine wave generator myself, so I cannot use signal source. Does using an amplifier at the output of the opAmp provide a good solution? Nov 27, 2019 at 18:43
• I wasn't suggesting using an external source; I used the term "signal source" to separate the oscillator function from the speaker driver function. There are other simple sine circuits (like the phase-shift oscillator) that are inherently stable with little distortion. Nov 27, 2019 at 20:30
• For an output amplifier, look into the LM386. Nov 27, 2019 at 20:32

simulate this circuit – Schematic created using CircuitLab

Here I use two LED's at low current as a zener. Red & White or whatever just to get Vcc/2 roughly.

Note Falstad's Simulation uses ideal Caps, Op Amps etc so Zout =0 and goes rail to rail so you cannot expect simple Op Amps to drive < 100 Ohms as they often have 200 Ohms series R and/or active current limiting.

An Op Amp is limited in current and needs to be buffered. But with high open loop gain the Vbe drop is compensated with high OA Gain to reduce voltage error for this simple current buffer.
Av+ = (1+ R2/R3) * Ratio of Pot (0.5% to 100%)

• That's a dangerous circuit. A complete newbie will assume that the op amp can be powered by the same battery as the output stage, and that will not end well. Nov 27, 2019 at 20:17
• Yes the battery resistance and effects on Op Amp may affect bias and stability Nov 27, 2019 at 20:33
• Sorry, but I was referring to the use of a common ground point for signal and power, resulting in clipping of negative half cycles, rather than establishing a virtual ground. Which you have now done via your LED and diode. Although you should probably specify LED color, since a (for instance) red LED won't give you the voltage you're suggesting. Nov 27, 2019 at 20:37
• On falstad I have chosen non ideal opamp and one of the two options was lm741. Nov 27, 2019 at 20:39
• @muyustan very good but need to buffer current Nov 27, 2019 at 20:42