simple protection for low wattage speakers?

First: do you have a DC blocking cap in series with the speaker? If not add one.

Asduming you do have a blocking cap in the circuit:

For playing around use 5 volts Vcc for the low wattage speakers, and put parallel protection1 (shown below) across the .25W speaker and the same for the .5W speaker EXCEPT this time use parallel protection2 (shown below). The idea is to keep the maximum voltage the speaker gets within its ohms law computed ratings. Use 1 amp rated diodes.

parallel protection1 parallel protection2

-+-[<D1]-[<D2]-+- -+-[<D1]-[<D2]-[<D3]-[<D4]-+- | | | |

-+-[D3>]-[D4>]-+- -+-[D5>]-[D6>]-[D7>]-[D8>]-+-

You might want to add something to the process of playing around. For example as you increase the amplitude of the signal applied to the speaker toward a level beyond what is "healthy" for that speaker, distortion will increase. That distortion is a good clue that you are approaching the soa for the speaker.

For the 3W speaker you can use back to back 4.7 volt 5W zeners.

See above. Zeners won't do it for the .25 and .5 watt speakers.

Well, if your 386 is blowing the 3W speaker something is surely amiss, in my opinion, beyond just protecting the speakers. I don't think a properly operating 386 can do that.

Ed

They are called fuses. They are cheap and effective. Look into it.

For 1 watt @ 8 ohms, for example, you want .35A in-line fuse or the next size up from there.

For 8 ohm speakers, Min fuse size = sqrt(twattage/8)

It would be easy to reduce the power supply voltage to the amp.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

One of the benefits of a single-supply rail audio amplifier is that you have to couple through a capacitor, usually of some 1000uf, to the speaker to remove the half-rail DC. The capacitor also protects the speaker since it does not pass DC.

I can still remember, from around 50 years ago, what happens when a fault in a dual-supply rail amplifier connects the speaker directly across one of the rails. It's a kind of crinkly sound followed by a burning smell.

When doing anything experimental with a speaker you don't want to lose, always include a series capacitor, even if the output is 0v with no signal.

You can still damage a tiny speaker with enough drive but it's harder.

Most things that you do AT the speaker will have some effect on sound quality (you've not stated how important that is).

Why not limit the POWER to your circuit so it can't be coerced into delivering more power to the load than the load can handle?

In case you're not already aware, LM386 is not designed for direct connection to a speaker. Note where the data sheet says "Self-Centering Output Quiescent Voltage". Note also that all the example circuits have an output series capacitor.

This is not like your average op amp where the output DC level is the same as the inputs (usually 0V with + and - supply rails but doesn't have to be).

The other thing to remember at least with bigger speakers is that their power handling is for music and you can comfortably kill one by applying less than rated power pure sine wave at certain tricky frequencies where there is a mechanical resonance in the speaker coil and cone suspension.

Same mechanical resonance effect that allows a good opera singer to destroy a wine glass although potentially a lot more expensive.

Smaller speakers tend to be remarkably resilient unless you do something silly to them. The suggestion of a fuse might work but it needs to be a low voltage drop automotive one to stand a chance of working at such low powers. I'd be inclined to put an extra resistance in series or make the amplifier logarithmic to limit the maximum possible current.

Thank you, would you mind please explaining a little more about these parallel protection circuits, i.e., how the protection works?

It's depressingly simple-minded. The current through a silicon diode is an exponential function of the current though it.

Stick two strings of diodes in parallel with your load, one string to limit the positive voltage swings and the other to limit the negative voltage swings, and if you apply an over-high voltage, the bulk of the current flows through the diodes rather than the load.

Get the data sheet for whatever diode you use and you can work out the rough numbers. Most diodes aren't all that tightly specified, but some are, and if you insist on using cheap diodes at least fix on a single part number and measure what they actually do. It is temperature dependent, but nothing all that dramatic.

While true (if we correct the error which is so obvious it doesn't matter), it's probably about the last thing I'd think of when trying to protect a loudspeaker.

I'd suggest that the OP puts a resistor in series with his capacitor. Better to watch a resistor go up in smoke rather than a speaker. If it's an 8 ohm speaker try a beefy 10 ohm resistor and if you can still hear what you want through the speaker then you have greatly reduced the chances of damaging your speaker.

That wrecks the "damping factor" but most people won't hear the difference.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Sure. The diodes act as simple voltage limiters.

General purpose diodes conduct when the voltage across them rises to about .7 volts. 2 diodes in series will conduct when the voltage across the series pair is about 1.4 volts. Therefore a speaker with a series pair of diodes across it won't be exposed to more than roughly 1.4 volts, provided of course, that the diodes are installed in the "right direction". Since a speaker usually is exposed to an AC voltage, you need two diodes in one direction and two in the other direction, so no matter what polarity signal is applied to the speaker one or the other pair of diodes will limit the voltage.

Two pairs of Si diodes will limit the audio (sine wave) to the speaker to 1.4 V peak to peak or 1.0 V RMS. Into an 8 ohm loudspeaker, that's a fabulous: Power = E^2 / 8 = 0.13 watts RMS My home entertainment system claims it can deliver 50 watts[1], which will probably vaporize a simple diode voltage limiter.

Hint: I haven't seen any commercial hi-fi amplifier schematics that include such diode clipper speaker protection.

On 6/03/2026 4:12 am, Christopher Howard wrote: On 6/03/2026 4:12 am, Christopher Howard wrote: >>

It's depressingly simple-minded. The current through a silicon diode is an exponential function of the voltage across it.

Stick two strings of diodes in parallel with your load, one string to limit the positive voltage swings and the other to limit the negative voltage swings, and if you apply an over-high voltage, the bulk of the current flows through the diodes rather than the load.

Get the data sheet for whatever diode you use and you can work out the rough numbers. Most diodes aren't all that tightly specified, but some are, and if you insist on using cheap diodes at least fix on a single part number and measure what they actually do. It is temperature dependent, but nothing all that dramatic.

One could use a JFET as input limiter driven by some diode voltage detector at the output. Like mike auto gain circuits do.

First answer (circuit example) from:

A bridge rectifier has 4 diodes in one can.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

The limiter I described addresses the OP's application, not yours: "I feed in a small signal, from my signal generator or from my analog computer, into my LM386 amplifier (20x gain) and then to the speaker."

Nor I, but irrelevant to the OP.

Ed

Trouble is, you can't use all 4 diodes in the bridge as a back to back diode bi-directional shunt limiter.

Ed

Of course you can.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics