14

Do these two input signals have different effects on a speaker?

  • An AC signal with some value of DC offset.
  • Same AC signal but centered at 0 V.
joribama
  • 2,187
  • 7
  • 12
muyustan
  • 2,046
  • 20
  • 54
  • 6
    Yes, they will have different effects. For example, a DC bias will pre-load/tension the speaker in some direction. This might be required for some types of speakers, but not for others. – DKNguyen Nov 29 '19 at 22:43
  • 8
    DC reduces the linear range of coil and wastes heat – Tony Stewart EE75 Nov 29 '19 at 22:43
  • don't know if that is state-of-the-art nowadays: HiFi amplifier designs used to include a series capacitor to protect speakers from final stage misconfigurations leaking DC, and sometimes a delay circuit to suppress any output ~5 seconds into power-up to prevent surges with significant DC from reaching speakers. – dlatikay Nov 30 '19 at 11:26
  • "a delay circuit to suppress any output ~5 seconds into power-up" isn't it just for better comfort? You cannot actually burn an adequate speaker with a few seconds of unpleasant noise on power-on (well, you can burn an inadequate one). – fraxinus Nov 30 '19 at 13:58
  • 2
    DC causes unwanted heating in the voice coil and may cause burnout or a fire. – user207421 Dec 01 '19 at 09:11

2 Answers2

34

Yes.

enter image description here

Figure 1. Cutaway view of a dynamic loudspeaker for the bass register. 1. Magnet, 2. Voicecoil, 3. Suspension, 4. Diaphragm. Source: Wikipedia.

A DC component in the signal will cause:

  • Bias of the speaker cone (4) and the suspension (3) from the neutral position. This means that the cone will reach the limit of its travel - or at least where it becomes highly non-linear - in one direction in a shorter distance compared with the other.
  • Maximum undistorted volume will be decreased as a result.
  • The coil (2) will be heating even when no audio is playing.
Transistor
  • 168,990
  • 12
  • 186
  • 385
  • 4
    Note that the DC current will be determined my voltage and speaker resistance not speaker nominal impedance which assumes inductive effects and back emf. Usually the speaker has a dc resistance of about half the nominal impedance. One sometimes useful trick is to introduce a deliberate small DC component and then use this to measure the voice coil temperature. – Dan Mills Nov 30 '19 at 13:16
  • 5
    The last point about heating brings up also the problem of power loss, which is specially important if the device is mobile, uses batteries or if it's always on. – Ronan Paixão Nov 30 '19 at 19:19
  • 2
    I've had more than one speaker go up in smoke from an accidental DC bias. That heating is not only a nuisance! – Hearth Dec 01 '19 at 20:38
6

There is no modern (read: solid-state, no transformer at the output) audio amplifier that doesn't leak at least some DC. It may be in mV range for a 100's W rated output at 4 or 8 ohm, but is measurable even when there is a capacitor at the output (it is usually electrolytic and usually leaks some dc).

Keep in mind that some high power low frequency speakers actually need the movement in order to cool its coil. A 100W rated "woofer" will probably burn from as low as 5W DC.

The other answers regarding wasted power and shortened travel of the cone are also pretty much true.

Edit: important to say, we are talking about the traditional speaker consisting of a coil glued to a cone and a permanent magnet. Piezo, for example, is way more tolerant to DC voltage (it still can be burnt or pushed outside of it's linear operation, but you have to try harder and apply a lot of DC, generally above the AC rating of the device).

Edit2: It is perfectly acceptable and widely used to drive a small and rarely-used speaker (like an interface beeper) with AC + DC (actually, turning it on and off with the frequency needed, so it gets either Vcc or 0). Just remember to leave it in "off" state when not needed or the the battery will go down. The speaker gets roughly Vcc/2 DC + Vcc/2 peak AC - as inefficient as it gets, but the beeping should be rare and short anyway. You can even use PWM or some A class amplifier and out some complex sound this way - that's how some kids toys speak and play music. In these cases, your "amplifier" is really simple - a single transistor connected to some GPIO of the controller.

This is also how the early wired telephones worked - generally, a speaker and a microphone (graphite, so it needs DC bias) connected in series. The DC bias for the microphone goes thru the speaker, but the speaker is designed to tolerate it. (Actually, no design efforts were ever made - these speakers operated well below their thermal limits and no one really cared about the sound distortion).

fraxinus
  • 8,726
  • 11
  • 34