Protection on headphone amps

[counter culture]

I'd start by looking at a low pass filter to read the OP DC levels and then compare this with positive and negative reference voltages, using the result to open a DP relay to interrupt the outputs. You could generate the references with forward biassed diodes.

47k + 3u3 capacitance will give ~1Hz and won't load the output significantly, or you can go to 1M + 160n.

Of course you have to power the relay all the time when the amp is on, but you have no choice because you don't know you're going to have a reliable supply in a fault condition.

You need a comparator or inverter (per channel) to give a normally high output going low if the offset is negative going and obviously one per channel for positive going. One low-power 4-device chip that will tolerate the output being short circuited (LM2902?). Lastly you need to clamp the outputs so they don't drive negative into the next stage. You arrange the outputs so that they’re normally high so that you’re not wasting a lot of current through the clamps.

I'd probably use a PIC12F508 to bring all the control signals together, you need to implement a 4-input AND gate, and it could do a switch-on delay too. If you use a PIC it will tolerate -0.3V on the inputs which ordinary logic won't like, but you need to pick the right Schottkys for the clamps. Much easier in SMT, but opamps don’t drive a lot of current anyway. One pin as output to control a transistor with the relay coil in the collector circuit, don't forget the freewheeling diode. If you pick the point where you take the relay power from you can probably make it disconnect before any switch-off transient occurs, in fact if it's a serious failure of either power rail the relay should drop out anyway.

You’d really need a +/- 5V supply for all of this or be forced to clamp the OP highs with zeners where they go into the PIC.

All a bit belt-and-braces, and power hungry, if it's a portable, although the majority of the current will be in the relay.

The problem is, will it switch quickly enough to save your headphones? If you set the filter's cutoff at 1Hz then the time constant (tau) is 159mS, that is:- a DC step at the input will take 159mS for 63% of it to appear at the output. You could pick 20Hz with a tau of 8mS plus a millisec or so for the relay to operate, but is a hundredth of a second quick enough to save your headphones from 9V DC?

¿Quién sabe?

The other issue is that the 20 Hz number only indicates the 3dB point of the filter, so a large amplitude bass note may trigger the cutoff, depending on where you set the threshold.

I've looked for information on this subject, I found a speaker protection circuit that triggers at 2V offset after 5 secs. I found a thread on headwize that said 1V for half a second was probably OK for phones. The good news is that these circuits are used all over the place so they're probably not without value.

[Mooly]

Originally Posted by jaycee

I also remembered a comparator circuit of sorts from an Arcam amp. I simplified it a bit and this is how it turned out. It seems to work fairly well.

Notice the relay is across both supplies - so if one supply fails, the relay will shut off anyway!

Looks workable Very similar to the scheme I have in my main amp.

[nereis]

Just a thought, your headphones might not yet be beyond repair if you'd care to open them up.

I had similar thing happen to me once and was ready to send it back for repairs, however the price I was quoted for out of warranty repair was simply too much to bear (could have gotten new phones with that money)

I decided to open them up and it turns out that the large DC bias had just "pushed" the drivers to a non-linear operating position without actually damaging the coils, so VERY CAREFULLY I applied some pressure to the driver and snap the transducers sounds fine again!

I suppose this might not work with every kind of DC induced trauma depending on polarity, but about half of the time this could remedy your phones. I would think a bit of scotch tape might save a "pushed-in" transducer, however I have not personally tried such repairs.

Another thing I notice is that it may be beneficial to feed your phones with a low frequency sine wave while attempting this trick, let some of that electromagnetic energy help you setting the transducers back in place!

DISCLAIMER: This may fudge up your phones to beyond repair, send it back for repair if still within warranty and ONLY try this as a last resort if none else worked!!!

Originally Posted by jaycee

Yeah I thought that. In sim I got a dramatic rise in THD with the 120 ohm resistor, and the amp sounded better without it. The Beyers are rated for 50mW max.

Mine got hit by 1.8W if I've calculated correctly! I'm surprised the voice coils are even intact - the 'phones still work but they sound absolutely awful now, especially the right channel. I knew I was screwed when i turned it on and they went CLICK.