Speaker Protection, new method?

[gmphadte]

Will this work?

Considering the speaker expects 2 signals
1. the DC offset at the output when the amp is healthy and idling
2. an audio signal that is crossing zero.

By this I understand that 2 simple circuits should do the job
1. A low (safe) voltage detector
2.A zero crossing detection every 50mS or less.

One of the two should be happening, then the speaker is safe.

This method should protect the speakers in any case.
Am I correct. shooooot...

Gajanan Phadte

[AndrewT]

Hi,
that does not sound right.

Sampling at 50mS intervals might miss the event. Worse you could end up sampling in between a series of catastrophic events

Low DC offset is not a severe problem. High DC is a problem that can only be tolerated for a relatively short time.

[gmphadte]

AndrewT wrote

Sampling at 50mS intervals might miss the event. Worse you could end up sampling in between a series of catastrophic events

looks like I have not framed it in right words

sampling will not be done. Only zero crossing detection and the absence of this zero crossing will raise an alarm. Considering the lowest frequency to be 20Hz, if the crossong does not occur for more then say 60mS or more, then this raises an alarm and activates the protecion whatever it may be. The zero crossing will keep happening and varying depending on the audio content and will be much faster for 20KHz.

Coming to a tweeter protecion - If the lowest freq is 4KHz the alarm should be raised if the zero crossing does not occur within the desired time period. Raising alarm be done little later. In this way a tweeter protection can be actvated much faster.

The circuit will keep watch on every consecutive zero crossing in the audio content.

In case of no audio input

...speaker expects 2 signals
1. the DC offset at the output...

I should have written
...speaker expects 2 signals
1. the LOW DC offset at the output...

Gajanan Phadte

[wintermute]

What happens when you have a quiet passage with effectively no signal??? won't the protection cut in?

Also what happens if the amp is clipping severely?? you will still have a zero crossing but your tweeters will most likely be toast

Tony.

[CBS240]

Kind of thinking on the fly here, but couldn't you sense output collector current, reference this as a voltage and compare to output voltage? Set up comparison limits based below SOA limits, of course taking load reactance into account. This way if you have a speaker short, lots of current with little voltage would trigger the circuit, or if biasing fails and saturation occurs.


I will have to think of how this could be done....

[anatech]

Hi CBS240,
That sounds like what some protection circuits do, except we look at derived emitter current. Emitter current = collector current + base current. The error isn't large so this is the current art.

-Chris

[bjornagain]

Protecting the speaker is one thing, protecting the amp is another, and both may be needed. For protecting the speaker I think a zero-dector can work. Output must be zero (within a few mV) or pass zero frequently. Simple and nice!

[anatech]

Hi bjornagain,
I would place speaker protection as a #1 concern. Protecting the amp is best done by shutting off the power supply. Latching is the best.

Then the amp can be repaired and it is born again!

-Chris

[bjornagain]

Agree, and current sensing can be done at the power supply, which will protect the supply from a faulty amp. A fuse may be the simplest of things. A eletronic fuse can interact with the speaker protection.

[anatech]

Hi bjornagain,
Yes, and we are back to the way many manufacturers do it. Simple and effective.

-Chris