Chip amp Speaker Protection Ideas

[sksengupta]

Hi:

I am new to chip audio amplifiers and I am trying to build some circuits with LM1875s, LM3886s and LM4780s. My test speakers are some old EPI110s that I am rather fond of and would not want to destroy accidentally. Any suggestions for fairly simple speaker protection? I understand the National chips have fairly decent built-in protection, but I have also found out from the various forums, that the chips themselves may fail catastrophically under certain conditions, so I would like to add something external.

Any and all suggestions, advice, sharing of experiences will be very much appreciated.

Thanks in advance.

Regards,

Sandip

PS- Getting other test speakers is not an option

[EBM_dude]

A DC protection circuit like this may be sufficient for that application. Another useful curcuit could be a clipping detector.

[Sebastiaan]

Here you go: http://www.datasheetcatalog.org/data...ec/UPC1237.pdf

With kind regards,
Bas

[pacificblue]

The simplest DC protection is a capacitor in series with the speaker. Many people in the hi-fi realm shy away from that, because they fear it might do something bad to the sonics due to the higher group delay and additional parasitic components. PA speakers however often come with that kind of protection.

Here are two design options.
1) Make the capacitor as big as possible, so that its influence on AC is as small as possible. That may lead to very big capacitance and you may not find bi-polar capacitors of that size. Two normal electrolytics back to back will do the trick then. 2 x 22000 µF for 4 Ohm or 2 x 10000 µF for 8 Ohm speakers should be a good starting point. Pretty much what you would normally use in the power supply and the same voltage rating, too.
2) Make use of the interaction between capacitor and speaker to achieve a flatter and deeper roll-off. The German magazine Hobby Hifi published the following approach in its issue 1/2006 for closed boxes. C = (265000 x Qts) / ( Re x fs) [µF] leads to f3 = (0,7 x fs) / Qts. It often works with bass reflex speakers as well. While it protects those also from extreme strokes below their resonant frequency, the addition of two long group delays (BR and cap) may sometimes turn out to be too much.

[consort_ee_um]

The old TDA2030 has built-in over-current and temperature protection but when this is active the output is rumoured to go to the negative rail. If so it would put a big DC offset on the speakers when the protection kicked in.
Hence Pacificblues suggestion is a good one.
Newer chip amps collapse to ground (hopefully)

My blog:
Consort3′s Blog

[Leglandu]

A funny way would be to use a tranny at the exit of the amp (1 by side).

A standard torroidal, with 2 secondaries of the max voltage as the amp can provide. Use only the 2 secondaries as 1 primary and 1 secondary.

As there is no DC courant, the job is easy for the tranny and the response curve can be very wide.

For a 50W amp, a 100VA (@50hz or 60hz) tranny will do the job at 25Hz/30Hz but 150VA would be perfect.

Not so much expensive and perfect protection, what else ?

Disclaimer : I never tested that solution by myself and I found no link about it ; so be careful...

[nigelwright7557]

I use a PIC microcontroller on my designs.
It holds off teh speaker relay for 3 seconds on power up.
It also detects DC on the out and switches off the speaker relay if DC is spotted for more than 500mS. So far it has saved me a fortune in speakers.

[AndrewT]

Originally Posted by pacificblue

The German magazine Hobby Hifi published the following approach in its issue 1/2006 for closed boxes. C = (265000 x Qts) / ( Re x fs) [µF] leads to f3 = (0,7 x fs) / Qts. It often works with bass reflex speakers as well.

I wonder if this is what KEF do with their capacitor coupled bass drivers?

[Mooly]

Originally Posted by Leglandu

A funny way would be to use a tranny at the exit of the amp (1 by side).

A standard torroidal, with 2 secondaries of the max voltage as the amp can provide. Use only the 2 secondaries as 1 primary and 1 secondary.

As there is no DC courant, the job is easy for the tranny and the response curve can be very wide.

For a 50W amp, a 100VA (@50hz or 60hz) tranny will do the job at 25Hz/30Hz but 150VA would be perfect.

Not so much expensive and perfect protection, what else ?

Disclaimer : I never tested that solution by myself and I found no link about it ; so be careful...

Test it yourself

It might be OK for general purpose use but it can only degrade the output, probably substantially. There would also be a significant shock hazard from the unterminated "real" primary.

For testing what could be simpler than capacitor coupling... and sonically it's far more transparent the transformer... and far cheaper too.

And ultimately you should have confidence in your designs to be able to DC couple in confidence.

[Ted205]

Originally Posted by nigelwright7557

I use a PIC microcontroller on my designs.
It holds off teh speaker relay for 3 seconds on power up.
It also detects DC on the out and switches off the speaker relay if DC is spotted for more than 500mS. So far it has saved me a fortune in speakers.

interested to know how you have interfaced the pic to the speaker (DC detection)