Amp supervisor circuit latency

[rborer]

I am desinging a supervisor circuit for a MOSFET power amp arround a PLD device.

The circuit will be responsible for power-on delay to connect the speakers and DC and over temperature monitoring. The circuit will also indicate, via a LED, the status of the system.

I would like to use a low frequency clock to drive the supervisor circuit, arround 4 Hz. This is to simplify the PLD circuit and to be able to flash the LED at split second intervals.

Because of the synchronous nature of the PLD, any stimuli will translate to an output (like disengaging the output relays) on the next clock tick. Since the clock is slow, I can expect at most a 250 msec latency.

Question is: will a large DC at output, due to MOSFET blow, e.g., be able to damage the speakers in 250 msec? After that the supervisor will disconect them amp to the speakers.

[tiroth]

Yes.

[Nelson Pass]

Maybe.

[halojoy]

Originally posted by rborer
Question is: will a large DC at output, due to MOSFET blow, e.g., be able to damage the speakers in 250 msec? After that the supervisor will disconect them amp to the speakers.

Why not make a test?

[Bakmeel]

Originally posted by halojoy

Why not make a test?

Yes... Lets do that and make it a good old game of "catch the woofer"

[traderbam]

There is a limit to how fast you can disconnect the speakers because your dc detection needs to test for at least one cycle of your lowest signal frequency to avoid false triggering. 10Hz is 100mS.

You can do better than this by detecting the voltage differential between input and output, adjusted for the gain of the amp. Then you are limited only by the opening time of the relay.

You might consider putting reverse biased diodes from the relay output back to the supplies to prevent arcing.

[rborer]

Okay, seems like 250 msec is too slow and may damage the speakers. And anything faster than 100 msec is too fast, and may lead to false triggering unless I use a more elaborate detection system, which I am not inclined to.

I will make the circuit clock @ 10Hz and satisfy both conditions. And will add reverse biased diodes to the relays.

Once I have it all built I can take volunteers for my field test.

[mrfeedback]

250 mSec is probably quite ok, 100 mSec will be ok.

[b] You might consider putting reverse biased diodes from the relay output back to the supplies to prevent arcing. [/b}

Some Yamaha pro amplifiers have 2 small magnets glued to the outside of the output relay to help quench any arcing across the contacts, but this is required only for high power applications.

If you use double throw contacts, with the moving contact connected to the speaker, and the NC contact grounded you are safer too.
Old Accuphase amps sensed load resistance and would not operate the output relay if the speaker lines were shorted by making use of the NC contact.

Eric.

[Nelson Pass]

A small capacitor, say .22 uF, across the contacts would
be helpful for this. You can buy line voltage rated ones
from Digikey.

[Morello]

Hello sirs,

My speakers have, by misstake, been subject to more than 30 volts DC for approx one or two seconds without any damage.


Anyhow, DC are best avoided at the speakerterminals

Best regards\Morello