Amplifier DC protection

[FdW]

All,

This is a very simple thing that I am proposing.

For DC protection in an amplifier, that has an DC-servo circuit using an OpAmp, it would be possible to connect an reed relay (like the one below) to the output of the OpAmp, when DC is detected in large quantities the reed-relay will be activated and with the output of the relay you can deactivate the amplifier.

http://www.cotorelay.com/images/stories/2900/2900.pdf

The questions are:
Has anyone tried this before?
Did it work?
And if not, why not?

Regards,
Frans.

Post #22 Getting near the theoretical limit Amplifier DC protection
Post #17 The new current design can be found at Amplifier DC protection
Post #12 Current design can be found at Amplifier DC protection

[tauro0221]

Hi,
I am working something similar what your are suggesting. See this link to my tread and see it is the same.
Protecting the speaker output using a microprocessor

[LineSource]

FdW,

A window comparator which can detect preset lo/hi DC limits on the output is sometimes added to the DC servo circuitry. Any DC outside of the window can trigger a relay to either open(isolate speaker) or close(shunt 1 ohm in parallel to speaker), in addition to playing a role in a total power up logic.

Safe power up logic can include:
1) short input to ground until safe power-up is proven
2) allow input front-end circuits enough time to power up before the high current output power supply slow-start resistor is shorted by a relay.
3) either open up a relay in series with the speaker, or shunt a 1 ohm safety load in parallel with the speaker for protection
4) if DC is detected during operation, the output power supply relay can be opened
5) when power is turned off, shunt input to ground and shunt speaker to 1 ohm resistor to avoid "THUMP"
etc...

[Bigun]

some of these look pretty complicated.

You only need a few transistors.... TGM5 - all-BJT Simple Symmetric Amplifier

[FdW]

This is what I was thinking about:

[FdW]

Originally Posted by LineSource

FdW,

A window comparator which can detect preset lo/hi DC limits on the output is sometimes added to the DC servo circuitry. Any DC outside of the window can trigger a relay to either open(isolate speaker) or close(shunt 1 ohm in parallel to speaker), in addition to playing a role in a total power up logic.

Safe power up logic can include:
1) short input to ground until safe power-up is proven
2) allow input front-end circuits enough time to power up before the high current output power supply slow-start resistor is shorted by a relay.
3) either open up a relay in series with the speaker, or shunt a 1 ohm safety load in parallel with the speaker for protection
4) if DC is detected during operation, the output power supply relay can be opened
5) when power is turned off, shunt input to ground and shunt speaker to 1 ohm resistor to avoid "THUMP"
etc...

This is what I am (about) planning to do, de main differences are:
I will attach a small relay directly to the comperators output.
I did not add a second stage intergrator (only a small RC network).
For the input of the DC protect I am using the existing servo ouput.

Any comments on that?

[FdW]

Can anyone tell me how fast a DC protect would need to act. Lets say the amplifier fails and produces 10 Volts DC is can not act in mSecs due to the fact that it is viable that it under normal conditions may produce 10 Hz at 10 Volts. Also when you drop the stylus on a record this produces an very low frequency 'thump' or a warped record spinning at 33 RPM will produce an .5 Hrz signal.

Also what would be the threshold for reacting, is 100 mV DC allowable, or even 500 mV. What is the lower threshold to be detected, around 200 mV (or less)?

Regards,
Frans.

[dadod]

Look here, complete amp protection, a loudspeaker protection and an capacitor multiplier in one bouard: ThermalTrak+TMC amp
dado

[FdW]

Originally Posted by dadod

Look here, complete amp protection, a loudspeaker protection and an capacitor multiplier in one bouard: ThermalTrak+TMC amp
dado

Thanks for responding, and yes I know there are more than one ways to solve the overcurrent and DC problems. Still I would like to see answers to my questions, and also I do not want to have anything interfering with the power lines. As my amplifier is not referenced (in any way) to the power supply (is uses an separate internal shunt supply for the driving and gain stage) I opted to use the window comparator, to validate my design (on going) it would be nice to know a bit more about the timing and thresholds desirable for such a circuit.

Regards,
Frans.

Originally Posted by FdW

Can anyone tell me how fast a DC protect would need to act. Lets say the amplifier fails and produces 10 Volts DC is can not act in mSecs due to the fact that it is viable that it under normal conditions may produce 10 Hz at 10 Volts. Also when you drop the stylus on a record this produces an very low frequency 'thump' or a warped record spinning at 33 RPM will produce an .5 Hrz signal.

Also what would be the threshold for reacting, is 100 mV DC allowable, or even 500 mV. What is the lower threshold to be detected, around 200 mV (or less)?

[AndrewT]

I set up my DC detection using a low pass filter as the input.
This filter automatically outputs a variable response depending on the amplitude and frequency of the Fault Signal.

A low frequency is necessary. It won't react to a 20Hz full amplitude signal no matter how long it is applied.
It will react to a 20Vdc signal within about 30 to 40ms.
It will react to a 3Vdc signal in about 100ms
It will react to a 5Hz signal in about 250ms for 10Vac.

Unfortunately it resets on that 5Hz signal after a second or so and then may retrigger again in 250ms. I have not yet found a way to cure this anomaly without adding components.