DC protection for leach

[audiofan]

I have made no simulation or test yet but I would like to have opinion on this circuit configuration

[audiofan]

here is the picture

[audiofan]

Me again.
Sorry inverting & non inverting input of op-amp are reversed

[audiofan]

I did no test or make simulation but I found found that the integrator with no close loop will always fire after a period of time. So I am going back to the drawing table.

[Perry Babin]

Have you tried connecting a 1 meg resistor in parallel with C3?

It works at higher frequencies. As long as the other resistors are relatively low in value (less than ~50K ohms), it may work for your circuit.

[jaycee]

I guess this is a crow-bar circuit to blow the fuses should DC appear.

The detector seems overcomplicated to me. You could achieve the same effect with 4 diodes, an R/C filter and a few transistors to fire the crowbar. This would have the advantage that you can run it from the poweramp rails without needing any +-15v supply for the opamps.

[audiofan]

A resistor across C3 is a great idea, I did not see this one by myself but after a quick analysis it is clear that he resistor will limt the gain at low frequency and will solve the run away problem. tanks Perry

+15V an -15V could be a problem for most but my amp is build with old transformer from few IBM PROPRINTER stuff and there is 2 10 volts AC tap in the powersupply with 2 diodes and 2 caps I will have a non regulated supply for the OP amp. Thanks Jaycee for the hint

[audiofan]

I made a few simulation and the new circuit is below

triac optocoupler led & stuff tig at 3.5 volts

Response of filter including op-amp stage

dc + 14 db
3Hz + 10db
1.6 Hz 0 db
4Hz -10db
8.5Hz -20db
16Hz -30db
30Hz -40db

system should trig at 0.7 Volt DC

[Quattor]

This type of circuit has to perform in two quite opposite directions. First, it should detect only DC voltage - thus removing everything AC at it's input. Secondly, it should react as fast as possible since you don't want to turn your speakers into heaters and melt them down in a matter of miliseconds. Improving one, degrades the other one. Of course, you want to react on rather low DC levels - 0.6 - 0.7 V (matches with the need for speed).

Attached circuit has performed quite well in all sorts of situations. Don't mind +33 V reference there. You can use different voltages as well. My suggestion is to use +12 V and to tie up CMOS 40106 or CMOS 4093 (Schmitt-triggers) at it's output. Works like a charm every time

Best regards,
Nick