Dc-detection, is this correct?

[fixation]

Hi everybody, i've been looking for a solution to my problem i want to build an uC controlled Dc-detection..


I came by this site URL=http://www.borbelyaudio.com/eb292132.asp]Borbely Audio[/URL]

and thought i could use this 'window detector"for my purpose
i've been modifying some parts and would like to know if someone have a few ideas or suggestions??

circuit explanation:

Q1A and Q1B, a dual, FET-input opamp, is configured as a window detector. The +input of Q1B is connected to a reference voltage of approx. +2V, while the -input of Q1A is referenced to -2V. The -input of Q1B and the +input of Q1A are connected together and form the DC sensing input. As long as the DC input is within the "window, defined by the +2V and -2V references, both outputs are positive. When the DC input moves outside the "window, i.e. more positive than +2V or more negative than -2V, Q1B or Q1A output will change to -12V

The output of your power amp has audio signals present, which have to be removed from the input of the DC sensing circuit. This is done with a single-pole RC network, consisting of R1 and C1/C2. The -3dB point of this network has to be selected low enough to avoid triggering the circuit by low frequency audio signals. However, it should not be too low, because it makes the circuit too slow. If you select C1=1uF, the circuit will activate at 50Hz with 100W into 8 Ohm. This is not acceptable in most applications. With 6.8uF I measured full power down to below l0Hz, without triggering the circuit. With C1 equal approx. 11uF (two 22uF caps in series) the -3dB point is at 0.15Hz and full output is available down to 5Hz. The circuit stays permanently shot down when you drive the amp with 2Hz at full power. You might want to experiment with this: if you want to prevent subsonic signals from reaching your speakers you might want to increase the -3dB point.

The purpose of D1 and D2 is two-fold: It prevents the caps from charging to a very high voltage, which would take a long time to discharge. More importantly, it is a protection for the opamp inputs: it prevents the input voltage to rise above the +l2Volt rail or below the -12V rail, which would destroy the ICs. A series resistor R2 limits the input current to the opamps.

When you look at the output you will see that after D3 & D4 a resistor and a led are. This led should be in a optocoupler but is drawn for explanatory reasons. The other site of that optocoupler will be put to Uc input pins.

text is mostly copied from www.borbelyaudio.com

is there someone that can give me suggestions, errors in the circuit, or maybe a total different circuit for DC-detection

Kind greetings Fixation

[fixation]

nobody any comments??

[AndrewT]

Hi,
I have seen similar circuits and they appear to work.
The RC timeconstant to filter out the audio takes time to react to low frequency signals. A small overvoltage signal will take a while to trigger the opamps. A large signal will trigger more quickly. This may take a bit of experimentation to set up. Some designers give guidance on selecting components for the RC.

The quick action on large signals is just what you need.

I do not think your LED will work. I think it is on all the time?
If the opamps are settling at 0V output (=ground) then the battery will push current through the opamps.
I think you can delete the battery and connect LED to ground. When either opamp output goes to negative then they sink current through the LED triggering the opto.
You could also put a transistor here to switch a relay and control other loads.

[richie00boy]

Surely the entire point of using a uC to do DC detection is so that no circuits such as that above are used - it's all done in software. If you build that circuit I can't see what a uC will add to the system apart from cost, real estate usage and EMC/EMI issues.

What did you think you were going to do with the signal from that circuit once it was in the uC?

[fixation]

i was thinking of a GLCD and 2 max6754 display( to control some kind of matrix led display) a temperaturesensor relays etc

maybe you have some idea how i could solve the dc-detection in software AndrewT?? i don't know how i could do that?

thanks for the input!

[AndrewT]

Hi,
your discrete circuit (as posted) does not need any software.

[fixation]

Hi AndrewT

well actually those battery's are dc-sources in my simulation program. Thanks for correcting the circuit, you're right the led should be connected to gnd.

although i'm still curious about ways in software to handle this?

i'll update the circuit soon as possible

[lineup]

Originally posted by fixation The purpose of D1 and D2 is two-fold: It prevents the caps from charging to a very high voltage, which would take a long time to discharge. More importantly, it is a protection for the opamp inputs: it prevents the input voltage to rise above the +12Volt rail or below the -12V rail, which would destroy the ICs. A series resistor R2 limits the input current to the opamps.[/B]

Two diodes in series?
Most often such input voltage limiters are connected
parallel across the input.

Are you sure D1 and D2 will work ......

[jaycee]

Rod Elliot has a DC protect circuit utilising a Microchip PIC on his website.

http://sound.westhost.com/project111.htm

Obviously no source code but enough to give you an idea. The detector is just 1 resistor/capacitor lowpass filter, a fullwave rectifier and detector transistor. Alternatively instead of the detector transistor you could perhaps use the ADC inputs you get on some uC's

It's only worth using a uC if you want to add other features that would use a uC also, such as an LCD display, or remote control, IMHO. Otherwise, you might as well stick with a simple 3 transistor circuit.

[jaycee]

You asked how the software would do it...

As you can see in Rod's schematic, the detector transistor is connected to one of the PIC's IO pins via a 10K pullup resistor. Normally this would mean that IO pin would read as a logic 1. If the detector activates, the transistor would shunt to ground causing the pin to go to logic 0. You detect that in software, and there's your DC fault trigger.

Rod uses the ADC feature of the PIC so that he can use the same pin for DC detect *and* a temperature sensor, but if DC detect is all you are after, any IO pin will do.