Input Protection

[Devil_H@ck]

Hi,

A few hours ago as I was reading a bit on the net, I stumbled across a page concerning input protection for ICs like opamps and the likes. Here is the image that was included on the webpage:


Now, my basic understanding tells me that if the input is either higher than (Vcc + diode drop) or lower than (Gnd - diode drop) that amount of voltage will go to Vcc/Gnd.

So far, no problems (I hope ). However, in the SSM2404 datasheet, there's another type of input protection shown. Here's a screenshot of it:


I'm guessing this works as follows: if the input's higher/lower than (diode drop) the 'amount of voltage' exceeding that (diode drop) gets shorted to ground. If that's correct, that means that no input signal greater than +- 0.7v is possible (well, unless you start putting diodes in series of course). Two anti-parallel diodes let all voltage about their voltage drop pass, right? Is that correct? If it is, wouldn't that be a mistake in the SSM2404 datasheet? I can't imagine someone's willing to implement such a sort of "protection", since it will ruin just about every audio signal you put at the input.

Am I correct? I suppose not, so could someone please explain it to me ?

[Nappylady]

I notice that I've never seen such a protection in any amp schematic I've looked at. This seems to me a pretty good indicator that you're right...

[Devil_H@ck]

Well, I'm going to use it for a preamp, using it for an amp would be kind of silly, because to get the protection to work, you'd need an input voltage almost equal to the amp's working voltage. If the amp would then try to amplify the signal, it would of course fail & start clipping, because there wouldn't be a high enough voltage available.

Anyway, I'm still in doubt about the lowest circuit. Should't it be: input - diode - +V & -V - diode - input? Why are they using the ground as the 'reference' voltage, to which the input voltage is passively compared?

[janneman]

Originally posted by Devil_H@ck
Hi,

A few hours ago as I was reading a bit on the net, I stumbled across a page concerning input protection for ICs like opamps and the likes. Here is the image that was included on the webpage:


Now, my basic understanding tells me that if the input is either higher than (Vcc + diode drop) or lower than (Gnd - diode drop) that amount of voltage will go to Vcc/Gnd.

So far, no problems (I hope ). However, in the SSM2404 datasheet, there's another type of input protection shown. Here's a screenshot of it:


I'm guessing this works as follows: if the input's higher/lower than (diode drop) the 'amount of voltage' exceeding that (diode drop) gets shorted to ground. If that's correct, that means that no input signal greater than +- 0.7v is possible (well, unless you start putting diodes in series of course). Two anti-parallel diodes let all voltage about their voltage drop pass, right? Is that correct? If it is, wouldn't that be a mistake in the SSM2404 datasheet? I can't imagine someone's willing to implement such a sort of "protection", since it will ruin just about every audio signal you put at the input.

Am I correct? I suppose not, so could someone please explain it to me ?

The first one (diodes to Vcc etc) works as you say. Actually, the "excess voltage" doesn't go to ground but appears across the series resistor. This resistor is REQUIRED for it to work.

In the second case (parallel diodes) the voltage across the diodes is essentially zero: they are on the inverting input, the virtual ground. They are just there to limit the input in case of problems, but will never see any appreciable voltage in practise.

Jan Didden

[Devil_H@ck]

Originally posted by janneman
The first one (diodes to Vcc etc) works as you say. Actually, the "excess voltage" doesn't go to ground but appears across the series resistor. This resistor is REQUIRED for it to work.

In the second case (parallel diodes) the voltage across the diodes is essentially zero: they are on the inverting input, the virtual ground. They are just there to limit the input in case of problems, but will never see any appreciable voltage in practise.

Jan Didden

Ah, thanks, now I see! Stupid of me not to take a look at the opamp connected to the SSM2404.

The series resistor you're talking about is R1 (the 100K/M10) resistor, right? The 10K hasn't got anything to do with it and is just there to keep the input of that NAND gate high, right?

Now, the 2nd circuit. The signal is at virtual ground, so no signal would ever reach a high/low enough voltage, right? Could you give an example in which case those diodes might be usefull?

[troystg]

Originally posted by Devil_H@ck
Hi,

A few hours ago as I was reading a bit on the net, I stumbled across a page concerning input protection for ICs like opamps and the likes. Here is the image that was included on the webpage:


Now, my basic understanding tells me that if the input is either higher than (Vcc + diode drop) or lower than (Gnd - diode drop) that amount of voltage will go to Vcc/Gnd.

Arise you dead thread.... You gotta love the search function!!

Ok, so I am researching this particular solution for limiting the input voltage / signal into a chip amp I made for a simple paging amp driven system... I want to remove the volume control from the front of the amp and have only the mic volume to adjust. So when it is distorted there is only ONE knob to adjust...

So the first drawing would suffice to limit the input voltage to +/- .7 volts max? Anything over is "drained" to ground?