I know that most operational amplifiers can tolerate the output being shorted to ground. Many can even do that for unlimited time and regardless of the output voltage according to many data sheets.

What if the output is shorted to something else than ground? An example would be that the output of the op amp is 0 V and it is shorted to either the positive or negative supply rail. What will happen then?

I guess that the answer depends on the op amp output stage and current protection circuit. I need to find a high precision op amp that can tolerate this situation - or make an external circuit to solve the problem. Any ideas are appreciated. Thanks in advance

 

I don't think there's much difference in sourcing or sinking to a "short", the data sheet should address this. You may need to focus on the power dissipation of the package, though. I think that is more likely the limiting factor that whatever current is cited in the data sheet.

Reducing the sink current could be as simple as a resistor. Hard to comment without a schematic.

 

Thanks for the reply. If the output protection really is bidirectional, then I wonder why the data sheets usually says "Output Short-Circuit Duration to GND" in the "ABSOLUTE MAXIMUM RATINGS" section. However some does not specify "GND"... I guess I have to test this IRL and see what happens with a particular op amp.

I do not have a schematic since this is an "abstract" question and not part of a real circuit.

Thanks

 

What is the load impedance? If it's high you could add an output series resistor to limit the short circuit current. You can put the resistor inside the feedback loop so its effect on the output voltage would be negligible.

 

The data sheet specifies short to ground because that is the most logical place for a short to occur. Shorts to external supplies may or may not kill the op amp depending on impedance, the supply voltage, supply current, etc., and your testing is not conclusive. Talk to the factory apps engineer.

 

In the application I have in mind, the op amp output will be protected with a resistor in series and a diode/zener network in order to limit both the voltage and the current through the op amp. So of course my circuit could still be dangerous to the op amp but not as dangerous if no protection is used.

The diode network could be the classic two zener clipper like this:
http://www.hqew.net/files/Images/Article/Circuit_Diagram/zener_cliping.jpg

If I limit the voltage to around 70-80% of the supply rail and limit the current to be within the continuos output current specification (well below the short circuit limit), I expect the op amp to be fine. Or should I still be reaching for the fire extinguisher?

 

Thanks for the reply. If the output protection really is bidirectional, then I wonder why the data sheets usually says "Output Short-Circuit Duration to GND" in the "ABSOLUTE MAXIMUM RATINGS" section. However some does not specify "GND"... I guess I have to test this IRL and see what happens with a particular op amp.

I do not have a schematic since this is an "abstract" question and not part of a real circuit.

Thanks

Protection from short to GND is quite common - short to Vcc (and/or a negative rail) less so.

Usually the data sheet should make this clear - when available, an internal schematic can be helpful in predicting what abuse it might tolerate.

 

The diode network could be the classic two zener clipper

You haven't said what frequency you need to operate at. Keep in mind that Zener diodes can have a significant amount of capacitance. For instance a 1/4 watt, 3.3 volt Zener can have 450 pF of capacitance! A 1/4 W, 12 volt Zener would be less at maybe 85 pF. Higher wattage Zener's have higher capacitance than lower wattage ones.

You might be better off clamping to the positive and negative power supplies using high speed diodes. These have much smaller capacitance.

 

It's common to add reversed biased diodes to the positive and negative rails from the output for protection.

 

Usually the data sheet should make this clear - when available, an internal schematic can be helpful in predicting what abuse it might tolerate.

Yes, some data sheet includes a (simplified) schematic of the op amp design that could be used to figure out the type of output protection circuit. However the op amps I want to use do not include a schematic in the data sheet.

You haven't said what frequency you need to operate at. Keep in mind that Zener diodes can have a significant amount of capacitance. For instance a 1/4 watt, 3.3 volt Zener can have 450 pF of capacitance! A 1/4 W, 12 volt Zener would be less at maybe 85 pF. Higher wattage Zener's have higher capacitance than lower wattage ones.

That's right. I find the four-diode and one zener clipper usually to be much faster that the one only using two zeners. When I need fast and precise clipping I only use 10V+ zener diodes. I usually amplify the signal before the clipper and then attenuate afterwards just to get "sharp" limits and fast operation, even though the circuit gets more expensive and the extra amplifier might introduce noise and things like that.

This situation is not "signal processing" but rather protection of op amp output and I cannot use amplifiers here. Luckily the voltage limit is going to be like 12V which allows the use of faster zener diodes.

You might be better off clamping to the positive and negative power supplies using high speed diodes. These have much smaller capacitance.

Interesting... do you have an example? I have not worked much with clamping before...

The reason I am asking these questions is that I am probably going to use the op amp as an output of some kind of low power DC generator that is going to provide stimuli to an external component which is to be measured. I want to protect the op amp output from external circuits that potentially could put a voltage on the output terminals. The limiting circuit is going to protect the op amp output and the same time provide an indication of overload to a microcontroller that will disconnect the output fairly quickly in order to save both the limiting circuit and the op amp. In other words, if the limiter is sensing an overload, the output relay will disconnect the external circuit within a few milliseconds. The op amp will be protected as long as the limiting diode network is not overheated.

An example of what I want to do is found in many newer laboratory equipment such as a function generator. The output of this generator could be connected to an external circuit, usually some kind of input. In the case that the user by accident connects the function generator to something else, or if the external circuit malfunctions, the function generator will disconnect and warn the user that the output has been overloaded.