Diode from input to output in buffer op amp?

[Gwow]

I was wondering what the effect of the diode in the circuit attached is.


Will this act as over voltage protection? If i input 30 volts and my op amp power supply is +-12 volts, will the op amp be destroyed??

 

[keith1200rs]

I don't know where you got the idea from, but as drawn it means nothing. With some more circuitry around the opamp, particularly to the input, some function could be determined.

All that can be said with what is drawn is the opamp output will clash with the input voltage. Supply more circuitry if you want a useful answer.

Keith

 

[Gwow]

Youre right. Heres the circuit. My question is how can i add overvoltage protection (+-30 volts) when my op amp only has a voltage supply of 12 volts?

Basically we here have an input buffer op amp which has very low input leakage current (in the fA) and we have similar jfets which have low leakage current at the offset voltage of the op amp. So, how do i add voltage protection without 'adding' to the current error (the capacitor current is in the 80 fa range). Or is there already protection built in?

The jfet diodes on the op amp allow us to recharge the capacitor by flipping the switch. The resistor right by the capacitor is air wired to the op amp.

 

[keith1200rs]

Again, I would need more information to answer fully. The opamp might have protection diodes built in, but as you don't say what it is, I cannot say. However, some opamp data sheets don't make it clear what is inside whereas others show a pair of cross coupled diodes on the inputs, which is helpful.

The drains of the JFETs appear to be floating, so I am assuming you are trying to use them as diodes? One snag I can see with that is that it can still allow the input of the opamp to exceed the supplies depending on the opamp output voltage range, assuming that you were hoping to use them as protection as well as to discharge the capacitor.

I think I would add some low leakage diodes from the inputs to new voltages generated around 1V below the opamp supply rails. Something like the PAD1 maybe. There are probably better low leakage diodes, although I am not sure you will get guaranteed femto amp ones.

Keith

 

[FvM]

Picoampere JFET OPs possibly don't have protection diodes, e.g. dielectrically isolated OPA129, or more exactly, at least
no protecion for one input polarity. You should check, if the datasheet has explicite information. High speed OPs mostly
have protection diodes that are capable of carrying 5 or 10 mA continuous and considerably higher pulse currents.

I understand, that the purpose of the tracking mode in your circuit is minimization of leakage current. For effective overvoltage
protection, you may want to add a low impedance zener or diode clamp to the tracking node.

 

[Prototyp_V1.0]

I was wondering what the effect of the diode in the circuit attached is.

Since both input and output will have same voltage, the diode won't do anything, as stated before.

Only situation the dioda may do anything at all, is if the opamp is very sliw, and the diode is very fast. Then a sudden voltage rise will cause the signal source (whatever it is) to supply the output for a very short period of time, as the opamp output soon will follow and stabilize the voltage.

 

[Gwow]

Woops, forgot the part numbers. The parts of relevance for this schematic i found are:

The op amps are
LMC6042

**broken link removed**

And the jfets are
J112

http://www.fairchildsemi.com/ds/J1/J111.pdf

R4, R8 are 1k resistors.

So, as i understand it, when the input of the op amp reaches above 12 volts the output will approximately be 'clamped' to 12 volts, since this is power supply. Then, once the input gets to 12.6 volts, one of the jfets will start conducting. Now, i know that the typical way to protect against overvoltage is to wire a diode to the supply, but in this case its like wiring a diode to the 1k resistor (R4), whose other side is at 12 volts.

Now, it doesn't sound like this will act as a clamp, because all we have is a varying voltage over the 1k resistor, from 0 volts to 30-12-.6 = 17.4 (30 is max possible op amp input voltage). So, no clamping action, instead the current is directed into the op amp output (it has nowhere else to go??)

So, if the op amp input cant handle 30 volts input (which it cant i believe) the op amp will be fried? And thus for protection i would have to add a low leakage diode??

 

[keith1200rs]

Strictly speaking you will exceed the ratings which say only 0.3V from the rails. The JFET will clamp above that I think. In reality, you might get away with it. My preference would be to use diodes (or JFETs) to clamp to something like +/-11V. That way you can keep within the opamp ratings.

Keith.

 

[FvM]

LMC6042 as a CMOS amplifier has input protection diodes, the maximum current rating is specified as +/- 5 mA. If you can
keep this limit with an input series resistor, no additional protection should be necessary.

 

[keith1200rs]

LMC6042 as a CMOS amplifier has input protection diodes, the maximum current rating is specified as +/- 5 mA. If you can
keep this limit with an input series resistor, no additional protection should be necessary.

That isn't an approach I would like to risk. I have seen devices do strange things when you start putting current through the protection diodes. It may not damage the device, but you might not get the output you expect.

Keith.

 

[Gwow]

So all CMOS amplifiers have built in protection diodes (or maybe im missing it in the datasheet)? Good to know! Thanks everyone!

 

[keith1200rs]

So all CMOS amplifiers have built in protection diodes (or maybe im missing it in the datasheet)? Good to know! Thanks everyone!

You need to be careful to distinguish operating conditions from 'absolute maximum ratings'. Outside the operating conditions you will not damage the device if you are within the absolute maximum, but it is not guaranteed to work. I have seen logic devices 'flip' their outputs when one input is overdriven so the protection diode operates,

Keith