opamp Rin when Vcc is floating

[shaiko]

I'm looking for an OPAMP IC that has very high input impedance when its Vcc is switched off.
Please advice on a component.

 

[crutschow]

That will be difficult to find since even high input-impedance CMOS op amps generally protect the inputs with diodes to the power supply rails. You thus wouldn't be able to apply more than perhaps ±0.5V to the inputs when they are unpowered without starting to forward bias the protection diodes.

How high an impedance do you need?

 

[shaiko]

more than 500Kohm

 

[FvM]

The question doesn't make much sense without talking about an input voltage range.

FET switches might be necessary to disconnect the input. If the problem is about DC impedance only, a RC parallel circuit can possibly isolate the inputs of a high impedance amplifier.

 

[shaiko]

FvM, the PCB is already manufactured so isolation transistors are out of the question. I have to find a replacement OPAMP istead of the one the original designer used.
The input voltage is around 3.3V and yes, I'm interested only in the DC impedance.
A schematic is attached.

 

[FvM]

Too bad...

 

[shaiko]

1. Please explain your suggestion about an RC parallel circuit.
2. Do you think an OPAMP with a constant high input impedance exists (even when unpowered)?

 

[FvM]

2. I'm not aware of any, to elaborate my comment in post #6.
1. A >500k series resistor limiting the OP input current, bypassed with a capacitor to reduce resistor noise.

 

[shaiko]

So, you're saying that in this case a redesign is the only option?

 

[crutschow]

So, you're saying that in this case a redesign is the only option?

Likely unless you can find such an op amp. The fix could be just adding a resistor in series with the input. Perhaps that could be added to the board with a cut of the input trace and soldering a small SMD resistor across the break.

 

[shaiko]

Wish I could do that - pretty sure it won't pass an IPC check with such a fix.

 

[crutschow]

Wish I could do that - pretty sure it won't pass an IPC check with such a fix.

Don't know who IPC is but at the aerospace firm I worked, they got such patches approved to fly in military space satellites. Of course to meet the vibration requirements, the resistor was epoxied to the board since the solder pads were obviously below size specification.

 

[shaiko]

https://en.wikipedia.org/wiki/IPC_(electronics)

 

[FvM]

Pretty pointless discusson.

Design faults need to be fixed, either in a hardware patch (as far as applicable) or in a PCB redesign. That's it.

 

[crutschow]

Pretty pointless discusson.

Design faults need to be fixed, either in a hardware patch (as far as applicable) or in a PCB redesign. That's it.

I was under the impression that the subject of this "pointless discussion" was just that.

 

[keith1200rs]

I am not aware of the IPC doing any checks. With regards to modifications to the circuit/PCB, that is perfectly acceptable as has already been stated. There are IPC standards for making modifications and I remember their were British Standard covering modifications years ago. There are probably military standards for it as well.

Keith

 

[mtwieg]

In my experience, bridging a cut trace with a SMT component is not only difficult, but has poor durability. Especially if the trace is thin (like under 20mil) and there are copper pours on both side. If that's the case then I would recommend greenwiring instead, in which you don't solder to the trace, but rather at two existing pads to which the cut trace connects. Though that will require epoxy to make it robust, it will probably work better.