Or should the non-inverting pin be tied to the output?

Which is correct? Left? Or Right?

 

Right or not, commercial designs I've seen use the one on the left, output strapped to inverting input and non-inverting input grounded.

 

Connecting both inputs to ground will anplify the input offset by the full gain oh the opamp, resulting in an unknown output. Connecting the output to the inverting input and the non- inverting input to ground will output as near to ground as possible, definitely the preferred option.

Bob

 

The right side is definately wrong, but the left COULD be wrong. The general deal is the lowest power dissipation and inability to oscillate. Having at least one resistor for Ib to drop across is also preferred as well. So, the left configuration with the a zero ohm resistor in the FB loop and a resistor from pin 6 to ground.

So, if the OP amp is not unity gain stable, neither configuration is a good idea. Say the minimum gain is 10. So, set the op amp for a gain of 10 with an input of 0 with a place for Ib to go.

The cool part is that if you need an OP amp later, it's there.

 

The figure on the left has the inputs backward.

The feedback should go to the "-" inverting input, with the non-inverting tied to GND or other suitable reference voltage.
The idea is to connect the opamp as a voltage follower, with a known and predictable input and output voltage.

The opamp configuration on the right might go into positive or negative saturation, depending on the input offset and the type of device.
You want to "stow" this unused amp in a known and predictable state.

 

Or should the non-inverting pin be tied to the output?

Which is correct? Left? Or Right?
View attachment 220549

Neither.

OK, let me soften that a bit, because the goals are prevent unwanted interactions with the rest of the circuit, and not damaging the unused device. Neither one will damage the device, but there is a better way to meet the goals.

Start with the left image, but connect the output to the inverting input (voltage follower) and the non-inverting input to GND. No out-of-range voltages on any pin, no input offset voltage issues, no noise bursts. The output will be mildly saturated. To prevent this, connect the + input to some other DC voltage in the circuit that is within both the input and output operating voltage ranges.

ak

 

Start with the left image, but connect the output to the inverting input (voltage follower) and the non-inverting input to GND.

I missed the transposed connections.

 

@Tonyr1084 Quit causing trouble with philosophical questions. Kidding, of course.

Most everyone is half asleep. I should be.

 

 

Don't forget the 25120 WOM https://www.baldengineer.com/wp-content/uploads/2018/02/25120-bw.pdf

 

How about if the common mode range doesn't include the negative supply?

 

I typically use the following method for unused op amps.
When the problem first presented itself to me, I did some research. This method was recommended several times and I believe I first found it on a manufacturers web site (LT?).

It does take a couple of extra components (resistors) and some PCB space, but it has worked for me.

 

From TI:

https://e2e.ti.com/blogs_/archives/b/thesignal/archive/2012/11/27/the-unused-op-amp-what-to-do

https://www.ti.com/lit/an/sboa204a/sboa204a.pdf

 

From TI:

https://e2e.ti.com/blogs_/archives/b/thesignal/archive/2012/11/27/the-unused-op-amp-what-to-do

https://www.ti.com/lit/an/sboa204a/sboa204a.pdf

The solution I posted is documented in Figure 2 in the second link @upand_at_them posted.

 

How about if the common mode range doesn't include the negative supply?

To prevent this, connect the + input to some other DC voltage in the circuit that is within both the input and output operating voltage ranges.

Note that there is no requirement that an unused component (opamp, NAND gate, whatever) be tied to a DC potential. If it is more convenient in the pc layout, tie the opamp input to a nearby signal voltage, or tie the gate inputs to a nearby digital signal. Yes, the output will move, but in the vast majority of situations this will have zero consequences.

If each of us can save just one via, we *can* make a difference.

ak

 

The feedback should go to the "-" inverting input, with the non-inverting tied to GND

connect the output to the inverting input (voltage follower) and the non-inverting input to GND.

Last night as I was watching the World Series (what a game!) in the back of my mind I thought of this option and wondered if it was going to be the best way to go. From your comments (and others) I could tie the non-inverting (+) to anything stable and the output will follow that level. Space is not going to be a concern and this is a one-off PCB. No point in doing a gerber file and have someone make a board; or to go through the process of etching my own board. Does anyone do that anymore?

Anyway, from the comments received - this is the approach I'll take. Not all will agree 100% but I just don't want the thing to fly off at some unknown frequency and either self destruct or more simply waste battery power. This is just a part of a freezer door alarm I'm building. The other day we put stuff in and closed the door. Some of that stuff shifted and pushed the door open slightly and the next morning we had to do a lot of smoking (sausages and hot dogs that had thawed on the door).

This is what I'm hearing: (below). A voltage divider on pin 5 will act as a slight drain, so I will be tying pin 5 to ground.

 

Last night as I was watching the World Series (what a game!) in the back of my mind I thought of this option and wondered if it was going to be the best way to go. From your comments (and others) I could tie the non-inverting (+) to anything stable and the output will follow that level. Space is not going to be a concern and this is a one-off PCB. No point in doing a gerber file and have someone make a board; or to go through the process of etching my own board. Does anyone do that anymore?

Anyway, from the comments received - this is the approach I'll take. Not all will agree 100% but I just don't want the thing to fly off at some unknown frequency and either self destruct or more simply waste battery power. This is just a part of a freezer door alarm I'm building. The other day we put stuff in and closed the door. Some of that stuff shifted and pushed the door open slightly and the next morning we had to do a lot of smoking (sausages and hot dogs that had thawed on the door).

This is what I'm hearing: (below). A voltage divider on pin 5 will act as a slight drain, so I will be tying pin 5 to ground.
View attachment 220592

Its important to note that the common mode voltage ranges needs to be considered. This configuration is OK for the LM358 because the LM358 common mode range includes ground.

 

Hello,

Have a look what TI and MAXIM have to say about the issue:
https://e2e.ti.com/blogs_/archives/b/thesignal/archive/2012/11/27/the-unused-op-amp-what-to-do
https://www.maximintegrated.com/en/design/technical-documents/app-notes/1/1957.html

Bertus

 

Hello,

Have a look what TI and MAXIM have to say about the issue:
https://e2e.ti.com/blogs_/archives/b/thesignal/archive/2012/11/27/the-unused-op-amp-what-to-do
https://www.maximintegrated.com/en/design/technical-documents/app-notes/1/1957.html

Bertus

The second link describes the solution I posted in #12

 

The schematics from that Maxim document are cringe worthy:

Connection dots and humps???