What opamp nonideality is this?

[hazuki]

Hi all. I have an opamp circuit acting as a diff amp. I'm using an LT1496 rail to rail (input and output) opamp. I have both an LTSpice model and the real thing on my bench. I noticed that there was a really high error voltage (100mV) when both inputs are at the rail voltage (24V). I was playing around in spice and I noticed that if I drag the negative supply of the opamp down just a little bit, that offset goes away! I verified this with the real thing and the offset dissapeared!

Can anyone tell me what effect this is? I know that it has something to do with the inputs being right at the rails. Why does making the negative supply on the opamp more negative fix this issue? I've attached my LTSpice model.

I've also attached a picture of the simulation showing the effect that sweeping the negative rail has on the opamp output.



Thanks!

[miceuz]

It's called Vol - voltage low output. The output of the real as opposed to an ideal opamp cannot really go down or up to a rail. You are always left with some offset. As you lower the rail, this offset disappears refered to ground as the lower limit for the output gows lower 

In datasheets those sometimes are called output swing.

[hazuki]

Thank you miceuz. The datasheet "advertises" this opamp as a Rail to rail input and output opamp. This kind of behaviour is obviously not rail to rail! Why is that so misleading. What makes a rail to rail opamp a rail to rail opamp?

Thanks!

[Len]

There are two lessons here:

1. There's no such thing as a rail-to-rail op amp.

2. Never believe any specs on the front page of the datasheet.

[jebcom]

And then it gets a bit more complicated. The output will probably swing from Vol to Voh, but it won't necessarily be linear. If you need the output to be linear with respect to the input, you'll need to look a but further to find that range. It might be specified as the output voltage range for the open loop gain spec, Aol. This will most certainly be a smaller range than Vol to Voh. But they like to push Vol and Voh because it looks better. A lot of these semiconductor marketing people are very well suited to selling sugary soft drinks to teenagers.

[miceuz]

One of the solutions to this problem is to opperate the opamp well within the rails - design the circuit in a way that an opamp does not have to hit a rail for normal operation.

[hazuki]

Thanks guys. You've been very helpful. I now know not to get stuck in that trap!

[Bored@Work]

There are two lessons here:

1. There's no such thing as a rail-to-rail op amp.

2. Never believe any specs on the front page of the datasheet.

To summarize: Rail-to-rail is a marketing term.

[free_electron]

rail to rail means within a few 100mV off the rails.

a classic opamp cant get to within 1.2 or 1.8 volt off one of it rails due to the current pump for the differential pair sitting there.

there is rail to rali input and rail to rail output.

rail to rail output opamps can be a s good as a few millivolts off the rails , provided you dont draw more than a few milliampere.... they use mosfets in the output stage and you only deal with rdson of the output stage. if rdson is 1 ohm and you draw 1 ma you are 1 mv off the rail....