14

There are a large number of opamps in the so-called "rail-to-rail" category: these either have an input common-mode range that approaches or exceeds the power supply rails, or they have an output range that approaches the power supply rails.

I'm aware that selecting the RRI/RRO requirement restricts yourself to a smaller selection of opamps, that will cost somewhat more (but not greatly so).

But are there other reasons to be careful w/r/t RRI/RRO opamps? I had a former coworker who avoided them like the plague and I want to know what to watch out for. In particular, are there any performance specs (gain-bandwidth, slew-rate, etc.) that degrade as the input and/or output approach the supply rails?

Jason S
  • 14,038
  • 3
  • 42
  • 70

3 Answers3

9

The only thing I think is constant across almost all rail-to-rail op-amps is they can source or sink only very small amounts of currents near the rails.

E.g. a RR OA may be able to swing \$Vcc-0.1\$ to \$Vss+0.1\$, but only with a \$100K\Omega\$ load on the output. With a \$1K\Omega\$ load it may only swing to \$Vcc-0.5\$ to \$Vss+0.5\$.

I haven't seen very many RR OAs that this is not true for.

Connor Wolf
  • 32,168
  • 6
  • 79
  • 138
4

Each opamp does what the datasheet says it does. If it meets the specs, what does it matter whether some of the specs could have been different if it weren't rail to rail input or output? Opamps have such a large range of commonly used parameters that rail to rail is just one of many in the mix.

In general, rail to rail output means CMOS, or at least CMOS output stage. It used to be that MOS opamps were very good at input impedance at the expense of offset voltage, but there are some with quite good offset voltage. Again, I don't think it's worth trying to guess what the parameters of a opamp will be from a few of the other specs. Some things are related, like slew rate and bandwidth, but even then there is nothing to rely on in that regard without looking at the datasheet.

So to answer your question, no, there aren't tradeoffs from the circuit designer's point of view because each opamp is a point case that has to be evaluated individually. There may well be tradeoffs when designing the opamp, but even with common general trends, there is still nothing to design a application circuit to.

As always, READ THE DATASHEET.

Olin Lathrop
  • 313,258
  • 36
  • 434
  • 925
  • 7
    I see your point, but this does not answer his question. First, because he wanted to know which specs, in general (not for a specific device), will degrade as we approach the rails, in a RRI/RRO opamp, so that he can know not only how behaves what he already chose, but more importantly, how can he choose better and quicker in the future, and second, because most datasheets don't cover all possible corners of the N-dimensional system they try to charaterize (and this may even be on purpose. There are some very good datasheet samples of "specmanship"). – Telaclavo Apr 28 '12 at 14:00
  • If a manufacturer claims rail-to-rail input/output OpAmp, can I really rely on that fact and provide a signal that is exactly 0V with respect to ground? My input signal is 0 to 3V, so I am trying to choose from AD8027, MAX4231, TLV2370, and TLV2450 the one that will provide best performance at the lower rail (powering single 5V supply) Which do you recommend – Nazar Aug 28 '14 at 17:10
  • Hopefully this link will work – Nazar Aug 28 '14 at 17:12
3

RRI "complementary differential input amplifiers achieves VICR exceeding the power supply limits, but there is a penalty to pay in input bias current, input offset voltage, and distortion." ref http://www.ti.com/lit/ml/sloa090/sloa090.pdf sect 18-4

"If the input voltage exceeded the power supply rails, the output stage might invert phase (it sometimes latched in the inverted position causing control problems) or the IC might self destruct on high Vcc with low ESR/SCR mode"

RRO means CMOS output which may be more sensitive to transient ingress on output causing latchup from ESD than BJT.

THey are Single supply OA's so you need a mid V+ Reference point .

But with care, they work well.

Tony Stewart EE75
  • 1
  • 3
  • 54
  • 185