Portable analog design needs rail-to-rail op amps

By Ron Mancini -- EDN, 11/23/2000

Portable-equipment design dictates single-supply, low-voltage design techniques. Rail-to-rail op amps are mandatory in portable designs because only they satisfy the design criteria of low noise, high dynamic range, signal sensing at the input rails, and rail-to-rail output-voltage swing. They satisfy the design criteria, but why do they cost more than conventional op amps?

Transducers are invariably connected to a supply rail rather than to a V_CC/2 bias voltage; thus, the transducer output voltage swings above and below a supply rail. Conventional op amps have a differential-amplifier input structure, a "long-tailed pair," and they use either npn or pnp input transistors. The pnp transistors can function at the lower but not the upper rail and vice versa for the npn transistors. Rail-to-rail op amps have parallel npn and pnp long-tailed pairs that cross over at V_CC/2, and these op amps can also function 200 mV above and below the rail voltage. They function at both rails, so they can amplify low-level transducer signals connected to either rail without clipping.

Low rail voltage means lower generated noise voltage because logic swings are lower and actuators and transient currents are smaller. Op-amp noise influences the overall noise performance more in a low generated noise than in a normal environment. Hence, rail-to-rail op amps require improved noise performance. The reasons for the cost difference between rail-to-rail and conventional op amps are that rail-to-rail op amps double the number of input transistors to obtain rail-to-rail performance, and they increase the size of the input transistors for better noise performance. In addition, low-noise manufacturing is more difficult.

The dynamic range of an op amp is proportional to the op amp's output-voltage swing (see "Defining an op amp's dynamic range," EDN, Aug 17, 2000, pg 30). The rail-to-rail op amp's output voltage must swing within millivolts of the rails to achieve an acceptable dynamic range. With conventional op amps, the output transistor's IR drop, offset voltage, and required bias headroom limit the output-voltage swing. The rail-to-rail-op-amp output stage uses FET transistors to eliminate the offset voltage, and rail-to-rail op amps employ special transistors to eliminate the bias headroom. Using large FETs reduces the IR drop to a few millivolts, so the rail-to-rail output-voltage swing is usually from ground +150 mV to V_CC–150 mV.

Rail-to-rail output-voltage swing requires large special transistors that cost more than conventional output transistors.This requirement is another reason that rail-to-rail op amps cost a bit more than conventional op amps. The op amp's input offset voltage, input current, and other error contributions reduce dynamic range, so designers take special care designing rail-to-rail op amps to minimize errors. Special care translates to more product cost, thereby driving rail-to-rail op amps' cost a little higher.

This situation seems like a paradox; I always emphasize minimum-cost design, but I recommend switching to rail-to-rail op amps for medium- and low-speed sockets. The other factors in rail-to-rail-op-amp cost are manufacturing volume and experience. Currently, conventional-op-amp volume far exceeds rail-to-rail-op-amp volume, but that situation is changing quickly. Rail-to-rail-op-amp volume is growing much faster than conventional-op-amp volume, and, over time, manufacturing will gain the experience it needs to reduce cost. I predict that rail-to-rail-op-amp cost will decrease to within a few cents of the price of conventional op amps, and then the performance advantages of rail-to-rail op amps will make them the most cost-effective op amps.

Furthermore, as rail-to-rail-op-amp volume increases in new-product designs, conventional-op-amp volume decreases. I wonder whether some of the conventional op amps will be available in a few years. Switching now to rail-to-rail op amps may be a prudent decision.

Author info