The LM108 op amp is an interesting chip to examine under a microscope because it uses special superbeta transistors for high performance.
Photos of the die reveal the tiny circuitry of the chip as well as unused components that make the chip more complex than necessary.
Surprisingly, these extra components allow the same die to be reused for two totally different chips!
In this article I examine the internals of the LM108 in detail, explain how it works and reveal how the die can take on two roles.
I’ve written about the famous 741 op amp, which came out in 1968.
A year later, the improved LM108 op amp was invented by eccentric analog IC design genius Bob Widlar.
The main claim to fame of the LM108 is it uses a very small input current, orders of magnitude smaller than the 741.
To attain this low input current, the LM108 contains special transistors called “superbeta” transistors, with about 25 times the amplification of a regular transistor.
The downside is the superbeta transistors are delicate and require special circuitry to protect them from damage.
The photo above shows the LM108 op amp in a metal can.
(The LM308 is the commercial-grade version of the LM108.
I opened up the can and photographed the die (below).
The chip’s metal layer is clearly visible, with thin metal traces connecting the different parts of the chip.
The square bonding pads around the edge of the chip are connected by thin wires to the chip’s external pins.
Under the metal layer, you can see the silicon that forms the basis of the chip.
To form transistors and resistors,
a process called doping
treats regions of the silicon with elements such as phosphorus or boron.
In the die photo, these regions have a slightly different color, which makes the structure of the chip visible under the metal.
While the chip seems incomprehensible at first, close examination reveals the different components and their connections.
By carefully studying the die photo, I reverse engineered the circuit for the op amp.
Surprisingly, this chip has an unusual circuit design, more modern than National Semiconductor’s “classic” LM108 design.
Although the package has the National Semiconductor logo, the internal circuitry matches the Motorola LM308 datasheet.
You might expect that LM108’s would all be the same internally, but as with many ICs, the part number doesn’t indicate as much as you expect. Different manufacturers have widely differing implementations of the chip, so you can’t expect two chips to behave the same just because they have the same name.
Even so, it’s puzzling that a National Semiconductor chip doesn’t match the National Semiconductor schematic.
The function of an op amp is to take two input voltages, subtract them, multiply the difference by a huge value (100,000 or more), and output the result as a voltage.
If you’ve studied analog circuits, op amps will be familiar to you, but otherwise this may seem like a bizarre and pointless device. How often do you need to subtract two voltages? And why would you want to amplify by such a huge factor? Would amplifying a 1 volt input result in lightning shooting from the op amp?
It turns out that op amps are extremely useful and versatile, making them a key component in analog circuits.
With simple feedback circuits, you can use an op amp as an amplifier, a filter, integrator, differentiator, or a variety of other circuits. When an op amp is in use, the voltages on the two inputs will normally be almost identical, so multiplying by the huge amplification factor yields a reasonable output of a few volts.
The point of the high amplification is it improves accuracy, even if the amplification of the overall circuit is small.