Solid State History

A work in progress. This is by no means much of a "history" yet. But check back as details will be added and expanded upon.

1947 is the well known year that a working point-contact transistor was developed, but it would take nearly two decades longer for transistor power amplifiers to make a significant impact in the audio reproduction world.

Early transistors were almost exclusively made of Germanium, and only came in the PNP variety. This led to a confusing period where the most negative voltage would be written at the top of a schematic, and the most positive voltage would be at the bottom of a schematic, and thus the transistors would be drawn in an "emitters-up" orientation that can be disorienting to modern eyes.

Solid State Designs with Output Transformers

Since early transistors only came in a single polarity, early power amp designs followed conventional tube amp designs. (Tubes also only come in one polarity.) Such designs would include output transformers to couple the transistor amplifiers to loudspeakers. This designs are relatively rare. Study of tube amp designs in concert with transistor theory should work to understanding this type of amp.

George Sziklai

In 1953, George Sziklai wrote a paper titled "Symmetrical Properties of Transistors" which outlined the idea of an transistor power amplifier that used PNP and NPN transistors to drive a loudspeaker. Unfortunately, NPN power transistors would not be powerful enough or fast enough for to achieve this design for many years. Sziklai's patent from the same year, US 2762870 shows a PNP/NPN pair driving an output transformer.

Since the PNP/NPN "complementary" design could not be realized yet, practical amplifiers had to rely solely upon the available PNP devices made of germanium.

Sziklai's name is used for a transistor configuration know by several names including, "Sziklai Pair," "Complementary Feedback Pair" (CFP), or "compound transistor."

https://en.wikipedia.org/wiki/Sziklai_pair

A low or medium power NPN transistor can drive a high power PNP output transistor so that the signal output of the PNP mimics the output of the NPN. Although this invention bears Sziklai's name, it was another engineer, Hung Chang Lin, who would develop the output section that took advantage of the Sziklai pair.

Lin Topology

In 1956, Dr. Hung Chang Lin built and demonstrated the "quasi complementary" output stage. Lin's stage uses a PNP output device to source current to the load, and another PNP in a Sziklai pair to sink current from the load. Transistors are low enough impedance in both sink and source modes to drive a loudspeaker directly in this way. Lin's original design used a single power source, so the output was at 1/2 the voltage supply, and thus required a dc blocking capacitor between the amplifier output and the loudspeaker. If a symmetrical power supply is used, the amplifier output may sit very near 0.000V, and thus allow for the capacitor to be eliminated. (The capacitor has many negative attributes: it must have a large Farad value to allow for a good low frequency response, thus it will be physically large, probably expensive, and of the electrolytic type which is more prone to failure than other types.)

Lin's design was also a bit ahead of its time. Transistor amplifiers were largely confined to applications where saving weight was more important that high fidelity or maximum power. This meant portable and automobile radios, but not the home "high fidelity" systems. Lin's transformerless design would not become common in high power systems for some time.

http://www.semiconductormuseum.com/Transistors/RCA/OralHistories/Lin/Lin_Page7.htm

Totem Poles

A very popular output design is nick named the "Totem Pole" output stage. This is a "stack" of transistors of the same polarity, usually PNP germanium devices in the '50s and '60s, and later NPN and silicon devices came along in the '60s and '70s.

Totem Pole Style Output Stage

In the above simplified schematic, Q1 is sourcing current to the loudspeaker, while Q2 is sinking current. Q1 is working as a common collector amplifier (emitter follower), and Q2 is working as a common emitter amplifier. This causes significant distortion issues. Not shown is the negative feedback path that will work to correct these distortion problems, but even large amounts of negative feedback cannot completely eliminate the problems from this asymmetrical design.

This amp worked well enough to be very popular for many guitar amps designed in the 1960s. R1 and R3 provide drive to "turn on" the bases of the transistors. Conversely, R2 and R4 will be relatively small and they are there to "turn off" the transistors. R1-R4 are the "bias resistors" and they set the idle current. When all of the resistors and the transistors are perfectly matched, the idle voltage across C1 will be 1/2V-. If a symmetrical power supply is used, such that R6 and R4 would connect to a V+ instead of 0V, then a perfectly matched circuit would idle at 0V at the "output node."

Totem Pole Output Stage with Transformer Driver Stage

The driver transformer T1 provides the two 180º out of phase signals for the output stage. Q4 is the driver for the primary coil, and this transistor is usually attached to a heat sink as this is similar to a class A output stage. Q3 provides an input point at the base, and also a negative feedback node at the emitter. There is feedback from the final output via R7, and feedback from the Q4 stage via R9. R8 provides a dc path to bias Q3, and this can also provide feedback for an earlier stage. Just as with the earlier example, if a symmetrical power supply is used, then C1 may be eliminated, as long as the output stage is reasonably balanced (i.e. near 0.0Vdc across the speaker).

The '60s

In 1965 many musical instrument amplifier companies brought out solid state models. Even Fender jumped on the bandwagon in 1966. Fender's amps were a major failure for the company. The designs are a mix of totem pole and driver transformer types (Bassman 100) and quasi-complementary (Lin) types. The amps were unreliable. In Fender: The Inside Story, Forrest White attributed the failure of the line to the new engineer who designed the amps to be difficult to service in the field. These amps were also constructed using circuit boards and solder baths: methods new and unfamiliar to the factory workers who allegedly produced failure prone amps due to lack of training and understanding of the new equipment. It is rare to see a working early solid state Fender amp today.

Vox is another company known for early solid state amps. U.K. made Vox amps are completely different from the U.S. made Vox amps. In the U.S., the amps were made by the Thomas Organ company. These amps are notorious for being difficult to service. The circuit boards are usually tied down with many solid core wires that break easily. Boards often require some wires desoldered to service, and then others will break unintentionally. U.K. made Vox amps are much nicer. In both cases, they are usually the totem pole with driver transformer type.

The Acoustic and Sunn amps of the 1960s and 1970s are much more common to see still in service than the Fender or Vox amps. These use totem poles with driver transformers.

The '70s

The transistor was finally widely accepted in the late 1960s. In recording studios, solid state mixing consoles were being installed. In the retail stores, high fidelity solid state stereo systems were pushing the old tube units aside.

Designers had embraced the Lin topology and interstage transformers faded away from power amps. Germanium disappeared and complimentary silicon devices proved to be powerful and reliable.

The '70s introduced many designs relying on integrated circuits, which had become widely available in the late '60s. Operational amplifier design and power amplifier design became very similar. While their are many ways to make a power amplifier, the most common variety is essentially a large operational amplifier.

Modern Trends

By the 1990s, technology was completely woven into the fabric of society, and things like audio amplifiers were simply taken for granted as a ubiquitous element of life. No longer were they exotic and expensive contraptions that needed care and attention; the amp was just something that provided a utilitarian purpose and it better work right and cause no harm (like destroying speakers). They also needed to get smaller and lighter.

Pretty much all powerful amps from the 1990s onward seem to employ at least one or more subcircuits devoted to "circuit protection." The most common scheme is to employ output relays that only connect the output to the speakers when the "protection circuit" thinks all is well. A thermal switch that monitors the temperature of the heatsink can be used to interrupt the output relay or the power supply if overheating is detected. Modern amplifiers may employ microcontrollers that are programmed to detect amp failures. These protection circuits usually shut down the amp and illuminate a LED to inform the user that a trip to the repair shop is necessary.

Class D amplifiers, while available in theory as far back as the 1940s, really didn't appear on the market until the 1990s, but they became widely accepted in the 21st century. They have the advantage of size, weight, and efficiency. They have always had the drawback of being very complicated and almost always requiring wholesale replacement in the event of a failure, of course wholesale replacement is the modern standard method of "repair," so this is of little consequence today. Modern Class D amplifiers are typically driven by custom integrated circuits, which may or may not be available when it comes time to repair them.

Links:

http://lib.dr.iastate.edu/cgi/viewcontent.cgi?article=3560&context=rtd