The term complementary describes a relationship between two different transistors. One transistor is an NPN type, and the other is a PNP type. These in-family relationships extend to MOSFETs and NMOS technology, where one device is an N-channel, and the other is a P-channel. Like the symmetry of human hands, the complementary semiconductors share similar structures and performance. The true nature of the complementary semiconductors is revealed when we observe that the flow of current is reversed between the NPN and PNP devices.
One of the most common complementary examples is the 2N3904 (NPN) transistor. The complement to the 2N3904 is the 2N3906 (PNP) transistor. Both transistors were introduced in the mid-1960s. The proximity of their part numbers implies that they were developed and then registered with JEDEC at the same time. Another classic international example is the BC548 (NPN) and PNP BC558 (PNP) pair. Once again, these transistors were designed together and then registered with European Pro Electron now ESIA. Two additional complementary transistor examples are shown in Figure 1 including the larger TO-220 packaged TIP41 (NPN) / TIP42 (PNP) pair and the smaller TO-92 MPSA06 (NPN) / MPSA56 (PNP) pair.
Figure 1: This circuit features the complementary TO-220 packaged TIP41 (NPN) and TIP42 (PNP) transistors. The smaller TO-92 packaged MPSA06 (NPN) and MPSA56 (PNP) are also complementary.
Frequently Asked Questions
The remainder of this article is dedicated to answering common questions about complementary semiconductors.
What is the complementary transistor to the 2N2222?
Many people recognize the famous 2N2222 (NPN) transistor, but few recognize the 2N2907 (PNP) complement. Note that there are many JEDEC numbers between 2222 and 2907. Both transistors were released in the 1965 time frame.
How can I identify a complementary pair in a schematic?
The complementary pair is often found in the output stage of an amplifier. An example is shown in Figure 2. Here we see that Q1 will pull the R_Load (speaker) up towards the positive 30 VDC rail. Transistor Q2 would pull the R_Load down toward the -30 VDC rail. Working together, they develop the +/- 30 V peak output signal. Similar schematics can be found for MOSFETs and NMOS output channels. In all cases the complementary semiconductors work together to drive the output signal.
The term complementary does not always imply a push-pull output stage. Instead, the term complementary implies transistors with similar performance characteristic between the NPN and PNP types. Circuit designers will often use complementary types in a given circuit. For example, the input stage, constant current source(s), voltage amplifier, and drivers, may all be selected from the complementary family.
Figure 2: The NPN and PNP transistors in this audio output stage are a complementary pair.
Why are complementary transistors important?
Complementary devices were not always available. For example, vacuum tubes are limited, as they only allow current to follow in one direction. Likewise, the earliest power transistors were generally limited to PNP Germanium types.
The answer to our question is found in audio amplifiers from the early 1960. Recall that most amplifiers, from vacuum tube to the latest linear op amp, are built using a push-pull pair. One transistor is responsible for the positive half-cycle (pulls up) while the other is responsible for the lower half-cycle (pulls down).
Before complementary transistors were available, designers had two options:
-
Output transformer: This was a costly solution for push-pull operation. It is also difficult to make a transformer linear over the entire audio range.
-
Quasi-complementary: Given two PNP transistors, one can be made to look like a NPN by including an inverting transistor. Unfortunately, these quasi-complementary PNP and NPN pairs do not match each other in terms of linearity and overload characteristics resulting in sub-optimal performance.
Many of the earlier amplifiers also used a driver transformer to level shift and bias the output devices into proper conduction. An example of non-complementary amplifier is included as Figure 3. This 25 W amplifier published in RCA’s Transistor manual (SC-10) features two PNP Germanium type 2N2147 transistors and an input transformer. A period amplifier is shown in Figure 4. This Sylvania stereo featured Germanium output transistors in a quasi-complementary circuit. The drive transformers are located to the left of the heatsinks. This video describes the setup and bias scheme.
With the advent of complementary NPN and PNP transistors, many of these problems were eliminated. The biasing networks are simplified, direct signal coupling is used, and there is no need for transformers. The resulting circuits are more reliable and cost less. On a related note, the designers could now increase the amount of global negative feedback. These changes improved linearity, lowered output resistance, and reduced power supply noise.
For other applications, we recognize that designers often use a variety of NPN and PNP transistors in any given circuit. With complementary transistors, the circuit designer has confidence that all transistors have the same performance in terms of frequency, gain, thermal, and frequency response characteristic.
History The PNP Germanium power transistors are generally windowed in the JEDEC 2N100 to 2N3000 range. Silicon power transistors and NPN types generally have numbers greater than 2N3000. By this measure, the famous 2N3055 would have been among the earliest NPN silicon power transistor on the market. The 2N3055 appears in the RCA’s 1964 edition of the Transistor Manual (SC-11). Note that the first small-signal silicon transistor was the 2N696. It was introduced in the late 1950s by the “traitorous eight” who defected from Schokley Semiconductor to form Fairchild.
Figure 3: Schematic of a first-generation transistor audio amplifier featuring a quasi-complementary output stage using the RCA 2N2147 PNP Germanium transistor.
Figure 4: Image of a Sylvania amplifier featuring quasi-complementary output stage. The driver transformers are immediately to the left of the heatsinks.
How can I locate complementary transistors within the DigiKey system?
Identification of a transistor’s complement is not a straightforward process as the JEDEC and ESIA number are rarely consecutive. The 2N3904 and 2N3906 pair is a good example. We note that the 2N3903 (NPN) / 2N3905 (PNP) pair and the 2N3904 (NPN) / 2N3906 (PNP) family members were developed together and submitted to JEDEC at the same time. JEDEC assigned the lesser sequence numbers to the NPN types and the higher sequence numbers to the PNP types. As a result, there is a gap between the complementary types. The classic 2N2222 (NPN) and 2N2907 (PNP) is more challenging, as there were many JEDEC assignments between the complementary pairs.
Years ago, it was easy to identify complementary transistors as they were usually listed side by side in a table. After all, they have the same specifications for V_CE, I_C, P_D, and current gain. Unfortunately, paper catalogs are a thing of the past. DigiKey’s last paper catalog was issued back in 2011.
Assuming you know the designation for one transistor, you can find the complement by:
-
Direct reference in a datasheet
-
Reverse parametric lookup as shown in Figure 5. We cast a wide net by selecting as few boxes as possible. Yes, the net will catch unrelated products, but the desired part can be spotted. In this example, we find the complement of the MPSA06 (NPN) of the MPSA56 (PNP)
-
A general internet search will often identify the complement.
-
Use of an AI.
Figure 5: This DigiKey reverse lookup method may be used to locate the complement of the OnSemi MPSA06 transistor.
Do all components have a complement?
No, we should not expect to find a complement for all transistors. This is especially true for specialized devices such as high voltage transistors or the high-power MOSFETs used in power supplies and motor drives.
What are the two 2 types of transistors?
Bipolar Junction Transistors (BJT) are available in NPN and PNP types. We start by recognizing that there are two types of crystalline structures. The N material have been modified (doped) to have an excess of electrons, while the P crystal structures have a deficit commonly referred to as holes.
A transistor may be modeled as a sandwich constructed from the two different types of semiconductor materials. The NPN transistor sandwich is constructed using N material as the bread with a thin marshmallow P material in-between. The resulting NPN sandwich is then connected to the external world using bond wires. The PNP is similar except we reverse the bread and marshmallows.
This marshmallow (collector, base, emitter) sandwich is a good and memorable starting point. However, it is just a model. When compared to the engineering challenges of making a transistor, it’s like using the concentric ring model to study quantum mechanics. It’s not wrong, it’s just incomplete.
With regards to complementary devices:
-
Recognize that the NPN and PNP have similar characteristics.
-
The flow of current is reversed. For example, the current for a NPN transistors flow from collector to emitter while the PNP transistor’s current flow is from emitter to collector. This assumes a conventional current flow from a positive to negative.
Tech Tip: The mnemonic Not-Pointing-In is an easy way to remember the schematic symbol for the NPN transistor.
How to match complementary transistors?
Generally, this step is not required. The complementary transistors are inherently matched. Also, most circuits include negative feedback to automatically correct for minor variations in the transistors. For example, an audio amplifier may include 30 dB or more of negative feedback to improve linearity.
For select applications, it may be desirable to use an array were both NPN and PNP transistor are integrated into the same package. An example is the Nexperia PBSS4140DPN,115 6-TSOP part.
Tech Tip: Transistors are sensitive to changes in temperature. Any matching will be undone if the temperature is different between the devices. This can be mitigated by using arrays that integrate multiple transistors into the same package. This is the same principle as mounting power transistors on a common heatsink.
You may also wish to use a transistor curve tracer such as the Digilent Analog Discovery coupled to the Transistor tester module. This video provides a brief introduction to matching NPN transistors:
What is a Darlington pair?
A Darlington pair may be described as:
-
a circuit configuration with one transistor driving another. Together the circuit acts as a single transistor with high gain.
-
a specially designed transistor with two transistors in one package. The resulting transistor has high gain.
With respect to complementary transistors, recognize that complementary Darlington transistors are available such as the TIP120 (NPN) and the TIP125 (PNP). These are considered jellybean components as they relatively common and are produced by multiple manufacturers.
Tech Tip: Jellybean parts may be difficult to locate, as most modern components contain a suffix which obscures the legacy part number. We can overcome this problem by searching using the keywords “TIP120 BJT”. See this article for additional information about expanding DigiKey parts searches.
Parting thoughts
Complementary transistors are a core technology. They shine when used in push-pull applications such as an audio amplifier or the output stage of an op amp. They are also a convenient way of identifying similar performing components to use in a circuit. For example, a manager who desires to minimize the shop’s line count may state that the BC548 (NPN) transistor is used for all projects unless you have good reason to do otherwise. To me, this implies that the complementary BC558 (PNP) is also available.
Please share your questions and comments. Stories, like the manager insisting on the BC548 are especially welcome.
Also, let us know if something is missing from the FAQs.
Best wishes,
APDahlen
Related Information
Please follow these links to related and useful information:
About this author
Aaron Dahlen, LCDR USCG (Ret.), serves as an application engineer at DigiKey. He has a unique electronics and automation foundation built over a 27-year military career as a technician and engineer which was further enhanced by 12 years of teaching (interwoven). With an MSEE degree from Minnesota State University, Mankato, Dahlen has taught in an ABET-accredited EE program, served as the program coordinator for an EET program, and taught component-level repair to military electronics technicians. Dahlen has returned to his Northern Minnesota home and thoroughly enjoys researching and writing articles such as this.