A Planet Analog article, “2N3904: Why use a 60-year-old transistor?” (Ref. 1) by Bill Schweber, inspired some interest in this old transistor and how it’s commonly used, and if any uncommon uses might exist. Here’s one we played around with.
The Linear Technology Application Note 47-D: “High Speed Amplifier Techniques” (Ref. 2) by Jim Williams offers an interesting side road to usual transistor use, where a typical fast pulse transistor is utilized in avalanche collector-to-emitter breakdown VBCEO to create sub-nanosecond pulses. The 2N3904 will work in this configuration, but requires a high voltage (>100 V) like the pulse transistor to reach the VBCEO breakdown, and produces a slower pulse, being a slower GP transistor.
A while back, I had measured the reverse breakdown of the 2N3904 base-emitter junction and noted the small area of negative resistance where the junction current reduces as applied reverse voltage increases (Figure 1).
 |
|
| Figure 1. | Measurement of the reverse breakdown of the 2N3904’s base-emitter junction, showing a small area of negative resistance. |
This base-emitter breakdown is much lower than the collector-emitter breakdown and might serve as a lower voltage version of the avalanche pulse generation method described in App Note 47-D.
A simple circuit was created with the 2N3904 emitter connected by a 100-kΩ resistor to a variable supply set to ~14 VDC. A shunt capacitance of 10 nF from the emitter to ground and a 50-Ω resistor from the collector to ground. Just two resistors, a capacitor, and the 2N3904 are all that’s required to create a simple relaxation oscillator (actually, the 50-Ω resistor isn’t required).
Figure 2 shows the result with the DSO AC-coupled blue trace, the relaxation voltage at the transistor emitter, and the DC-coupled magenta trace, the voltage across the 50-Ω resistor from the collector to ground (remember the NPN is upside down or inverted!).
 |
|
| Figure 2. | Waveforms of the simple relaxation oscillator circuit with the AC-coupled blue trace and DC-coupled magenta trace. |
The pulse across the 50-Ω resistor in Figure 3 shows the avalanche current in more detail, where this current is ~ 2 V peak across the 50-Ω resistor, or ~40 mA peak. This isn’t fast, however, the 2N3904 is a general-purpose (GP) transistor that is not intended for speed.
 |
|
| Figure 3. | Avalanche current shown in more detail on the DSO, showing a ~40 mA peak. |
Utilizing faster transistors such as the 2N2369 should produce narrower pulses with faster rise times. Whether these produce faster rise times and narrower pulse widths than in the collector-emitter avalanche breakdown method from App Note 47-D remains an experiment waiting for those interested. Intuition indicates the “normal” avalanche collector-emitter mode will be faster, though!
Anyway, I hope folks find this simple and unusual use of these old standby 2N3904 transistors interesting, I certainly did!!
References
- Schweber, Bill. "2N3904: Why use a 60-year-old transistor?"
- Williams, Jim. "AN47 - High Speed Amplifier Techniques."