I'm experimenting with RF feedback amplifiers (this topic is well covered in "Experimental Methods in RF Design" book, if anyone would like to know what that is). Particularly I'm interested in finding inexpensive transistors available in local shops with high fT to build amplifiers with flat frequency response. Well-known 2N3904, 2N2219A have fT = 300 Mhz and it's hard to build a broadband feedback amplifier using them.
I've discovered that 2N2369A and it's Russian equivalent KT3142A both have fT = 500 Mhz, don't cost much and work well as a substitution for 2N3904.
As a substitution for 2N2219A I decided to try 2SC5694 (Ft = 330 Mhz, datasheet [PDF]) and 2SC3953 (Ft = 400 Mhz, datasheet [PDF]). The following circuit was used for testing:
(As a side note, I know it's not matched well to 50 Ohms. It's input and output impedance is closer to 25 Ohms and is a little reactive. However this circuit is used as a second stage of three-stage QRP power amplifier, similar to one used in Summit Prowler 6B and Summit Prowler 7 transceivers by Paul Taylor, VK3HN. The third stage is not quite 50 Ohms either, and 25 Ohms impedance of this stage gives a better frequency response of the entire PA than a "right" 50 Ohm stage. I've checked.)
Here are frequency responses of this stage with 2N2219A (yellow), 2SC5694 (purple) and 2SC3953 (blue):
2SC3953 flattened the frequency response because of the higher fT, just as I was expecting. But surprisingly 2SC5694 worsened the frequency response. Also the SWR rised from about 2.0 (which is OK for a 25 Ohm stage) to almost 10. I though maybe the transistor is damaged or I made a mistake during soldering. So I resoldered the circuit with another 2SC5694 and got the same result. I also tried to drive this stage with -5 dBm, almost as it would be in a real PA, instead of -20 dBm, but it made no difference. After changing the transistor back to 2N2219A everything works fine.
I'm not that long into homebrewing and apparently I don't know something about NPN transistors, or missed something while reading EMRFD. This is the first time I've encountered such behavior. Could you please explain what causes it? Which parameters of the transistor I should be more mindful of next time?
Best Answer
The first difference I notice when comparing the datasheets is that the 2SC3953 has an output capacitance of 2.8pF while the 2SC5694 is 28pF. The higher output capacitance goes along with the much larger current capability. This larger output capacitance may also be causing the mismatch you observe.