# BJT Topologies - Design Principles at RF

Here are some rules of thumb I've learned when designing single-transistor BJT amplifiers:

1. High input impedance is good
2. Low output impedance is good
3. Biasing resistors should draw ~10% of the collector current
4. Base terminal is ideally at >2VDC (for temperature stability)
5. Degenerate slightly to provide better linearity

How would these design principles change when working with HF, VHF, or even UHF frequencies?

I imagine that #1 and #2 above would no longer work, because we need to match all input/output impedances to 50Ω (or whatever characteristic impedance is being used). I also expect that degeneration would be a lot more important, to avoid miller capacitance. Is there anything else major I am missing?

• Temperature is always a good consideration when dealing with any type of transistor for any application especially for stability over the operating range. Sep 5, 2014 at 3:52

Single-transistor BJT amplifiers won't come up trumps for all the "rules" you have written at any frequency. For instance, you require an amplifier so I'm assuming you want voltage amplification. OK so far?

The collector delivers a current output signal and not a voltage signal. This means that this current signal only becomes a sensible voltage signal when fed through a load resistor i.e. the collector resistor. This means that the output impedance of a common emitter amplifier is largely dictated by the collector resistor and if you want a small output impedance then you must use a small collector resistor.

A small collector resistor also means smaller gain so, you have a quandary - rule 2 is tending to fight against the word "amplifier" but let's say you can have a 50 ohm resistor in the collector for now and look at the implications.

If you have 50 ohms in the collector then you have an output impedance that is 50 ohms but to get gain you'll need a small value emitter resistor. Let's say you go for a 5 ohms resistor in the emitter - this will give you a gain of ten (you may well be wondering how I jumped to that conclusion of course). Here's why - the AC signal voltage on the base drives through the forward biased base-emitter junction therefore that input base AC signal is also largely seen on the emitter. The current thru collector and emitter are nearly the same hence if the collector-emitter resistor ratio is 10, it's a fair bet that the circuit amplifies by approximately ten.

This leads to another contradiction of the rules you specified - the input impedance is going to be limited to just a few hundred kohms with an emitter resistor of 5 ohms - basically you can make a rough calculation of input impedance by using Hfe and emitter resistor - if Hfe is 100 and emitter resistor is 5 ohms then input impedance is 500 ohms - it's a decent approximation method not an exact method so don't hang me on this.

Given also that your biasing resistor rule implies 10% of colelctor current flowing thru them your input impedance is not high - it's a messy-low-to-middling couple of hundred ohms.

Now that I've assasinated your rules, please consider that a two transistor solution with a common emitter followed by a common collector stage is much more practical to the rules you have written and, for RF amplifiers a common base tolology is quite often implemented followed by a common collector circuit.

• The above answer from Andy aka clearly demonstrates that - in analog electronics - in most (if not in all) cases a trade-off is necessary between several important but conflicting requirements. With other words: Improving one parameter leads to a deterioration of another important parameter. That´s life! And it is one of the most challenging tasks of a good engineer to find the "best" trade-off for each particular application.
– LvW
Sep 5, 2014 at 9:39
• Err, yes, I guess I meant to say "two-transistor amplifiers." I was kind of assuming a common-collector final stage, but I should have clarified that.
– Ryan
Sep 5, 2014 at 14:51
• @Ryan - for UHF you pick a transistor with current gain that extends to the several GHz and, using a tuned circuits in the collector overcomes problems such as miller capacitance because m-cap can be shunted when using a common base configuration - input is then at the emitter (low impedance) but that's not a problem for 50 ohm systems. Wideband UHF amplifiers are pretty much a difficult thing to design and I applaud those who can... maybe one day!! Sep 5, 2014 at 15:10