RF, general concept of impedance matching

[wighou]

Hello,

Matching concept is not very clear for me yet.
I know that if you match your load impedance with complex conjugate impedance of your source, you obtain a maximal power transfer. OK.
In audio applications, when load impedance is fixed, you will obtain the maximal power transfer lowering output impedance of your source.
In RF, this audio concept is not relevant because of the use of transmission lines. This fact would lead to the creation of standing wave / reflected wave that could damage the source. This concept of wave is not very clear. I would prefer work with tension / current.
And I made simulations with various impedance, phases of transmission line, I always obtained, for a fixed output impedance (50 ohms), the most output power when lowering source impedance !!! So, why matching PA output for example and not lowering its output impedance as it is done for audio amplifier for example ?
I hope my answer is clear.

 

[jiripolivka]

The principle of conjugate match works at any signal frequency, be it AC, audio or RF.
I do not know about "simulations" but the impedance is always NORMALIZED to transmission line impedance.

 

[flanello]

With impedance matching you get an efficiency of only 50%.
But for a source with a fixed output voltage and a fixed impedance you'll get the most power into your load if its impedance matched.
For example if you have DC circuit with a 10V source with an internal resistance of 5Ohm and a load of 5Ohm, each the load and the internal resistance gets 5W.
If the load is changed to 10Ohm it gets 4.44W. And if the load is 3Ohm it gets 4.69W. This is less then the 5W in the matched case.

 

[wighou]

Yes, it is true.
But, if you have a fixed load impedance (as said 50 ohms). You can obtain a better power efficiency (> 50%) if you lower your source impedance (bridging). But it is not used for power amplifiers in RF applications because of the use of transmission line (as mentioned in my first post). But, it is a little bit fuzzy for me.
I found interesting posts on this topic : https://www.designers-guide.org/Forum/YaBB.pl?num=1244262521/4 but for the opposite channel (reception).

 

[chuckey]

Conjugate matches are BAD, effectively you have a tuned circuit, so that narrows the bandwidth. You can't inject a signal or measure one with a conjugate match. At more then moderate power you will get flashovers or burnt section due to excessive current.
You say that you can get a better power match if you lower the source impedance, this is true but the circuit will already be optimised for maximum power. So redesigning it for a lower output impedance, will take you back to where you were in power terms- more I, less V.
The impedances chosen were worked out for the minimum amount of material for the maximum power output, 50 ohm for coaxial cable, 600 for open feeders.
Frank

 

[mtwieg]

Yes, it is true.
But, if you have a fixed load impedance (as said 50 ohms). You can obtain a better power efficiency (> 50%) if you lower your source impedance (bridging). But it is not used for power amplifiers in RF applications because of the use of transmission line (as mentioned in my first post). But, it is a little bit fuzzy for me.

I understand what you're thinking, and it can be confusing because there are some unsaid assumptions in play. For example, for an audio amplifier's output is usually an equivalent voltage source with a series output impedance. In this case, if your load impedance is fixed then you should obviously try to minimize the output impedance (it's real part, anyways) to deliver maximum power and get good efficiency. Setting the amp's output impedance to be equal to the load impedance wouldn't make any sense.

Now consider another case where the amplifier's output impedance is fixed but you can control the load impedance. Now conjugate matching might make sense if you just want maximum power delivery. On the other hand if you also want good efficiency then having a higher load impedance might be better. It becomes a tradeoff.

Now consider what happens when your amplifier is effectively a current source with a parallel output impedance. Now everything changes. For good efficiency, the amp's output impedance should be much higher than the load impedance.

However, all of this becomes somewhat irrelevant when you realize that the amplifiers aren't magic sources, they're usually made from transistors making low efficiency amplifiers like class A, which is inherently limited to less than 50% efficiency regardless of the source/load impedance. When that's the case, your best course is to just try to get as much power out of the output as possible, which is why we conjugate match. This applied to all class A (and I think class B) amps, regardless of whether they are RF or audio.

 

[wighou]

When that's the case, your best course is to just try to get as much power out of the output as possible, which is why we conjugate match.

OK. It is true from the point of view of the load. Conjugate matching will bring max power. But, I want to work on output impedance of power amplifier to optimize power eficiency.

 

[vfone]

The maximum power transfer concept was for the first time discovered on the early days of electric motors, when was found that the most efficient transfer of power from the battery into the motor required that the resistance of the different parts of the circuit be the same, in other words matched.
The RF PA matching network affects not only the output power, but also efficiency, linearity, and gain of the amplifier.
Particularly for efficiency optimization, it was determined that the load at the second harmonic is crucial to improving efficiency.
The characteristics and optimization of the collector/drain choke (or λ/4 feeding transmission lines in case of microwave PA's) has also a major influence on the amplifier efficiency.