# Understanding impedance matching

Discussion in 'General Technical Questions and Answers' started by VU2TVE, Oct 15, 2020.

1. ### VU2TVEHam MemberQRZ Page

Hi, I have a noob RF theoretical question as I try to understand the design of RF amplifiers. As an example of an amplifier I am considering to build, I am looking at the MRF101AN LDMOS device, and looking to build an amplifier in the 88-108MHz range. From the datasheet, when I look at the impedance numbers they are listed as below.

Code:
```f(MHz)           Zsource                 Zload
87.5               8.52 + j12.46        13.15 + j5.4898
98                  10.59 + j14.03      13.12 + j5.2110
108                812.21 + j15.02    10.74 + j5.52```

Next, when I look at the reference circuit, it looks like this:

Let's say I am looking at the input matching network. From the datasheet, the values of C1, C2 are 200pF, and L1 is 36nH.

Using this calculator, if I punch in the following numbers (frequency 98e6 i.e. 98MHz, source impedance: 10.59 + j14.03, load impedance: 50 + j0.0), the values I get for hi-pass T-matching network are these: C1: -35pF, C2: 31pF, L1: 42nH) which are clearly different than the values used above in the datasheet.

I'm trying to understand the circuit and the numbers, so any pointers, or clarifications would be much appreciated.

2. ### WB2UAQHam MemberQRZ Page

I have been interested in making a PA myself having some RF power mosfets in my junk box. I have so many other things to study and work on that I have not arrived there yet However, I have worked on many existing designs in terms of testing and tweaking.
I looked at the "calculator" and I believe the circuit config you chose can not achieve the match with realizable capacitors and/or inductors.
Chose a different circuit or topology. The bottom line is to get the RF gate voltage to swing to the level needed to achieve the output power. You are driving a fairly low resistance and capacitive reactance looking into the gate (150 pF).
I'm sure others with more design experience will jump in and I am interested in what they have to say

3. ### KB1CKTHam MemberQRZ Page

Been forever since I was in college and looked at these things. But my recollection was this: design the output filter for 50 ohm in/out. Then the filter can be altered for impedance transformation, like say 5 ohm in and 50 out. I don't remember at all how to do this, sorry. I can't find my college textbook on filters, which is where I think I saw the technique. But I think that is where you want to dig into: first look at filter design, then after that there may be blurbs about how to alter the filter for impedance in/out.

I think wide-band amps, particularly if push-pull, may use a balun to step up the impedance. For a single band application, and for a one-off application, it may be easier to design an un-un here, or an L matching network, then your classic 50 ohm filter. It may be "inefficient" with parts count and space used, but if you are just experimenting and playing around, it may allow you to finish a project, understand the big picture, and then allow you to come back around to the issue at a later date. It may also allow you to modularize your design, for ease of debug.

4. ### VU2TVEHam MemberQRZ Page

Since I am new to this as well, I thought of first understanding an existing design, rather than trying to design a new one myself. The design I am discussing is from the official NXP website, which I am trying to understand. So, the working assumption is that "the circuit is good". I am trying to understand how is it good?

5. ### VU2TVEHam MemberQRZ Page

The design being discussed above is probably a non-wideband design which I am trying to understand. For my application I need a bandwidth of 10MHz max in the VHF band, so I can probably get away with the complexity involved in wideband designs. Just trying to understand how the existing NXP official design for the MRF101AN works, as in, how do the numbers makes sense.

6. ### INTEGRALQRZ Member

Hi,
Let’s first clarify terminology. The “Device Under Test” wants, per datasheet, to see Zsource of (10.59 + j*14.03)Ohm when looking (from the right side) into the Input Matching Network. This Zsource is the output impedance of the Input Matching Network, which we want to design. (Note that this Network output impedance can be different from the DUT input impedance, which actually we don’t know.) The input of this Matching Network is assumed to be connected to a source that has an output impedance of 50 Ohm. So, the Network needs to convert 50 Ohm into Zsource, at 98 MHz.

I apply a Smith-Chart (Smith V4.0) showing the matching. We start at 50 Ohm (DP1, center of the graph). The capacitor C1=200pF brings us to TP3. From there, the inductance L1=36nH brings us to TP4. And finally, C2=200pF moves us to TP5 which is reasonably close to our desired destination Zsource (DP2). Bingo!

Cheers,
L.

7. ### AA3EEHam MemberQRZ Page

The ZSource and ZLoad are the impedances the device wants to see looking outward. The networks in the datasheet will transform 50 Ohms resistive to those impedances (or close anyways).

In broadcast use, there are two types of amplifiers you'll find in data sheets: narrowband and wideband or wideband/fixed tuned. Narrowband is obvious, the amp is designed for a small range of frequencies. Small means bandwidth compared to the actual frequency, such as 144 to 148 MHz is 4 MHz out of 146 (center). The UHF TV spectrum is VERY wide, so the base design will cover the whole band, but with certain components chosen for optimum performance within that huge span. So you might see the same design but with a component selection chart for low, middle and high channels.

Dave