3
\$\begingroup\$

I'm trying to create a basic RF power meter using the LTC5532.

I want to add a SMA connector, and termination, so that the device has a 50ohm input impedance. I'll be using it primarily at 2.4Ghz, but if I can make it work down to 900Mhz, or up to 5.8Ghz that would be nice.

I'm having trouble figuring out what termination parts I'll need for the impedance matching. There's a table of input impedance with columns for reactance (-70.13ohms at 2.5Ghz), and resistance (34.82ohms at 2.5ghz).

How do I use these to choose a termination resistor?

There is a schematic for the dev board available here.

\$\endgroup\$
3
\$\begingroup\$

First, +1 for linking to the datasheet and the schematic.

It depends how you want to use this (reception off air, or cable connection from a transmitter), what accuracy you're expecting, which package the dev board uses, and what the layout of the dev board is.

You'll notice from the datasheet that the real part of the input impedance at 2.4GHz and up is already lower than 50 ohms. Adding a simple shunt resistor will make things worse. A shunt R here would be useful below 1GHz.

You'll also notice that the resistance changes with power level, so you will not achieve a consistent match as your level varies.

However, the input impedance is not far from 50 ohms, and the trajectory of S11 round the Smith chart is fairly tame, so there's nothing exciting going on that might trip you up at a certain frequency. These development boards are built intending to be used with 50 ohm connectors and equipment. Although it's not perfectly matched, simply using with a 50 ohm source will not give you not much level error, maybe a dB or two, and over your extended range from 900M to 5.8G.

If you want to make your dev board 'better' over a broad band and range of input powers, then adding an input pad is probably the easiest way to go about it, if you can afford the drop in level. Adding (for instance) a 6dB pad will only lose you 6dB of input power, and will theoretically improve your S11 by 12dB. This is known in some circles as 'masking' the input impedance. If the layout permits, you could use the R1 site to place a shunt resistor, and then cut the track with a scalpel immediately before and after it to add two series resistors to make a T pad. You should be able to find T pad values with a simple search. However, getting a pad to work well at 5.8G is no mean feat, layout is everything. While at 900MHz, a pad should make things better, at 5.8GHz, it could easily make it worse in practice.

\$\endgroup\$
  • \$\begingroup\$ These are excellent tips, thank you. I didn't know about the masking option. But assuming I only care about 2.4ghz, is there not some kind of LC network I can use to match a 50ohm antenna "perfectly"? Related question, is the resistance and reactance in these charts in parallel or series? \$\endgroup\$ – Drew Feb 19 at 9:18
  • \$\begingroup\$ As you can see that the input match varies with the power, obviously no, you cannot match perfectly, even at a single frequency. However, if you want to pick a spot power level, and a spot frequency, then yes, you can tune it to perfectly 50 ohms with an LC tuner. The techniques to do so are beyond the scope of this sort of QA forum to teach. Read up about how to use a Smith chart to design matching circuits. \$\endgroup\$ – Neil_UK Feb 19 at 10:21

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.