Measuring Power - Attenuators and 50 Ohm Termination

Question

I'm attempting to measure the output power of my transceiver + linear amp combination and I would like to confirm that my measurement approach is correct. I need to get the output power down to a level that's safe for my spectrum analyzer.

In particular, I need to confirm that I should terminate my attenuation chain with a 50 ohm load and it's OK to measure across the dummy load with my oscilloscope probe in high-impedance (x10) mode (my scope doesn't have a 50 ohm mode).

I get confused sometimes and reason that 50 ohms in parallel is 25 ohms and that doesn't seem correct. However, I'm guessing the amp "sees" 50 ohms and as I understand it, the attenuator must be terminated in 50 ohms to achieve its labelled attenuation. So, maybe it's when you look back into the 50 ohm loaded attenuator chain that you see 25 ohms (i.e., from the scope) and maybe that's OK. Or maybe it's not? That's my confusion.

EDIT

I believe the large attenuator performs well in the HF frequency range based on previous measurements. Simple tests at DC show it's certainly a 50 ohm in/out attenuator.

The purpose of all this is to understand the impact of placing a homemade low-pass filter after the linear. For that, I intend to use the Tiny SA, but since I'm not sure what the output power of my amp is, I'm cautious to measure the attenuated output with an oscilloscope, which has a high impedance input and an overall better tolerance for gaffs as I experimentally reduce the signal power to a desired level.

As I understand it (but I may be incorrect), the Tiny SA would like a -25dBm input signal, so I'm aiming for this level.

Finally, I'm not looking for pin-point accuracy here. I realize the oscilloscope has its own tolerances and the cabling and all the rest of it will introduce errors. Not to mention the fact that the attenuators themselves probably aren't perfect. What I'm more concerned about is making a measurement that's fundamentally wrong and I end up calculating substantially more or less power than I actually have.

Here's a schematic picture of the exact measurement setup. Frequency is around 7MHz or 40m band.

Answer 1

You should ideally terminate an attenuator output with 50 ohms, when measuring with your high impedance scope. The behaviour of attenuators is undefined when operating into an open circuit.

Attenuators can be Pi or T networks but the principle is the same. The values depend on the attenuation. For example:

^{simulate this circuit – Schematic created using CircuitLab}

The attenuator is symmetrical, and when you connect another 50 ohm load to one side, the other side will present 50 ohms. It's easy to work out the resistor values and attenuation, just use DC analysis, or use a calculator.

You can see from the right-hand diagram that the voltage on the output will be halved by connecting a 50 ohm load, and it won't depend on the rest of the circuit. You could safely measure this without a termination, just divide by two.

The left-hand attenuator basically passes the load through almost unmodified, so the voltage you see depends on the source impedance of the circuit. And filters are not 50 ohm devices in the stop band, so measurements taken assuming this will be all wrong.

Your high power attenuator looks great, its VSWR will be fine to 100 MHz. If it's 30 dB (800 series, 53 parallel) then it's OK not to terminate it, just know that the voltage is doubled. So

Power_into_TinySA = $10\log_{10}({{V_{RMS}}^2\over50}) - 33$ dBm