How much attenuation is enough?

Foundations of Amateur Radio Recently I had the opportunity to use a piece of professional equipment to measure the so-called unwanted or spurious emissions that a transceiver might produce. In describing this I finished off with the idea that you could use a $20 RTL-SDR dongle to do these measurements in your own shack. I did point out that you should use enough attenuation to prevent the white smoke from escaping from your dongle, but it left a question, how much attenuation is enough? An RTL-SDR dongle is a USB powered device originally designed to act as a Digital TV and FM radio receiver. It's normally fitted with an antenna plugged into a socket on the side. I'll refer to it more generically as a receiver because much of what we're about to explore is applicable for other devices too. Using your transceiver, or transmitter, as a signal source isn't the same as tuning to a broadcast station, unless you move it some distance away, as-in meters or even kilometres away, depending on how much power you're using at the time. Ideally we want to connect the transmitter output directly to the receiver input so, at least theoretically, the RF coming from the transmitter stays within the measuring set-up between the two devices. Assuming you have a way to physically connect your transmitter to your receiver we need to work out what power levels are supported by your receiver. For an RTL-SDR dongle, this is tricky to discover. I came across several documents that stated that the maximum power level was 10 dBm or 0.01 Watt, but that seemed a little high, since an S9 signal is -73 dBm, so I kept digging and discovered a thoughtful report published in August 2013 by Walter, HB9AJG. It's called "Some Measurements on DVB-T Dongles with E4000 and R820T Tuners". There's plenty to learn from that report, but for our purposes today, we're interested in essentially two things, the weakest and strongest signals that the receiver can accommodate. We're obviously interested in the maximum signal, because out of the box our transmitter is likely to be much too strong for the receiver. We're going to need to reduce the power by a known amount using one or more connected RF attenuators. At the other end of the scale, the minimum signal is important because if we add too much attenuation, we might end up below the minimum detectable signal level of the receiver. Over the entire frequency range of the receivers tested in the report the minimum varies by about 14 dB, so let's pick the highest minimum from the report to get started. That's -127 dBm. What that means is that any signal that's stronger than -127 dBm is probably going to be detectable by the receiver and for some receivers on some frequencies, you might be able to go as low as -141 dBm. At the other end of the scale the report shows that the receiver range is about 60 dB, which means that the strongest signal that we can use is -67 dBm before various types of distortion start occurring. For comparison, that's four times the strength of an S9 signal. So, if we have a 10 Watt transmitter, or 40 dBm, we need to bring that signal down to a maximum of -67 dBm. In other words we need at least 107 dB of attenuation and if we have a safety margin of two, we'll need 110 dB of attenuation, remember, double power means adding 3 dB. So, find 110 dB of attenuation. As it happens, if I connect most of my attenuators together, I could achieve that level of attenuation, but there's one further issue that we'll need to handle and that's power. As you might recall, an attenuator has several attributes, the most obvious one is how much attenuation it brings to the party. It's specified in dB. My collection of attenuators range from 1 dB to 30 dB. Another attribute is the connector it comes with, I have both N-type and SMA connectors in my collection, so I'll need some adaptors to connect them together. One less obvious and at the cheap end of the scale, often undocumented, aspect of an attenuator is its ability to handle power. Essentially we're turning an RF signal into heat, so an attenuator needs to be able to dissipate that heat to handle what your transmitter is throwing at it. I said that from a safety perspective I'd like to be able to handle 20 Watts of power. Fortunately we don't need all our attenuators to be able to handle 20 Watts, just the first one directly connected to the transmitter. If we were to use a 20 Watt, 30 dB attenuator, the signal through the attenuator is reduced to 0.02 Watts and the next attenuator in line only needs to be able to handle that power level and so-on. To get started, find about 110 dB of attenuation, make sure it can handle 20 Watts and you can start playing. Before you start keying up your transmitter, how might you handle a range of different transmitters and power levels and can you remove an attenuator when you test on a different frequency? On that last point, let me say "No", you cannot remove the attenuator when you're measuring a different frequency. I'm Onno VK6FLAB