# Measuring dBs with software defined radio

If the "floor" of the graph in my SDR is at -110dB, and when I key up, it goes to -70dB, and the wattage is 5W....is there a direct correlation between dBs and wattage?

## 2 Answers

Decibels express ratios. SDRs are often not calibrated to any reference, so when it says "-110 dB", that's relative to some arbitrary reference power. Often the reference is "full scale", meaning the full range of the ADC. Still, that doesn't directly relate to power.

If the measurement goes from -110 dB to -70 dB, that's a difference of 40 dB. That's a ratio of:

$$10^{40/10} = 10000$$

So whatever you were measuring, the power is now 10,000 times greater.

If you have a signal source of some known power, you can calibrate the SDR. Say you have a signal generator that you know outputs 1 mW, and the SDR reads this as -10 dB. Now adding 10 dB to any measurement from the SDR yields decibels relative to 1 mW, or dBm.

Keep in mind when measuring powers on a waterfall display the total power must be integrated over the bandwidth it occupies. For an unmodulated carrier with very narrow bandwidth this may not matter, but if the signal is modulated to have 4 kHz bandwidth, the transmitter power is now spread across multiple bins in the waterfall. The power represented in each bin must be summed to calculate the total transmitter power.

• Note that many SDR receivers are pretty sensitive receivers: for example, for most USRPs, -15 dBm (that is, less than 1/30 of 1mW) is the "maximum damage free input power", which means that at that power the device is already pretty deafened by the sheer loudness. So, if you have a signal source of known power, try that, and maybe add a small bit of attenuation (e.g. 3dB attenuator), to check if you're in a linear region. – Marcus Müller Feb 15 '18 at 23:01

I use this as a very practical A/B comparison "measurement". Not very scientific, but to give me an indication on antenna performance.

Example:

I have two 2m dipoles, which I call my "reference antennas". They are as identical as possible.

I place them 2m appart, with an SDR on one end, and a HT 2m TX on the other.

I will key up the HT, and I get a X dB rise in the spectrum. This is my reference 0 point.

I will then replace the TX antenna with another home brew antenna. Key up again, and this will result in Y dB rise in the spectrum.

Suppose that Y is better than X, then I can conclude that the Y antenna is Y-X dB better in this particular setup

This itself does not give me any dBi or dBd or any other scientific measurement other than one is better than the other.... I can use this to further improve the antenna's.

A simple, and very rough, "comparison tool"

[EDIT based on comment]

So lets same my reference is 40 db....i test somewhere else and its 45db....somewhere else is 5db better

The trick to doing an A/B comparison is to only change 1 element of your reference setup.

So to answer your comment question: if you change location, then ensure that only the location changes, and the rest of the setup remains the same. The "measurement" is now a comparison in relation to the "location change", not a change of equipment.

Example: If I am using my own reference antenna's which I normally place at a specific point in a field behind my property, about 20 meters away from any building, I will get a particular spectrum response. If I move this whole setup to another location, about 1 meter away from a big steel garage door, I get another measurement. Either better or worse, but the location and the surroundings have changed the properties of the setup and is influencing the "measurement"

This would be a test to see how the big steel garage door (and other environmental differences) is influencing the received RF in that particual setup.

So yes; if by moving your test to "somewhere else" gets you a +5dB difference, then the new location is +5dB "better" on RX. You just need to figure out why that is. It might not be actually "better" at all, just an influence of environment. And the "5 dB" would be questionable as well.

I would not use such a setup to test "locations" I would more be inclined to use it as comparison of equipment such as: transmitters, receivers, antenna's, coax... Furthermore I would use it as far away from property, buildings, trees and the like to minimize influence from that.

As already given in a very good answer, there is a relation between dB's "measured" and "Watts", but to actually derive one from another using a simple SDR would be very rough and open to influence of other environmental factors. Unless you have calibrated equipment, and a very (RF) clean room.

In this answer I have used the word "measured" as quoted "measured", as I would not consider this a true scientific measurement. It really is a comparison using a particular type of equipment, which is far from calibrated.

Who says that if the spectrum has a 40 dB response on your computer screen, that it is a real 40 dB... what is the influence of the LNA in the SDR ? Is the software the A/D converters, the LNA, the tuner, filters, all calibrated ? --> no, they are not.

My advice; don't try to use this [setup] to try to calculate "Watts" based on "dB's".

• So lets same my reference is 40 db....i test somewhere else and its 45db....somewhere else is 5db better? – FACTORY909 Feb 13 '18 at 18:32
• @FACTORY909 answer updated – Edwin van Mierlo Feb 14 '18 at 8:08
• "Who says that if the spectrum has a 40 dB response on your computer screen, that it is a real 40 dB" A lot of work goes into making the receiver response linear. If it were not, intermodulation and harmonic distortion would render the receiver useless. So when the receiver indicates a 40 dB change, and you're not operating the receiver outside it's linear range for example by overloading it, you can be assured it really is 40 dB to a precision sufficient for practical engineering. – Phil Frost - W8II Feb 14 '18 at 13:43
• @PhilFrost-W8II fair enough, but the OP did not specify any particular software, and therefor it is still to be taken with some caution. Furthermore, most SDR's come with a built in LNA, and the 40 dB on screen may be less on the receive before LNA, as you don't know if the hardware is linear... eitherway I would not use and "unspecified SDR" with "unspecified software" with "unspecified settings" as a refernce, nor in this answer... I was just being cautious to take "values" as basis for any calculation(s). – Edwin van Mierlo Feb 14 '18 at 13:55