Using SA to measure broadband instantaneous power??

[rs20]

First, my problem: I'm interesting in using a ZX60-83LN amplifier to amplify the signal coming out of an 5~6GHz (WiFi) antenna. The amplifier specifies a output-power-at-1-dB-compression of 19.4 dBm. Now, I can only assume that this 19.4 dBm is measured as a total power across the amplifier's entire bandwidth. So I have the antenna set up in its final position, and I'm interested in measuring the power coming out of the antenna across the entire 5~6 GHz range. (Acknowledging of course that taking a single measurement doesn't predict everything the antenna might ever see, but let's ignore that for now.)

The only relevant equipment that I have is the lovely Signal Hound BB60C Signal Analyzer, which has 27 MHz instantaneous bandwidth.

Now, WiFi is very spiky/bursty/packety, and that the amplifier very much cares whether spikes/bursts at (for example) 5.1GHz and 5.9GHz are simultaneous (because the powers should be summed if and only if they're simultaneous), but the SA can only look at one band at a time and therefore has no idea if the packets line up or not. Therefore, my instinct is that it is, at least at a pedantic level, it's impossible for me to perform the measurement I want to perform using this SA.

I don't really have a particular question, but some discussion points:

• Am I right in my reasoning?
• What would be the most ideal tool for this situation?
• How would you proceed if you only had my SA?
• To what extent would it be considered acceptable to just assume that the different bands are operating in a completely independent and uncorrelated way? (Obviously this depends on the situation/industry/usecase, but still)

[awallin]

use a sine-wave of known amplitude (from the SA TG, or external siggen) to characterize your amplifier (gain, compression point etc.).

it's *very* unlikely that your antenna is outputting 0 dBm which would get amplified to +20dBm (unless your antenna already has lots of gain...)

for looking at the average power of WiFi in some particular band I guess you could just let the SA average for a minute or two, and then move the band to some other place, and stitch together a wide-band average power picture?

[rs20]

Thanks for the response!

use a sine-wave of known amplitude (from the SA TG, or external siggen) to characterize your amplifier (gain, compression point etc.).

I could, but a) I don't have the amplifier yet (still selecting) and b) it seems like I can just trust the datasheet for those figures anyway.

it's *very* unlikely that your antenna is outputting 0 dBm which would get amplified to +20dBm (unless your antenna already has lots of gain...)

Thanks, that's reassuring to know, but I'm still curious about the answers to my questions because I'll be moving on to designing the rest of the system (mixers, possibly preceded by further amplification) so the same issues will still apply.

for looking at the average power of WiFi in some particular band I guess you could just let the SA average for a minute or two, and then move the band to some other place, and stitch together a wide-band average power picture?

As I explained in the OP, average power is of no interest to me. I'm looking to measure the peak instantaneous power, which isn't so easy to figure out from separate measurements across the bandwidth (without making the questionable assumptions mentioned in the OP.)

[radioactive]

As I explained in the OP, average power is of no interest to me. I'm looking to measure the peak instantaneous power, which isn't so easy to figure out from separate measurements across the bandwidth (without making the questionable assumptions mentioned in the OP.)

In my experience, measuring WiFi power accurately is difficult without a WiFi analyzer, but you can probably get in the ballpark.  I assume you are dealing with OFDM / QAM64/256/etc.  One way you might get close is to measure the power spectral density with peak-hold (wait for peaks to develop) at a lower RBW,  maybe like the FCC test for this  (+8dBm max in any 3kHz BW).   So, if you measured  8dBm max with RBW of 3kHz,  and occupied channel bandwidth is 16 Mhz,    10 * log10(16e6)/log10(3e3) = 20.7 dB.    So,  28.7 dBm  average.  For OFDM / QAM modulations,  you have to account for the PAPR  (peak to average power ratio).  Recommended seems to vary, but seems to be a good rule of thumb to shoot for about 7dB headroom on the PA for OFDM/QAM modulation.  The PA should be very linear ( up to to 35.7dBm for the example) to keep EVM as low as possible.  Doesn't look like the software for your analyzer can do OFDM WiFi analysis from what I can see, so seems like you might have to resort to some ballpark method like this.

Also,  instantaneous peak power for these types of modulations with very high PAPR might be more accurately described with CCDF:  http://rfmw.em.keysight.com/wireless/helpfiles/n7617/Content/Main/Understanding_CCDF_Curves.htm