amplifier measurements using dBuVemf units

[Leiser]

Hi all,

I have to do some lab measurements for an amplifier, LNA, for AM and FM bands. Reading the customer´s specs AM 1dB compression point and AM 2nd harmonics, AM 3rd IMR and FM 2nd IMR measurements are specified in dBuVemf. For example:

AM 2nd harmonic test: Signal 100dBuVemf input frequency 810KHz at FET Probe load. Measurement should be min 130 dBuVemf.

I do not understand what input power should I use and how to read the result as I am using a spectrum analyzer in dBm or dBuV. I found the relationship between dBuV and dBuVemf. Shall I just apply that? which I believe is 6dBs.

thanks in advance

 

[zappedagain]

If that is the same amp as your previous post, that looks correct. It appears they use dBuVemf because the high impedance doesn't load the port. If you have a real 'open circuit' than your 6dB conversion is correct. Your S11 measurement may show it is different.

I've never seen this dBuVemf technique before; I'm used to tuned circuits where you are always dealing with a fixed load (and working for maximum power transfer by matching your source impedance to the S11 of the preamp). I guess if you want your circuit really, really tiny you'll skip the matching at these low frequencies and take the performance hit.

 

[VEBill]

They're specifying the input stimulus as a voltage.

By "100dBuVemf" (for example), they simply mean 100,000uV (0.1 volts) - a good strong signal - into the high Z input.

They're specifying a FET probe to make sure that the stimulus level doesn't change (much) when the probe is disconnected from the UUT input.

As discussed in the other thread, using voltage (not power) makes perfect sense for a UUT with a very high-Z input where 'power' measurements are at least one-step less clear.

The 6dB factor mentioned is based on the assumption that the open circuit voltage halves when loaded with a matching value (half voltage, thus half current, therefore one-quarter power, which equals -6.02 dB).


(Free advice follows

Personally, I'd simply skip the entire 'power' confusion and stick with a crystal-clear measurement of the applied stimulus voltage using the specified FET probe and an oscilloscope (read as RMS).

QA folks usually don't readily accept different methods than those "...specifically specified on the specification..."

Oh sure, you can try (if you have a day to kill), but it's often much cheaper to not even bother trying to apply engineering logic to these sorts of QA situations.

Do you want to try to prove that "6dB" assumption? And that this adjustment factor isn't actually 5.27dB with your gear on that particular day?

The ONLY way to prove it to their satisfaction would be by comparing it to the direct measurement that they want to see in the first place. So you can't win.

Time to break out the oscilloscope and signal generator...

PS: Make sure that all of your calibration stickers are up-to-date.