Hi all,

i'm designing a receiver signal strength indicator (RSSI) in analog, and i've got a question. One way to do the RSSI is to have a log amplifier followed by peak detection. The log amp usually consists of a series of limiting amplifiers with each of their outputs added together. So I noticed that for N limiting amplifier stages, most designers used N peak detectors and then added the DC signals together. However, to save power, I want to add the AC signals together and have one peak detector. Does anyone know why this isn't normally done?

thanks,
Aaron

i agree with didac, also most of the limiting amp structures give you a simple peak recified level (for no or next to no extra power consumption) which when added together gives a good piecewise linear rssi.
out-of-interest, how would you propose to "add the AC signals together" ?

The way i want to add the AC signals together is quite simple. You have a differential amplifier, and instead of having only one input pair you divide it into several input pairs each one fed by one of the limiting amps. So you add the currents. In the actual design I combined the adder with the peak detector.

Also as for the point about phases i guess you are right. Actually I already took that into account. In my design, the phase shift from beginning to end is only 10 degrees which results in an error somewhere around 1-3%. Now that you mention it though, i had to force the limiter bandwidths to somewhere around 30 times the operating frequency which resulted in higher power consumption. So i guess maybe the standard way is better.

BTW Jerry, I don't understand your point about the limiters in limiting region not contributing to usable signal amplitude information. When you add the limiter outputs up, the limiting outputs contribute the largest part to the overall amplitude...

thanks everybody,
Aaron