WARP Project Forums - Wireless Open-Access Research Platform

You are not logged in.

#1 2007-Jul-11 22:13:22

From: Vienna, VA
Registered: 2007-Jan-24
Posts: 121

Idle noise level calibration

We try to determine the best idle noise level (in other words, the default rx gain settings) related with our packet detection algorithms. For this purpose, we did the following experiment and measurement.

We put WARP with antenna inside a shielded room. (We actually have a facility dedicated for this purpose.) Without a transmitter, we set the rx LNA gain at highest and baseband gain with different values. Using chipscope, we record I/Q values, RSSI values, digital power estimation, and several other values. We further changed the center frequency among channel 1, 3, 6, 7, 9, 11 and 14. We recorded results in each frequency channel. Based on these measurement, we found ---

1. Given the same gain settings (LNA 03, BB 23 or 22), the two radio cards have about 2 dB difference in idle noise power.

2. Given the same gain settings (LNA 03, BB 23 or 22), the average idle noise power and RSSI are different at different channel. In particular, channel 9 gives us the worst results. The idle noise power in channel 9 is more than 6dB higher than those in channel 3, 6, 7. The average RSSI value in channel 9 is around 180, while in channel 3, 6,7 around 120. This is consistent over two radio cards we have.

Anyone have comment on our experiment and findings? Thanks.



#2 2007-Jul-11 22:46:17

From: Mango Communications
Registered: 2006-Jul-03
Posts: 5159

Re: Idle noise level calibration

These are very interesting results- thanks for posting them!

We haven't tried this characterization ourselves (mostly because we don't have an RF anechoic chamber handy).

Are your noise power measurements derived from just RSSI? If so, it's possible the different levels you observed are actually artifacts of the ADC reference level. The RSSI ADC uses a voltage divider to set its midpoint reference at 1.5v. The ADC encodes 1024 steps in a 2Vp-p span around this center voltage. If the resistors on your two radio boards are slightly different (even still in the resistor's tolerance), it might be enough to cause a shift in the ADC's analog input range. This could explain two different digital levels for the same analog input on two boards.

If you have a good scope available, try measuring the voltage at pin 21 on the AD9200 (highlighted in the radio board users guide) on both your boards. You can see the full RSSI ADC circuit on pg 5 of the radio board schematics.



Board footer