AD8302 - Phase difference between two 2.4 GHz signals

Hi Matt,

The AD8302 will take two single-ended singles - we'll call them Signal A and Signal B - and output the difference in magnitude and phase of the two signals, in a DC voltage format.  So if we call magnitude of Signal A: MAG_A and the magnitude of Signal B: MAG_B, the VMAG pin output will be proportional to the log ratio of the magnitudes: MAG_A:MAG_B  (see equation 8a on page 16 of the datasheet; you can extend this in terms of power with equation 8b).  Similarly, the voltage on the pin VPHS is proportional to the linear difference in phases of the two signals, equation 9.  

Referencing TPCs 6 and 29, we see the VMAG and VPHS responses at 2200 MHz (the closest frequency we have data on).  Each plot shows the actual voltage output response along with the error.  For your application, you will have to take into account the large amounts of phase error that is centered at 0 degrees and +/- 180 degrees.

Let me know if you have any specific questions on this.

Thanks,

Joel

Edited title from "New here and need some help1" to "AD8302 - Phase difference between two 2.4 GHz signals".

Moved from Hittite Microwave Products from ADI to RF Components community.

Hey Joel, thanks for the help and reply! How clean does the signal need to be to use this? Ideally I would like some way to do this receiving the rf from the air. I know doing this at this frequency is difficult but I would really like to try. If I can get it to be within 5-10% error then I would be happy. As for the phase error, what does that mean for my application? I may be miss understanding terminology.

To give you a better idea of what I am doing, I am trying to use an adcock antenna to do direction finding. So comparing the received RF needs to be and the phase level (why I'm not using a frontend).

Thanks!

Matt

I'm guessing you will need at least an LNA and filter to get the levels to something in the neighborhood of -30 dBm and clean enough so that the amplifiers can discriminate the zero crossings.  You may need more amps before the inputs of the AD8302 - it just depends on the expected amplitude of the received signal.

Phase error is defined as the difference in the actual phase and the phase reported.  For example, if your signals are completely in phase with each other, 0 degrees phase difference, and the AD8302 was reporting 5 degrees phase difference, then the error would be 5 degrees.  Does that make sense?

Joel

Yes it does, I think I understand now. How high is the error at 0 and 180? Could you give a page number in the datasheet? Sorry about all the questions but I just looked through the datasheet and didn't find it. Also what would I need to measure the output from the ad8302? ADC I'd think but with a high sample rate?

Thanks for your help!

Matt

Actually could I use a frontend before the 8302 to isolate the signal?

Hi Matt,

Below is TPC 29 on page 10 of the datasheet.  This shows the VPHS output response (the triangular plot that references the "phase out -V" y-axis on the left and the phase error, in degrees, which references the right y-axis.  At 0 degrees and +/- 180 degrees, the error is off the plot; something more than +/- 10 degrees.  Low phase error occurs between 30 degrees and roughly 155 to 160 degrees.

At the end of the day, if your signals are an unknown phase that can be anything between 0 and +/- 180 degrees, then the AD8302 might not be the best choice for your system, given the large phase error.

An ADC with a low sample rate would work fine to measure the VMAG and VPHS outputs.  They are both DC-ish outputs, in that once they are settled they do not move until either the phase or magnitude changes.  They only have about 30MHz of video bandwidth each.

Yes, I would recommend some sort of front end receiver before the AD8302.

Joel, it makes sense now, thanks! The only thing I'm unclear on is how to isolate a signal. The 2.4 ghz band may have several devices on different frequencies. I know usually at this point it goes into a ADC and a DSP resolves that, but what about here? I have seen a few band-pass filters that are variable, but none that would do this.

Could I use a mixer and bring it down to an IF and keep phase info? I'm not really sure how mixers work in this regard. Then I could just sample with an ADC and negate the AD8302 entirely as it seems difficult to isolate an RF signal in this system.

Hi Matt,

Usually on a receiver, isolation of a signal is done with filtering of some sort.

You may be able to down-convert the signal and then measure phase, but I'm not familiar enough with that approach to say one way or the other.

Regards,

Joel