Mixer IQ phase
A complex (IQ) mixing step is used to move the
signal from IF (100-400MHz) to baseband. Two mixers are used:
The lo's are programmable.
- 100-250 MHz with LO at 175 MHz
- 250-400 MHz with LO at 325 MHz
IQ mixing needs the phase bewteen the I,Q output
signal to be close to 90 degrees. If not, then unwanted images will
appear in the opposite sideband.
The phase of the IQ mixers was tested on 24oct07 and
The setup was:
The plots show the results of the processing:
- The red noise source was sent to Box 103.
- Separate measurements were done for each 150 MHz subband and
- The signal goes through the IQ mixer, 75 Mhz bandpass
filters, spliters, amps , and then two the digitizers.
- The I signal was sent to Re(polAinp). 0 was sent to Im(polAinp)
(done digitally in the pdev device).
- The Q signal was sent to Re(polBinp). 0 was sent o Im(polBinp).
- Full stokes data was taken and then angle between polA, and polB
(atan(U/V)) was plotted. This is the phase difference between the I and
Q inputs.. with one minor detail....
- The minor detail..
- Let Enp be the signal in channel LO + n
- Let Enm be the signal in channel LO - n
- At baseband, the individual I and Q spectra have the + and -
frequencies folded on top of one another.
- In = (Enp - Enm), Qn=(Enp90 - Enm90).. and Enm90 has a
negative rather then positive going 90 degree shift.
- The output signal is not a function of the gain variation in
the band between the Plus and Minuse image.
- To make this work we need to filter out 1 of the image bands
(below or above the LO) prior to the downconversion.
- For the High band (250-400) we used an IF filter covering 220-320
and an lo of 325. This gave data from 240-325 (the lower half).
- For the low band (100-240) we used an IF filter covering 140-200.
Data sets were taken with an LO of 205 and then 215 to separate
out the phase offsets in the 75 Mhz baseband filters and the I,Q mixer.
Hi band 240-325 (.ps) (.pdf):
Lo band 140-205 LO=205 and 215
- Page 1: The "Stokes" parameters for the I,Q spectra.
- Page 2: The phase angle between "polA,polB" which is measuring
the phase difference between the mixer I and Q outputs. The top plots
are versus IF frequency while the bottom plots are at baseband.
- The phase in both cases is within 2 degrees of 90 degrees with a
deviation acrros the band of about a 1 degree.
- The 120 to 140 MHz portion of the lower band had little
power (because of the IF filter) so the excursions from 90 deg should
angle between I,Q for the lo band has both the 205 and 215 Mhz LO
data. The structure in the phase vs frequency at
baseband is the same for both LO's so this is coming from the baseband