I'm developing a receiver using the ADRF6806 evaluation board. All seems as expected except the noise performance. Since I'm interested in low IF (100Hz plus, maybe down to DC) I've disconnected the I and Q baluns. I'm feeding each balanced pair of I and Q into an INA163 low noise instrumentation amplifier set for a gain of 100. I'm then feeding the output of one INA163 into a spectrum analyser looking from DC to a few hundred KHz.
The gain, signal levels, VCO synthesiser and mixer operation are all fine and doing exactly as they should. However the noise floor of the spectrum analyser plot shows more noise in the 0 to 100KHz region. The noise slopes, high noise at low frequencies and lower noise towards 200KHz. There is an approximate 18dB difference in noise density between 200Hz and 200KHz. I'm confident this noise is coming from the I and Q outputs, as I can see a significant fall in noise when I power down the ADRF and short the INA163 input together.
Note that this output is present with no RF input, indeed it is still the same when I power down the VCO. It seems clear the I and Q output stage is noisy at low frequencies.
Is this to be expected and can it be improved upon as it severely affects the noise figure of the device depending upon the IF frequency chosen. At >few hundred KHz noise figure is as quoted (about 11dB) however at 200Hz IF the noise figure is closer to 30dB!
My question is:- Can this be improved upon or is it an IC design issue. Can I improve some decoupling or perhaps it is a noisy supply. I have tried to modify these but to no avail so I'm fearing that it is just high flicker noise with a high corner frequency.
Can anyone shed any light on this issue?
Pity, because I like the chip.....