2-bit+gps+1-bit

I saw many GPS frontend with the IF is sampling at 2-bit ADC. Is 2-bit resolution is enough for IF sampling? I even saw 1-bit output. Please help me on how it work with only 1-,2-bit resolution.

GPS Frontend's amplifier is a AGC amplifier.so the 2bits ADC always have 2 bits valid data.
GPS is a Spread Spectrum system. it do not need high precision

I think 2 bits ADC is used in GPS system just cause it is a BPSK modulated system. 2 bits is enough to recover the info.
The 1 bit output might be a plus result of 2 bit original "mag" and "sign". But I never saw this in product datasheets.
Where do u see it?
What puzzles me is how could we judge whether the ADC works well.
Any ideas?

Please refer some paper about GPS baseband algorithm. 1bit A/D leads to about 2dB SNR demodulation loss, while 2bit A/D only leads to 0.6dB. For interference rejection performance, 2 bit A/D is better than sinlge bit(>5dB). So many ICs use 2bit A/D
for GPS BB with acceptable implement loss while keeping hardware simplicity.

Thank you very much guys. I saw it from many GPS frontend ICs. For example http://assets.zarlink.com/DS/zarlink_GP1010_JAN_01.pdf or http://www.sige.com/

I know that there is a technique in multirate system that you can use only 1-bit you can represent the data with the precision of 16-bit by oversampling it like crazy.. It is ised in 1-bit technology CD. I am not sure whether it is same technique. ?

for bpsk demodulator, it is no need to use 16 bits precision, 4 or 8 bit may be enough.

for dsss demodulator, the process is:

IF input(2bit) ---> after filtering and decimating(3-4bit) ----> correlator out(12-16bit) ----> bpsk demod

so, the 2bits input is very enough.

1-bit is the sign bit and 2-bit includes sign and magnitude. I have designed SW receivers just using the single bit. And it works.

You may check Philip Mattos paper ....

The basic idea is quantization noise very much ignorable when compared with the AWGN.

Thank you guys. As brmadhukar mentions, increasing number of bits will not gain any advantage since the signal level is already below noise level. There is a paper in detail. Does anyone has it?

Chang H, Presampling Filtering, "Sampling and quatization effects on digital match filter performance", Proceeding of the International Telemetering Conference., San Diego. CA, 1982, pp. 889-915

Many book mention about this but none of them go in detail.

I also would suggest that you look at the IEEE Proceedings special issue on GPS. There was a very good paper which discussed the trade off between sampling rate and the number of bits ...

Hope this helps.

Hi,
how do u test the ADC?