Quadrature error correction

DETAILED DESCRIPTION It is an advantage, according to this invention, to provide an adaptive system and process for correcting quadrature signal errors for amplitude, phase and bias errors, thus eliminating the need for expensive channel matching, repeated manual calibration and other time-consuming and costly techniques.
It is a particular advantage, according to this invention, that the system or process is adaptive because quadrature errors are constantly monitored and corrected.
Briefly, according to this invention, the quadrature signal errors are continuously corrected by continuously generating an error function signal E having a value equal to the sum of the squares of the normalized values of each quadrature signal (S and C) minus a reference constant, say one.
The signal corresponding to the error function E at the zero crossings of the quadrature signals is saved to produce instantaneous error value signals E(0), E(.
pi.
/2), E(.
pi.
), E(3.
pi.
/2), and in a preferred embodiment, the value of the error function is saved when the difference between the absolute values of the quadrature signals cross zero to produce instantaneous error value signals for E(.
pi.
/4) and E(3.
pi.
/4).
These four or six instantaneous error value signals are used to produce signals corresponding to the values of the base and amplitude errors for each quadrature signal and the phase error signal as follows:
TABLE I
______________________________________
dkse = [E(.
pi.
/2) + E(3.
pi.
/2)]/2
(SIN channel amplitude error)
Soe = E(.
pi.
/2) - dkse
(SIN channel bias error)
dkce = [E(0) + E(.
pi.
)]/2
(COS channel amplitude error)
Coe = E(0) - dkce (COS channel bias error)
.
phi.
3 =
E(.
pi.
/4) - E(3.
pi.
/4)
(spatial phase error)
______________________________________
These quadrature error signal values are then used to construct normalized or corrected quadrature signals S and C as follows:
TABLE II
______________________________________
S = [1/(1 + dkse)] SIN - Soe, and
C = [1/(1 + dkse)] COS -