Introduction

As mobile communications systems continue to develop, higher carrier frequencies and more complex modulation schemes are being used. Since the next generation of such systems is still under development, and much of this development work is proprietary to the companies funding it, there have been few reports of modulators that can be used directly at these higher operating frequencies. Such modulators are necessary since the conventional type of passive modulators in use are unsuitable for operation at higher microwave frequencies. This paper details a design procedure that can be applied to the design of I/Q modulators operating directly at microwave frequencies. The procedure is shown by detailing results obtained from an I/Q modulator operating directly at 2 GHz.

Motivation

Many I/Q modulators currently in use at frequencies of 450 MHz or 900 MHz are based on quaddiode arrangements with ferrite core hybrids or combiners. Unfortunately, such techniques cannot be used at microwave frequencies as the permeability of these cores decreases rapidly above 1 GHz. Similarly, while modulators employing a Gilbert cell are commercially available in surface-mount packages, their operating frequency range is also limited to 1 GHz. An attempt to overcome this restriction has been made by upconverting mixing on-chip. However, this arrangement not only requires a second local oscillator signal to upconvert the modulated data, but also generates many spurious mixing products, which necessitate accuracy filtering after the device, and thereby inevitably increase system component costs.

The described I/Q modulator operates directly at microwave frequencies and has the advantage of providing conversion gain rather than causing conversion loss. The design methodolgy employed in this hybrid design approach has the advantage that it may be readily applied to MMIC devices.

I/Q Modulator Operation

The operation of an I/Q modulator is shown in Figure 1. The carrier (or local oscillator signal) is passed through a 90 [degrees] hybrid that produces two signals of equal amplitude but differing in phase by 90 [degrees]. These signals are then mixed with two low frequency modulating signals (I and Q) that are also 90 [degrees] out of phase with each other. Finally, both mixed signals are combined at the output to provide a composite modulated signal.

Using conventional notation, the split LO signals may be written as

[S.sub.A] = Sin ([[Omega].sub.1]t)

[S.sub.B] = Sin ([[Omega].sub.1]t + 90)

If the I and Q signals are given by

I = Sin ([[Omega].sub.r]t)

and Q = Sin ([[Omega].sub.r]t + 90)

then the mixed signals at points C and D are given by

SC = Sin ([[Omega].sub.1]t) x Sin ([[Omega].sub.r]t)

= 1/2[Cos ([[Omega].sub.1] + [[Omega].sub.r])t + cos ([[Omega].sub.1] - [[Omega].sub.r])t]

and

[S.sub.D] = Sin ([[Omega].sub.1]t + 90) x Sin ([[Omega].sub.r]t + 90)

= 1/2[Cos ([[Omega].sub.1] +[[Omega].sub.r] + 180)t + Cos ([[Omega].sub.1] - [[Omega].sub.r])t]

Thus, when these signals are passed through the output combiner, the modulated signal is represented by the expression

S = Cos ([[Omega].sub.1] - [[Omega].sub.r]) t

+ 1/2[Cos ([[Omega].sub.1] + [[Omega].sub.r]) t

+ Cos ([[Omega].sub.1] + [[Omega].sub.r] + 180) t]

= Cos ([[Omega].sub.1] - [[Omega].sub.r]) t

Therefore, in principle, the spectrum...