Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                

Frequency Doubler

Download as pdf or txt
Download as pdf or txt
You are on page 1of 4
At a glance
Powered by AI
The paper presents a CMOS frequency doubler circuit that can convert a 0.6 GHz signal to a 1.2 GHz output using a time-delay element and XOR gate. It aims to achieve good fundamental suppression, compact layout and low power consumption.

Frequency multiplier circuits are used in communication systems to design oscillators at lower frequencies and convert them to higher frequencies, simplifying design and improving phase noise. However, existing CMOS frequency doubler circuits have various drawbacks like producing many spectral components or requiring input/output matching and filtering.

It uses an integrator to delay the input square wave by T/4 and feeds it along with the original input into an XOR gate to generate an output square wave with twice the frequency. This exploits the principle that XOR-ing a signal with a T/4 delayed version of itself produces a doubled frequency output.

An L-Band CMOS Frequency Doubler using a Time-Delay Technique

Brad R. Jackson and Carlos E. Saavedra


Department of Electrical and Computer Engineering
Queens University, Kingston, Ontario, Canada, K7L 3N6
Abstract In this paper, a frequency doubler circuit is
presented that converts a 0.6 GHz signal to a 1.2 GHz output
using standard CMOS 0.18 m technology. The proposed
circuit uses a time-delay element and an XOR logic gate to
perform the frequency multiplication and is implemented
entirely on-chip. Advantages of this topology include good
fundamental suppression, compact layout, and low power
consumption.
Experimental results show a relatively
constant output power of approximately 4 dBm with an input
power from -3 dBm to 10 dBm, fundamental and third order
harmonic suppressions of up to -30 dBc and a power
consumption of 9 mW. The phase noise of the output signal
is -117 dBc at a 500 kHz offset.

(a)

Index Terms CMOS frequency doubler, multiplier,


mixer, frequency conversion.

I. INTRODUCTION
Frequency multiplier circuits are used in a wide range of
applications in communication systems. Using frequency
multiplication, oscillators can be designed at lower
frequencies and then converted to higher ones, which can
simplify the design of the oscillator and improve the phase
noise of the resulting signal.
Many CMOS frequency doubler circuits have been
demonstrated using various methods [1]-[5]. A common
technique is to use the non-linearities of a transistor with a
large input signal such as in [2]. With this method, the
output has many spectral components, among which is the
desired doubled frequency signal. Input matching at the
fundamental frequency is generally used, along with
filtering and matching for the doubled frequency
component at the output. An alternative frequency
doubler circuit was proposed in [1] and demonstrated at
baseband frequencies.
That topology can also be
employed monolithically and at much higher frequencies.
Shown in Figure 1 is a block diagram and waveforms that
describe the principle behind the frequency multiplication
in this paper.
An input square wave, Vin, is delayed by T/4 where T is
the input signal period, and is fed into an XOR gate along
with the input. The output is a square wave with twice the
frequency of the input. To implement the T/4 delay
element, an integrator circuit is used along with a

SiRF 2006

(b)
Figure 1 (a) Frequency doubler block diagram and (b) Waveforms

comparator. The integrator converts the square wave to a


triangular wave and the comparator is then used to
compare the triangular wave with a reference voltage.
The result is Vd, a square wave that is delayed by T/4 with
respect to the input. This signal is then used as the second
input to the XOR gate to generate the doubled frequency
at the output.
This topology can also be used as an analog multiplier
by inputting a sinusoid, converting it to a square wave,
and then filtering the odd-order harmonics of the square
wave at the output.
In this paper, a 0.6 GHz to 1.2 GHz CMOS frequency
doubler is presented based on the approach in Figure 1 and
experimental results are shown. Using this technique,
frequency multiplication can be achieved with low power
consumption, small chip area, and good inherent rejection
of undesired frequency components. In particular, a

131

0-7803-9472-0/06/$20.002006 IEEE

significant advantage to this topology is that excellent


fundamental suppression can be achieved at the output
without filtering.

The XOR gate was constructed using transmission gates


[7] and inverters as shown in Figure 2. The output of this
circuit is a square wave at twice the frequency of the
input. If a sinusoidal output is desired, a simple low pass
filter could be used to attenuate the higher order
harmonics in the square wave. The layout for this circuit
can be quite compact and in this implementation required
2
an area of approximately 200 m x 75 m (0.015 mm ) for
the core of the frequency doubler circuit. The full chip
area was 400 m x 400 m including bonding pads.

II. CIRCUIT DESIGN


A simplified schematic of the frequency doubler circuit
is shown in Figure 2. Since the input to the circuit is a
sinusoid and the doubler circuit requires a square wave,
the first step is to make this conversion. To do this, a DClevel shifting circuit was used, followed by a series of
inverters. The DC level shifting circuit offsets the
incoming signal to the threshold voltage of the inverters.
The inverters then create a high quality square wave for
the core doubler circuitry.
The integrator was implemented using a passive RC
integrator. The RC network approximates an ideal
integrator when the period of the input signal is small with
respect to the RC time constant. Note, however, that as
the RC integrator becomes closer to an ideal integrator the
amplitude of its output triangular wave becomes smaller
and it consequently becomes more difficult for the
comparator in the next stage to determine the correct state.
To implement the comparator circuit, first a commonsource amplifier is used to amplify the triangular wave and
then an inverter is used. Using this configuration for the
comparator adds a small additional delay to the T/4 and as
such could not be used at higher frequencies where this
delay becomes significant.

III. EXPERIMENTAL RESULTS


To measure the performance of the proposed frequency
doubler, an input sinusoidal signal at 0.6 GHz was used.
The input power levels were varied and the output
spectrum was observed. Shown in Figure 3 is the output
spectrum with an input power of 4 dBm. The desired
signal at 1.2 GHz is at approximately 4 dBm and is clearly
the strongest at more than 30 dB above the fundamental at
0.6 GHz and the third harmonic at 1.8 GHz. The fourth
harmonic at 2.4 GHz is approximately 14 dB below the
1.2 GHz signal. The presence of this harmonic is due to
the uneven duty cycle of the output square wave which is
caused by the two signals entering the XOR gate not being
exactly 90 out of phase. To avoid this, the passive RC
integrator could be adjusted, or a more sophisticated high
speed comparator could be employed. Since this signal is
1.2 GHz away from the desired signal and already 10 to

Figure 2 Simplified circuit diagram of the frequency doubler

SiRF 2006

132

0-7803-9472-0/06/$20.002006 IEEE

the input do not produce a well-defined square wave.


Consequently, in this low input power region the second
harmonic signal strength is close to that of the
fundamental and other harmonics, thus the circuit is out of
its operational range.
The phase noise of the 1.2 GHz output signal was
measured to be -109 dBc at a 100 kHz offset and -117 dBc
at a 500 kHz offset. The DC voltage used to power the
circuit (Vdd) is 1.8 V and the circuits DC current draw is
5 mA, resulting in a power consumption 9 mW for the
circuit. A microphotograph of the chip is shown in Figure
5.
Figure 3 Output spectrum from 0 3 GHz with 4 dBm input
power at 0.6 GHz

15 dB below it, it could easily be removed by filtering.


Furthermore, since the output is a square wave, there are
odd-order harmonics also present which could be easily
filtered if a sinusoid is desired.
Shown in Figure 4 are the fundamental, second, third,
and fourth harmonics at various input powers. It is
apparent that the power of the frequency-doubled signal is
relatively constant at approximately 4 dBm for input
powers greater than -10 dBm. This constant output power
is expected because of the digital circuit components with
logic gates functioning from rail-to-rail producing a
constant amplitude square wave at the output. At input
powers higher than 0 dBm the fundamental and the fourth
harmonic are significantly rejected and the third harmonic
is at least 10 dB below the desired signal because the input
is strong enough to generate a high quality square wave
after propagating through the series of input inverters. At
power levels below approximately -3 dBm the inverters at

Figure 5 Microphotograph of the frequency doubler chip

IV. CONCLUSIONS
A microwave CMOS 0.18 m frequency doubler circuit
has been designed using a time-delay element, comparator
and XOR gate. This circuit converts a 0.6 GHz input to a
1.2 GHz output. Experimental results show very good
fundamental and third order harmonic suppression. A
further advantage of this topology is the relatively
constant output power of approximately 4 dBm over a
wide range of input power levels.
The power
consumption of the circuit is 9 mW and the core circuit
layout area is approximately 0.015 mm2. This circuit
could either be used as an analog multiplier by employing
a low-pass filter to suppress the higher order harmonics of
the output square wave or as a clock frequency doubler.

10
5
0

Pout (dBm)

-5
-10
-15
-20

600 MHz
1.2 GHz
1.8 GHz
2.4 GHz

-25
-30
-35
-40
-3

-2

-1

10

Pin (dBm)

Figure 4 Harmonic output power levels with 0.6 GHz input signal

SiRF 2006

133

0-7803-9472-0/06/$20.002006 IEEE

ACKNOWLEDGEMENTS
The authors would like to thank Mr. You Zheng for
assistance with the test and measurement of this circuit.

[4]

REFERENCES

[5]

[1] C. E. Saavedra and Y. Zhang, "A Clock Frequency Doubler


using a Passive Integrator and Emitter-Coupled Comparator
Circuit," 17th IEEE Canadian Conference on Electrical and
Computer Engineering, pp. 137-140, May 2004.
[2] F. Ellinger and H. Jackel, Ultracompact SOI CMOS
frequency doublerfor low power applications at 26.528.5
GHz, IEEE Microwave Wireless Compoent. Letters., vol.
14, no. 2, pp. 5355, Feb. 2004
[3] J. Wong and H. Luong, "A 1.5-V 4-GHz Dynamic-Loading
Regenerative Frequency Doubler in a 0.35-um CMOS

SiRF 2006

[6]
[7]
[8]

134

Process," IEEE Trans. on Circuits and Systems II: Analog


and Digital Signal Processing, Vol. 50, No. 8, pp. 450-455,
Aug. 2003.
M. Yang, S. Oh, and S. Lee, "Low Power Fully Differential
Frequency Doubler," IEE Electronics Letters, Vol. 39, No.
19, pp. 1388 1389, Sept. 2003.
F. Cheng, C. Chen, O. Choy, A 1.0 um CMOS all Digital
Clock Multiplier, Proc. IEEE 40th Midwest Symposium on
Circuits and Systems, Vol. 1, pp. 460-462, August 1997.
Y. Li, Delay Line Implements Clock Doubler, EDN
Magazine, July 18, 1996.
S. Brown and Z. Vranesic, Fundamentals of Digital Logic
with VHDL Design, McGraw Hill, 2000, ch. 3.
Y. Lee, S. Choi, S. Kim, J. Lee, K. Kim, Clock multiplier
using digital CMOS standard cells for high-speed digital
communication systems, Electronics Letters, Vol. 35, No.
24, pp. 2073-2074, November 1999.

0-7803-9472-0/06/$20.002006 IEEE

You might also like