Part 3. Analog signal processing

Document technical information

Format doc
Size 146.4 kB
First found May 22, 2018

Document content analysis

Category Also themed
Language
English
Type
not defined
Concepts
no text concepts found

Persons

Organizations

Places

Transcript

Resonant Cavities
as Beam Position Monitors
Part 3. Analog signal processing.
Laboratory measurements.
A. Liapine
This part explains the basic principles used in the electronics for the BPM’s signals processing. Two basic
methods of the downconversion, heterodyne and homodyne, are discussed. Simple tests which can be done to
the electronics and to the system BPM+electronics are described briefly.
1. Electronics
Signals coming from a cavity beam position monitor usually oscillate at a very high frequency
– from a few GHz up to a few tens of GHz. These signals must be downconverted before they
may be digitized for the following analysis, because the capabilities of the digitizers available
on the market are still limited to a few hundreds of MHz. Basically, two methods are used in
the radio technics – heterodyne and homodyne. They also can be used for the conversion of
the signal to a higher frequency, but we are not interested in this possibility and furthermore
– need to suppress the high harmonic. We will start with the heterodyne receiver and then
continue with the homodyne one (which could be treated as an extremely simplified version
of the heterodyne unless it would not have some special properties).
1.1
Heterodyne Receiver
The (super-)heterodyne receiver is widely spread one. It is used everywhere starting from
personal communication devices to radio and TV tuners. Figure below illustrates the main
principle of heterodyne receiver. Radio frequency (RF) signal is first amplified in a frequency
selective (but usually broadband) low noise stage, then translated to a lower intermediate
frequency (IF): fIF = fRF – fLO. After a significant amplification and an additional filtering on
the intermediate frequency it is finally downconverted to the baseband (i.e. to the original or
desired signal frequency).
Band select
filter
Mirror frequency
rejection filter
"I"
Channel select
filter
LO
LNA
RF
LO +900
"Q"
IF
LO
Band select
filter
Channel select
filter
f
fIF
f
This technique is flexible in changing the receiving frequency, because the only change to be
done is the frequency of the first local oscillator (LO) in the way that at the frequency of the
signal at the output of the mixer would not change. The rest of electronics is free from
additional readjustments, what is very important in the applications of the heterodyne
receiver in the radio, TV, satellite and other communications.
The 2-stage downconversion of the BPM signal is sometimes needed to reduce the losses
introduced by long cables carrying the BPM signals out of the accelerator tunnel. In that case
the first stage of downconversion is placed to the BPM as close as possible, the signal is
downconverted to some intermediate frequency, cable losses for which are considerably
lower. The second stage is then placed outside of the tunnel, in the region, which is not as
strongly loaded with radiation of every kind.
The main problem in heterodyne technique is so-called image frequency. To understand the
problem let's first discuss in more detail the principle of the mixer’s functionality.
The RF-signal at the front end of the electronics oscillates with some frequency RF,
U RF (t )  a cos( RF t )
(1)
The signal coming from the local oscillator has some fixed frequency LO
U LO (t )  b cos( LO t )
(2)
These two signals are multiplied by a non-linear element of the mixer. As a result two signals
with frequencies RF - LO and RF + LO are produced.
U RF (t )U LO (t ) 
ab
[cos( RF   LO )  cos( RF   LO )]
2
(3)
Taking the first term we have a down-conversion. The second term corresponds to upconversion. Usually we are interested in the downconverted term and  LO   IF . In this case
a frequency of  IM   LO   IF is downconverted to a negative IF frequency. In this way the
signal at this frequency joins the signal of interest. Therefore any noise or interference signals
at this frequency have to be rejected. The frequency  IM is called image frequency.
To reject the mirror frequency signal an additional filter is often applied in front of the mixer.
ωIF
ωIF
f
f
ωIM
ωLO
ωRF
-ωIF
RF, IM
IF
LO
0
ωIF
1.2 Homodyne receiver
Nowadays in a close connection to the growing mobile and wireless communications more
and more attention is paid to direct conversion receivers. Direct conversion reception, also
referred to as homodyne, or zero-IF conversion, is the most natural solution to detect
information transmitted by a carrier in just one conversion stage.
In this case the electronics has at least one mixer stage less then in a heterodyne system. But
from the other hand the flexibility to change receiving frequency is reduced significantly,
because the narrow band filtering is not possible before downconversion and the only narrow
band filter is the output lowpass. Though, in case of the BPM signal processing there is no
need to change the input frequency. That means a narrow band filter can be used already at
the front-end of the electronics.
The principle is illustrated below. The signal is first amplified at a low noise stage and then
directly converted to the baseband or even to a direct current signal. When the frequencies of
the RF and the LO signals are equal, this scheme works as a phase detector. In some
literature, only when the local oscillator is synchronized in phase with the incoming carrier
frequency, the receiver is called homodyne.
Low-pass
filter
“I”
Band select
filter
LO
LO +900
LNA
Low-pass
filter
DC
“Q”
Band select
filter
Low-pass
Filter
f
f
Suppose that the IF in a heterodyne is reduced to zero. The LO will then translate the centre
of the desired channel to 0Hz, and the portion of the channel translated to the negative
frequency half-axis becomes the “image” to the other half of the same channel at the positive
frequency half-axis.
0
f
To achieve maximum information, we should take both parts of signal. It’s done by a method,
which is called quadrature downconversion. The principle of this method is that the signal is
at first divided into two channels and then downconverted by an LO signal, which has a phase
shift of 90o in one channel with respect to another. The vector of the resulting signal is
described as:
arg( Signal )    arctg
Signal  I 2  Q 2
Q
I
The main problem in homodyne technique is an offset caused by the LO signal leakage to the
RF port of the mixer. The propagated signal reflects from the components in the front-end of
the receiver and goes back to the mixer, where it is mixed down to DC. The offset can be
considerable with respect to the signals to be measured. This leads to a narrower dynamic
range of the electronics, because the active components get saturated easier than it would be
in case of a zero offset.
For example, let’s take a mixer with the LO-drive equal to +10dBm and RF/LO isolation equal
to 40 dB. In this case the offset can be as high as –30dBm, or about 2mV. In case of a high
sensitivity this can be a large number (this is the signal level at the output of the mixer, some
amplification stages follow!).
Band select
filter
Low-pass
filter
DC
LNA
LO
Another problem of the homodyne receiver, or, more concretely, of the I/Q (inphase/quadrature) mixer, is mismatches in its branches. Assuming a mismatch of  for the
amplitude and  for the phase, we can estimate the error, caused by these mismatches. In
this way we get:
S
 S miss  1

EIQ  ideal
     2 .
S ideal
2

For typical values of   0.3 and   30 this gives an error of 1.5 10 3 .
“I”
a cost
a(1   ) cos(t   )
b cos(t   )
b sin( t   )
“Q”
In the processing of the BPM signals it is important to know the phase with respect to the
phase of the reference cavity. That means that an I/Q mixer becomes very attractive for this
application. On the other side, it has too many specific hardware problems. That’s why it is
desirable to apply a digital downconversion at the last stage of the downconversion. That
means that a processing usually applied to the signal in an analog form, is applied to a
digitized signal in a digital form. In that case an I/Q mixer and other elements realized in
forms of idealized models have no mismatches or reflections etc. Surely, the initial signal still
has to be downconverted and prepared for the digitizing. The entire scheme looks then like a
heterodyne scheme in the picture at the very beginning of this part with the difference that
the second downconversion stage is not real but realized in digits.
2. Electronics Tests
As the entire electronics relays on mixers, it makes sense to test the mixers before they are
inserted into the electronics. The most important properties of the mixers are the
transmission characteristic (linearity, 1dB compression point, conversion loss), isolation and,
perhaps, higher harmonics.
The transmission characteristic can be measured with two oscillators, attached to RF and LO
inputs of the mixer and any device measuring the signal level – spectrum analyzer,
oscilloscope, ADC, even simple voltmeter. In case of oscilloscope the form of the signal can be
controlled visually for any distortions, while a spectrum analyzer shows distortions due to
non-linearity directly as a set of harmonics.
G1
RF
IF
G2
SA
LO
The signal level of the oscillator attached to the LO port of the mixer is kept constant and
equal to the nominal level required for the mixer’s operation. The signal level at the RF port is
changed and the level of the output signal is controlled. This test allows selecting the best
mixers available for use in the electronics – ones with the best linearity, lowest losses, and
highest input signal limit. In a very similar way the entire electronics can be tested as well.
The isolation between RF and LO ports can be measured directly by measuring the
transmission from the LO to the RF port with a network analyzer. Isolation between LO and
RF is very important for the mixers working in low frequency stages.
RF
NA
IF
LO
50Ω
G1
X
"I"
1,5GHz- "Q"
Elektronik "Ladung"
Referenz
G2
Oszilloskop
×

Report this document