Philips Semiconductors Product data
SA8027
2.5 GHz low voltage, low power
RF fractional-N/IF integer frequency synthesizer
2001 Aug 21
10
1.5 Main Output Charge Pumps and Fractional
Compensation Currents (see Figure 8)
The main charge pumps on pins PHP and PHI are driven by the
main phase detector and the charge pump current values are
determined by the current at pin R
SET
in conjunction with bits CP0,
CP1 in the C-word (see Table 1). The main charge pumps will enter
speed up mode after the A-word is set and strobe goes High. When
strobe goes Low, charge pump will exit speed up mode. The
fractional compensation is derived from the current at R
SET
, the
contents of the fractional accumulator (FRD) and by the program
value of the FDAC. The timing for the fractional compensation is
derived from the main divider.
1.6 Principle of Fractional Compensation
The fractional compensation is designed into the circuit as a means
of reducing or eliminating fractional spurs that are caused by the
fractional phase ripple of the main divider. If I
COMP
is the
compensation current and I
PUMP
is the pump current, then for each
charge pump:
I
PUMP_TOTAL
= I
PUMP
+ I
COMP
.
The compensation is done by sourcing a small current, I
COMP
, see
Figure 9, that is proportional to the fractional error phase. For proper
fractional compensation, the area of the fractional compensation
current pulse must be equal to the area of the fractional charge
pump ripple. The width of the fractional compensation pulse is fixed
to 128 VCO cycles, the amplitude is proportional to the fractional
accumulator value and is adjusted by FDAC values (bits FC7–0 in
the B-word). The fractional compensation current is derived from the
main charge pump in that it follows all the current scaling through
external resistor setting, R
SET
, programming or speed-up operation.
For a given charge pump,
I
COMP
= ( I
PUMP
/ 128 ) * ( FDAC / 5*128) * FRD
FRD is the fractional accumulator value and is automatically
updated.
The theoretical values for FDAC are: 128 for FMOD = 1 (modulo 5)
and 80 for FMOD = 0 (modulo 8).
SR02359
REFERENCE R
MAIN M DIVIDE RATIO
CHARGE PUMP OUTPUT
ACCUMULATOR VALUE (FRD)
FRACTIONAL COMPENSATION
CURRENT (I
COMP
)
I
PUMP–TOTAL
N N N+1 N N+1
241
3
0
PULSE
WIDTH
MODULATION
PULSE LEVEL
MODULATION
mA
µA
GRAPHS NOT TO SCALE.
NOTE: For a proper fractional compensation, the area of the fractional compensation current pulse must be equal to the area of the charge pump output.
Figure 8. Waveforms for NF = 2 Modulo 5 fraction =
2
/
5
SR01800
f
RF
MAIN DIVIDER
FRACTIONAL
ACCUMULATOR
f
REF
I
COMP
I
PUMP
LOOP FILTER
& VCO
Σ
Figure 9. Current Injection Concept
Philips Semiconductors Product data
SA8027
2.5 GHz low voltage, low power
RF fractional-N/IF integer frequency synthesizer
2001 Aug 21
11
1.7 Charge Pumps
The PHP and PHI charge pumps are driven by the main phase
detector, while the PHA charge pump is driven by the auxiliary
phase detector. The I
SET
value (refer to Table 1) is determined by
the external resistor attached to the R
SET
pin.
The charge pump, by default, will automatically go into speed-up
mode (which can deliver up to 15*I
SET
for PHP_SU, and 36*I
SET
for
PHI), based on the strobe pulse width following the A word, to
reduce switching speed for large tuning voltage steps (i.e., large
frequency steps). Figure 10 shows the recommended passive loop
filter configuration. Note: This charge pump architecture eliminates
the need for added active switches and reduces external component
count. Furthermore, the programmable charge pump gains provide
some programmability to the loop filter bandwidth.
The duration of speed-up mode is determined by the strobe pulse
width following the A word. Recommended optimal strobe width is
equal to the total loop filter capacitance charge time from state 1 to
state 2. The strobe width must not exceed this charge time. The
strobe width is controlled by the CPU (× number of clock cycles).
In addition, charge pumps will stay in speed-up mode continuously
while Tspu = 1 (in D word). The speed-up mode can also be
disabled by programming T
dis-spu
= 1 (in D word).
SR02356
VCO
C3C2
R2
C1
R1
PHP[PHP–SU]
PHI
Figure 10. Typical passive 3-pole loop filter
Table 1. Main and auxiliary charge pump currents
CP1 CP0 I
PHA
I
PHP
I
PHP–SU
I
PHI
0 0 1.5xl
SET
3xI
SET
15xl
SET
36xl
SET
0 1 0.5xl
SET
1xl
SET
5xl
SET
12xl
SET
1 0 1.5xl
SET
3xl
SET
15xl
SET
0
1 1 0.5xl
SET
1xl
SET
5xl
SET
0
NOTES:
1. I
SET
= V
SET
/R
SET
: bias current for charge pumps.
2. CP1 is used to disable the PHI pump, I
PHP–SU
is the total current
at pin PHP during speed up condition.
1.8 Lock Detect
The output LOCK maintains a logic ‘1’ when the auxiliary phase
detector (AND/ORed) with the main phase detector indicates a lock
condition. The lock condition for the main and auxiliary synthesizers
is defined as a phase difference of less than "1 period of the
frequency at the input REF
in+,
. One counter can fulfill the lock
condition when the other counter is powered down. Out of lock
(logic ‘0’) is indicated when both counters are powered down.
1.9 Power-down mode
The power-down signal can be either hardware (PON) or software
(PD). The PON signal is exclusively ORed with the PD bits in
B-word. If PON = 0, then the part is powered up when PD = 1. PON
can be used to invert the polarity of the software bit PD. When the
synthesizer is reactivated after power-down, the main and reference
dividers are synchronized to avoid possibility of random phase
errors on power-up.
Philips Semiconductors Product data
SA8027
2.5 GHz low voltage, low power
RF fractional-N/IF integer frequency synthesizer
2001 Aug 21
12
2.0 SERIAL PROGRAMMING BUS
The serial input is a 3-wire input (CLOCK, STROBE, DATA) to
program all counter divide ratios, fractional compensation DAC,
selection and enable bits. The programming data is structured into
24 bit words; each word includes 2 or 3 address bits. Figure 11
shows the timing diagram of the serial input. When the STROBE
goes active HIGH, the clock is disabled and the data in the shift
register remains unchanged. Depending on the address bits, the
data is latched into the selected working registers or temporary
registers. In order to fully program the synthesizer, 3 words must be
sent: C, B, and A, in that order. A typical programming sequence is
illustrated in Figure 12. Table 2 shows the format and the contents of
each word. The D word is used for testing purposes and should be
initially set to 0 for normal operation. When sending the B-word, data
bits FC7–0 for the fractional compensation DAC are not loaded
immediately. Instead they are stored in temporary registers. Only
when the A-word is loaded, these temporary registers are loaded
together with the main divider ratio.
2.1 Serial bus timing characteristics (see Figure 11)
V
DD
= V
DDCP
=+3.0 V; T
amb
= +25 °C unless otherwise specified.
SYMBOL PARAMETER MIN. TYP. MAX. UNIT
Serial programming clock; CLK
t
r
Input rise time 10 40 ns
t
f
Input fall time 10 40 ns
T
cy
Clock period 100 ns
Enable programming; STROBE
t
START
Delay to rising clock edge 40 ns
t
W
Minimum inactive pulse width 1/f
COMP
ns
t
SU;E
Enable set-up time to next clock edge 20 ns
Register serial input data; DATA
t
SU;DAT
Input data to clock set-up time 20 ns
t
HD;DAT
Input data to clock hold time 20 ns
Application information
SR01417
CLK
DATA
STROBE
LSB ADDRESS
t
SU;DAT
t
HD;DAT
t
f
t
w
t
r
t
SU;E
t
START
T
cy
MSB
0
Figure 11. Serial Bus Timing Diagram

SA8027DH,512

Mfr. #:
Manufacturer:
NXP Semiconductors
Description:
IC N/IF FREG SYNTHESIZER 20TSSOP
Lifecycle:
New from this manufacturer.
Delivery:
DHL FedEx Ups TNT EMS
Payment:
T/T Paypal Visa MoneyGram Western Union

Products related to this Datasheet