Datasheet
LMX2541SQ2060E, LMX2541SQ2380E
LMX2541SQ2690E, LMX2541SQ3030E
LMX2541SQ3320E, LMX2541SQ3740E
www.ti.com
SNOSB31I –JULY 2009–REVISED FEBRUARY 2013
Table 10. Change in Current Consumption in Bypass Mode as a Function of VCOGAIN and OUTTERM
VCOGAIN
3 6 9 12 15
3 -26.0 -22.3 -18.6 -15.1 -11.8
6 -18.5 -15.5 -12.6 -9.7 -6.9
OUTTERM 9 -11.1 -9.0 -6.9 -4.7 -2.5
12 -3.8 -2.6 -1.4 +0.0 +1.5
15 +3.3 +3.7 +4.0 +4.5 +5.3
Table 11. Change in Current Consumption in Divided Mode as a Function of DIVGAIN and OUTTERM
DIVGAIN
3 6 9 12 15
3 -24.4 -21.7 -18.7 -15.9 -13.3
6 -16.2 -14.6 -12.6 -10.1 -8.0
OUTTERM 9 -8.3 -7.6 -6.8 -5.0 -3.2
12 -0.5 -0.7 -0.7 +0.0 +1.3
15 +7.1 +6.0 +5.2 +4.9 +5.6
Loop Filter
Loop filter design can be rather complicated, but there are design tools and references available at www.ti.com.
The loop bandwidth can impact the size of loop filter capacitors and also how the phase noise is filtered. For
optimal integrated phase noise, choose the bandwidth to be about 20% wider than the frequency where the in-
band PLL phase noise (as described in PLL Phase Noise) and open loop VCO noise cross. This optimal loop
bandwidth may need adjustment depending on the application requirements. Reduction of spurs can be achieved
by reducing the loop bandwidth. On the other hand, a wider loop bandwidth may be required for faster lock time.
Note that using the integrated loop filter components can lead to a significant restriction on the loop bandwidth
and should be used with care. 2 kΩ for R3_LF and R4_LF is a good starting point. If the integrated loop filter
restricts the loop bandwidth, then first try to relieve this restriction by reducing the integrated loop filter resistors
and then reduce the capacitors only if necessary.
Configuring the LMX2541 for Optimal Performance
1. Determine the Channel Spacing (f
CH
)
– For a system that has a VCO that tunes over several frequencies, the channel spacing is the tuning
increment. In the case that the VCO frequency is fixed, this channel spacing is the greatest number that
divides both the VCO frequency and the OSCin frequency.
2. Determine OSCin Frequency (f
OSCin
)
– If the OSCin frequency is not already determined, then there are several considerations. A higher
frequency is generally, but not always, preferable. One reason for this is that it has a higher slew rate if it
is a sine wave. Another reason is that the clock for the VCO frequency calibration is based on the OSCin
frequency and in general will run faster for higher OSCin frequencies.
– Although a higher OSCin frequency is desirable, there are also reasons to use a lower frequency. If the
OSCin frequency is strategically chosen, the worst case fractional spur channels might fall out of band.
Also, if the OSCin frequency can be chosen such that the fractional denominator can avoid factors of 2
and/or 3, the sub-fractional spurs can be reduced.
3. Determine the Phase Detector Frequency (f
PD
) , Charge Pump Gain (K
PD
) and Fractional Denominator
(FDEN)
– In general, choose the highest phase detector frequency and charge pump gain, unless it leads to loop
filter capacitor values that are unrealistically large for a given loop bandwidth. In this case, reducing either
the phase detector frequency or the charge pump gain can yield more feasible capacitor values. Other
reasons for not using the highest charge pump gain is to allow some adjustment margin to compensate
for changes in the VCO gain or allow the use of Fastlock.
– For choosing the fractional denominator, start with FDEN = f
PD
/f
CH
. As discussed previously, there might
be reasons to choose larger equivalent fractions.
Copyright © 2009–2013, Texas Instruments Incorporated Submit Documentation Feedback 55
Product Folder Links: LMX2541SQ2060E LMX2541SQ2380E LMX2541SQ2690E LMX2541SQ3030E
LMX2541SQ3320E LMX2541SQ3740E