Re: Wide loop filter considered (not so) harmful
- Hi All,
The input bias current in a PLL op amp will be the main determiner
of PDF spurs. Trace, board cleanliness is usually secondary. If both
were perfect, there would be no PDF spurs. This is also assuming the
loop capacitors have absolutely no leakage (yeah, right).
I have used both of these amplifiers for the PLL 1, with these
results: (The PDF is at 668 KHz)
part BW Slew Rate Voffset InpBiasCur V Noise I Noise PDFspur
OP27 8MHz 2.8V/µs 30µV 15nA 3.2nV/rtHz 1.7pA/vHz -31 dBc
AD8671 10MHz 4V/µs 20µV 3nA 2.8nV/rtHz 0.3pA/vHz -54 dBc
The AD8671 is an extremely close substitute for the OP27 and has good
results for the PDF spur due to much lower Input Bias Current.
Do not worry about the Input Voltage Offset. This is critical, only
when designing fixed gain amplifiers. The PLL op amp is essentially
open loop at DC and the op amp doesn't care about the offset voltage.
The other two Op Amp specs to look for are lowest possible Input
Voltage noise and Input Current noise. These would be important in
very low frequency PLL's but not in a 1000 MHz PLL. The 20LogN noise
is way over any noise the op amp creates.
If you can find any op amp that beats the above specs, by all means,
use it. Just make sure it will handle the high VCC voltage (+20v).
--- In firstname.lastname@example.org, "swetterlin"
> Hi John
> Input bias current is the current in/out of the inputs; input offset
> current is the difference between the Input bias current for the two
> inputs. You are right, they are both whoppingly huge on the TLE2141.