Hello:
So I am building an application specific oscillator, and have attached a partial (work in progress) schematic. It is a straightforward N=128 frequency multiplier synced to the AC line, with an output freq of 7.68 kHz. I am really interested in anyone's experiences with the low pass filter for the phase locked loop. I have looked into numerous app notes and schematics off the web for simple equations to help me design the LPF without having to go back to feedback control systems.
I have found examples of resistors in the LPF that vary from 1 Meg to 5k, with similar valued capacitance and similar frequency ranges. It appears there is not a lot of consistency. I have used the equations provided in Philips 74HC4046A http://www.datasheetcatalog.org/datasheet/philips/74HC4046A.pdf, and came up with the attached results. Do they look reasonable? I don't need extremely fast settling time; just a stable lock to 60 Hz.
VCO output = 70 Hz at 4.1V
VCO output = 50 Hz at 0.9V
Kp = phase comparator gain = 5/(4pi) = 0.4
Kv = VCO gain = (2fL * 2pi)/(4.1-0.9) = 5026
Kn = 1/N divider ratio = 0.0078125
choose settling time = 0.25 sec, damping ratio 0.7
wn = 5/settling time = 20
t1+t2 = Kp*Kv*Kn/wn^2 = 0.039 sec
select C2 = 1.5 uF
R4 = [(t1+t2)*2*wn*DF - 1] / (Kp*Kv*Kn*C2) = 3.9k
R3 = t1/C2 - R4 = 18.2k
So I am building an application specific oscillator, and have attached a partial (work in progress) schematic. It is a straightforward N=128 frequency multiplier synced to the AC line, with an output freq of 7.68 kHz. I am really interested in anyone's experiences with the low pass filter for the phase locked loop. I have looked into numerous app notes and schematics off the web for simple equations to help me design the LPF without having to go back to feedback control systems.
I have found examples of resistors in the LPF that vary from 1 Meg to 5k, with similar valued capacitance and similar frequency ranges. It appears there is not a lot of consistency. I have used the equations provided in Philips 74HC4046A http://www.datasheetcatalog.org/datasheet/philips/74HC4046A.pdf, and came up with the attached results. Do they look reasonable? I don't need extremely fast settling time; just a stable lock to 60 Hz.
VCO output = 70 Hz at 4.1V
VCO output = 50 Hz at 0.9V
Kp = phase comparator gain = 5/(4pi) = 0.4
Kv = VCO gain = (2fL * 2pi)/(4.1-0.9) = 5026
Kn = 1/N divider ratio = 0.0078125
choose settling time = 0.25 sec, damping ratio 0.7
wn = 5/settling time = 20
t1+t2 = Kp*Kv*Kn/wn^2 = 0.039 sec
select C2 = 1.5 uF
R4 = [(t1+t2)*2*wn*DF - 1] / (Kp*Kv*Kn*C2) = 3.9k
R3 = t1/C2 - R4 = 18.2k
Well, never mind I guess 🙄
Got it dialed in very nicely. Fast settling time, very little overshoot, and extremely clean sine wave output for the application.
PLL equations only get you in the ballpark; you have to adjust by test for optimal performance.
Now on to the power inverter.
Got it dialed in very nicely. Fast settling time, very little overshoot, and extremely clean sine wave output for the application.
PLL equations only get you in the ballpark; you have to adjust by test for optimal performance.
Now on to the power inverter.
Old project, but works like a champ.
Latest schematic attached. I used a switched capacitor filter to convert the square into sine. After this circuit, I feed a Class-D power amp and step-up transformer to 120V, so after those components I have no purpose in ultra-low distortion of the SCF.
The addition of MC14557 allows variable phase shift, so you can how tune not just the lock, but the phasing from 0-360 degrees from the original input.
Latest schematic attached. I used a switched capacitor filter to convert the square into sine. After this circuit, I feed a Class-D power amp and step-up transformer to 120V, so after those components I have no purpose in ultra-low distortion of the SCF.
The addition of MC14557 allows variable phase shift, so you can how tune not just the lock, but the phasing from 0-360 degrees from the original input.