13-11-2012, 09:12 PM
I can think of several reasons that might explain that behavior.
1. Each LM317 needs a 0.1 µF cap. from its input terminal to 0v.
2. The 2nd. LM317: the junction of the 390 Ω and the 4k7 variable resistor needs a 10 µF capacitor to 0v. (Incidentally, the manufacturer's datasheet recommends a 240 Ω instead of that 390 Ω when an LM317 is used in this configuration).
3. The arrangement of the earthing between all the components may not be optimal. Wiring from rectifier -ve to reservoir cap. -ve needs to be short and fairly thick. Then all other 0v. returns, plus the wiring to '0v. out', need to be taken only to the -ve of that cap. This ensures single-point earthing and thus eliminates 'hum loops'.
4. Your 'scope earth may be picking up noise and hum on account of the changes that may need to be implemented as in 4 above.
5. Connect a 0.1 µF cap. in parallel with the main reservoir cap. This reduces the high-frequency impedance of the main reservoir cap., leading to reduced noise on the unreg. HT line.
6. The unusual rectification scheme you're using could be making a significant contribution to the noise, in particular the additional half-wave rectifier diode. If it is temporarily removed, does the noise performance substantially improve? A bleeder resistance across the final reservoir capacitor might help.
Not related to noise performance, but to add some reliability to the PSU, I would add two protection diodes: one from input to output of the 2nd. LM317 (wired with cathode towards the reservoir capacitor) and one in parallel with that aforementioned 240 Ω: cathode to the LM317 O/P pin. 1N4004 / 5 / 6 / 7 are just fine for those diodes.
Plenty of ideas to experiment with in that lot!
Best of luck -
Al.
1. Each LM317 needs a 0.1 µF cap. from its input terminal to 0v.
2. The 2nd. LM317: the junction of the 390 Ω and the 4k7 variable resistor needs a 10 µF capacitor to 0v. (Incidentally, the manufacturer's datasheet recommends a 240 Ω instead of that 390 Ω when an LM317 is used in this configuration).
3. The arrangement of the earthing between all the components may not be optimal. Wiring from rectifier -ve to reservoir cap. -ve needs to be short and fairly thick. Then all other 0v. returns, plus the wiring to '0v. out', need to be taken only to the -ve of that cap. This ensures single-point earthing and thus eliminates 'hum loops'.
4. Your 'scope earth may be picking up noise and hum on account of the changes that may need to be implemented as in 4 above.
5. Connect a 0.1 µF cap. in parallel with the main reservoir cap. This reduces the high-frequency impedance of the main reservoir cap., leading to reduced noise on the unreg. HT line.
6. The unusual rectification scheme you're using could be making a significant contribution to the noise, in particular the additional half-wave rectifier diode. If it is temporarily removed, does the noise performance substantially improve? A bleeder resistance across the final reservoir capacitor might help.
Not related to noise performance, but to add some reliability to the PSU, I would add two protection diodes: one from input to output of the 2nd. LM317 (wired with cathode towards the reservoir capacitor) and one in parallel with that aforementioned 240 Ω: cathode to the LM317 O/P pin. 1N4004 / 5 / 6 / 7 are just fine for those diodes.
Plenty of ideas to experiment with in that lot!
Best of luck -
Al.
