I was tracing the schematic of this preamp/tone control board I bought from Aliexpress and found to my surprise an unusual (to me) filter on the power supply rails. Can someone give me a quick overview, or just some comments, on the effectiveness (or lack of) of this arrangement? The IC used in this position is "D358" with no other visible markings.
Trying to understand this 7812/7912 filter cct - it llooks like the D358 IC is attempting to smooth the AC ripple at the output of the 7812/7912, by comparing the AC output signal at the output of the 7812/7912 to the power supply's AC GND. That's as far as I can get so far.
This has the same principle of operation as the D-Noizator and variations made by Elvee and described in this long thread: https://www.diyaudio.com/community/...-retrofit-upgrade-any-317-based-v-reg.331491/
D358 is likely to be a LM358. How is it powered ? pin 8 of D358 is connected on V+ or before the 7812 ?
D358 is likely to be a LM358. How is it powered ? pin 8 of D358 is connected on V+ or before the 7812 ?
Yep. Though the 7812 and 7912 are not designed to be floating regulators. I wonder how this circuit behaves on start-up. Too bad they didn't want to "splurge" on a pair of LM317/337.
Tom
That appears to be the general idea they had in mind. It responds to reduce ripple by superimposing the inverse on to the ground pin of the 7812/7912. Additional to the situation which Tom mentioned, this is supposed to have 0.1uF between the regulator output and actual return path for that reg to remain stable. That might be capacitor “473”, but it should be checked.
I only see three op amps on the board, so IC3a/IC3b for ripple regulation appear to be in the same package, D358.
Technics did use this kind of regulator enhancer circuit, but they provide a resistor to ground connected to the ground pin of the regulator, not only the opamp output. That may be to handle a startup issue related to the quiescent current.
https://www.diyaudio.com/community/...tor-for-dac-clock.304013/page-10#post-6361772
I'll sim both circuit If I find some time today, I've seen unofficial transistor model of 78 reg on Internet.
I've got these, source unknown, don't know how good they are as I don't think I've ever used them.
Cheers,
Cabirio
Edit: 1 is input, 2 is common, 3 is output
Thank you for the files Cabirio.
I have made some quick tests with them, and I was not convinced by the result. After some Internet search, I found that 78xx model was available into the example/Education folder of LTSpice distribution ! (LM78XX.asc file).
I have made sim for 7815 alone, for @Elvee DxNoiser, for OP circuit, and for Technics circuit.
I removed the 100uF output cap, because it produces massive peaking and reduce reject band size.
Circuits :
Vout/V2 :
Using much better opamp on ALi only slightly expand reject band, so no need to replace the old LM358... which produce, at least on this sim with, the best result. DxNoiser is very good to has a major advantage : it doesn't require a negative supply.
I'll try to simulate transients at startup.
I have made some quick tests with them, and I was not convinced by the result. After some Internet search, I found that 78xx model was available into the example/Education folder of LTSpice distribution ! (LM78XX.asc file).
I have made sim for 7815 alone, for @Elvee DxNoiser, for OP circuit, and for Technics circuit.
- DxNoiser : https://www.diyaudio.com/community/...y-317-based-v-reg.331491/page-33#post-6138183
- Technics : https://www.diyaudio.com/community/...tor-for-dac-clock.304013/page-10#post-6361772
I removed the 100uF output cap, because it produces massive peaking and reduce reject band size.
Circuits :
Vout/V2 :
- Green : 7815
- Light blue : Technics
- Red : Ali
- Dark blue : DxNoiser
Using much better opamp on ALi only slightly expand reject band, so no need to replace the old LM358... which produce, at least on this sim with, the best result. DxNoiser is very good to has a major advantage : it doesn't require a negative supply.
I'll try to simulate transients at startup.
Startup :
Now use a symetric supply :
- Light blue : Vin
- Pink, blue, red : 7815 ,Technics, DxNoiser
- Green : Ali
Now use a symetric supply :
- Green : Vin
- Pink : 7815 GND pin current
- Dark blue : Vout
- Red : 7815 GND pin voltage
With the LM358, I don't think you need a negative supply.
Opamps can certainly provide an excellent PSRR and output impedance, but none can match the noise performance of a humble transistor. If the priority is PSRR and Zout, that's perfectly fine, but if you also need the noise reduction feature, a discrete is mandatory
Opamps can certainly provide an excellent PSRR and output impedance, but none can match the noise performance of a humble transistor. If the priority is PSRR and Zout, that's perfectly fine, but if you also need the noise reduction feature, a discrete is mandatory