D-Noizator: a magic active noise canceller to retrofit & upgrade any 317-based V.Reg.

Which circuit is the canonical solution for the LM317?
This one:
798268d1574972003-noizator-magic-active-noise-canceller-retrofit-upgrade-317-based-reg-denoisatorf-png
]
 
Tested OK.

I take the opportunity to rectify one or two mistakes:

attachment.php


C5 needs to be reversed
R3 should be returned to the collector of Q2 to increase the Vce
R3 could be larger than 180K for the same reason

A compensation cap is required, as for other versions; a few nF is sufficient. 10nF is safe.
 

Attachments

  • DienFix.png
    DienFix.png
    12.6 KB · Views: 2,246
Tested OK.

I take the opportunity to rectify one or two mistakes:

828551d1585426367-noizator-magic-active-noise-canceller-retrofit-upgrade-317-based-reg-dienfix-png


C5 needs to be reversed
R3 should be returned to the collector of Q2 to increase the Vce
R3 could be larger than 180K for the same reason

A compensation cap is required, as for other versions; a few nF is sufficient. 10nF is safe.

Thx for the great thread I am also asking for a group buy and some test result?
What about sound quality?
Could it be a great upgrade for my cheep volume with lm317/337 feeding PGA2310?
Thanks again
 
Thx for the great thread I am also asking for a group buy and some test result?
What about sound quality?
Could it be a great upgrade for my cheep volume with lm317/337 feeding PGA2310?
Thanks again

Hi Kimschips
I designed a dual rail board using lm317/337.

Post 518 page 52.

There are some gerber files that can be used to order pcbs from a PCB house like JLCpcb.

There is also a mouser BOM.

I can verify that the board works but we don't yet have noise figures. I'm waiting on Carlmart to do some noise measurements on one of my boards.
 
Tested OK.

I take the opportunity to rectify one or two mistakes:

828551d1585426367-noizator-magic-active-noise-canceller-retrofit-upgrade-317-based-reg-dienfix-png


C5 needs to be reversed
R3 should be returned to the collector of Q2 to increase the Vce
R3 could be larger than 180K for the same reason

A compensation cap is required, as for other versions; a few nF is sufficient. 10nF is safe.

Thanks Elvee for testing it and correcting the polarity of C5.

Have you tried R2 at 820 ohms, since that would be the recommended value for Vout = 15 V?. That would involve testing with the connection of R3 in the original schematic position and with the value at 330K, although, perhaps, it will reduce VCE (Q2) to near saturation values.

Possibly need to be implemented with a single transistor, similar to the denoiser (with less performance).

Regards
 
Last edited:
What about sound quality?
Could it be a great upgrade for my cheep volume with lm317/337 feeding PGA2310?
Thanks again
The sound quality will very much depend on the context: for example, a circuit using only good opamps will probably see no improvement, but many discrete circuits are very sensitive to the supply quality.
I have no idea about the PGA2310, and the datasheet does not give a PSRR spec

Elvee,

Do you have an LTSpice model for the 7815?
No, I don't, but it is probably available on the Yahoo LTspice group

Thanks Elvee for testing it and correcting the polarity of C5.

Have you tried R2 at 820 ohms, since that would be the recommended value for Vout = 15 V?. That would involve testing with the connection of R3 in the original schematic position and with the value at 330K, although, perhaps, it will reduce VCE (Q2) to near saturation values.

Possibly need to be implemented with a single transistor, similar to the denoiser (with less performance).

Regards
As I said, I just made a functional test and no measurements, but resistors values could (and should) be suited to the supply voltage.

It would work with a single transistor, but the 100 ohm would shunt much of the gain away, and I think 2 transistors are preferable