please advise MUCH better chip than 7815 for 15V

Hi
I am looking for the chip stabilizers similar to std 7815/7915 but with much better performance,
for example for +5V stabilizer LM2940CT is much much better than old 7805 - I am looking for similar chips but from the range +-12 to +-18V to power two TL072s, so current consumption is very small.

The voltage at the input of the stabilizer is +23,8Vdc and -25,6Vdc.

I do not have a room for good discrete designs.

will 317/337 with 2 resistors to set the voltage and some capacitors be better than 7815/7819?

maybe simply classic circuit zener +transistor + capacitor will be better than 7815/19?

maybe it is no sense for looking for better chips, maybe it is enough to use OSCON SEPC 330u/25V capacitors after 7815/7819 chips for extraordinary quality?

thank you in advance for any suggestions!
 

Mark Johnson

Member
Paid Member
2011-05-27 3:27 pm
Silicon Valley
Volume 4 of Linear Audio magazine contains "A comparative overview of power supply regulator designs with listening tests" by John Walton. It included the LM317/LM337 pair of regulator ICs, and also the Linear Technology LT1963A/LT3015 pair of regulator ICs, but not the LM7815/LM7915 pair.

Some of the measurement results (but not the full text of the article itself) were posted on the Linear Audio website:

DIYaudio member Fred Dieckmann posted a simple circuit to improve the performance of the LM317/LM337 (HERE)
 

FoMoCo

Member
2012-12-04 10:04 pm
What exactly are you looking for when you say "better"? The 7815/7915 is good enough for almost every application. There are indeed pricier options with better tolerances, PSRR, etc. but for most applications it doesn't better.
 
Last edited:
Hi
I am looking for the chip stabilizers similar to std 7815/7915 but with much better performance,
:confused:
Define "better performance"
Stabilizers "stabilize", that's their function.
You are feeding 2 TL072, which are very happy and will work/sound *exactly* the same with any rail voltage within, say, +/-14 to +/-16Vwith any rail voltage voltage.

for example for +5V stabilizer LM2940CT is much much better than old 7805 - I am looking for similar chips but from the range +-12 to +-18V to power two TL072s, so current consumption is very small.
An "improved" regulator may have lower minimum voltage dropout, and may dissipate less ... if you do your homework and feed it with a correspondingly lower raw voltage.
Otherwise, it will dissipate exactly the same.
Being low dropout is an advantage (no semicolons used because it's real) but you may choose not to use it, which seems to be the case.
The voltage at the input of the stabilizer is +23,8Vdc and -25,6Vdc.
What I said above.
You are working with an about 9 to 10V dropout.
Any of those chips will dissipate and behave exactly the same.

And NOW we may have a problem:
can't speak for the hypothetical regulator you can't find, but as the Rabsperry Pi guys say:
to replace the 7805 battle integrated can use the LM2940-5.0 as it is a direct replacement in terms pinouts and power delivery, but bear in mind that your tolerance maximum input voltage of 26V is against the 35V of the other,
and you are working *dangerously* close to it.
Just don't be carried away by the "better/improved" label, first find out what it means exactly, it may even be bad news in your intended application.
And as I said before, it will *not* affect sound.

Besides, you are feeding 2 TL072 , which use around 3 mA each , 6 or 7 mA total, so your regulator, which can supply up to 1000mA isn't exactly struggling running at 0.7% capacity.


I do not have a room for good discrete designs.
Nor need it.

will 317/337 with 2 resistors to set the voltage and some capacitors be better than 7815/7819?
Again, define "better".
Although the answer will probably be "no difference at all".

maybe simply classic circuit zener +transistor + capacitor will be better than 7815/19?

maybe it is no sense for looking for better chips, maybe it is enough to use OSCON SEPC 330u/25V capacitors after 7815/7819 chips for extraordinary quality?
Do you think your preamp will somehow sound better if you use $60 capacitors instead of $1 types?
Personally I would invest $$$ and effort in improving some *sound* related paramater, but maybe that's me.

By the way, regulator datasheet suggests .1uF to 10uF caps after the regulator; 330uF means stressing it, both at turn-on because it obsessively tries to get to designed voltage *NOW*!! , even a microsecond is "too long" , and at turn off, because if raw voltage discharges faster than the regulated one (quite possible here because of very low preamp consumption), the internal pass transistor gets reverse voltage across it and believe me, it does not like that *at all*.

Yes, you can add a reverse biased protective diode across it, but it's bad engineering having to add protection because of self-created problems.
Much better trying to do things right from the beginning.
 
Its kind of funny: Before Regulators (were cheap, ubiquitous, varied, etc), designers were happy with CRC configurations. R's were cheaper than L's (still true!). If the goal was way less noise and space was of little concern, then CLC. Or, even CLCLC. Noise - from diode switching that bad old mains A/C - was always a problem. Hence... in the design of amplifiers why so much trouble was gone to, to ensure PSRR (power supply ripple rejection ... or ... rejection ratio) was good.

Along come regulators. Voilá - with a tiny heat sink, one could replace that nasty "R", or that bulky "L", and you would get clean, regulated voltage. Next was worry of injected noise. Uh... did we forget something? Noise reduction remains easily in the domain of ... C, R and C. 1 uF across the output of a bog-standard 7815, a resistor (E=IR, what's the expected I?) to drop about 0.5V ... and a nice BIG C (like 1000 uF or larger) to hold plenty of juice for transients. Output impedance is additionally dropped to near-zero (which no regulator is all that good at doing - without attendant noise).

I've been designing just such supplies for years (like 41!) and my "Go To" design (for low-voltages) is nearly always the same: bridge to 100 to 560 uF, to REGULATOR, to 1uF to (R=0.5/I typically 1 to 47 ohms), to 220 to 2200 uF paralleled with a handful of 1uF polyesters. Their non-linearity is actually an advantage in power supply final-reservoir application.

Good Luck!

GoatGuy
 
Last edited:
Dear Guys,
thank you for your replies,
I cannot answer to all of the posts, I do not want BIG arguing in here why the sound of moded cd is extraordinary with LM2940CT and realy bad with 7805... and why power amps sound better with stabilized power supply vs unstabilized (some claims that smps is the best choice fo power amps).
I asked in the first post about the solution because my last experiments with decoupling power rails with ~10u MKP 400Vdc caps made astonished effects in amps sound quality.
Of course, better means better: lower noise, lower output impedance, etc.
Some of you answered like very experienced in this matter, thank you once again.