Capacitor topology after Regulators...

[weissi]

Hi there!

I´m facing the following problem: I have dozens of Panasonic FC 680µF/16V.... Tested them in my PC-Soundcard(12V/-12V filtering, after regulators at the DAC(WM8770) and at the opamp´s) With them I got really fantastic results with the "unreal" RMAA-test...
I don´t give much about this test (SNR improved by 4dB?? THD+N also...) BUT: It also does sound fantastic!

So, the Problem: you know, all datasheets tell us to use "small" caps after regulator, ~<10µF, to keep the dynamic stabilisation...

Do you think there´s a problem when putting them after regulators (7812/7912/TL413)
I plan to build another TDA1541A DAC with plenty of them, as I have them :-) My approach is: [100nF 0812] || [680µF FC] at all decoupling points.... and ~10 of them paralleled at every rail in the PSU

P.S.: Are they better than OS-CON in analog systems? OS-CON´s made the sound really harsh...

Regards,
max weissi

[weissi]

OK, the more you (I) read the more you (I) know....

[wazzup]

What did u read?

[jean-paul]

Using ultra low ESR caps directly *behind* the regs is not wise to do. Not with 78xx and 79xx series anyway. Do a Search here for details ( keyword : ESR ) as it has been discussed even more than cable directionality...

[weissi]

I read some app´notes from ONsemi.....

So, it primary depends on the output stage of the regulator... In the case of 78xx, and 317 it´s a NPN transistor that should be unconditionaly stable - no matter what type of cap is used....
LDO´s require more investigation for choosing the right capacitor, because they typicaly work in a more unstable area of the output stage (therefore typical a PNP).

[weissi]

But I think the app´note describes only DC-~100Hz
Don´t know how that is at MHz... when a 78xx runs out of gain...
The main thing is that it works... I´ll try some other cap´s aswell, anyone using Rubycon ZL? they are much cheaper than FC

[EE_Mark]

The regulaors don't like to see a lot of capacitance on their outputs or they can oscillate because of the inductance of their outputs. You can put a small value resistor, like 1 ohm (how much current is flowing?) between the regulator output and a giant cap and you won't have any problems.

[Jocko Homo]

All of those 3-terminal jobs have a feedback loop, and just throwing any old cap......especially with low ESR is asking for problems. It affects phase margin, and can easily make an oscillator. As you have found out.

Those of you with access to a spectrum analyzer.....here's what you do.

Look at the noise vs frequency with several different value and types of caps. You will see how the noise will peak at some frequency if you pick the wrong one.

If you don' t have an analyzer, but do have a 'scope and a square wave source.............drive the output with the square wave, and see if it rings. If it rings a lot, you have stabilty problems

Jocko

[weissi]

Well, I do have access to something that has a FFT function... But it might bre that it hasn´t enough bandwith for really serious measurements....

http://www.rapid-tech.com.au/dl708e.htm

But that thing looks quite prominent to someone (actually, this was one of a reason why I choosed my actual occupation, hehe, beside the metcal´s and the microscopes for soldering SMD...)

[wazzup]

So what values are we talking here? Is a low esr 1000uf Not a good thing behind a regulator? Or are even smaller values causing trouble?

I have build my ps, for my cd pro2 mechanism, using this schematic, using lm317t`s for regulators.

I don`t have acces to a scope, so I have to just build it correct, if anyone have advise on the caps behind the regulator, plz share it!