The simplistic Salas low voltage shunt regulator

[iko]

IMHO, even your first image has too much noise for this regulator; might be from your scope probes. Also, you have to be very careful what capacitors you use at Vout. You might want to try removing all those smd caps, and just use one larger size electrolytic, and test it first with a resistor load.

 

[Salas]

His first image is the rail if the clock is connected to its receiver. If not its flat. The second pic its his clock itself. Maybe with an earth rake at probe's end would show different, and could be the crock earth cable picking?

 

[canvas]

ikoflexer,

I did that with one 630uf cap before. My probe is a Tek P5050 model. I guess it might be the ground loop problem. Once the clock was connected to next stage, other power lines was infected, too!

 

[Salas]

After a power line is flat and tame locally to the reg with a dummy load, who knows what lurks in the digital corridors.

 

[iko]

I see. Have a look at this classic article.

What's All This Ground Noise Stuff, Anyhow?

 

[Rtop84]

Try change decoupling network from FB-C-FB-C-FB-C to simple C1-FB-C-C ( C1= electrolytic cap at output regulator then ferrite bead +100nf+10n+1nf very close XO power pin). Separating HF decoupling caps by ferrite bead is bad idea.

 

[canvas]

Do you think this is the problem of my scope? I heard many scopes have difficulty on measuring low-level signal.

ps. My scope is powered from an isolated transformer.

 

[Salas]

Does the system loose points subjectively after you see the contamination? Does it give clues it goes less resolution, noisier?

 

[canvas]

Salas,

The strange thing is the sound did improved on 1.2R. There is no sign of resolution loss, noise or anything. I suspect my previous power supply may have this problem, I just didn't notice before. I will switch back to old PS tonight and see how it goes, thanks.

 

[Salas]

Weird contradictory clues. Investigate I say. Good luck.

 

[canvas]

Salas,

The old PS has been re-installed. Still, the problem remains, so the cause is clearly not 1.2R. This is beyond my knowledge, I wish someone could show me some directions. I've tried several combinations of bypass technique without any sign of relief. Anyhow, 1.2R is definitely a great PS for digital. A big improvement over my previous PS.

 

[Salas]

Read well the Bob Pease article Ikoflexer linked in the previous page.

 

[stormsonic]

Supply_decoupling.pdf and a tons of other good articles can be found with Google

 

[iko]

Here's another Bob Pease pearl that might be of help:

Bob Pease describes circuit solutions that solve common-mode noise problems.

 

[schultzsch]

hello

i have this regulator Salas shunt reg. +5V to +3.3V modificaton - My Photo Gallery

i want to get from it 300ma.what value should r1 have? do i need to do other modifications?
i want to use it to power some digital circuits

thanks

 

[Salas]

Vr1/r1=ICCS. See that the small sinks will be comfortable to touch for 5secs when connected to its intended load.

 

[schultzsch]

who is vr1? voltage at the input or output of the regulator?

 

[Salas]

Voltage drop across r1.

 

[schultzsch]

thank you

 

[arche13]

hello,

wich is the last diagram for shunt regulator Salas?

thanks