In an effort to create cleaner DC power rails (off a 12vdc wall wart)

Thread Starter

Mahonroy

Joined Oct 21, 2014
406
Hey guys,
I am wanting to create cleaner DC power rails for some circuits I have been developing, and I wanted to run it by you guys to see if this is a good method or not. I understand that generally you design some of this based off the maximum allowable ripple required, however I do not have those values, so I am trying to make it as clean as reasonably possible.

Basically I need a clean 5V rail (600mA peak tops -> 1 amp will do), and a clean 3.3V rail (80mA peak tops -> 200mA will do). Power input is a 12VDC power adapter. I am using various microcontrollers, sensors, RF, and other breakout boards. Some are susceptible to noise and others are not.

My idea was to have the 12VDC power input come into the circuit, then immediately go through a pi filter (how do I determine appropriate values for this?). I was then thinking of using a adjustable LM2575 switcher and set its voltage output to 7.5 volts, then use a 7805 style linear regulator to get a nice 5V off that. (such as the MC7805AB).

Are these good parts to accomplish this? I was a little uncertain about the 7805 and the heat it may generate (I really do not want to have to use a heatsink if I don't have to), this is why I was trying to get the switcher to output just above the 7805 threshold input voltage.

I had another for instead of using a 7805 style part, to instead use a 5V LDO that has a high PSRR rating - however there are so many parts I didn't really know what I was looking for). This would also allow for me to bring the LM2575 switcher voltage down closer to the 5V rail. I don't know...

I also searched around on the net for a while of possible pi filters that people use for dc power adapter inputs, but couldn't really find much info. I am inclined to use 2x 10uF ceramic capacitors, and 1x 30uH inductor (amperage rating equaling the max power input amperage).

What are your thoughts on this?
 

blocco a spirale

Joined Jun 18, 2008
1,546
You seem to be attempting to address a problem that you may not even have.

I would just use a good quality 5V linear or switching PSU and a 3.3V linear regulator and pay attention to the wiring topology and local power decoupling for the various devices.

If there's a problem in practice, address it, but don't waste effort solving non-problems. With a few exceptions, any well designed electronics should be tolerant of some PSU noise.
 

Thread Starter

Mahonroy

Joined Oct 21, 2014
406
You seem to be attempting to address a problem that you may not even have.

I would just use a good quality 5V linear or switching PSU and a 3.3V linear regulator and pay attention to the wiring topology and local power decoupling for the various devices.

If there's a problem in practice, address it, but don't waste effort solving non-problems. With a few exceptions, any well designed electronics should be tolerant of some PSU noise.
Thanks for the response.
12VDC to a 5V linear regulator @ 1 amp will dissipate a ton of heat, too much heat.
I have been using RECOM R-78E5.0-1.0 switchers in the past, but this outputs a decent amount of ripple - even with a filter. So instead of putting in all of this filterering, I thought it would be better to just start at the source and move down?
 

blocco a spirale

Joined Jun 18, 2008
1,546
Why not start with a 5V PSU rather than 12V?
Or, if the RECOM R-78E5.0-1.0 is too noisy, find a better one, quieter alternatives must exist.
The problem is, no one knows what "too much ripple" is for your application. Is power-supply ripple/noise actually causing problems?
 

AlbertHall

Joined Jun 4, 2014
12,345
This might help along the way. I know it as a 'super filter'.
(Clearly something more substantial than 2N4401 is needed, but you get the idea)
 

cmartinez

Joined Jan 17, 2007
8,218
The 78XX series of linear regulators are not too bad for low ripple purposes. But if you want the best supply you can get, with the lowest ripple and the lowest drift, I strongly suggest you take a look at these circuits.

Capture.JPG

Both use a precision voltage reference, and boost its current output to more acceptable levels. I don't know if you could make them output a higher current by doing some tweaking, but it's possible.
 

Attachments

AlbertHall

Joined Jun 4, 2014
12,345
This looks pretty interesting. Have you used it before? Would you mind explaining more about it?
A product our company made used them to smooth the supply to a radio receiver's sensitive parts.
The resistor is sized to drop a small voltage ≈ 0.5V-1V. This resistor and the capacitor form the filter and then the transistor gives a low output impedance. It allows you to use a bigger resistor and smaller capacitor than you could if they were directly in line with the supply.
 

OBW0549

Joined Mar 2, 2015
3,566
This might help along the way. I know it as a 'super filter'.
(Clearly something more substantial than 2N4401 is needed, but you get the idea)
This is a good circuit, though I always called it a "capacitor multiplier." But no matter.

With an emitter follower circuit like this, it's usually a good idea to insert a small resistor (50-100Ω) or ferrite bead in between the transistor's base and the junction of R1 and C1 to prevent high-frequency oscillation, especially if the emitter follower is driving a capacitive load such as any downstream filter capacitors or decoupling caps. Emitter followers are notorious for these spurious oscillations (often at VHF or UHF frequencies), as explained here.
 
Top