Linear post regulator for SMPS, what's your take?

[Calvin]

Hi,

even bad SMPS feature quite high clock-frequencies (>>10kHz). So a simulation should be done at more realistic ripple-freqs.
If You add a parallel resistor to C1 in post #9, the gyrator features voltage regulation too. The output will be the resistor-divider´s -2x Ube.
Add a electrolytic/ceramic cap combination from Q1s emitter to gnd to cope with the loads current demands.
Optionally a ferrite bead ahead of the gyrator may impove matters also.
Imho this is sufficient for audio supplies and You can safely omit with the series regulator, without loss of performance.

jauu
Calvin

[wwenze]

Would it be easier if I used AC analysis instead?





(the 1.75k resistor should've been 2.4k)

I think I'll settle for the third one since I already ordered the parts for the top right for a small 3-5A~ish amp. But the gyrator will be very useful when I want to power a computer.

So glad I made this thread! Learnt a lot of things!

[agdr]

I like Calvin's idea of the ferrite bead ahead of the gyrator! Seems like that or a choke would go along way to nailing the switching spike in particular. The article in martin clark's post mentioned maybe using 20 turns of magnet wire around a resistor for an input choke.

Here is the same circuit but with a 6mA load instead of 6A. Note the (regulated) output is still the same, but look where the gyrator output voltage is at now.

That's without the resistor Calvin mentioned though. That is a great suggestion he has. It might be enough regulation for what you are doing and would avoid the added regulator drop. Or, conversely, add Calvin's resistor and then feed that to the regulator. Might be even flatter.

[Calvin]

Hi,

a coil wound from wire exhibits some capacitive coupling between each winding. Its like as if a small cap were paralleled to the coil. HF-spikes will pass the coil nearly unaffected. Take a good ferrite bead instead (see the datasheets of FBs).

jauu
Calvin

[agdr]

Thanks, Calvin!

So, maybe something like this, now 12V out at 4A load.

I've adjusted the voltage sources to more-or-less produce the scope waveform from the laptop power brick in that article. A 30mV p-p switching spike at a 50kHz rate, plus the scope seems to show 1 mHz background noise from somewhere, if I'm reading the divisions correctly.

Going from 4A to 4mA load the output of the gyrator now only goes from 17.2V to 18V with that parallel resistor added to C1.

[AndrewT]

What is causing that spike in supply voltage at Vin? ~45mVpp
It ripples right through to the output. Yes it does get attenuated, but the spike should not be there!!!!

[agdr]

Hi AndrewT,

Good question! I would have expected the ferrite bead to do more (in terms of supression). I probably have the value too low - 10uH (?) I'm most likely doing something wrong with the model there. I found this:

Ferrite Beads

"Ferrite beads are often used in gigahertz filtering applications. Ferrite beads come in two flavors: high-Q, resonant beads and low-Q, nonresonant beads, also called lossy, or absorptive beads. The high-Q type has no place in a digital circuit. These beads are used to construct RF oscillators and filters and other circuits that need highly resonant circuit elements. In a digital power-filtering application, the last thing you want is resonance. The low-Q type is commonly used for power-supply filtering, in series with the power connection. Most often, this style of filter also has a capacitor to ground on either side of the inductor (Figure 1)."

I'm pretty out of date - my last 15 years have all been finance and operations. I'm still a coil guy with not a lot of knowledge about ferrites. Hopefully someone will post a better model for wwenze to try. I'm probably going to build one myself for fun. A laptop brick - to - audio-worthy supply filter would be a handy thing to have. I'm going to go crawl back under my rock for now, lol.

[wwenze]

Will this type help?



Also I'd believe almost all power bricks have that thing in the cable itself right before the connector already.
Typically, what values would those have?

[darkfenriz]

When it comes to SMPS regulation, it may turn out easier to regulate the dropout voltage between linear reg's output and input to th whole linear stage.
Typical smps regulation voltage to ground measurement can become problematic when you think about dropout margin, tolerance, parameters wander etc. while not sacrificing efficiency too much.

[AndrewT]

Originally Posted by agdr

I probably have the value too low - 10uH (?)

forget the ferrite bead.
Look at your capacitance multiplier.

Some of your capacitance multipliers have omitted the necessary lower resistor. They cannot work properly.
None of your capacitance multipliers have a two stage RCRC holding the base at a ripple free voltage.
Sort the capacitance multiplier, to give a very low ripple at the input to the regulator.
Then the regulator only has to cope with meeting the variable demand of the load, without having to (poorly) attenuate input interference.