SMPS vs 50Hz Transformer noise

"Phil Hobbs"

SED has been destroyed by these total fools.

You proud of that ?

..... Phil

SED is an unusally good ng, as ng's go. If you set your spam filters up properly.

I commonly see, say, 250 new headers, of which Agent pitchs all but 10 or so.

John

George, I didn't used 1Ohm in case of 10A I used 0.1 Ohm + 1000uF. Well about my new design 10uV, It is not just a power supply. It is device that has an power amplifier, various feedbacks,... but in power supply section I want to make it quite as much as I can and as I said I also consider 50Hz Transformer.

regards

Nah, it's just that every time I prod D from BC he starts talking about religion, and off we go for another thousand bad-tempered, unthinking, and very ill-informed posts.

This guy showed enough glimmers of cluefulness that I thought it might be worth showing him the ropes. We shall see.

Cheers

Phil

Actually, I'd expect it to be in the double-digit nanoHenry ballpark. Much depends on the details of manufacturing.

Jeroen Belleman

Ebrahim, Check out the spec sheet for the 1000uF cap and look at it's impedance at the frequency of interest. (whatever the SMPS switching frequency is.)

If you can live with a bit of voltage drop Phil H's suggestion of a capacitance multiplier will work. You'll need a beefy transistor to pass the 10 A. (you're already throwing a way a volt for the resistor after all.)

George H.

Before Phil gets totally carried away (and I wish he would), it depends on the design of each power supply as to which gives out more noise. SMPSs do NOT give out pure dc. Far from it.

With a switched mode power supply, you can choose what frequency you use, hence can shift the frequency of the interference away from any spectrally sensitive area of your circuit. Naturally, as they have a higher switching frequency than 50Hz, the reactive components are smaller (magnetics and capacitors), so very physically small designs can be achieved with an SMPS. Compare a switched mode in a mobile phone (or a laptop cable) with that of a

Comparing regulation, you can get an SMPS to regulate very accurately and very efficiently with very small components with output ripple of only 50mV or so. To get the output ripple with a 50Hz linear power supply would mean a huge output cap.

However, 50Hz transformers are easier to design.

If you want to design a mains rated SMPS, look at Power Integrations. They make SMPS design simple.

For general SMPS design, keep the output caps good quality (tants, OSCONs etc), keep the inductor close to the switching pin and keep your output cap close to the inductor and you should be OK with radiated and electrical noise. Put an input cap close to the input switch to provide a low impedance path for the input current surges.Shielded inductors also reduce radiated emissions.

Hope this helps

Keep the inductor close to the switch in the chip, Keep the output cap close to the inductor. Use a shielded inductor to minimise radiated noise and decouple the input.

A good ground also helps

General good rules for SMPS design

-- Bill Naylor

"Electronworks.co.uk"

Generally well over 100 times more.

Filter electro values in SMPS are typically larger than for a similar size

F*ck off - fool.

... Phil

Either scheme can achieve good 'noise' (apparently you really mean ripple when you say noise).

The SMPS frequency is higher, so stray capacitance is more important; the 50 Hz transformer is larger, so inductive coupling can have significant range.

When the output power is good enough, it doesn't matter what the source design is.