Why most synchronous rectification DC power switcher runs at hundreds Khz or higher ?

[BravoV]

For DC to DC conversion like buck, boost or both, I think majority of them either controller with external power switch, or the monolithic with internal are running above hundreds of Khz up to Mhz range.

I'm aware that the inductor and the capacitors might be physically larger at lower frequency, other than these factors, are there other concerns like efficiency maybe if running < 100 Khz ? or maybe others ?

Please share some of your thought or experience regarding this.

Btw, maybe my Google-Fu is so lame, is there any synchronous rectification product that runs < 100 Khz ?

TIA

 

[Blueteeth]

I'm not seasoned expert on DC/DC, but I've built a few buck, boost, flybacks, both with specialist chips and discrete designs from scratch. There are some which run at a lower frequency (LM2675 runs at 52kHz I believe). You've already stated the reasons for a higher switching frequency: smaller inductor/caps, as well as better transient response and more easily filtered for a clean DC output. But it comes at a cost of switching losess, and tighter layout specifications to minimise RFI/noise.

I'll admit, I haven't seen any SMPS chips using synchronous rectification that run at <100Khz. There could be many reasons, perhaps because modern MOSFET's can have sufficiently low gate charge to be switched at a higher frequency without significant power losess (especially with internal switches, as the IC manufacturer can control all aspects of the silicon). Also, what with many digital and analogue electronics sensitive to noise in the 10-100kHz, it may not be commercially viable to produce 'efficient sychronous DC/DC' chips, what with the expense of adding the sync switch and its control method, that might prove harder to filter than one jsut as easily produced that works at a higher frequency.

One more thought (I'm just thinking out loud here), the addition of a sychronous switch - or its control - in DC/DC controllers is a fairly recent thing, and may just coincide with better manufacturing that allows for higher switching frequencies at high efficiency. Not directly related, but both results of better technology that hit the market at a similar time.

 

[bountyhunter]

Since switchers were first used (with vacuum tubes BTW) the drive has always been to push the frequency as high as possible to reduce the size and cost of the filter components (caps and inductors).

The down side of higher freqs is increased switching losses in the magnetics, FETs, caps, and also: much worse EMI (a really nasty problem).

Also: going above about 300 kHz FORCES the use of either ceramic/film caps because alums and tantalums are no good up there. The higher sw freqs commonly used today are a direct result of improvements in FETs and caps.

Engineering is all about tradeoffs.

 

[Dean Huster]

Synchronous rectification implies that you're interested in high efficiency and typically, higher frequencies means better efficiency, including weight and size savings. What's the point in a synchronous rectifier if you're going to drive a 60 Hz transformer (assuming you're looking at line isolation as well)?

 

[bountyhunter]

Synchronous rectification implies that you're interested in high efficiency and typically, higher frequencies means better efficiency

Actually, it's a lot harder to keep the power conversion efficiency high as you go up in frequency because losses go up dramatically.

What's the point in a synchronous rectifier if you're going to drive a 60 Hz transformer (assuming you're looking at line isolation as well)?

Sync rec has it's greatest advantage at low output voltages where the typical "free wheeling diode" would be on a lot, giving a VBE frward drop. The sync rec substitutes a FET for the diode and drops the power dissipation in that element. Biggest power savings in conversions where VIN/VOUT ratio is high.

 

[ronsimpson]

PWM usually run faster than 20khz so the noise can not be heard.
The higher the speed the smaller the coils, transformers and capacitors.

I have seen synchronous rectification used at 60hz but not often. It is used with low voltage and high current. Recently I made a 10 amp 1.1 volt supply for a FPGA. With a 1.1 volt supply you do not want to waste 0.7 volts in a diode.

 

[simonbramble]

If you want to look at the maths behind dc/dc converters (it is quite simple) have a look at this:
http://www.simonbramble.co.uk/dc_dc_converter_design/dc_dc_converter_design.htm

This will explain that as the frequency goes up, the inductor gets smaller and you can get away with using a lower value output cap (which is usually smaller). It is similar to filling up a pool with a bucket of water. you can use a smaller bucket with less water if you top the bucket up more often.