There's more to consider than just switching frequency. Like the mode of control. Is it PWM or variable frequency? Hysteretic? Note the feedback control loop is right in the audio band. Having a switching frequency of 1MHz or some other high number guarantees nothing. You have to design the control loop for minial impact to audio.
yes true. Most modern DC-DC converters are now PWM driven though. Not all designs need feedback at all. a normal PWM output can be tuned to give the desired voltage/current without regard to actual load up to the point of pulling the whole system down but I'm not sure where you see that the feedback is in the audio band, the feedback is generally current sense via resistor drop, essensially sensing voltage level, not frequency. If the output is filtered enough then there should be no ripple on the feedback at all, but if there is then you might get a type of oscillation.
Other aspects that help are using proper ferrite core inductors or transformers. With shielding. Winding your own will likely lead to a lot of magnetic spray (to other circuits).
very true, this is why computer PSUs are in their own cases separate from the motherboard components.. BUT if you have enough distance(deemed by the switching freq, mag field and other such things) then you can be unshielded. computer powersupplies are also operating under huge loads and power densities. we are talking about less than 200ma here. ferrite cores were mentioned in an earlier post of mine in this thread.
Of course it can be done, and wouldn't cost too much. But it takes more than what you see in a typical dc-dc application note. Extra filtering on both input and output will help. You could even do a hybrid: dc-dc step-up folowed by a linear regulator.
this would be a cheap AND easy solution. App notes/datasheets are for the circuit in question, not supporting circuitry usually. so yes you are correct in stating that there is more to it, but this *should* be obvious to those interested even though it seems like it's more complicated. to me it's almost a matter of understanding what you are getting into. I could read anatomy books all day long and understand what they say but does that make me think i can go do surgery? nope. but on the other hand, we could easily overdo the situation with excessive design. do we really need all those extra Ls, Cs, and Rs on the input and output? likely not. In DIY this is not much concern, but in the real world this is cost and possible failure modes.
People seem to assume that because it's switching that there will be noise on the output for some reason. well yes there will be noise on a improper design but if you follow the data and do a little O'scoping you can make a switcher output like a linear reg and save some heat too.
we have to think about what's happening in the world too.. china has produced a billion different types of switchers and a lot of them are crap. this has flooded the market with noisy, poorly efficient, poorly regulating powersupplies that are fine for digital devices, but really suck for anything analog. this has given switchers a bad name and the stigma seems to have stuck very well even for those who understand the technology and especially for those who don't.
I'll get some ICs and drop a small board and see what I get in real life testing, but it might take me a few weeks to get it together with my schedule.
:thumb:
PS: thanks for the discussion!
