[sdiy] Switched-mode power supply module in synths... good or bad?

[Mattias Rickardsson]

Thanks for the many very good advice & answers to this question.
SMPSU is definitely an everyday technique by now, but in my case I
guess the potential (sic!) problems would come from the fact that the
SMPSU wasn't designed there from the start. I'll have to try it out I
guess. Twitching supply next to the mic preamp, this is gonna get
interesting... :-)

/mr  -  bzzzt!

On 26 January 2012 01:55, Veronica Merryfield <veronica at merryfield.ca> wrote:
> Thought I'd add my experience to the mix.
>
> For the last 2 years I have been working on contracts for pro-audio gear, mostly electronic percussion, and have a number of products powered by cheap wall wart SMPSU. FCC regs, as far as I can tell, don't do much below about 30Mhz or so, so one expect a ton of emissions from them. I sampled plenty and none were that great. In all cases the output noise and regulation was well outside of the spec but my clients did not want to bring the power supply inside the units (cost, time, FCC approval) and did not want to spend a large amount on better units.
>
> Each of the projects required multiple power rails requiring more power supplies on the boards. Typically these were a +Ve/-Ve pair for audio and a logic rail, although in one design I had 3 pairs of +Ve/-Ve rails and a couple of logic rails. I used SMPSUs for the bigger current rails and linear regs for the low current rails.
>
> The input noise had to be suppressed with inline inductors and capacitors. Taking a decent look at the noise on a scope and tinkering allowed me to use a mix of capacitors but basically 3 across three decades 0.5uF and up did the trick.
>
> The on board SMPSUs were laid out on double sided boards (cost driven) and by a careful choice of thru hole and SMT parts, I managed to have decent ground planes top and bottom. Part choice and location is key - it is worth putting money into low ESR caps for instance. In one incarnation, a mod wire was needed and depending on how this was fitted, the units were quiet or exceptionally noisy. There can be a lot of radiated energy and the current flow can be high. Grounding and path (route, size) are important. Locating the SMPSUs at one end of the board is extremely useful and being able to separate the PSUs from the audio with ground traces also helps. The audio nearest the PSUs in these cases is the low impedance part (output drivers, midi etc).
>
> Long and short, the units are quiet. In fact, the problem now is the delta-sigma DAC noise - this is on a unit that is bought in. The supplier did not think I would be able to produce a quiet enough audio and PSU board using two layers. Whilst 4 layers or more would certain make life easier, it is not the panacea.
>
> So, if the OP is thinking of a mix, buy in a wall wart and make up the other rails, it is possible and it can be very quiet, but it is not a simple proposition.
>
> For what it is worth, my on board SMPSUs include a mix or buck, flyback, continuos and discontinuous mode supplies, 120Khz to over 1Mhz, and a mix of manufacturers controllers. Discontinuous mode switchers ad noise way below their switching speed but sometime one has no real choice but to use them.
>
> In short - passive filtering, ground plane, proper attention to layout, and they are clean. Linear regs will not clean up a dirty rail. Current spikes can be huge therefore think about radiated emissions.
>
> Veronica
>
> _______________________________________________
> Synth-diy mailing list
> Synth-diy at dropmix.xs4all.nl
> http://dropmix.xs4all.nl/mailman/listinfo/synth-diy