[sdiy] Powerline LC filters etc.

ASSI Stromeko at nexgo.de
Wed Feb 29 22:57:39 CET 2012

On Wednesday 29 February 2012, 16:00:23, Tom Bugs wrote:
> For the standalones - 9v DC operated boxes powered by wall-warts - now
> that many wall-warts are switchers.
> - but I have a feeling that audio noise I was experiencing is actually
> down to my circuit approach rather than due to power-line born noise.
> (Douglas Self's book 'Small Signal Audio Design' is useful for these
> areas - very good readability too)

I have yet to see a low-wattage wall-wart that has decent RF filtering (or 
any at all) and most of them also have rather high ripple.  I'm not saying 
better ones don't exist, I've just never come across any.  If you've got 
some non-linearity around in your circuit (like a rectifier), you could very 
well be demodulating the RF and have all sorts of unwanted sound-effects 
going on.

So it would be very wise to filter their output and the BNXxxx ones you've 
mentioned look to be good candidates.  The schematic is not very clear, but 
the only way this makes sense to me is if the two inductors closest to the 
source are in fact a common-mode (current compensated) choke and the rest of 
the filter is a pi section for dealing with the RF.  This gives a pretty 
good filter provided that the switching frequency of the source is around 
2MHz, which is what they seem to be optimized for.  It could very well be 
much lower in reality, which would then require different (and more bulky) 
inductive components.  The best place to put this filter would actually be 
directly at the output of the switcher and then follow with a twisted pair 
to the input to your circuit, where you'd put a decoupling / buffer cap 

> http://bugbrand.co.uk/images/PowerPlan.jpg - first half is a module with
> DC-DC converters which then has output headers to go to distro boards
> which each have LDO Linear VRegs
> - external wall-wart supplying 24V DC
> - ? input DC filter ? Murata BNX016-01
> - two 20W DC-DC converters (Cincon) with trims to take the output
> voltage up to c.15.5V

Doing a cascade of AC/DC -- DC/DC is not very efficient at those close 
voltages, but it may be unavoidable if you need to use a wall wart.

Pairing up two DC/DC in tandem has some (probably unintended) consequences: 
as Harry already mentioned, if they are in close proximity they might beat 
against or lock onto each other, which would produce wholly unexpected 
voltage variations.  Even if they don't do that, you must consider that the 
regulation on either side is completely independent of the other and large 
asymmetric load transients might provoke excursions on the other side that 
the LDO might have a hard time catching.  This depends a lot on what modules 
you will connect and how symmetric the load is in the first place, of 

> - output DC filters - note that the negative side one is connected the
> other way around as detailed in the Murata datasheet
> - system ground banana connection comes post DC filters
> - headers to take the +/- 15.5V to the distros
> - distros then have a pair of LDO Linear V-Regs (a positive and a
> negative type) to regulate down to +/-15V

You should probably give the regulators a bit more headroom and just use a 
16V or 17V output from the DC/DC, in any case trimming the output is not 
something I'd want to do.

> Does this sound like a sound approach? I'm keen to try to make DC-DC
> converters work, but have had my doubts so far.. They would bring many
> benefits (compactness, worldwide voltages, no mains worries, etc) but
> obviously I don't want to be opening a can of worms.

Before settling on that solution, ask yourself again if not dealing with 
mains in the cabinet is worth the added trouble with the wall-wart.  Even if 
you supply a good one, inevitably someone else will decide that they plug 
their super-noisy model into your modular and it's still your problem.  All 
things considered I'd personally prefer a filtered IEC inlet connected to a 
good closed-frame dual-output supply.  Another possible solution might be 
wall-wart into a really good filter and then point-of-load generation of 
voltages, i.e. instead of having a beefy +-15V supply you'd go for many 
small ones that can just supply a single module or a handful.  If the input 
range of the POL converter covers 9-40V (or at least 20V-36V), you can just 
plug them into about any power distribution you might encounter in a modular 
with a simple adapter.

+<[Q+ Matrix-12 WAVE#46+305 Neuron microQkb Andromeda XTk Blofeld]>+

Wavetables for the Waldorf Blofeld:

More information about the Synth-diy mailing list