High-load SMPS for a recording console

[Jasnevo]

I'm currently re-vitalising an analog recording desk and during process I decided to replace the linear power supplies with switch-mode ones. Based on otherwise good experience, I turned to Meanwell first.

Here are the power requirements:
1) Audio circuits: +16V/35A, -16V/35A
2) Motorized faders +12V/50A
3) Control circuits: +5V/30A

I decided for HRP-600-15 (two in series for the +/- 16V), HRP-600-12 and HRP-300-5.
https://www.meanwell.com/webapp/product/search.aspx?prod=HRP-300
https://www.meanwell.com/webapp/product/search.aspx?prod=HRP-600
For the 48V phantom power I’ll use the existing linear supply.

The power supplies are remote and require to be wired to the console bridging a distance of ~8m. My plan is to have a central GND for all supplies (as the individual voltages go on a central GND copper rail anyway). In order to have a constant supply voltage, I'm going to connect the sense input from the power supply directly to the relevant power rail within the desk.

So far so good. While I don’t very much care for SMPS noise and ripple with regard to the fader motors and the control circuits, I do care for the audio circuits. Since my knowledge is limited in that regard, I first had a very naive idea: a simple RC low-pass filter with the 8m cable as the R and a suitable C for a cut-off frequency of about 30kHz. – assuming the cables R=0,01 Ohm, the C accounts to ~530µF – at 16V not hard to accomplish. In order to care for the typical spikes at around 10-100MHz I’d place some ceramics in the pF range right at the incoming DC lines at the desk.

Why I’m here… just for absolution? ) Nope, I came across a second approach that was heavily commented from members of this forum: a LC filter design using two common-mode chokes and parallel C’s for a dual output supply – so exactly what I’m going for. I was skeptical because of the 35+A going through the L but I just “mousered” and “reichelted” any there are suitable chokes for that amount of current available.

Does anyone here who has some experience with filter designs under such load? There are still lots of discussions SMPS vs. linear supplies but I’ve seen (and heard, of course) top-class amplifiers with SMPS and didn’t hear a thing without an input signal and full gain – so it’s obviously doable.

Any thoughts and comments are much appreciated – and don’t bother with critics, I’m neither fragile nor resentful!

Christian.

[mediatechnology]

Christian - Thanks for joining us!

I guess the question that comes to mind is whether or not the common mode chokes will saturate with a DC load.
I haven't looked at the data sheets yet.
Large CM chokes usually are for switcher primary circuits, and with most switchers being designed for either AC or DC input I would think they would.
Does the datasheet derate for DC?

[Jasnevo]

Good point - the DC-Bias was another uncertainty of mine, besides the current!

The data sheet doesn't mention any derating, however, in the usual introductory blah-blah they mention suitability till 600VAC and 1000VDC.
https://uk.rs-online.com/web/p/common-m ... es/1705165

Christian.

[mediatechnology]

I just read this:

...but they are as well applicable in DC power lines of photovoltaic installations or similar applications up to 1000 VDC.

Being common mode, and thinking about it further, it would seem that the DC fields would cancel so that may be why there is no derating. On the bipolar rails the CM choke would see the difference in currents which is low relative to the load.

I noticed something recently on the vintage Amber 3500 test set power supply which made perfect sense. There were LC filters with inductors in the supply and in the ground return to make an LCL filter.

Don't know how you were going to configure it but for the bipolar output would 2 CM chokes be appropriate with their 0V tied after the CM filter?

The logic and illumination supplies seem pretty straightforward.

[Jasnevo]

I opened I case with Meanwell (hope dies last) in order to get their opinion with regard to my approach. If haven't heard anything back by the end of this week, I'll go with the two HRP-600-15 and a double common-mode choke with caps to GND - so a typical LC filter by the theory. That's unfortunately the only thing I can go with because of my lack of experience. I'll measure out different scenarios to get a better picture.

I'll keep you posted.

[mediatechnology]

Take a look at this filter board from MPower Systems



There appears to be relays for interlocking followed by a C and two cascaded LC filters followed by two more load control relays.
The two pairs (per rail) of common mode chokes may be series-connected to increase value.