[David Gravereaux]

Hi,

Was wondering about papers/articles written about using switchmode power
supplies and class-d amps and the proper line filtering required. I
don't mean FCC requirements above 150kHz, just stuff about neutral
abuse, cross-talk effects and stuff specific to use in audio systems.

[Chris Hornbeck]

On Tue, 19 May 2009 20:35:32 -0700, David Gravereaux
wrote:

Was wondering about papers/articles written about using switchmode power
supplies and class-d amps and the proper line filtering required. I
don't mean FCC requirements above 150kHz, just stuff about neutral
abuse, cross-talk effects and stuff specific to use in audio systems.

Your post is sophisticated and sufficiently self-contained to remove
any generic responses, so you may need to be more specific. IOW, what
's the gig? It's an amazingly broad topic, with lots of answers from
lots of perspectives, some of which aren't very pretty. Technical,
legal, moral, and it gets worse... arf. But God dwells in the details.

If I were to say: don't contaminate the grounds, even on a truly
balanced interconnection, you might well ask why. Like that.

Simple answers to complex questions are just BS. Your question
is sufficiently complicated to require a complicated answer.


All the best fortune,
Chris Hornbeck

[David Gravereaux]

Chris Hornbeck wrote:
On Tue, 19 May 2009 20:35:32 -0700, David Gravereaux
wrote:

Was wondering about papers/articles written about using switchmode power
supplies and class-d amps and the proper line filtering required. I
don't mean FCC requirements above 150kHz, just stuff about neutral
abuse, cross-talk effects and stuff specific to use in audio systems.

Your post is sophisticated and sufficiently self-contained to remove
any generic responses, so you may need to be more specific. IOW, what
's the gig? It's an amazingly broad topic, with lots of answers from
lots of perspectives, some of which aren't very pretty. Technical,
legal, moral, and it gets worse... arf. But God dwells in the details.

If I were to say: don't contaminate the grounds, even on a truly
balanced interconnection, you might well ask why. Like that.

Simple answers to complex questions are just BS. Your question
is sufficiently complicated to require a complicated answer.


Ok, was afraid of that 10 units capable of 500W peak output power.
Probably all on two 20amp circuits. The amps themselves use a variable
switching frequency. The supply runs at 45kHz. Next chance I get I
want to monitor instant line input current vs. supply output current vs.
driver (load) current just to see how literal things really are.

[Scott Dorsey]

David Gravereaux wrote:

Was wondering about papers/articles written about using switchmode power
supplies and class-d amps and the proper line filtering required. I
don't mean FCC requirements above 150kHz, just stuff about neutral
abuse, cross-talk effects and stuff specific to use in audio systems.

Well, my first thought is to say that switching supplies are bad news in
audio systems in part because it is impossible to really clean up the inputs
and outputs completely.

But.... one advantage of the switching supply is that it is pretty immune
to incoming noise on the supply lines... so your job becomes only to try
and make the power factor nice and prevent it from dumping too much trash
on the incoming line.

Trying to make the input really clean is futile anyway, because someone is
always going to be plugging in some crappy laptop whose power supply has been
carefully engineered to just barely meet FCC requirements at the lowest cost
possible, on the same circuit.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[Scott Dorsey]

David Gravereaux wrote:

Ok, was afraid of that 10 units capable of 500W peak output power.
Probably all on two 20amp circuits. The amps themselves use a variable
switching frequency. The supply runs at 45kHz. Next chance I get I
want to monitor instant line input current vs. supply output current vs.
driver (load) current just to see how literal things really are.

Don't just do that, look at current and voltage waveforms. The current
waveform will get _really_ weird-looking as the load on the supply changes,
with most switching controllers.
--scott
--
"C'est un Nagra. C'est suisse, et tres, tres precis."

[David Gravereaux]

Thanks guys. I'll let you know what I find.

[David Gravereaux]

Scott Dorsey wrote:
....
Well, my first thought is to say that switching supplies are bad news in
audio systems in part because it is impossible to really clean up the inputs
and outputs completely.
....

Time to bump this old thread.

First run of the DSP and what we call the integration board (iboard)
didn't turn out well. The noise floor of the DACs on the DSP board
contained clock hash at 370Hz and all harmonics related to it. 370Hz is
the bulk processing rate of the DSP when being fed a 24b/96k source.

http://www.facebook.com/photo.php?pid=500864&id=1045841527

This is the test jig I made for testing the electronics outside the speaker.

http://www.facebook.com/photo.php?pid=434242&id=1045841527

This is the noise floor of the amps themselves with a shorted input

http://www.facebook.com/photo.php?pid=500865&id=1045841527

The reason for the hash noise is pretty simple:
V = I * R (Ohms law)

Any current drawn across a resistance (the traces) results in a voltage
drop. The digital grounds and analog ground on the DSP are common, thus
the processor's current activity is modulating the analog reference
locally... and there we have it.

With all the DC-DC converters in here, there's no clean ground. The
layout for the iboard has no star grounding point. The current carrying
return for the +7VDC for the DSP is also the ground reference. The DACs
and the volume control chip (PGA4311) separate the digital and analog
side references within the chips themselves, but the guy that did the
layout missed the opportunity to use it. He just connected all AGND
pins to the DGND ones.

I sent him a large change set and also a change set to the person who
did the iboard for how to take advantage of star-grounding. Both don't
believe me and nor does my boss. The DSP guy is blaming the iboard for
"not having enough current supply". What a load of BS.

I offered to redo the layout for both boards as this concept isn't new
to me at all. I've seen this many times before. I'm glad I know that
I'm working with a bunch of failures now who drink each other's kool-aid.