switchers and amps = bad?

[Michael (Micksa) Slade]

While at the local parts shop doing my routine skulk around I asked about
making an RIAA preamp and powering it with a switched (single-rail)
supply, and one of the guys there went "nonono, bad idea, switchers are
too noisy for audio applications".

Is this true? I'm no veteran but as far as I know any supply noise should
be able to be filtered out with sufficiently large/numerous capacitors,
should it not?

Mick.

 

[Frank Bemelman]

While at the local parts shop doing my routine skulk around I asked about
making an RIAA preamp and powering it with a switched (single-rail)
supply, and one of the guys there went "nonono, bad idea, switchers are
too noisy for audio applications".

Is this true? I'm no veteran but as far as I know any supply noise should
be able to be filtered out with sufficiently large/numerous capacitors,
should it not?

There are plenty audio amplifiers using switched power supplies, but
a RIAA preamp does not need much power and using a switcher seems a
lot of hassle. I'd go for a simple linear regulator.

 

[Dave VanHorn]

While at the local parts shop doing my routine skulk around I asked about
making an RIAA preamp and powering it with a switched (single-rail)
supply, and one of the guys there went "nonono, bad idea, switchers are
too noisy for audio applications".

Is this true? I'm no veteran but as far as I know any supply noise should
be able to be filtered out with sufficiently large/numerous capacitors,
should it not?

True to some degree.. It's possible to make a REALLY noisy switcher, but
it's also possible to make a really quiet one. I did an application a while
back, reading the signal from a MICR head (microvolts) with an amplifier
powered by a switcher, and sitting next to the controller for a 24W stepper
motor with chopped current regulation.. No problem at all. It just took
careful attention to design.

 

[Mark Jones]

True to some degree.. It's possible to make a REALLY noisy switcher, but
it's also possible to make a really quiet one. I did an application a while
back, reading the signal from a MICR head (microvolts) with an amplifier
powered by a switcher, and sitting next to the controller for a 24W stepper
motor with chopped current regulation.. No problem at all. It just took
careful attention to design.

Thanks Dave. Read: switchers can make all kinds of nasty
electro-magnetic interference, from the high frequencies used in
switching to the magnetic fields used in the inductors. I agree that a
RIAA preamp should use very little power, there is no need to use a
switcher.

Still, if you must use a switcher, you can get really clean power out
of it (at small currents) by using four RC lowpass filters. i.e.,


[in]---\/\/\--+--\/\/\--+--\/\/\--+--\/\/\--+---[out]
R1 | R2 | R3 | R4 |
+ | + | + | + |
--- C1 --- C2 --- C3 --- C4
--- --- --- ---
| | | |
| | | |
--- --- --- ---
- - - -

You'd want to select the R/C values to reject the lowest frequencies
present in the source. You could use say 1uF/3.75K and it will block
out almost all frequencies, but the voltage risetime will suffer as
each stage slowly charges up. And (3.75k x 4 = 15k) of DC impedance
would leave a +15v supply able to provide only 1mA. But this might be
acceptable for an RIAA preamp.


-- "The problem with participating in a rat-race is that everyone is a
rat." MCJ 200312

 

[[email protected]]

I would tend to agree with the guy at the parts shop. Discrete changes
in voltage or current generally spell trouble with line and radiated
noise that can couple onto your signal unless great care is taken. Why
not make it easy on your self an build or pick up a linear supply.

 

[Dave VanHorn]

Still, if you must use a switcher, you can get really clean power out
of it (at small currents) by using four RC lowpass filters. i.e.,


[in]---\/\/\--+--\/\/\--+--\/\/\--+--\/\/\--+---[out]
R1 | R2 | R3 | R4 |
+ | + | + | + |
--- C1 --- C2 --- C3 --- C4
--- --- --- ---
| | | |
| | | |
--- --- --- ---
- - - -

Funny thing is, at low currents, the Rs are more effecitve than inductors,
especially if you consider the cost.

You'd want to select the R/C values to reject the lowest frequencies
present in the source. You could use say 1uF/3.75K and it will block
out almost all frequencies, but the voltage risetime will suffer as
each stage slowly charges up. And (3.75k x 4 = 15k) of DC impedance
would leave a +15v supply able to provide only 1mA. But this might be
acceptable for an RIAA preamp.

Or scale the Rs down by 10, and the caps up, and add a 0.1 on the input
side.


On most of my designs, where I have analog and microcontrollers together, I
use a single stage RC on the analog supply, right at the chip. As long as
you watch the current demands, it works nicely.

 

[Pooh Bear]

While at the local parts shop doing my routine skulk around I asked about
making an RIAA preamp and powering it with a switched (single-rail)
supply, and one of the guys there went "nonono, bad idea, switchers are
too noisy for audio applications".

Is this true? I'm no veteran but as far as I know any supply noise should
be able to be filtered out with sufficiently large/numerous capacitors,
should it not?

Don't forget that switchers generate plenty of HF magnetic field too.

It concerns me more that in today's world you might be making an RIAA pre-amp
( any pre-amp in fact ) using a single supply rail.

The switcher I've just designed for a semi-pro audio mixer generates a mere
couple of hundred microvolts audio band noise on the supply rails btw.
Slightly quieter than the previous linear regulator based design but I did
apply extra filtering compared to a stock swicher. The magnetic field still
needs a small amount of taming - but you'd get even worse problems with a line
frequency transformer when the PSU is mere centrimetres from the pcb.

A line frequency supply is a lot easier to design though.


Graham