how to "clean up" 5V power supply?

[Michael]

Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
that I'm working on - as the onboard sound card is
horribly noisy. But I'd really need to use the onboard
5V supply as it would see way too much usage to have to
constantly be putting in new batteries.

So how difficult would it be to clean up the supply
signal? Now if it matters - the power supply for this
is a 5V,5A brick, about 1"x2"x4", Delta Electronics
ADP-25EB (though I can't seem to find a datasheet for
it)

Should I expect this to be a fairly noisy supply?



Well thanks for anyone's help with this.

-Michael

 

[Ian Stirling]

Michael <nleahcimathotmaildotcom@nowhere.com> wrote:

Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
that I'm working on - as the onboard sound card is
horribly noisy. But I'd really need to use the onboard

Do you care especially about price?

The simple way would be a 1W or so 5V-12V converter, a 7805 linear
regulator, and do it that way.

 

[Michael]

Ian Stirling <root@mauve.demon.co.uk> wrote in news:3x9qc.4463$wI4.513793
@wards.force9.net:

Michael <nleahcimathotmaildotcom@nowhere.com> wrote:
Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
that I'm working on - as the onboard sound card is
horribly noisy. But I'd really need to use the onboard

Do you care especially about price?

The simple way would be a 1W or so 5V-12V converter, a 7805 linear
regulator, and do it that way.

Price - not so much. But size and power consumption are very important. I
have next to no space to work with (hence why I'm so interested in this
teeny USB soundcard!) and I don't know exactly how much power I have to
work with - but not much at all! So I would really like as efficient of a
circuit as possible.

So with that circuit wouldn't I be seeing about 40% efficiency at best?

 

[Tam/WB2TT]

"Michael" <nleahcimathotmaildotcom@nowhere.com> wrote in message
news:Xns94EC96C2CBC8Fnleahcim@63.240.76.16...

Ian Stirling <root@mauve.demon.co.uk> wrote in news:3x9qc.4463$wI4.513793
@wards.force9.net:

Michael <nleahcimathotmaildotcom@nowhere.com> wrote:
Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
that I'm working on - as the onboard sound card is
horribly noisy. But I'd really need to use the onboard

Do you care especially about price?

The simple way would be a 1W or so 5V-12V converter, a 7805 linear
regulator, and do it that way.


Price - not so much. But size and power consumption are very important. I
have next to no space to work with (hence why I'm so interested in this
teeny USB soundcard!) and I don't know exactly how much power I have to
work with - but not much at all! So I would really like as efficient of a
circuit as possible.

So with that circuit wouldn't I be seeing about 40% efficiency at best?

Michael,
Since the thing runs off batteries, it probably requires something other
than 5V. If it takes 2 AAs, you could just use an LDO 3.3V regulator.

Tam

 

[Terry Given]

"Michael" <nleahcimathotmaildotcom@nowhere.com> wrote in message
news:Xns94EC86F762DEAnleahcim@216.148.227.77...

Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
that I'm working on - as the onboard sound card is
horribly noisy. But I'd really need to use the onboard
5V supply as it would see way too much usage to have to
constantly be putting in new batteries.

So how difficult would it be to clean up the supply
signal? Now if it matters - the power supply for this
is a 5V,5A brick, about 1"x2"x4", Delta Electronics
ADP-25EB (though I can't seem to find a datasheet for
it)

Should I expect this to be a fairly noisy supply?



Well thanks for anyone's help with this.

-Michael

Hi Michael,

dc-dc bricks tend to be dreadfully noisy, the extent of which is governed by
the bypassing/filtering that YOU add. dc-dc's usually have little or no
filtering (it makes the brick cheaper), and at the very least need good
bypass caps (ceramic||low ESR tant/electrolytic). caveat emptor

Terry

 

[Ian Stirling]

Michael <nleahcimathotmaildotcom@nowhere.com> wrote:

Ian Stirling <root@mauve.demon.co.uk> wrote in news:3x9qc.4463$wI4.513793
@wards.force9.net:

Michael <nleahcimathotmaildotcom@nowhere.com> wrote:
Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
that I'm working on - as the onboard sound card is
horribly noisy. But I'd really need to use the onboard

Do you care especially about price?

The simple way would be a 1W or so 5V-12V converter, a 7805 linear
regulator, and do it that way.


Price - not so much. But size and power consumption are very important. I
have next to no space to work with (hence why I'm so interested in this
teeny USB soundcard!) and I don't know exactly how much power I have to
work with - but not much at all! So I would really like as efficient of a
circuit as possible.

So with that circuit wouldn't I be seeing about 40% efficiency at best?

Yes.
Hmm, you can get 5-6V small converters, which would probably be the
easy way.
A 5V-6V converter (if the device does not use lower than 5V internally),
followed by a 5V low-dropout regulator.
Should be around 70-80% efficient, and filtering becomes easy.
Total should be under $15 or so.
Filtering would be the better way, but may require large inductors/capacitors.

 

[Ben Bradley]

On Mon, 17 May 2004 21:49:12 GMT, Michael
<nleahcimathotmaildotcom@nowhere.com> wrote:

Ian Stirling <root@mauve.demon.co.uk> wrote in news:3x9qc.4463$wI4.513793
@wards.force9.net:

Michael <nleahcimathotmaildotcom@nowhere.com> wrote:
Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
that I'm working on - as the onboard sound card is
horribly noisy. But I'd really need to use the onboard

Do you care especially about price?

The simple way would be a 1W or so 5V-12V converter, a 7805 linear
regulator, and do it that way.

That IS way inefficient. The other suggestion of 5V to 6V converter
and a 5V LDO regulatur would be much better.

Price - not so much. But size and power consumption are very important. I
have next to no space to work with (hence why I'm so interested in this
teeny USB soundcard!) and I don't know exactly how much power I have to
work with - but not much at all! So I would really like as efficient of a
circuit as possible.

So with that circuit wouldn't I be seeing about 40% efficiency at best?

ISTR (from reading here) that USB has a max current spec of 100mA,
so you don't have a lot to work with. OTOH, if everything is
reasonably efficient, there's still enough power available for most
headphones to cause ear damage.

What's AT LEAST as important as filtering and/or linear regulators
is adequately large grounding and running the signal and ground
connections to appropriate places, so that (for onee example) the
switching power supply return current doesn't show up in the ground
connection of the headphones.

From your original message:

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply.

There could have been several reasons they did this. Batteries are
definitely a lot quieter, and they may have just not put enough design
time into filtering and such.

I'm very interested in this circuit for a SBC

"SBC" took me a moment - Single Board Computer.

that I'm working on - as the onboard sound card is
horribly noisy. But I'd really need to use the onboard
5V supply as it would see way too much usage to have to
constantly be putting in new batteries.

So how difficult would it be to clean up the supply
signal?

It may be diffucult and time-consuming if you've not done something
like this before, but if you have time to put into it and especially
if you plan on making a quantity of these, it's the way I'd go. Read
through the articles on this page - they are about grounding-related
hum and buzz in interconnecting equipment, but the ideas should help
you with connecting different stages to a 'bad' noisy power supply.

http://www.jensen-transformers.com/apps_wp.html

Now if it matters - the power supply for this
is a 5V,5A brick, about 1"x2"x4", Delta Electronics
ADP-25EB (though I can't seem to find a datasheet for
it)

Knowing it's a "switching supply for a PC" is enough to see the
magnitude of the problem.

-----
http://mindspring.com/~benbradley

 

[Ian Stirling]

Ben Bradley <ben_nospam_bradley@mindspring.com> wrote:

On Mon, 17 May 2004 21:49:12 GMT, Michael
nleahcimathotmaildotcom@nowhere.com> wrote:

Ian Stirling <root@mauve.demon.co.uk> wrote in news:3x9qc.4463$wI4.513793
@wards.force9.net:

Michael <nleahcimathotmaildotcom@nowhere.com> wrote:
Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
snip
The simple way would be a 1W or so 5V-12V converter, a 7805 linear
regulator, and do it that way.

That IS way inefficient. The other suggestion of 5V to 6V converter
and a 5V LDO regulatur would be much better.

<snip>

ISTR (from reading here) that USB has a max current spec of 100mA,
so you don't have a lot to work with. OTOH, if everything is
reasonably efficient, there's still enough power available for most
headphones to cause ear damage.

500ma.

 

[Spehro Pefhany]

On Tue, 18 May 2004 19:38:37 GMT, the renowned Ian Stirling
<root@mauve.demon.co.uk> wrote:

Ben Bradley <ben_nospam_bradley@mindspring.com> wrote:
On Mon, 17 May 2004 21:49:12 GMT, Michael
nleahcimathotmaildotcom@nowhere.com> wrote:

Ian Stirling <root@mauve.demon.co.uk> wrote in news:3x9qc.4463$wI4.513793
@wards.force9.net:

Michael <nleahcimathotmaildotcom@nowhere.com> wrote:
Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
snip
The simple way would be a 1W or so 5V-12V converter, a 7805 linear
regulator, and do it that way.

That IS way inefficient. The other suggestion of 5V to 6V converter
and a 5V LDO regulatur would be much better.


snip
ISTR (from reading here) that USB has a max current spec of 100mA,
so you don't have a lot to work with. OTOH, if everything is
reasonably efficient, there's still enough power available for most
headphones to cause ear damage.

500ma.

One "unit load" is 100mA. Quoting from the spec:

---
A device may be either low-power at one unit load or high-power,
consuming up to five unit loads. All devices default to low-power. The
transition to high-power is under software control. It is the
responsibility of software to ensure adequate power is available
before allowing devices to consume high-power.
---

You may not have 500mA available from a laptop, but you will have at
least 100mA.

Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[Tom Bruhns]

If you pay attention to layout as Ben suggested, that's a good part of
the battle. Then you may be able to use a power-supply cleaning
circuit, instead of full-blown re-regulating, to get your power clean
enough. The idea is to filter out the "bad stuff." You can do this
with passive components, but they tend to get really big at low
frequencies. Bob Pease has had some articles on cleaning up an old
tube-type power supply using active filtering, and Charles Wenzel has
a really nice circuit idea on
http://www.techlib.com/electronics/finesse.html. There's a
significant difference between the two: Bob's circuit provides always
a very low output impedance, but Charles' adds a resistance in series
with the power supply. It can be a small resistance and still be
practical, but it will have a drop that depends on current supplied to
your circuit. In other words, it's best for things like RF
oscillators that draw essentially constant current. (You want clean
power for very low phase noise...) A key concept here is that your
power supply may be several millivolts (or even tens of millivolts)
noisy, and that's bad in a sensitive audio circuit, but there's not
actually so much _power_ involved in that noise, and it doesn't take
very much power to cancel it out. I've done Wenzel's circuit using an
op amp and 1% resistors to get 40dB noise cancellation. You could
tweak it to get even better.

Note also that if you have room for enough capacitance and inductance,
you can do a pretty effective job cleaning things passively. If your
"brick" has noise mainly at switching frequencies, that's not so
difficult to get rid of, so long as it's not magnetically coupled into
your circuit. But if it's at 100-120Hz and harmonics thereof, you'll
need L*C in the neighborhood of 0.001 henries*farads or more to be
effective. The L you have to use in such a filter may have more R
than the R you'd use in Wenzel's circuit!

Another way to attack this is to use power amplifiers with very good
power supply rejection, and put the main effort into
filtering/cleaning the relatively low current going to the lower-level
components.

Cheers,
Tom



Michael <nleahcimathotmaildotcom@nowhere.com> wrote in message news:<Xns94EC86F762DEAnleahcim@216.148.227.77>...

Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
that I'm working on - as the onboard sound card is
horribly noisy. But I'd really need to use the onboard
5V supply as it would see way too much usage to have to
constantly be putting in new batteries.

So how difficult would it be to clean up the supply
signal? Now if it matters - the power supply for this
is a 5V,5A brick, about 1"x2"x4", Delta Electronics
ADP-25EB (though I can't seem to find a datasheet for
it)

Should I expect this to be a fairly noisy supply?



Well thanks for anyone's help with this.

-Michael

 

[Michael]

k7itm@aol.com (Tom Bruhns) wrote in
news:3200347.0405181535.e885e65@posting.google.com:

If you pay attention to layout as Ben suggested, that's a good part of
the battle. Then you may be able to use a power-supply cleaning
circuit, instead of full-blown re-regulating, to get your power clean
enough. The idea is to filter out the "bad stuff." You can do this
with passive components, but they tend to get really big at low
frequencies. Bob Pease has had some articles on cleaning up an old
tube-type power supply using active filtering, and Charles Wenzel has
a really nice circuit idea on
http://www.techlib.com/electronics/finesse.html. There's a
significant difference between the two: Bob's circuit provides always
a very low output impedance, but Charles' adds a resistance in series
with the power supply. It can be a small resistance and still be
practical, but it will have a drop that depends on current supplied to
your circuit. In other words, it's best for things like RF
oscillators that draw essentially constant current. (You want clean
power for very low phase noise...) A key concept here is that your
power supply may be several millivolts (or even tens of millivolts)
noisy, and that's bad in a sensitive audio circuit, but there's not
actually so much _power_ involved in that noise, and it doesn't take
very much power to cancel it out. I've done Wenzel's circuit using an
op amp and 1% resistors to get 40dB noise cancellation. You could
tweak it to get even better.

Note also that if you have room for enough capacitance and inductance,
you can do a pretty effective job cleaning things passively. If your
"brick" has noise mainly at switching frequencies, that's not so
difficult to get rid of, so long as it's not magnetically coupled into
your circuit. But if it's at 100-120Hz and harmonics thereof, you'll
need L*C in the neighborhood of 0.001 henries*farads or more to be
effective. The L you have to use in such a filter may have more R
than the R you'd use in Wenzel's circuit!

Another way to attack this is to use power amplifiers with very good
power supply rejection, and put the main effort into
filtering/cleaning the relatively low current going to the lower-level
components.

Cheers,
Tom

OK so then I guess first I need to figure out what kind of noise I'm
dealing with, correct? Like what frequency and what apmlitude, right? Is
there any good way to do this? I can get access to an oscilliscope - if
that would work. I'm sorry my questions on this topic are so simple - I'm
just completely not used to having to do anything more than put a 1uf
tantalum in parallel with the supply and ground lines.

 

[Michael]

Spehro Pefhany <speffSNIP@interlogDOTyou.knowwhat> wrote in
news:4aqka091ojl5k2b341fqe9p46956inebhu@4ax.com:

On Tue, 18 May 2004 19:38:37 GMT, the renowned Ian Stirling
root@mauve.demon.co.uk> wrote:

Ben Bradley <ben_nospam_bradley@mindspring.com> wrote:
On Mon, 17 May 2004 21:49:12 GMT, Michael
nleahcimathotmaildotcom@nowhere.com> wrote:

Ian Stirling <root@mauve.demon.co.uk> wrote in
news:3x9qc.4463$wI4.513793 @wards.force9.net:

Michael <nleahcimathotmaildotcom@nowhere.com> wrote:
Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
snip
The simple way would be a 1W or so 5V-12V converter, a 7805 linear
regulator, and do it that way.

That IS way inefficient. The other suggestion of 5V to 6V
converter
and a 5V LDO regulatur would be much better.


snip
ISTR (from reading here) that USB has a max current spec of 100mA,
so you don't have a lot to work with. OTOH, if everything is
reasonably efficient, there's still enough power available for most
headphones to cause ear damage.

500ma.

One "unit load" is 100mA. Quoting from the spec:

---
A device may be either low-power at one unit load or high-power,
consuming up to five unit loads. All devices default to low-power. The
transition to high-power is under software control. It is the
responsibility of software to ensure adequate power is available
before allowing devices to consume high-power.
---

You may not have 500mA available from a laptop, but you will have at
least 100mA.

Best regards,
Spehro Pefhany

A-ha! So I read more closely and it is estimated that the circuit will draw
about 45ma, at 2.4V. So does that make any solutions more clear? Thanks!

-Michael

 

[Tam/WB2TT]

"Michael" <nleahcimathotmaildotcom@nowhere.com> wrote in message
news:Xns94EE4D7AE36Fnleahcim@216.148.227.77...

Spehro Pefhany <speffSNIP@interlogDOTyou.knowwhat> wrote in
news:4aqka091ojl5k2b341fqe9p46956inebhu@4ax.com:

On Tue, 18 May 2004 19:38:37 GMT, the renowned Ian Stirling
root@mauve.demon.co.uk> wrote:

Ben Bradley <ben_nospam_bradley@mindspring.com> wrote:
On Mon, 17 May 2004 21:49:12 GMT, Michael
nleahcimathotmaildotcom@nowhere.com> wrote:

Ian Stirling <root@mauve.demon.co.uk> wrote in
news:3x9qc.4463$wI4.513793 @wards.force9.net:

Michael <nleahcimathotmaildotcom@nowhere.com> wrote:
Hi - I'm very interested in a project I saw on head-fi:

http://www5.head-fi.org/forums/showthread.php?t=51870
&page=1&pp=20

It's a USB sound card designed specifically for
headphones - but the problem is the design that they
came up with uses batteries instead of the 5V USB
supply. I'm very interested in this circuit for a SBC
snip
The simple way would be a 1W or so 5V-12V converter, a 7805 linear
regulator, and do it that way.

That IS way inefficient. The other suggestion of 5V to 6V
converter
and a 5V LDO regulatur would be much better.


snip
ISTR (from reading here) that USB has a max current spec of 100mA,
so you don't have a lot to work with. OTOH, if everything is
reasonably efficient, there's still enough power available for most
headphones to cause ear damage.

500ma.

One "unit load" is 100mA. Quoting from the spec:

---
A device may be either low-power at one unit load or high-power,
consuming up to five unit loads. All devices default to low-power. The
transition to high-power is under software control. It is the
responsibility of software to ensure adequate power is available
before allowing devices to consume high-power.
---

You may not have 500mA available from a laptop, but you will have at
least 100mA.

Best regards,
Spehro Pefhany

A-ha! So I read more closely and it is estimated that the circuit will
draw
about 45ma, at 2.4V. So does that make any solutions more clear? Thanks!

-Michael
Makes it easier. Use some ferrite beads and an LC filter to get rid of the

high frequency noise, and a linear regulator to get rid of the noise within
the audio band. Note that an LC filter that works down to 20 Hz would weigh
several pounds, whereas, the noise rejection of linear regulators gets to be
poor above a few tens of KHz. For an inductor, look at the small molded
inductors that are about the size of a 1/2 W resistor, and pick one that
will carry 45 ma.

Tam

 

[Tom Bruhns]

Yes, a scope would be a reasonable place to start. Beware of ground
loops and the like. A spectrum analyzer would be even better, and in
fact you can feed the noise to a sound card (AC couple so you're just
looking at the audio signals, not DC; sound card probably already is
AC coupled, but you might want to use an external capacitor anyway.)
You should be able to find spectral analysis software that will work
with your sound card. Worst-case you can use a program like Scilab
and do an FFT on the data. There's probably a whole lot I'm taking
for granted here that I should be mentioning...I've worked with
low-level precision analog circuits and FFT-based spectral analysis
equipment for too long.

Cheers,
Tom


Michael <nleahcimathotmaildotcom@nowhere.com> wrote in message news:<Xns94EE1661197Dnleahcim@216.148.227.77>...
....

OK so then I guess first I need to figure out what kind of noise I'm
dealing with, correct? Like what frequency and what apmlitude, right? Is
there any good way to do this? I can get access to an oscilliscope - if
that would work. I'm sorry my questions on this topic are so simple - I'm
just completely not used to having to do anything more than put a 1uf
tantalum in parallel with the supply and ground lines.