About power supply noise

[[email protected]]

Hi,

I built a small LM386 amplifier and attached an antenna and an
envelope detector to it. For a few days i was able to listen to some
local radio through it but when i tried to power the circuit from my
ATX psu(5V), all i heard was a loud noise. It's fine while powering
from a 9V battery but when i switch to a power supply, this problem
arises. When the input is my PC's sound card, i can listen to anything
without noise, and also when i disconnect the antenna(long wire) from
the detector, the noise disappears. I have tried this with a linear
regulated supply, too. The result is the same noise.

What is causing this noise? The first think that comes to my mind is
that it is the noise from the AC line but then why it disappears when
i disconnect the antenna or when i use it as a normal AF amplifier? I
am totally confused.

What i tried;

Put an active decoupler between the circuit and psu,
used a 7805 with 100n caps on each side,
tried a linearly regulated power supply,
put a 10 uF bypassing cap between pin 7 and ground,
put two(470 uF and 100 nF) bypassing cap between the V+ pin and
ground,
moved the psu away from the circuit,
put a small 3.3 mH filtering choke between the psu and the circuit
and other countless attempts that i don't remember now. None of them
actually eliminated this loud noise.

I am waiting for your advices, thanks in advance...

 

[martin griffith]

On 4 Feb 2007 11:36:53 -0800, in sci.electronics.design

Hi,

I built a small LM386 amplifier and attached an antenna and an
envelope detector to it. For a few days i was able to listen to some
local radio through it but when i tried to power the circuit from my
ATX psu(5V), all i heard was a loud noise.
I am waiting for your advices, thanks in advance...

Computer power suppliescan be very noisy, use an external power
supply, preferably a linear one


martin

 

[[email protected]]

Computer power suppliescan be very noisy, use an external power
supply, preferably a linear one

martin

I have tried using one as i said in my previous message. It didn't
help. Also i am not talking about high frequency switching noise, it's
not at a specific frequency, it's just roaring. I hope you
understand me.

Thanks..

 

[Ben Jackson]

also when i disconnect the antenna(long wire) from
the detector, the noise disappears. I have tried this with a linear
regulated supply, too. The result is the same noise.

If the noise is coming in over the antenna then you have to consider that
it's not the output of the supply but the RFI it generates just by being
on. Have you tried powering from the 9V battery and then turning on the
supply independently to see if the noise starts?

 

[[email protected]]

Yes i have tried it, i ran the circuit from 9V battery and than turned
on the adjacent power supply, it didn't practically have any effect.
There were no noise. It is only present when i run it from power
supply.

 

[Eeyore]

Hi,

I built a small LM386 amplifier and attached an antenna and an
envelope detector to it. For a few days i was able to listen to some
local radio through it

Hey, I know LM386's can be unstable and oscillating but I've never heard of
anyone using them as radio receivers !

Graham

 

[Jamie]

Yes i have tried it, i ran the circuit from 9V battery and than turned
on the adjacent power supply, it didn't practically have any effect.
There were no noise. It is only present when i run it from power
supply.

you need an RF choke filter on the rails of your device.

 

[ehsjr]

Yes i have tried it, i ran the circuit from 9V battery and than turned
on the adjacent power supply, it didn't practically have any effect.
There were no noise. It is only present when i run it from power
supply.

The test may not be valid if there is no
load on the powersupply. Connect it to a
resistor that draws about the same current
as your circuit, then test with the battery
powering the 386.

Even if the supply is not generating RF, it
could be conducting into your circuit, so
the post about adding filters may be the answer.

Ed

 

[[email protected]]

I also tried it with a load the result is the same. Thanks anyway..

Should connecting two high inductance (higher is better, right?)
chokes on both negative and positive rails help? Or do i need
something more?

Thanks..

 

[ehsjr]

I also tried it with a load the result is the same. Thanks anyway..

Should connecting two high inductance (higher is better, right?)
chokes on both negative and positive rails help? Or do i need
something more?

Thanks..

I don't know if higher is better or not.
You described the noise as "roaring", so
it beats me what is in that noise. You
can try inductors in series with the power
supply leads, and a capacitor across them.

Can you back the gain off on the 386?

Ed

 

[MassiveProng]

Computer power suppliescan be very noisy, use an external power
supply, preferably a linear one


martin

If you had REALLY read the friggin' post, you would have read where
he stated that hew DID try a linear supply.

It isn't that you post was wrong... it is that YOU ARE.

By giving his post a cursory glance, and then drawing conclusions
within the first couple sentences of it, you screwed up royal.

If you are going to claim to being helpful, at least actually read
what you are replying to.

More likely, the difference is the difference yielded when
attempting to power something you had 9V going into with a mere 5
Volts.

 

[MassiveProng]

Hey, I know LM386's can be unstable and oscillating but I've never heard of
anyone using them as radio receivers !

To me, this indicates that it could be the difference between the
rail voltages he is presenting the circuit with.

The battery makes it work, but the supply is off voltage far enough
that the whole danged thing doesn't even get going.

Take several 1.2 or 1.5 V batteries and string them together in
series. Starting at the 9V point, work your way down the chain until
the problem manifests itself. The minimum rail voltage is somewhere
between that one and the battery voltage point that does work.

How's that for a diagnostic plan! A makeshift adjustable PS.

 

[[email protected]]

To me, this indicates that it could be the difference between the
rail voltages he is presenting the circuit with.

The battery makes it work, but the supply is off voltage far enough
that the whole danged thing doesn't even get going.

Take several 1.2 or 1.5 V batteries and string them together in
series. Starting at the 9V point, work your way down the chain until
the problem manifests itself. The minimum rail voltage is somewhere
between that one and the battery voltage point that does work.

How's that for a diagnostic plan! A makeshift adjustable PS.

Do you mean that the PSU voltage is too high for the circuit to get
working? I also tried this circuit with an adjustable linear
regulator(LM317). I started at the point where circuit doesn't
work(it's like 4 volts i think) and worked all the way up till 12V
which is the limit for the LM386 IC. The noise were present at all
voltages.

Even if it were like that, when i disconnect the antenna, whole noise
disappears. I think i should work on that filtering thing.

 

[MassiveProng]

Do you mean that the PSU voltage is too high for the circuit to get
working? I also tried this circuit with an adjustable linear
regulator(LM317). I started at the point where circuit doesn't
work(it's like 4 volts i think) and worked all the way up till 12V
which is the limit for the LM386 IC. The noise were present at all
voltages.

Even if it were like that, when i disconnect the antenna, whole noise
disappears. I think i should work on that filtering thing.

Just buy a supercap and see if that acts the same as the battery,
and pump the supercap "on demand" after/between listening sessions.

 

[[email protected]]

Just buy a supercap and see if that acts the same as the battery,
and pump the supercap "on demand" after/between listening sessions.

I tried it, i charge a 10000 uF cap with 5 volts from PSU and then
connect it to the circuit. It acts the same as the battery. Connecting
it between the power rails didn't help, too.

Thanks...

 

[MassiveProng]

I tried it, i charge a 10000 uF cap with 5 volts from PSU and then
connect it to the circuit. It acts the same as the battery. Connecting
it between the power rails didn't help, too.

Thanks...

10,000 isn't enough. That's why I suggested the supercap. It acts
like a battery but has a small package so you don't have some big
bulky can hanging on your project.

 

[[email protected]]

10,000 isn't enough. That's why I suggested the supercap. It acts
like a battery but has a small package so you don't have some big
bulky can hanging on your project.

I tried something a few minutes ago. I had a small high permeability
toroid salvaged from a fluourescent lighter circuit. I wound some 1mm
like wire on it till there is no winding surface left. I connected
circuit's power lines to the 10000 uF cap's terminals. Then i
connected the toroid in series with the negative power line and the
cap's negative lead. Then i touched the positive PSU lead to the cap's
positive side to charge it. When i disconnected it, the circuit went
on working as before. But the noise was gone! Before connecting this
choke, there were the same noise even with only negative PSU lead
connected to the cap. I used the cap here because i don't have enough
toroids for two power lines. When i get the second one, i will connect
it to the positive side and see if it cures the problem. I hope my
explanation was understandable.

Thanks for your interest...

 

[[email protected]]

I measured the inductance of this choke and it's only .7 mH. It isn't
as high as i guessed. I have another commercially made choke that
measures 3.3 mH. When i put this 3.3 mH choke in place of my 0.7 mH
toroid, it attenuates the noise but not as much as the toroid does. So
my previous "the higher inductance choke is better" theory is wrong, i
think. Can anyone explain this behaviour?

The second toroid will probably solve this noise problem but i have
another question now. It's fine with only one circuit attached to the
PS (i need only 2 chokes), but when i have more circuits working at
the same time, will i have to have chokes on all power rails? Is there
any practical way for this? Or isn't there anyone experienced this
problem before?

Thanks...

 

[MassiveProng]

I tried something a few minutes ago. I had a small high permeability
toroid salvaged from a fluourescent lighter circuit. I wound some 1mm
like wire on it till there is no winding surface left. I connected
circuit's power lines to the 10000 uF cap's terminals. Then i
connected the toroid in series with the negative power line and the
cap's negative lead. Then i touched the positive PSU lead to the cap's
positive side to charge it. When i disconnected it, the circuit went
on working as before. But the noise was gone! Before connecting this
choke, there were the same noise even with only negative PSU lead
connected to the cap. I used the cap here because i don't have enough
toroids for two power lines. When i get the second one, i will connect
it to the positive side and see if it cures the problem. I hope my
explanation was understandable.

Thanks for your interest...

LC filters do work well.

 

[MassiveProng]

I measured the inductance of this choke and it's only .7 mH. It isn't
as high as i guessed. I have another commercially made choke that
measures 3.3 mH. When i put this 3.3 mH choke in place of my 0.7 mH
toroid, it attenuates the noise but not as much as the toroid does. So
my previous "the higher inductance choke is better" theory is wrong, i
think. Can anyone explain this behaviour?

You could be saturating the smaller choke's core, even though it has
a higher measured inductance, or, if it has a larger form factor than
your choke, perhaps it is your choke that is saturating and that is
the desired operating mode.

OR, it could be related to the wire size used in each case.

The permeability of the core is/can be a big factor as well.

The second toroid will probably solve this noise problem but i have
another question now. It's fine with only one circuit attached to the
PS (i need only 2 chokes), but when i have more circuits working at
the same time, will i have to have chokes on all power rails? Is there
any practical way for this? Or isn't there anyone experienced this
problem before?

The single LC filter setup at the front end of the main rail should
be enough (my not even need two chokes). Perhaps some storage
capacitance near each sub-circuit, but not likely to need a full on LC
filter.