60 Cylce Hum Driving Me Batty

[K1ngR0y]

I have a circuit which takes a 15VDC supply from a wall adapter.
Most connections are made by shielded cables and I have employed the manufacturer's (and others) recommendations for hum reduction.
Is it possible to insert a 60-cycle notch filter or other high-pass that'll attenuate the hum?

 

[betwixt]

Possibly but what exactly is the device producing the hum?

Most cases are caused by ground loops or inadequate supply capability but we need to know what load your device is placing on the adapter and what else, if anything is also connected. Are you sure it is 60Hz which is a very low frequency, it could be you are hearing 120Hz or higher and that gives clues as to where the problem lies.

Brian.

 

[Easy peasy]

There is also 60Hz / 120Hz common mode hum due to the switch mode RFI caps ( and the switcher itself ) which can be very hard to overcome ( solder in some CM chokes in the 0 & 15V lines and also if necessary some 470nF 100V caps to earth on the load side of the CM choke )

you can measure this with a DVM on AC from either the +15 or 0v to earth

If you can buy or make a linear psu to give you the 15V a lot of your issues will go away.

 

[D.A.(Tony)Stewart]

Can you get PE gnd to shield your signals and shunt line ingress perhaps due to high Z or lack of balanced differential instrument amp use?

It's up to you to show layout, schematic and cabling.

 

[K1ngR0y]

Can you get PE gnd to shield your signals and shunt line ingress perhaps due to high Z or lack of balanced differential instrument amp use?

It's up to you to show layout, schematic and cabling.

Quite so . . . in actuality the grounding looks like a wire jungle. Perhaps, I've overdone it.
Anyway, I've uploaded the schematic.
You'll find that I have my own ideas about symbols, etc.

 

[D.A.(Tony)Stewart]

Ground is a symbol you define as 0V. Realizing one point is key and by design attempts to make all symbols the same 0V. But I said PE ground which is the wall socket 0V. Have you tried that? This lowers the high impedance floating DC supply and 0V to attenuate e-field coupling to unbalanced impedances.

--- Updated Jul 15, 2024 ---

Although audio equipment tends to be floating so to avoid that, we need to see how the hum is getting it. I assume it is a clean hum without a buzz.

1st step is it conducted or radiated, meaning can you make the hum better or worse by hand/finger locations or disconnecting signal cables.

If you can describe how you can affect the hum, will help or even if you can't and it is constant with, without sources connected.

But eventually we'll need to see supply noise and signals and layout.

 

[K1ngR0y]

First, are you suggesting that I run a wire from the chasis to the "house" ground, thus effecting a physical "earthing?"
Second (and I've attached a schematic), the hum becomes prominent when rotating the LEFT volume control. (The left side of the stereo signal is the one being inverted by the LM386.)

 

[betwixt]

Agreeing with Tony.

If you disconnect the inputs does the hum go away? May seem like a silly question but it will tell us whether the hum is actually from the wall adapter itself or whether it is coming in through the signal sockets. It is quite likely the hum comes from ground currents, not necessarily from inside the unit itself but because of differences between its ground connection and the signal sources ground connection. No amount of screening will help ground loops and it is doubtful whether filters will help either.

Brian.

 

[K1ngR0y]

When I disconnect the inputs the hum is unaltered.

 

[D.A.(Tony)Stewart]

Is the Chassis PE grounded?

First, are you suggesting that I run a wire from the chasis to the "house" ground, thus effecting a physical "earthing?"
Second (and I've attached a schematic), the hum becomes prominent when rotating the LEFT volume control. (The left side of the stereo signal is the one being inverted by the LM386.)

Yes for PE gnd. The chassis is acting as a capacitor to stray e-fields from house wiring. You are not using balanced diff. amps.
There are other problems. e.g. The Null signal amp should not be AC coupled at the inputs.

 

[K1ngR0y]

I connected a wire from the chasis to a known external ground and there was no change at all.

What do you mean by: "The Null signal amp should not be AC coupled at the inputs?"
Also, what else do you see that's wrong?

 

[D.A.(Tony)Stewart]

Sorry disregard that. I missed the 50k bias when I glanced at the schematic.
But you can measure to see if any DC drop is on the caps. These may be redundant.

I would start to use my finger to inject hum in the circuit and find out how it is getting in. The ground design is not self-evident from the schematic alone and the structure of the cables, PCB and chassis connections for common mode rejection need to be examined.

Tell us how the L pot affects the hum. Is it proportional or does it maximize near the 50% range?

Another test is to replace the +/-15 with external floating DC like a linear lab supply or two 9V batteries to test if the hum goes away.
That will reduce the number of paths hum can get in or identify the problem.
Also reverse the AC input polarity so if it is C coupling XFMR feedthru hum it may affect the result.

 

[betwixt]

Back to basics, if the hum is still there with no input the chances of it originating in the wall adapter are greater, it may simply be that it isn't giving a clean DC output and there is residual 'ripple' at 60Hz/120Hz on it's output that is being conducted to the output. You can check that by adding a capacitor (>100uF, preferably 470uF or more) across the power input. The extra reservoir capacitance will smooth out the ripple on the supply lines and the hum will drop.

It appears you can use either 8 Ohm output or line out through the attenuators. Are you hearing the hum from the 8 Ohm output loudspeaker or do you have another amplifier between your device and loudspeaker? It makes a difference, an external amplifier can introduce ground loops again but directly driving a loudspeaker shouldn't exhibit the problem.

You should be made aware that LM386 is a very old device, has quite high noise and distortion levels and isn't really intended for line level driving, it was developed as a low cost, low power amplifier in domestic radio and record player applications.

Technically, it isn't wrong to AC couple an LM386 because it has internal resistors to ground from both input pins. It would be wrong for most other amplifiers though.

Brian.

 

[barry]

As a simple test, you could put two 9V batteries in series to power the unit. Can it take 18V instead of 15V (depends on which version of the LM386 you have)?

It seems to me if "the grounding looks like a wire jungle" you've got a problem. Grounding should not look like a jungle, it should look like the plains (pun intended) or a star.

And no one has asked, and you haven't offered, what the physical structure is. Is it a PCB? How many layers? What does it look like? If you're using one of those crappy white breadboards, don't.

But just to echo others, why would you use a 40 year old part when I have to believe there are MUCH better parts available?

 

[K1ngR0y]

View attachment 192358View attachment 192359
Sorry disregard that. I missed the 50k bias when I glanced at the schematic.
But you can measure to see if any DC drop is on the caps. These may be redundant.

I would start to use my finger to inject hum in the circuit and find out how it is getting in. The ground design is not self-evident from the schematic alone and the structure of the cables, PCB and chassis connections for common mode rejection need to be examined.

Tell us how the L pot affects the hum. Is it proportional or does it maximize near the 50% range?

Another test is to replace the +/-15 with external floating DC like a linear lab supply or two 9V batteries to test if the hum goes away.
That will reduce the number of paths hum can get in or identify the problem.
Also reverse the AC input polarity so if it is C coupling XFMR feedthru hum it may affect the result.

Tony, are you saying DO NOT REMOVE the 4.7uF caps between the balance pot and inputs 2 and 3?
I tried to play antenna and touch various parts of the circuit. Nothing made a difference.
The hum is loudest when: 1. I lower the input pot to the center channel circuitry and 2. when i turn the inversion/NULL balance pot to favor the left (inverted) channel.
With a 9 volt battery there is no hum at all. (I also believe it's 120 Hz, not 60.)

 

[D.A.(Tony)Stewart]

I think now it does not make any difference for the Caps whether in or out.

As I suspected 50% pot position might be the loudest as that is the highest impedance for stray current to be amplified.

Compare all the grounds for L/R as there must be something missing or different.

When Vcm hum is conducted, the highest unbalanced impedance creates the largest voltage and usually gets partially rectified to double the frequency by the input stage protection diodes. grounding the supply to PE gnd at the source may help using a spade or ring lug.

--- Updated Jul 16, 2024 ---

The LM386 input impedance is only 50k and the source impedance of the pots varies around the 0 to 100k. This makes for very poor CMRR while a balanced high input impedance has good CMRR. But there has to be something different on the L vs R signal to ground path coupling to the signal.

Since batteries work as expected, raising the CM impedance is usually done by a Balun/ CM choke. But at 60 Hz this might look like a 100mH 1:1 transformer which shouldn't be necessary. There seems to be a ground loop or open shield.

You may have to re-route your passive input paths.

 

[Easy peasy]

When you remove the input the hum is unaltered - so psu then - see above for fixes.

 

[barry]

When you remove the input the hum is unaltered - so psu then - see above for fixes.

I don't think you can jump to that conclusion. If the layout sucks OP could be picking up radiated 60 HZ.

 

[Easy peasy]

connecting any sort of input should up the 60Hz if it was picking up any radiated - QED.

 

[FvM]

LM386 amp circuit has relative low impedance, unlikely to pick up hum by capacitive coupling. I guess 15VDC supply turns out as unregulated adapter with massive 120 Hz ripple.