LM3886 hum

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello,
I've build audio amp with 2x LM3886. Power supply is unregulated +-30V. The signal goes to the preamp first.
The problem is that I can hear quite audible 100 Hz noise. When I'm changing the volume the hum is the same so when playing loud I cannot hear it, but if want to play quietly I can hear the hum. I tried to disconnect LM3886 signal input wire (not signal ground, just signal +) and the hum disappears.

The blue waveform is power supply ripple voltage, pp voltage is 150mV.
The yellow waveform is the humming output of amplifier. It looks like there are small 100Hz squares which last during the rise time of power supply ripple voltage.
View attachment IMAG001.BMP

I amplified the humming output using inverting op amp and it looks like this:
View attachment IMAG003.BMP

Do you know where the hum is coming from and how to eliminate it? Could it be from ripple voltage of power supply or bad grounding? Thanks
 
Administrator
Joined 2007
Paid Member
The thing with grounding and wiring is to try and understand the problems. Saying to connect something here and this there isn't going work... you have to figure it out :)

All the problems come from the very small but ever present resistance of the wires and PCB print.

So if you have a wire or PCB track that carries the charging current for the PSU caps then that wire has a small volt drop along its length. The problems come when you connect different parts of the circuit to such a wire because you are connecting at different points along its length and so allowing a volt drop (the 'hum' voltage) to develop between those points.

If those points are for example the input ground and the feedback return ground then you have just introduced an unwanted 'hum' signal into the amplifier.

Try running your ground connections back to a clean 'star' point in the power supply.

This can be just a point taken off the connection between the reservoir caps, but it it must be a spur off that connection that is 'clean' and has no circulating currents flowing. That point then becomes your amplifier ground reference point.
 
My power ground connections are already in one point in PSU, but what about signal ground?
I tried to connect audio input directly to the LM3886 so the left and right audio ground are connected together at first and then they are divided to the left and right amplifier channel where they are connected to the power ground. This wiring makes the hum a bit lower.
 
Administrator
Joined 2007
Paid Member
The input ground and the feedback ground should be returned to the same point. If they are not then any voltage drop generated between the two will be amplified as a 'valid' signal and you hear it as hum.

Remember that the feedback point is actually 'the other input' to the amplifier... it has an inverting and non inverting input.

So the parts that connect to ground from those inputs have to go to a common and clean ground which is typically what we call a 'star point. This is a point with no contamination of power supply charging pulses flowing.
 
I have drawn a simple schematic of my ground connections:
grounding.png
The feedback capacitors are much closer to the signal ground than to the power ground connectors (although it does not look like that). So they should be at the same potential but it is humming.
But when I disconnect power ground from the preamp, the hum is much lower so I think there is high current flowing from LM3886 power ground through signal ground and preamp ground to the PSU.
grounding2.png
Another thing is that the right channel (the physically further than the left one) is humming more because of longer wires - larger voltage drop on them I guess.

Do you have any idea? I still do not know how to solve it. Thank you
 
Hum is usually caused by:
1/ More or less than one earth connection.
2/ Bad ground path with smoothing capacitors. Should be single ground wire into smoothing caps and single wire out. Dont mix in audio grounds with smoothing cap grounds.
3/ Bad smoothing caps or wrong value.
4/ Pickup from close AC signals or a magnetic item like a transformer.
5/ Input wire should be screened especially at pre amp end.

In your setup ground wires look good.
So look at (1) (3) (4) or (5).
 
Last edited:
Administrator
Joined 2007
Paid Member
Is the 'GND on PCB' a star point?

If it is not and if there are reservoir cap charging pulses flowing along that trace then that would cause the buzzing.

Also make sure you have no other ground connections such as input sockets not isolated from a metal enclosure as things like that can create other unwanted loops and returns.
 
Ground on the PSU is a star point. But I'm not sure about the amplifier:
amp.png
There are two big caps on the left between +Ucc GND and -Ucc GND. Between the two caps there are connectors for speaker and PSU ground. Is it possible that high current from LM3886 ground influences feedback caps and the signal on the right?
 
Administrator
Joined 2007
Paid Member
The LM3886 current would not produce the 100Hz buzz you hear. The 100Hz is the big clue that this is caused by reservoir cap charging pulses contaminating the grounds.

Why not start again with the wiring and build it up one section at a time. Wire just one channel up initially and get that noise free.

I would begin by only having one LM3886 board connected. Apply shorting plugs or links to the input and return the ground wire to a star.

Return the speaker negative back to the power supply at a point before the star... at least to begin with.

That channel should now be silent apart from slight hiss.
 
I wired one (right) channel as you said and I've found out:
If I put normal audio to the preamp input or if I short preamp input, it is humming.
But it is completely silent when either preamp power ground is disconnected from PSU or LM3886 audio input is shorted (or both).
 
Administrator
Joined 2007
Paid Member
If you still have the preamp in circuit then it could well hum.

It must be the just the LM3886 board with its input shorted. That board then connects to the star ground point along with its - and + supply. It should be silent.

---------------------------------------------------------------------

Does it all work correctly when you disconnect the preamp power ground as you mention? If so then your preamp is probably getting its ground under those conditions via the connection between pre and power amps.

You could try a series 10 ohm or lower resistor in series with that ground connection in that case to break the direct loop.
 
I've got rid of the hum finally!
The problem was that I was mixing two different grounds: PSU ground and ground in LM3886 PCB.

LM3886 is connected to PSU so due to high currents (100Hz charging pulses) their grounds are not the same. Then there is a preamp where audio input is connected to and at its output there is connection to the LM3886. So in LM3886 PCB and preamp PCB there is the same ground (different than the PSU ground). The problem starts when I connect power to preamp -> I connect PSU ground and LM3886 ground together.

As I wrote before, I could disconnect preamp ground from PSU but that would solve the problem if I used just one channel. The ground in right LM3886 channel was also different than the left one.

So I created a ground star point exactly between left and right LM3886 channel and connect it using one wire to PSU. The reason was to create one point which is as close as possible to both channels. This point is the place where connect preamp ground. The hum does not disappears completely but it is better. The hum will disappear completely if I do not connect preamp ground at all or if I put a small series resistor with that ground as Mooly said.

I have another question. If I disconnect preamp power ground, the current will have to go through signal ground. Although I cannot hear the hum, could it be a problem?

Thank you very much for your advice!
 
Administrator
Joined 2007
Paid Member
Good to hear you are getting there with it :)

It is always very difficult to give exact advice without seeing the bigger picture and having it all in front of you but not connecting the preamp ground could be a valid solution for your set up. The resistor is also a valid solution and many power amps actually incorporate this into their design to prevent hum issues.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.