Inexplicable Humming Noise In Self-built Amplifier

Status
Not open for further replies.
I built a simple stereo amplifier based on two boards, each one with 3 parallel TDA7293 ICs (TDA7293 255W in Parallel BTL Mono Power Amplifier Board | eBay). The power supply is based on a 2x30V~ 600VA toroidal transformer, a 20A/400V bridge rectifier and two 22000uF/63V capacitors (resulting a +/-43V DC power supply).

A secondary, completely separated +/-5V power source is used to supply current to a OPA2134-based preamplifier.

Now the problem: everything works as expected, apart from a very unpleasant humming that seems to be generated by the transformer, or in any case, related to the mains. The most intriguing part is that if I disconnect any of the two amplifier boards from the power supply, the other one sounds crystal clear, with no noise.

Before deciding to seek help on this forum I have read a lot online on this matter and I made sure that:

  • There are no ground loops between the preamplifier and the amplifier. The +/-5V and the +/-43V power sources are completely separated (their grounds are not directly connected). The audio signal is carried between the preamp PCB and the power amplifier boards by an audio shielded cable (shield is connected at both ends, being the only point where the two grounds meet).
  • The transformer is completely covered by a shield that is not directly connected to the central screw that keeps the toroid fixed.
  • All power cables are thick and short enough so their resistance is insignificant.

Other remarks and things I have noticed while trying to identify and fix the problem:

  • The oscilloscope shows a perfectly flat +/-43V with either one or the two amplifier boards connected, which tells me the capacitors are big enough to handle the drawn current. The same applies to the +/-5V power supply.
  • The humming goes away immediately after disconnecting the mains supply from the transformer (the amplifiers still work for another few seconds), which clearly indicates that it's related to the transformer/mains supply.
  • I have tried connecting the grounds of the two power supplies (5V and 43V) together, while leaving the shielded audio cable ground connected only at the preamp side; the effect was that the humming noise was even louder.
  • All experiments have been made with no actual sound source. Everything I could hear in the speakers, beside the painful 50Hz noise, was the (thermal?) noise of the (pre)amplifiers.

The idea that disconnecting one of the amplifier boards from the power supply removes the noise completely, just makes no sense to me. Does the extra current drawn by the secondary amp board create such an electro-magnetic field so that the noise gets induced into some signal path? What could possibly be the source of the noise here?

Thanks in advance for any help on this matter.
 
Guys, thank you all for your answers and suggestions. It now makes sense to me that ground loops can be created between the two channels.

I added a 10R series resistor to the ground of the shielded cable between the preamp and one of the power amps. It totally cancelled the humming in the opposite channel but then the other kind of humming (the one that you can hear when input ground is disconnected) could be heard on the first channel.

Then I replaced the resistors with trimmers and found out that a balance between the two types of noises could be achieved by setting the two trimmers (L and R) to about 2.5 Ohm. This is somehow acceptable but still noticeable.

Is there anything else that can be done to totally remove the two noises from both channels?
 
Guys, thank you all for your answers and suggestions. It now makes sense to me that ground loops can be created between the two channels.

I added a 10R series resistor to the ground of the shielded cable between the preamp and one of the power amps. It totally cancelled the humming in the opposite channel but then the other kind of humming (the one that you can hear when input ground is disconnected) could be heard on the first channel.

Then I replaced the resistors with trimmers and found out that a balance between the two types of noises could be achieved by setting the two trimmers (L and R) to about 2.5 Ohm. This is somehow acceptable but still noticeable.

Is there anything else that can be done to totally remove the two noises from both channels?

Put another identical transformer, one for each channel !
 
If you study the link I posted closely you will see that the hum breaking resistors are not in the signal return wire, if you can implement the proper application it should improve the situation

I carefully read and understood the article and I know the modification I made is different from theirs. The reason why I couldn't do exactly what they said is that my preamp schematic is a bit different: http://sound.whsites.net/p88-f2.gif.

That's why I thought I'd break the loops made by the shielded interconnection wires. Where do you suggest I place the resistor(s) in this schematic? Or does it belong to the power amp stage?
 
What works for me is to use a GLB (10ohm resistor+bridge rectifier+capacitor) per transformer/PSU to connect the power ground (0V) to chassis and connect PE to chassis only. Remember that the power grounds of each PSU are connected in your circuit via the interconnects and this connection is really only to make your build safe.

Modify the ground connections of each pre and power amp so that the feedback/input grounds connect to the power ground via 1ohm resistors plus parallel diodes.

None of the above will help much if the layout is bad. You have (ground) loops that pick up interference and you have transmitters like the transformer and its leads. Twisting leads and using coax for inputs will help, as will shortening the leads.

If you want a more detailed help, please post some photos of your build.
 
Last edited:
Did you look at fig. 5? It is almost identical to your schematic. The resistors can be used in either/both stages

Referring to my preamp schematics, I figured the 10R resistor should be placed between (R1L, R4L) point and the ground. So I did that and nothing changed.

Then I replaced R3L with a trimmer so that I can play with my preamp's gain. I so discovered the humming doesn't change when I alter the gain. Therefore I believe the ground loop has to be related to the power amp.

The schematic of the power amp is (roughly) given in TDA7293 datasheet, page 14, figure 11. I have three of these ICs, two slaves and one master.

How would you go about modifying this schematic so that it incorporates the ground loop breaking resistors?
 
What works for me is to use a GLB (10ohm resistor+bridge rectifier+capacitor) per transformer/PSU to connect the power ground (0V) to chassis and connect PE to chassis only.

Is this change going to eliminate ground loop humming even when all components and stages are fit into the same chassis? In my case, only the big power supply ground is connected to the chassis.

Modify the ground connections of each pre and power amp so that the feedback/input grounds connect to the power ground via 1ohm resistors plus parallel diodes.

This is what I'm hoping will fix the problem. I'm having a hard time though figuring out how to do that with my power amp schematics. What do parallel diodes achieve in addition to the small resistor?
 
Status
Not open for further replies.