Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Mains hum problem

Status
Not open for further replies.

yash gupta

New Member
Hello friends!

I made a power audio amp, using the popular tda 2822 ic in streo config. For power supply I used a 12v .5A transformer, requlated by a 9v 1A regulator. The speakers are 6Ohm 10cm dia.
the filter cap is 1000uf.
I don't think that it is becoz of unstable power supply, as many desktop stereo speakers don't use a regulator and give fairly clean sound on the same ic.
I get a lot of mains hum and noise from the speakers, though the sound is loud enough.

Does it have something to do with the PCB design or the cabels or both. Currently I'm using PVC insulated unshielded single core wire for all the connections. the volume pots are not on the pcb, but connected via wires.

Please help me.
 
Once I've met the same problem. One kind Indian in this fine forum gave me a basic idea which reduced the hum.

use any conductive material piece underneath your circuit with a good insulator in between, then connect it to your ground.
 
thanks.

Here's my pcb art work
**broken link removed**
 
The PCB layout is pretty poor, there are some tight bends which might be difficult to etch and the wire jumpers are too long and untidy.

All bends should have an angle >90° but this probably isn't causing the problem here.

The number of wire jumpers should be kept to a minimum.

Any wire jumpers used should be straight - I used tinned copper for jumpers.

The amplifier inputs should be kept as short as possible - especially the feedback network and any op-amp virtual grounds.

Please post a schematic.
 
Oh yes,the 1st idea worked a lot. the noise was damped many times.
Thanks..
Looking forward for more suggestions.

For the input:
omit the 10kOhm's (2 on the left side).
a 47kOhm, a 0.1uF and a 50k pot in series with the audio out from mp3 player and the input is taken from the center leg of the pot.
2 such things for L and R.
the power supply and audio input, output, all have a common ground.

The power supply:
(Vs) is 9VDC (12v, 100uf filter cap regulated by a 9v 1A regulator)
a switch for on off, and a switch to switch from battery to mains supply.

N.B. I haven't created the final schematic yet. I did all the things directly on the pcb artwork.( the initial one has been posted).

**broken link removed**
 
the power supply and audio input, output, all have a common ground.

This might be a major source of your noise issue. It's generally a good layout practice where/when a ground plane is not used to take separate common ground paths such as low current inputs, high current outputs and supply, back to a star ground at the power supply, a single common point, to avoid loop currents. From your schematic above, I would have placed a minimum of three ground returns, but possibly as many as five, and no flying lead jumpers.
 
Please, suggest me some good and free pcb layout soft, having auto route feature.
 
the power supply and audio input, output, all have a common ground.

It would be my guess as well that it might be due to the way the ground connections are run. In my experience a consequent star ground in most cases cure hum. Another reason for hum I have observed is oscillation but don't know if it is possible/relevant with this design.

Best,

Jesper
 
The best way would be use a double sided PCB. Leave one side a solid piece of copper for the ground plane and route all the connections on the other side. You'll need to countersink all the holes which you don't need connecting to 0V on the component side and solder all the legs which need to be connected to 0V, on the component side to 0V. With proper placement, you should then be able to get rid of all the jumpers.

If you don't have a double sided PCB, route the 0V trace first and use a start ground, i.e. rout all the connections back to a single 0V pad whenever practical.
 
route the 0V trace first and use a start ground, i.e. rout all the connections back to a single 0V pad whenever practical.

Oh yes, neither I have a double sided pcb, nor I have equipments for etching it.

And please tell me, can I run a long ground track and run all the groungs to it, meeting it from the sides, like this:
|| stands for the ground track ... for nothing and == stands for the tributaries coming from other points.


....||==
....||....
==||....
....||....
....||....
....||==
==||....

Should I do so?
 
Last edited:
The forum removes all the spaces from ASCII drawings so they don't make any sense.

Here's a application note which discusses star grounding.
https://www.electro-tech-online.com/custompdfs/2010/05/AN-42036.pdf

The audio inputs and the IC's 0V should go to to the same point of the star, the outputs and IC are less critical.

The layout is complicated by the reserve socket for the TDA2822 which you should consider removing.

How did you make your PCB? Did you get a company to make it for you?

Most PCB places don't charge much more for double sided.
 
Last edited:
I added the provision for tda2822 at the end time, so it resulted in inefficient layout.
No, actually, I bought a copper clad pcb board and used toner transfer to print the resist on it. Then I simply etched it by suing a mixture of dil. HCl and H2O2(hydrogen peroxide) as etchant.
the place where I live doesn't have enough facilities for getting a pcb made by a company. And if they are, I don't know about them as I'm a 10th grade student.
 
And, for a double sided pcb, I need a laminator (for the toner transfer thing), which I don't have and I can't buy one.
 
And, for a double sided pcb, I need a laminator (for the toner transfer thing), which I don't have and I can't buy one.
No you don't.

You only need to etch one side of the board.: perform toner transfer on one side, cover the other side, which will be the ground plane, in sticky backed plastic and etch.

The idea is the plain coper board is connected to 0V and all your tracks go on the other side.

You'll need to countersink the holes containing pins you don't want connecting to 0V, on plain copper side.
 
How do I countersink the holes which are not to be grounded?
 
With a countersink tool.

It you don't have one, use a larger drill bit to remove the copper immediately surrounding the pin. It doesn't matter as long as there are no pins connecting to 0V which you don't want connecting to 0V. The IC is probably the most critical part where there's a risk of it happening.
 
Exactly Hero, a few twist of a drill bit and you cleared the copper. In other words, I agree. :)
 
Last edited:
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top