Need help with 60hz hum/ground loop. Schematic attached.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
This is my first amp design. Im not new to electronics but I am new to audio amps.

This is just a simple class A amp (schematic attached). The common-emitter stage is a BC517 and the common-collector stage is a NTE261 as is the current source. The bias current is set by Rbias (obviously). Rails are +/-12V.

As you can see the amp is just has through hole components on a radio shack pcb and the transformer is just mounted on the wood. Earth is not connected since there is no chassis shielding (since it's constructed on wood).

Here are some things I noticed about the 60hz hum:

1. Hum amplitude increases as Ibias is increased.
2. There is no hum when nothing is connected to the RCA jacks. (input cap open circuit).

Please advise. I can take any measurement (voltage or current) or signal (I have an oscilloscope).
 

Attachments

  • schematic.png
    schematic.png
    18.5 KB · Views: 286
  • amp.jpg
    amp.jpg
    255.5 KB · Views: 281
Last edited:
I'm not sure if this is going to turn into my blog or whatever but after some further testing it seems like whenever there is any kind of DC path to ground at the input the hum is there. The higher the impedance, the lower the hum. Obviously if I short the input to ground the hum is at it's loudest.

What could be the cause of this? Input impedance too high? Too low? input AC cap value? IQ of input stage?
 
Administrator
Joined 2007
Paid Member
Do you get the hum with the input shorted (using RCA plug with shorted terminals) ?

If the amp is silent with the input open as you mention and hums with the input shorted then I would be looking at a wiring problem and the points you have.

Is it a hum or a buzz ? Pure 60Hz hum will be a very deep and pure tone. If there are harmonics present it will be harsher and at 120Hz.
 
Your "R bias " needs a cap to filter out any ac on the lower resistor. Then the current source will be just DC.

Easier to describe with labels on the components. Values would help, too.

Then look at ground current paths.

Where would AC come from on the rbias resistor? The gate of the bjt with have a very small ac signal from the opamp (to compensate vds changes), but I don't think that will couple to the resistor right? What cap value will I need? 0.1uf?

Also, I am looking at grounding paths by reading previous ground loop threads, but would also like some direction.

The component values are 10k/1k on the CC stage (for a gain of 10), the bias value for the opamp is about -10.5V to get a ~1.5V drop across the resistor.

The input stage I believe is 20k to ground but I forget the top resistor. It's biased to achieve ~1.2V at the output.
 
Do you get the hum with the input shorted (using RCA plug with shorted terminals) ?

If the amp is silent with the input open as you mention and hums with the input shorted then I would be looking at a wiring problem and the points you have.

Is it a hum or a buzz ? Pure 60Hz hum will be a very deep and pure tone. If there are harmonics present it will be harsher and at 120Hz.

Just taking the pin and shorting it to ground (not involving the RCA terminals) produces the hum. What what do you mean by "wiring problem"? Ground wiring? It's not a mistaken connection because the amp works, just the hum is also there.

And whether its 60hz or 120hz I don't know I'll measure it when I get home. Will the fix be different whether its 60 or 120?
 
Administrator
Joined 2007
Paid Member
The frequency is a good clue. Noise from the PSU is at twice line frequency (due to action of the bridge rectifier and all those half cycles :)) That why it sounds a bit harsh. A pure 60Hz hum at low levels can be pretty inaudible.

I'm going to point you to this thread and you should read it all as it will give you a good insight into the problems. Remember every piece of wire that passes current develops a volt drop across it. That applies even to the thickest wire. And that small volt drop matters because if you connect two points in a circuit along that wires length they actually have that voltage difference between them. Now imagine that wire is in fact the ground and those two points are your input ground reference point and the feedback return. When you short the input to ground the amplifier actually sees an AC voltage as being present and amplifies it. The AC voltage is caused by the huge (in the tens of amps range) charging pulses of the bridge rectifier which only conducts briefy for a few degrees on each cycle to bring the reservoir caps back to full supply voltage.

http://www.diyaudio.com/forums/soli...-lin-topology-nfb-tappings-2.html#post1624677
 
The frequency is a good clue. Noise from the PSU is at twice line frequency (due to action of the bridge rectifier and all those half cycles :)) That why it sounds a bit harsh. A pure 60Hz hum at low levels can be pretty inaudible.

I'm only doing half-wave rectification so it it's from cap charge pulses it would definitely be 60hz right?

I'm going to point you to this thread and you should read it all as it will give you a good insight into the problems. Remember every piece of wire that passes current develops a volt drop across it. That applies even to the thickest wire. And that small volt drop matters because if you connect two points in a circuit along that wires length they actually have that voltage difference between them. Now imagine that wire is in fact the ground and those two points are your input ground reference point and the feedback return. When you short the input to ground the amplifier actually sees an AC voltage as being present and amplifies it. The AC voltage is caused by the huge (in the tens of amps range) charging pulses of the bridge rectifier which only conducts briefy for a few degrees on each cycle to bring the reservoir caps back to full supply voltage.

http://www.diyaudio.com/forums/soli...-lin-topology-nfb-tappings-2.html#post1624677

I understand what you're saying but I'm not getting a general principle as to where the problem can be in my circuit or what ground connection I need to improve and where.

I've attached my physical tolopolgy. Please take a look.

Will it help to physcially seperate the amp stage and the power supply more? Will it help to use thicker vcc/vee/gnd wires (less resistance). Will it help to beef up the grounding on the power supply? Where is the best place to ground the RCA connectors and speaker connectors? Thicker wires there?
 

Attachments

  • diagram.png
    diagram.png
    34.7 KB · Views: 216
Last edited:
Its the class A, running dual supplies thats the problem.
Reference should be -ve rail, opamp is compounding the issue.

Edit: just remembered you are using half wave rectification. not good.
what idle current are you using. ?

Why is running dual supplies a problem for class A? And why would -ve be the reference? And how is the opamp compounding the issue? What that is bad about using half wave rectification other than larger caps are needed?

Bias current is about 400mA. As I said in the op if I lower the bias current the hum does decrease substantial.
 
Administrator
Joined 2007
Paid Member
I think whizgeek has the answer. When you had posted your first diagram (post #1) it looked like an AC coupled amp on single rail but its not :)

The problem as whizgeek says is that ripple (yes its 60hz for half wave) is getting into the amp. The base bias network for the first transistor needs to be very very clean and referenced to "ground" or "zero" volts point.


Dug I think mentioned the opamp based current sink. That could be an issue too for the same reasons.

These are great circuits for learning...
one idea might be to replace the current sink with a simple two transistor type circuit.
 
What does the top resistor need to be split in two and a signal ground cap (im assuming what this is) needed on the first stage bias network? What value cap (or cutoff freq) do I need?

I've replaced the opamp with a current mirror (not the wilson you posted, a simple diode connected device on the right) but the problem persist but it persist in a different way. It seems then the bias resistor (connected to the diode device) is cold the output is dead slient, then slow the output starts to hum again (maybe about 3-4 seconds after the power is turned on).
 
Administrator
Joined 2007
Paid Member
Here you go then. The error was in the bias network polarity.

I don't what values of resistor you have used but these are typical. The current sink of around 400ma is a limiting factor in how far negative the output can swing. This is nearly at the point of clipping. I used two individual transistors to make a darlington... not sure if that is what you have and I used a non darlington for the current sink.

I think the circuit could be rehashed significantly while keeping the basic topology but hopefully this will help with what you have.
 

Attachments

  • Simple Class A.JPG
    Simple Class A.JPG
    121.2 KB · Views: 105
  • Simple Class A Performance.JPG
    Simple Class A Performance.JPG
    116.9 KB · Views: 92
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.