Picking up radio in an amplifier..Why?

[TimNJ]

I know this isn't an audio forum, but in general, why does this happen?

I built an LM386 circuit to work as a small guitar amp. I took a lot of time trying to be neat about my wiring and component placement...but  I apparently wasn't neat enough. I get barely any amplification of the source AT ALL, but I do get a shit ton of 60Hz noise and then a random Spanish radio station. I think it's AM.

If you look at my picture of the underside of the board, the long thin copper wire that makes a horseshoe shape on the bottom of the board, is ground. Is it possible that it is too far away? Should all of the components just be closer together?

I didn't think for an audio frequency circuit that all of these things mattered a whole lot. Perhaps the LM386 is oscillating or something. Whatever it is, I'm pretty pissed because I'm going to have to start from scratch because I've debugged about as much as I can take...unless anyone has input.

Thanks.

[vk3yedotcom]

It's fairly common for unshielded high gain audio amplifiers to pick up RF, especially if near a radio station.  IF the audio is clear it's probably an AM station. If it's fuzzy it's probably an FM station.

Broadcast pickup is due to a non-linear diode effect somewhere in the amplifier circuit causing it to detect the signal.

Do the following:

1. Use shielded audio cable for the amplifier input.

2. Shunt RF to earth by wiring decoupling capacitors at various parts of the circuit. These should be around 10 or 100nF and have short leads to earth.  For the LM386 they need to be across pins 3, 6 & 7 to earth.

3. If there is a capacitor (typically 1uF or so) across pins 1 & 8 take this out to reduce gain and possible pickup.

4. Try various power supplies including batteries and see if the effect changes. 

5. Try much shorter speaker leads in case these are picking up the signal.

6. If all else fails mount in a shielded metal box (eg Spam or Altoids tin) preferably with good quality sockets and feedthrough capacitors to maximise shielding.

[TimNJ]

Thanks. I think I am going to rebuild the circuit (or a similar one). I will take as many of the tips you have provided in mind as I build it.

Amp ground and earth should be connected, right?

Maybe this requires a long explanation, but why did you decide pins 3, 6, and 7 need decoupling caps?

Do you think the way I did ground on the board is problematic? Too many long runs of wire perhaps?

[vk3yedotcom]

On the LM386

3 is the audio input
6 is the supply rail
7 is normally connected to earth via a capacitor though some circuits omit it
2 & 4 are normally earthed already
1 & 8 is for the gain adjustment cap (or left unconnected)

[TimNJ]

Thank you!

[TimNJ]

One final question,

I've stumbled across the idea that using a capacitor with too low of an ESR can essentially form an LC circuit that might cause some ringing. The traces and wire can serve as the inductor.  I'm not sure if that's relevant here. I'm not into audiophoolery, but do want the circuit to perform as best I can make it, seeing that I'm going to do it again.

But, do you reckon there will be any difference in using a polypropelene vs polyster vs X7R?

[c4757p]

Capacitors resonating with parasitic inductance is unlikely. First, you usually need a parallel LC circuit for oscillation.* Also, in an audio circuit you are not going to have the sort of quick, high di/dt pulses that will set off such a high frequency resonant tank. If you do see any HF/RF ringing anywhere, just poke your finger around a bit until you find a spot where you can stick a cap (or cap+resistor). And if you have any parasitic resonant tanks with frequencies in the audio range, well, stop winding your wire around ferrites....

*Though at high frequency, many things act 'parallel' that do not appear so, and many things do not act parallel that ought to...

You won't find any real difference between polyprop. and polyester in audio applications. X7R are pretty crap for most signal purposes, though they're fine for imprecise frequency compensation, very imprecise timing, and their usual purpose, decoupling. I wouldn't use them for audio.

[TimNJ]

Thanks very much. I'll just go with standard polyster caps then.

[christos]

same thing with me when i build an amp for a shop...to avoid that i used audio cables and in the case where i put that , inside i glued aluminium tape ..

if you play guitar..put some aluminium tape under the pickups and cables..cover all the wood with aluminium tape even the plastic cover  ..you can understand the differecne if you have a good amplifier(i got a line6 spider IV 75W)no buzzing sound

[TimNJ]

Interesting!

On the note of shielding...

Before I posted this thread I built up a prototype enclosure for the amp. The (proposed) electronics are encased in a 4 sided box that I made out of sheet metal. This box is within the wooden amplifier cabinet. If this "box" is earthed, do I get any significant amount of EMI/RFI shielding? I always hear of conductive gaskets etc....but my enclosure is the farthest thing from that! Two sides are missing completely, and there's not really a (simple) way around this for me.

Is it better than nothing?

[IanB]

I have never built an amp like yours, but I do have many memories of playing with amps and sound equipment in the past, and one thing I remember clearly is that touching an amplifier input with a finger tends to cause a loud mains hum "BRRRRR" noise from the speakers.

The key takeaway from this is that amplifiers will amplify anything they see at the input. So for clean sound it is vital to use shielded cable from the source right up to the amplifier input and to make sure the signal path from the input to the first stage is also well shielded (e.g. by putting the amp in a grounded case).

In light of the comments above it may also be useful to have a block/bypass filter on the input to stop any frequencies outside the intended input range (like RF frequencies) from getting through to the amplifier.

[TimNJ]

Thanks Ian.

On my prototype, which failed, I did not use shielded wire...I thought the run was short enough to not justify using shielded cable. I perhaps was wrong.

When you suggested using an AF bandpass filter, I'm trying to think about how that would help. I don't know too much about RF at all. The RF that would be present on the input jack/cable (let's call it the antenna?) would not make audible noise alone, however, if it were to interact in the right way with some semiconductor device (perhaps a diode?) and a capacitor/resistor as a low pass filter, perhaps the signal could be demodulated and produce an audible audio signal.

So thus if you wanted to prevent the RF from ever reaching the amplifier, you'd have to cut it off with a filter first. Is my reasoning at all justified?

I'm not sure if I intend on using a filter, but it doesn't sound like a terrible idea. If you just used a simple RC low pass filter, surely there would be some losses on the input signal, so I suppose active would be the way to go..and that adds another level of complexity.

Of course I could be talking out of my ass...but that's what I gathered.

[vk3yedotcom]

If RF is still a problem, even after using shielded wires, a low pass filter may be needed on the input.

An RC filter could work but an LC may be better (the inductance offering next to no attenuation at AF but significant loss at RF).

The values depend on the amplifier's input impedence. 

But as  a rough starting point consider inductance values of maybe 100uH - 1mH and capacitance values in the nanofarads.

If the problems are being caused by VHF RF signals then a ferrite bead on the base of the first transistor may help.

These fixes are common in radio transmitters where low power AF amplifiers (for the microphone) operate in proximity to high power RF amplifiers.

[TimNJ]

Awesome thank you for your help!

If the problem persists I'll give it a shot.