1
\$\begingroup\$

I built a small amplifier using an LM386 on a protoboard. The problem I have however is that it makes those typical cell phone sounds if my mobile phone is close to it.

Since I want to 3d print an enclosure for it I was wondering what the most efficient way of getting rid of this interference is?

Installing it inside a box with aluminum foil on the inside? Should connect the foil to ground?

Will the holes that I need to make in the box for the buttons and leds not cause any leakage?

This is what I built: http://www.instructables.com/id/Tales-From-the-Chip-LM386-Audio-Amplifier/

enter image description here

\$\endgroup\$
  • 2
    \$\begingroup\$ You could start with making the circuit and its pcb layout be less susceptible, which we can't help with since we don't know it though. \$\endgroup\$ – PlasmaHH Jan 17 '18 at 15:29
1
\$\begingroup\$

This will get a lot better once you lay out the circuit on a real PC board. Even with a two layer board and the bottom layer being mostly a ground plane, there is a good chance the interference problem will be below the level you care about.

I wouldn't try to go for a shielded enclosure before determining it is really necessary.

Use good layout, proper bypassing caps, and small caps to ground on all signals that go off the board. Include a chip inductor in series between the external connection and the cap on any line that can tolerate the extra few 100 mΩ in series.

For the caps to ground, look at capacitor datasheets carefully to make sure the capacitors have the low impedance at the RF frequency of interest that you expect. You'll probably end up with 100 pF or so ceramic, and then choose a particular model from a particular manufacturer. At these frequencies, there can be significant differences between 100 pF, 50 V, 0805 ceramic caps, for example, between models and manufacturer.

\$\endgroup\$
  • \$\begingroup\$ Excuse my ignorance, but what are "bypassing caps" and "small caps"? Capacitors I guess? I don't know what a "chip inductor" is either. As I said in my question (but someone decided to remove it while editing my post) I am a beginner, so this is all new to me. I have some very basic understanding, but a lot of limits too. Looking into following some decent online course to get to know this stuff better. I am a software guy. \$\endgroup\$ – Joris Mans Jan 17 '18 at 17:22
  • \$\begingroup\$ For my usage I don't know if I will need to do a PCB. This is not something I want to produce, I will only need one or two, so a proto board seems largely sufficient for this? \$\endgroup\$ – Joris Mans Jan 17 '18 at 17:27
  • 1
    \$\begingroup\$ @Joris: Does this "protoboard" thing have a ground plane? Good layout and grounding are the first steps to reducing electromagnetic susceptibility. You might be able to get away with a shield only, but clean design and layout are usually more effective. PCBs can be had cheaply from the other end of the internet nowadays. \$\endgroup\$ – Olin Lathrop Jan 17 '18 at 17:45
  • \$\begingroup\$ Maybe I've used the wrong name? It's one of these things: cdn.solarbotics.com/products/photos/… I just use the entire first row as ground. \$\endgroup\$ – Joris Mans Jan 17 '18 at 17:47
0
\$\begingroup\$

Perhaps this app note from Maxim will help you. Minimizing RF Susceptibility in Cell-Phone Headphone Amplifiers

Of course it pitches their audio part MAX9724 that claims low RF susceptibility.

TLDR quote from article:

When using external headphone amplifiers, two methods can ensure that RF noise does not become audible (Solutions 2 and 3 above): Shielding and shortening the input-signal traces to minimize the amount of RF energy that the amplifier sees; Choosing an RF-immune amplifier that internally rejects RF energy, minimizing the amount of noise coupled to the outputs.

Hard to offer specific advice without the pcb, but keeping traces short, using good gnd plane and decoupling. Maybe adding a can could all help you. You still have to watch coupling on the inputs from whatever your source is.

enter image description here

You can get shields you don't have to solder with shield clips and bendable shield material.

enter image description here

\$\endgroup\$
  • \$\begingroup\$ What do you mean by "decoupling"? \$\endgroup\$ – Joris Mans Jan 17 '18 at 17:20
  • \$\begingroup\$ Yeah, so that is why it's noisy. \$\endgroup\$ – MrGerber Jan 17 '18 at 17:39
  • \$\begingroup\$ look up decoupling capacitors for more information. Also this answer on how to improve an LM386 design might help you. electronics.stackexchange.com/questions/134131/… \$\endgroup\$ – Some Hardware Guy Jan 17 '18 at 17:43
0
\$\begingroup\$

Since I want to 3d print an enclosure for it I was wondering what the most efficient way of getting rid of this interference is ?

Installing it inside a box with aluminum foil on the inside?

Should connect the foil to ground?

Will the holes that I need to make in the box for the buttons and leds not cause any leakage?

You are on the right track, one solution for you would be to make a Faraday cage. This is what you are thinking about, I think.

Loosely speaking: It's essentially a conductive shield enclosing your apparatus. If you do decide that you want some holes through your shield, then that's fine because the wavelength of the RF waves that the cell phones are using are in the MHz range (I believe), which have wavelength's of several centimeters. This means that for them to pass through the hole, the holes need to be at least a quarter of the wavelength in diameter to pass it through. I am not 100% sure about the quarter wavelength part though... so someone will probably correct me

And yes, connecting it to ground will allow any currents that comes from noise to dissipate safely through ground, instead of your signal traces.


TLDR;

Install it inside a box with aluminum foil around it that is connected to ground. Make some small holes for wires to pass through, the smaller the holes the better.

The leakage will most likely not be noticeable due to the wavelength's of the RF waves that the cell phones are using.

\$\endgroup\$
  • \$\begingroup\$ I'm going to try this and see what happens \$\endgroup\$ – Joris Mans Jan 17 '18 at 17:16
0
\$\begingroup\$

Electrical shielding becomes gradually less efficient as frequency decreases. For audible frequencies, the skin depth for aluminum is of order of several millimeters, so half a millimeter of foil will not achieve anything.

A much greater effect can be achieved by eliminating the loops in your circuit which pick up EMI:

  1. Make all wires as short as possible.

  2. Reduce the area of the contours carrying signal. Run signal-carrying wires very close to the ground, or twist them with ground wires.

Building a proper PCB also helps a lot, as it improves both.

\$\endgroup\$
  • 1
    \$\begingroup\$ It's probably not necessary to target audio frequencies from the sheild, the audible noise comes from unintended envelope detection of the mobile signals by a slow amplifier. Mobile phones are 900MHz+. If you can block those frequencies the audible interference should go away. \$\endgroup\$ – Tom Carpenter Jan 17 '18 at 16:20
  • \$\begingroup\$ I'm not blocking audible frequencies I think? Or where do I miss something? \$\endgroup\$ – Joris Mans Jan 17 '18 at 17:23

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.