1. Home
  2. Hobby Electronics Basics
  3. Finding the noise-source in a headphone-amp

Finding the noise-source in a headphone-amp

Greetings

I found and built this project:

formatting link

After I finished and tested it out, I found lots of noise in the headphones. Since my skill of analog electronics is way too low to get a good guess what might have gone wrong, I took the construction to my lecturer at the university. He gave me a couple of advices how to find the errors in such piece of circuit.

Another week passed and after extensive searches, I think I have pinned the problem. Here's an explanation of what I did to test and pin:

The blocks of the circuit are transformer, rectifier, linear regulators, sound amplification. Using it plainly, I get lots of noise in the headamps. Bypassing the transformer and rectifier-bridge (using a DC-psu and feed the DC into the regulators), the noise goes away. Doing the same but feeding to the bridge (i.e. "rectifying" my DC), the noise comes back.

I take this means the rectifier bridge (four 1N4005) is broken in some way, but how? Before building, I tested each diode with an ohmmeter, and they all showed "infinite" resistance in the wrong direction, and low resistance in the correct direction.

So, what might it be that's happening at the diodes? Magnetic field inducing? Short circuits?

--
Rikard.
Reply to
Rikard Bosnjakovic
Loading thread data ...

Is it built on a pre-made pcb or have you made it on stripboard ?

Hiss (white noise) ? Does it sound 'swirly' at all ?

Doesn't sound much like it.

I suspect the op-amps are oscillating at VHF due to an absence of local supply decoupling. That tends to make them noisy. These 'project designs' on the net are often very amateurish.

Try adding some plastic film or ceramic 22nF capacitors from both V+ and V- on the op-amp power pins to ground.

You should really have something similar on the inputs to the LM317 and 337 too.

Graham

Reply to
Eeyore

It is built on a veroboard / stripboard.

I don't think it sounds swirly or noisy, it sounds more like a constant "clear tone" ringing all the time.

[...]

A lot other people have told me this as well, that the PSU in the schematic is really a waste of power. I'm too unskilled to have an own opinion, but running into problems like these are probably a good idea to learn more things that aren't in the books.

I will add some caps to the op-amps and the regulators and see what (if something) happens.

--
Rikard.
Reply to
Rikard Bosnjakovic

project:

formatting link

Hi, Rikard. You've pretty much proven that the issue is with your power supply. But I think it's possible the basic problem is with the zeners rather than the rectifiers. I'd use a standard resistive voltage divider (try replacing the zeners with1.2K and the 200 ohm resistor with a 120) instead of the 12V zeners for the 317 and 337. Any zener noise is just going to be copied directly onto the output line, with poor audio resuts.

Good luck Chris

Reply to
Chris

project:

formatting link

Which increases the noise of the 317 / 337 by ~ 10 times, the gain of the reference diode !

You're talking complete and utter rubbish.

A. The noise fron zeners isn't enough to be problematic. B. Op-amps have very significant supply voltage rejection. They don't care too much about in band supply noise.

Graham

Reply to
Eeyore

Like a whine, or a whistle? Then it sounds like you've built an oscillator. :-) Do you have access to a 'scope?

Good Luck! Rich

Reply to
Rich Grise

Each regulator should have a .1 uF cap from Vin to ground. Whether or not that fixes your problem, install them.

Ed

Reply to
ehsjr

In message , Rikard Bosnjakovic writes

More likely you've got one of the diodes in backwards. Check and double check.

--
Clint Sharp
Reply to
Clint Sharp

Not a chance. Putting in one or more diodes backwards would either make it buzz or stop it working altogether.

As everyone else has said, it's almost certain to be one or more of the chips oscillating from the lack of bypass chips - especially seeing as it's built on Veroboard. I'd solder a 100nF monoblock ceramic across the power input/output pins of each chip, *as close as physically possible*. (And maybe a 4.7uF tantalum to the input side of each LM317.)

Oh, & if I were building that circuit myself, I'd use a 78L12 & a

79L12 & replace each Zener diode with a pair of 1N4001 diodes to jack the power rails up to 13.2V. (Cheaper & smaller)
Reply to
Lionel

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.