Ground Loop Isolator Circuit for Sound

[TrickyNekro]

Hallo guys and girl,

I finally decided to make myself useful in the forum  .
From time to time I release youtube videos about electronics or other stuff that I like, but sometimes you need to work from your
PC / Laptop to make a video. With the integrated microphones of my laptop being rather inefficient for producing a pleasant sound,
I borrowed a Behringer mixer and microphone from a friend. Unfortunately every time I connected it to the laptop (either input/output),
it would become noisy. Reading around the internet and after some research I found out that it´s a grounding problem and a ground
loop isolator was needed. Unfortunately everything that has the word audio in it, and it is not a horrible Chinese product is expensive.
So, I said to myself... Why not design one myself! And I did!!! 

I have no interest in earning money from it so I release it here open hardware 

Here´s the specs:
Input ~3Vpk-pk (MAX)
Output ~3.5Vpk-pk (MAX)
Flat baseband from 10Hz to 20kHz with ~1.5db of gain.

I have NOT built it yet and not tested it in real world (only sim.), but should work nicely.
Another thing. I don´t know how lineal 4n27 is. I biased it at 5mA and it´s working pretty much like
a class A amplifier. Another good thing is that the design is working from a single 5V rail when you
want to experiment with it! 

I´ve built everything in just 4 - 5 hours so if I missed anything please have mercy!
I hope you like it and that someone might find it useful!

Best Regards, Lefteris
Thessaloniki, Greece

P.S. If there is someone who is willing to build it, I will upload the schematics and board. The project is built using Eagle 6.3.0..

[nickm]

You can get "high linearity" optocouplers which have two identical outputs, one of which you take your feedback from to compensate for the non-linearities.  LOC110 is an example.  Did you consider any other design topologies?  I'm curious because I too have been thinking about breaking ground loops in long runs of cable. 

[TrickyNekro]

To tell you the truth, the above circuit was a 4a.m. brain fart cause of insomnia. 
I´ve seen that kind of optocouplers before, like the LOC110. Vishay has something similar as far as I remember.

According to the datasheet of 4n27 I am indeed working in the linear range. I am also driving the led with current and
not with voltage, that should make a difference. Furthermore I am biasing the LED at 5mA with a max swing amplitude
of 4mA. Between 1mA and 10mA the graph seems linear enough. Also the sim showed promising results.

Also, I really only use it for PC - mixer and mixer - PC interconnection. The voltage output of the PC is 2Vpk-pk? Something
like that, maybe less. And I think that the mixer has a 2.5Vpk-pk voltage output max, something like that. That means
that the LED is going to operate in the 2mA - 8mA range. It should be fairly linear at that range.

I have to build the circuit first and test it, to see firstly if it eliminates the hum, and secondly if it sounds right. If it sounds
right, it´s going to be pretty much ok. I´m not looking for studio performance.

I´ll try this one for the moment, see how well it goes and then I´ll probably try one of this high linearity optocouplers. It´s
worth a shot, although compensating the feedback loop seems a little bit scary :-P


Best Regards, Lefteris
Thessaloniki, Greece

[HomerA]

Hi there,
i am just wondering if this circuit worked for you or not? I fancy making one.

[moffy]

You could use a high quality isolating transformer such as: M0705, M0706 or M0707 from: www.altronics.com.au.
They are very linear and have good bandwidth.