Ground loop problem in DIY pre-amp

[pfn]

Yay, a forum for audio DIYers

Here's my basic setup:

iPod into volume controller, volume controller into amp in trunk. Currently, I'm using a 100k dual-gang pot to control volume, and, with this, I have no noise. The pot is a cheap piece of junk that I would like to improve on by using something more advanced and robust.

I found this circuit online and would like to use it for performing volume control:

http://www.elecdesign.com/Files/29/2531/Figure_01.gif

Its workings is described by:

http://www.elecdesign.com/Articles/ArticleID/2531/2531.html

I've built up the circuit as diagrammed, and tried installing it into my system such that it replaces my current pot. It's powered by a LM2940T regulating car power down to 5V. With the car off, it sounds beautiful, no noise at all. Once the car is turned on, though, all sorts of noises associated with ground loops show through, alternator whine, ignition ticking, etc.

The preamp and my iPod are both grounded by the same connection (factory ground point connected to console sheet metal).

One thing I haven't tried is connecting pin 6 of the digital pots directly to audio signal ground, and not connecting pin 6 into circuit ground. I've tried isolating the audio ground previously by not connecting the shields of the audio lines.

The voltage at the digital potentiometer resistor elements must stay within the power-supply voltage rails. Therefore, the 47-kW resistors and associated coupling capacitors maintain a 2.5-V dc bias on these elements. The corresponding audio voltage range is ±2.5 V, in which normal audio levels fall.

Given this quote above, and looking at the specs on the digital pot, I don't understand why the voltage drop across the resistor network in the pot must be within rail voltage. Maybe I should bypass the 47K resistors altogether and wire pin 6 directly to the audio shield ground. Would this help, or should I look into filtering the power supply some more?

Any recommendations on how to eliminate this noise problem?

Thanks!

 

[pfn]

Hmm, so I looked at the specs on the digital pot and it seems that pin 6 needs the resistors on it to bias the voltage, since the the datasheet on the pot says voltage across the resistive elements needs to stay between Vcc and GND or bad things happen :-(.

What to do now.

 

[pfn]

Nobody has any input?

Anyway, it seems I misplaced the coupling capacitor on the Rl to audio GND pin of the pot. I'll try relocating it after I get back from vacation and see if it helps the situation any.

 

[djQUAN]

try adding a 5-10R resistor between the chassis GND and the circuit GND.

 

[tsmith1315]

Have you tried the normal filtering elements on the +12V input, and on the input/output of the +5V regulator?