Virtural ground current, power ground, audio noise, head room

I have been experimenting with the following parts in separate power supply experiments: DCH010512DN7, DCP010512DBP, and 2- MAX1044 charge pumps in parallel. All separate experiments have been powered by 9 volts regulated DC. The miniature isolated DC/DC converters have the 9 volts DC reduced to a power of 5 volts DC with 7805 voltage regulator.  My circuit is an audio circuit that draws a little over 30mA., well within the range of these devices.  My circuit has positive and negative voltage rails with ground.  The issues I am trying to resolve are audio noise, and distortion. I manufacture guitar pedals that hook to guitar amplifiers, so there is an issue of different ground potentials. In the past I have used TLE2426 to create a virtural ground separate from the power ground.  Using the TLE2426 used in this way has been extremely successful in eliminating audio noise.  The problem at hand is the TLE2426 is only rated at 20mA.  I am aware that a virtural ground can be buffered to create more current.  I got audio white noise when using the two above mentioned DC/DC converters, even when I used their grounds.  The noise was still an issue even after adding the Texas Instruments suggested filtering on the input and output of the DC/DC converters. My circuit is also filtered correctly with low ESR capacitors and .1uf bypass capacitors. I tried not using the grounds of the DC/DC converters and using the TLE2426, even though I was a little over the 20mA limit of the TLE2426. In using the DC/DC converter grounds and it seemed I was getting the same ground results as using the TLE2426.  Furthermore, these DC/DC converts are supposed to have "GALVANIC ISOLATION"---but I am not so sure that using the ground of these DC/DC converters does not conflict with the difference in ground potential when connected with a guitar cord to the grounding of a guitar amplifier. I got white noise.  Can I use these DC/DC converters and not use the ground, and just use the power rails?  If I did that I could creat a virtural buffered ground. If I use the DC/DC converters what can I do to totally eliminate white noise? The 2-MAX 1044 charge pumps hooked in paralled, and used with the TLE2426 as a virtural ground proved to have less noise than the DC/DC converters.  I did not even use the the power ground of the charge pumps.  I have been so discouraged with noise issues I even tried a straight 9 volt DC power supply, which gave me 4.5 volt rails.  I eliminated some current draw from the circuit, getting it down in the range 20mA of the TLE2426 as virtural ground. I did not use a ground buffer.  There was no noise, but not enough head room, which caused distortion when a guitar was driven. I did put a capacitor between the power rails of the TLE2426, which eliminated some of the distortion. I assume the distortion came from not enough ground current, which caused a lack of head room with only 4.5 rails. OK here are my questions: For production I would like a simple, noise free power system, that does not require a complex filtering system.  I need to need to know if I can use the ground of the DC/DC converters or should I build a separate virtural audio ground to eliminate all noise issues. It is obvious I need a virtural ground with more current.  I would like a circuit diagram with suggestions on grounding system that will be noise and distortion free. I would also like suggestions on a good simple buffered ground system that can handle a little more current.

  • Keith:

    Audio circuit s are very sensitive to E-field noise . This the e-field emitted by the  transformer .

    1. The filters on both input and output isolated buses  are PI filer with ferrite beads 2773019447, 73 material with ceramic capacitors (1uf) on all input and output pins .

    One possible addition is the 1000pf(2kV)  capacitor. It can be connected  between primary ground and secondary ground to reduce transformer emissions.

    Tom