Confusion about output driver and signal levels

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
...

If you want a "universal output" that works with both balanced and unbalanced downstream devices, choose an impedance balanced output.

See drawing 2.4 on this link. on page 3.

https://www.jensen-transformers.com/wp-content/uploads/2014/08/an003.pdf

or

https://tinyurl.com/37f9ks9f

Seeing this - I thought I'd share Bill Whitlock's (of this parish) reply to a question I had about the better CMRR at 60Hz with an unbalanced stage post transformer :
=======================================================================

With regard to AN003, specifically the paragraphs under schematic 2.2:

The circuit as shown (garden variety balanced input) will have a CMRR of about 55 dB at 60 Hz - and that CMRR will deteriorate with increasing frequency due to the inter-winding capacitances of the transformer (which are in the neighborhood of 10 nF or 10,000 pF because it's an output type transformer with no Faraday shield - as used in many cheap ground isolator products). Now, if the receiving input were an unbalanced (as in most consumer products), the effective CMRR could be 70 dB or higher at 60 Hz. If a normal RCA to RCA cable linked the two devices, hum and buzz is caused mostly by the ground-voltage-difference or GVD that exists between "ground" in the two devices - remember the shield conductor has resistance, so even tiny (such as "leakage" current of under 0.75 mA, as allowed by UL for two-prong powered devices) current flows through it will create a voltage drop that is directly added to the signal seen at the receiver input. By using the (output type) transformer, this noise voltage must flow through the inter-winding capacitance of the transformer (this and the input impedance of the unbalanced input effectively form a high-pass filter). This filter can easily have an attenuation of 70 dB at 60 Hz. The signal itself isn't affected by this filter since it is differential on both sides of the transformer. Of course, this attenuation of ground noise is very frequency dependent (it is a high-pass filter, after all), so the effective CMRR, say 70 dB at 60 Hz, will decrease at 6 dB/octave (20 dB/decade) ... 70 dB at 60 Hz, 50 dB at 600 Hz, and only 30 dB at 6 kHz.

I hope this helps your understanding. I know some of this stuff is not exactly intuitive.

Best regards,
Bill Whitlock
Jensen Tech Support
AES Life Fellow
IEEE Life Senior Member
 
Back
Top