Auditronics 110A phantom/grounding issue

I know everyone shows 6K81 resistors in their phantom schematics to imply the use of 1% tolerance parts. The general idea is for the values of the two resistors to be as close as possible for best CMRR. It is interesting to note that a 1% 6K81 resistor costs in the region of 10p and a 1% 6K8 resistor costs around 3p . 1% only guarantees the resistances will be within 68 ohms of each other. For better CMRR you need 0.1% resistors but these cost about 60p each.

An alternative is to buy a batch of regular 1% 6K8 resistors (say 100). I have found that resistors from the same batch tend to have very similar actual resistances. With a DVM it is quite easy to find pair of resistors that match within one ohm of each other which is better than 0.1%. It does not matter how accurate your DVM really is as long as it is repeatable. The actual values of the two resistors does not matter as long as they are the same and somewhere in the region of 6800 ohms.

Cheers

Ian

One must not forget that CMRR is also determined by the other CM impedances in the circuit. Not much concern for xfmr inputs, except at high frequency, but for transformerless inputs, the bias resistors are in parallels (AC-wise) with the phantom feed resistors. Since they are often in the 10kohm region, their effect is far from negligible.

abbey road d enfer said:

One must not forget that CMRR is also determined by the other CM impedances in the circuit. Not much concern for xfmr inputs, except at high frequency, but for transformerless inputs, the bias resistors are in parallels (AC-wise) with the phantom feed resistors. Since they are often in the 10kohm region, their effect is far from negligible.

Click to expand...

Very true. Often there are parallel small value film caps for RFI purposes whose tolerance can also have a significant effect at supersonic frequencies. Some circuits I have seen even include a trimmer cap to trim these out.

Cheers

Ian