> True servo balanced inputs, with no pin 1 problems usually have good CMMR
> specs...as long as the balanced lines are short.
Jim, I think you mean servo-balanced *instrumentation* inputs where three
single op-amp stages are used? In a good instrumentation design, CMRR stays
high regardless of the source Z and common-mode frequency. The function of
the "servo" (i.e., the 4th op-amp) does nothing for CMRR. Rather, it is
used to dynamically minimize DC offsets, allowing for DC inter-stage
> Make em longer, and the CMRR
> specs will drop off. In extreme cases, by using xfmrs on each end, the
> will improve by a whole bunch.
See comment above, but I'm with you on the use of high quality input
transformers (e.g., Jensen or CineMag). To get CMRR as consistent as those
transformers with an instrumentation input requires a design that maintains
high CMRR not only across the audio spectrum, but well above it to minimize
RFI/EMI common mode input noise. You can be much sloppier with a
transformer input design.
> Equipment makers will try and avoid the use of
> xfmrs like the plague. I can see why. At $75.00 a pop, you will have
> a lot of $$
> invested in xfmrs.
I've only seen a handful of really good instrumentation input designs and
they don't cost a lot more to design right -- and still well under the cost
of a Jensen or CineMag transfomer.
> problem you can think of. I think the icom 7800 has a fiber input on the
> rear panel.
Loved the '7800/'7700 S/PDIF connections. Wish everyone supplied those.
For a time, much of the shack audio was converted to light, and switched
with a Z-Systems router. Ended up selling the '7800 and '7700 and the
router to try other stuff.
TowerTalk mailing list