# Analysing signal unbalancing stage

I'm analysing a particular schematic, an audio compressor: https://www.soundskulptor.com/en/proddetail.php?prod=LA502

The schematic is here: https://www.soundskulptor.com/docs/la502-schematic-03.pdf

My goal is not a specific issue with a product (I'd just ask the creator), but rather finding out how it works and learning in the process of doing so. Because it wouldn't make sense to just ask about explanation in one go, I decided to analyse it on my own and just try to ask specific questions whenever I feel blocked.

My first question is on the stage where input signal (line level, balanced) is transformed to the unbalanced one:

What I understand of it:

• IN+ and IN- are the voltages carried by the input. They're essentially the same signal (sans noise) with opposite phases. That's just what balanced input is.

• The transformer T1 is used to go from balanced to unbalanced. Because the left coil is connected to both sources, this means the changes of the voltage across it are twice as big as the original signal (either IN+ or IN-) so the output of the transformer is double the voltage of the input (and here is where noise-cancellation happens).

My questions:

• What is the role of the capacitors C1, C2 and C5?
• What is the difference of the ground connected to C5 and the "empty" (unlabelled) triangle under right coil? If that's just some schematic peculiarity and not general knowledge then it's fine, I might reach author, but I think this has to do something with the fact the transformer isolates those grounds, so they're not common.
• What's the role of the R1 and R2 resistors? I'm only intereste in the "output" of the transformer stage, so the unbalanced signal - would that be a signal denoted by the voltage across R1 ? And then R2 would be just part of the further stages which are not the part of this question?

Hope this is fine to ask questions in such manner here!

1. A capacitive delta would be normal, but there's no reason not to use a Y as here. It's to help provide RF immunity. NOTE : Don't confuse "Class Y" a safety rating, with the Y (or star) topology seen here - 3 caps connected to a common point (as opposed to the more usual Delta : 3 capacitors in a triangle, one from IN+ to IN- and one from each input to GND. You're right the summing provides cancellation - but only to the extent that the inputs match perfectly, Which they never do at HF. Shorting the inputs to RF eliminates most of the problem before it reaches the inputs.

2. You don't dump the incoming RF noise from those capacitors onto the clean audio GND. Dump it to chassis GND (which will generally also be safety GND, especially in metal cased equipment where, if mains powered, exposed metalwork must be earthed.

3. R1 provides the correct load for the transformer. Among other things, it damps the resonance of the inter-winding capacitance with the secondary inductance which could otherwise provide a peak in the high frequency response.

There isn't enough circuit to reveal the purpose of R2 but the RH signal might be a virtual earth mixing bus.

• 2. Alright, that makes sense! So it means in circuits like these we have the "chassis ground" and the separate ground (wondering what's the common term for it). Nov 14, 2020 at 7:33
• 3. Yes, the role of R2 becomes apparent when considering further parts of the circuit. Nov 14, 2020 at 7:35
• 1. This still puzzles me. When I looked Y capacitors ive found the usage for connecting to AC mains safety. How are they filtering the RF? I thought RF noise is cancelled out when "summing" the IN+ and IN- ? Nov 14, 2020 at 7:37
• Sorry I could have spelled it out in more detail : Don't confuse "Class Y" a safety rating, with the Y (or star) topology seen here - 3 caps connected to a common point (as opposed to the more usual Delta : 3 capacitors in a triangle, one from IN+ to IN- and one from each input to GND. You're right the summing provides cancellation - but only to the extent that the inputs match perfectly, Which they never do at HF. Shorting the inputs to RF eliminates most of the problem before it reaches the inputs...
– user16324
Nov 14, 2020 at 12:51