Weird 600 ohm op-amp circuit I can't wrap my head around

[DavidAlfa]

Today I was repairing some proprietary equipment made by engineers who fled the company a long time ago.

The device is mounted in a cab that gets dark at night so the driver can see the road better, so the display is supposed to dim to reduce eye strain and avoid glare.
There's an LDR used for the display dimming, which I simply decided not to connect to save some annoying wiring while diagnosing an audio problem.
That put me in a situation we had never tested before: audio with a dimmed display.
We always tested each part separately, and well... the audio came out with a terrible square-wave-like sound!

Looking around, I found the noise was being generated at the microphone input, with nothing connected!
I spent quite some time trying to figure out where it was coming from, then I noticed the display PWM was running at 244 Hz, which to my ears was close to the noise pitch, so I decided to connect the LDR. The moment the display went to full brightness, the noise went away!
There was also a small 20 mV ripple on the 5V rail while the dimming was active... strange, since the dimmer is powered from 12V on a different board, but go figure, maybe a ground loop...

The real surprise came when I got the schematics out.
It looks like an inverting op-amp where the output loads the input...
The output is taken from the IN- pin...
And one side of the transformer goes straight to a 2.5V Vreg to make a VCC/2 bias.

I've never seen anything like this, and I really can't understand that circuit.

Is this a real thing, or an immensely botched job that somehow turned out to work?

[DavidAlfa]

I expected something like this...

[Conrad Hoffman]

I don't think that can work. Consider that any properly working opamp with feedback will have identical voltages at both inputs. One input is grounded. The other must be a virtual ground, so no output there to tap off of. Not to mention no bias network to get the thing into a linear region.

[PCB.Wiz]

Are you sure the circuit really is what was built?

[jomas]

The circuit wil work as a 2x inverting amplifier if you put a signal at  point 'out' on the schematic. The amplified signal then drives the 600 ohm to te transformer. So I think 'out' should be the input and the schematic is drawn in the wrong direction.

[DavidAlfa]

You might be very right!
I swear I saw the tx part very clear and obvious, whereas the rx had this weird thing.

I had a terrible headache last two days, the thinking bits didn't recover completely yet, they never were much but they've been much worse lately

I'll check again tomorrow

[jomas]

This is the same schematic as the topic start, drawn in an other way.

[Geoff-AU]

This is the same schematic as the topic start, drawn in an other way.

Maybe the original schematic was drawn by one of those "electron flow is the real way to define current" people 

[DavidAlfa]

Hmm, yeah, it was one of those situations where "Rx" means "goes to Rx", not "receives".
And a terrible layout, it looked like a tumbleweed!
Issue solved: I found that some Einstein used a VCC/2 resistor network to bias the mic amplifier on the auxiliary board, connected to the main 12V rail, which powers everything and has a shitload of noise.
The funny thing is: this device was certified a few years ago, and changing even a speck of dust means running the certification again.
You can imagine the eyebrows going up in panic around the office when I told them there was a minor issue in the audio circuit.
We'll have to live with it until the client complains someday!

[Conrad Hoffman]

@jomas has got it! It's bassackwards.

[vk6zgo]

This is the same schematic as the topic start, drawn in an other way.

Maybe the original schematic was drawn by one of those "electron flow is the real way to define current" people 

I'm one of "those people" & I draw my schematics like everyone else.