Falcon030 line-level audio mods

[dml]

Maybe the blurry thing with the blurry text is indicating the implicit ground connection you can get when you plug something grounded into the headphone jack? It doesn't look like a component to me.

...or maybe it is showing a real connection to GND via the shielding. Which means non-isolated audio ground.

Screenshot 2025-05-09 at 13.09.37.png (68.63 KiB) Viewed 143 times

[dml]

The RevG schematic doesn't show that ground connection.....

Screenshot 2025-05-09 at 13.11.34.png (89.14 KiB) Viewed 143 times

[dml]

One last update on this, while I'm looking at it...

Checked 3x Falcons now (RevA, RevD, Rev<?>) and they all have a DC path between AGND and the main ground / shield.

[exxos]

Checked 3x Falcons now (RevA, RevD, Rev<?>) and they all have a DC path between AGND and the main ground / shield.

They would have to otherwise stuff wouldn't work..

Analogue and digital ground are still connected but normally analogue ground is physically isolated other than one single connection point back to digital ground...


EG:

White area is "no copper". When AGND will have all analogue parts (quiet ground) and linked to a single point back to digital ground (system gnd)

gnd.jpg (18.49 KiB) Viewed 135 times

[dml]

Right, so it's just trying to avoid ground loops and not DC-isolating them fully. That makes more sense.

[exxos]

The 2 grounds are labeled differently.

GND will be system GND
AGND will be on its own chunk of copper and linked to main GND *somewhere*.

Aren't there some gerbers for a early falcon floating around somewhere ?

Capture.PNG (260.74 KiB) Viewed 129 times

[dml]

Looking at this again, I realised I was looking at a non-standard opamp configuration using the LM386's extra pins for external gain control.

Normally gain is set by a resistor ratio on the input and feedback. With this chip it can be additionally controlled via special pins 1 & 8.

It seems Atari combined the gain boost with bass boost and it's not obvious yet how to detangle the two without some experiments and the LM386 datasheet. Also, both the pre- and post- 'modded' versions of the circuit actually uses a configuration not described in the datasheet, where pin 1 is tied to the output instead of to pin 8.

So I'm rewinding my last 'remix' and going with my slightly earlier 'fix' where the line-level out and bass boost delete are one and the same mod - but the U21/10K part is still for the mono audio circuits only.


LOWER THE MONO SPEAKER & MONITOR/VIDEO OUTPUT AMPLIFIER GAIN:
Put a 10kohm resistor between pin 2 of U21 (@ R119) and AGND (@ R111).

LOWER THE STEREO OUTPUT AMPLIFIER GAIN & REMOVE BASS-BOOST:
Change C58, C59 from 100pF to 1000pF
Change R123, R130 from 10kohm to 15kohm
Change R59, R68 from 100kohm to 4.7kohm
Change C95, C90 from 0.033uF to 1kohm resistor

LOWER THE STEREO INPUT GAIN TO -10DB:
Remove R56, R57, R90, R92, C21, C81
Change R66, R87 from 100kohm to 1.8kohm
Put a 470pF capacitor in parallel with R67, R88
Put a 100pF capacitor in parallel with R66, R87

I'll come back to this sometime with a scope but will leave it here for now!

[exxos]

Where are you looking for c95 etc?

Normally bass boosts isn't what you think. I'd assume Atari increased the amp gain then just used a LPF.. just filtering the highs to boost the bass. Generally be a resistor and cap or something like that.

EDIT:
OK headphone amp not the image I posted earlier

[exxos]

Change C58, C59 from 100pF to 1000pF
Change R123, R130 from 10kohm to 15kohm
Change R59, R68 from 100kohm to 4.7kohm
Change C95, C90 from 0.033uF to 1kohm resistor

So theres 2 versions of the schematic then...

1.PNG (51.41 KiB) Viewed 41 times

2.PNG (324.83 KiB) Viewed 41 times

But unless I am misremembering something, increasing the resistors to 15K and 1K will increase the amplifier gain ?! Normally 0R is used for a buffer amplifier.. (no gain) that's what I did on the H5...

Also C122, R132 will act as a HPF and filter the bass anyway. I don't know how much that would be by, but it used to pi$$ me off when amps had 2.2uF and 22K as it rolled off around 200Hz and cut all the bass off! So I used to swap them out for 2.2K. I suppose it depends how good your headphones are about bass response as well.. But that input filter would really need checking and verifying..

Made myself chuckle..
You could end up with Atari noticing not good bass response because of the input rolloff filter, fixing it by boosting the bass... C-lab removing the bass boost.. then leaving lack of bass due to the input filter

I'm here all day......

[dml]

Yep - you'll notice though both versions (the 'G' rev is the modded rev) are using a nonstandard opamp in a nonstandard way.

The feedback isn't going to the -ve input - it's going to 'special' pin 1. Which is meant to be used with 'special' pin 8 or ground for a secondary/external (maybe more easily adjustable?) gain control.

Here pin 1 is connected in a kind of feedback config instead of using -ve input for that. Datasheet doesn't show that as one of the expected configs

But it might be deeper lore and I don't really feel like reading the opamp books again so happy to accept that it somehow works. -_-