Line level of the Amiga 600 (and possibly other models)

[ugrnm]

Originally Posted by Stedy

Hi,
To boost the output change R325 and R335 from 390R to either 470R or 510R. This will get you a peak voltage of 0.9408V for 470R resistor or 0.9702V for 510R.

I don't understand how you end up with these values. If I aim at something around 3.472Vpp, R325 and R335 should be much higher, no?

Is there something wrong with my reasoning? See attached.

[ugrnm]

I'm not sure it was clear what I was trying to do...

But...

In the end I changed all the components in the audio circuit path (new op amp and 1% resistors, new caps, etc). If anything it was a good way to learn about SMD package size, as well as resolder/solder these. The result is that my slight unbalance between L and R is gone and the overall audio signal is a bit louder.

Still not loud enough for my original plan, so I put a 1.5K at R335, and I'm getting close to "Pro" line level, all good.

I've also replaced the big 1ohm resistors on the video and audio power circuit with equivalent ferrite beads. I am not sure if it's snake oil, or if it's my maintenance work on the audio circuit, but now even at max volume/gain on the mixer, it's super silent. Before I had a slight hum, and could hear the optical mouse activity when moving it.

Been playing mods non stop for the past 2 weeks, everything is fine, stable, loud and clear

[pandy71]

Originally Posted by ugrnm

I've also replaced the big 1ohm resistors on the video and audio power circuit with equivalent ferrite beads. I am not sure if it's snake oil, or if it's my maintenance work on the audio circuit, but now even at max volume/gain on the mixer, it's super silent. Before I had a slight hum, and could hear the optical mouse activity when moving it.

You could record (PC card or external audio interface) noise level (silence sample) with active system (floppy, mouse activity) before and after modifications - this would give final answer for all doubt's (modern PC cards usually provide more than 80dB SNR so they can be used as decent objective audio analyzer and any modification can be verified in objective way).

Anyway as overall congratulations!

[ugrnm]

Thanks!

I know I should have measured things more seriously, I did not think of just recording with a soundcard... I was having way too much fun working on all this.

I did notice though that before I made all these changes, with all the gain up on the mixer, the Amiga noise floor was appearing on the meters at -30dB, bleeding a bit on the -20dB. And now, *nothing*

[pandy71]

Originally Posted by ugrnm

Thanks!

I know I should have measured things more seriously, I did not think of just recording with a soundcard... I was having way too much fun working on all this.

I did notice though that before I made all these changes, with all the gain up on the mixer, the Amiga noise floor was appearing on the meters at -30dB, bleeding a bit on the -20dB. And now, *nothing*

this is really nice - my assumption is that in original design 1ohm resistors together with some decoupling capacitance created low pass filter but i have impression that audio quality was never main goal for Amiga designers - 8 bit was anyway spectacular in those times so quality was less relevant but today our knowledge substantially increased so this is default scheme, ferrite beads, proper PCB layout to keep signal integrity, proper ground planes (in case of audio frequently local ground plane connected to rest of system also by ferrite bead).
Well done, if you share details perhaps other will follow your path.

[ugrnm]

Yes, my plan is to gather my notes and document the changes I made! I will bump this thread when done.

[amigasith]

I'd be interested in the details, too. Better sound never hurts

[Amiga1992]

thanks to pandy for pointing me to this thread, i am very interested by this topic too, since not only my amigas have really low audio volume, but removing C331 and C321 have made the A600s some times distort on the high end. I would like to replace C331 and C321 with the best possible values as well to apply any other changes that may make the filter AND line level better.

So, any updates, ugrnm?

[8bitbubsy]

A600 to A1200 low-pass filter modification:

C321 & C331: change from 0.1µF to 6800pF
R321 & R331: change from 360 ohm to 680 ohm

This will adjust the 6dB/oct low-pass filter's cutoff from a nominal ~4.4kHz to a nominal ~34.4kHz so that it's exact to that of A1200.
Do not attempt to just remove parts instead of replacing them, like some people are doing. The filter is there for a reason, even in the A1200 where the cutoff is well above human hearing. It filters away volume PWM artifacts.

If you want to increase the gain, you have to change other components too.

Also to the other people in this thread: Don't trust the schematics floating around on the internet. Almost all of them are wrong for the audio filter RC values. The nominal cutoff is never above 4kHz on rev2 A600, even though the schematics imply so. Always check your own motherboards.

EDIT:

Originally Posted by Amiga1992

[...] but removing C331 and C321 have made the A600s some times distort on the high end. [...]

Sigh, who told you to do this?