TL072 is comparable to RC4558, so it should work fine.
Regarding PSG level, its all about adjusting 1.5K resistors. PSG is too quiet/inaudible when you have not removed PSG from the original circuit in most cases. When it comes to clones I don't quite know which is which... PSG usually is the loudest output on a machine.
From what I can tell, PSG is the quietest source as back when I tried to use the audio out of the Gen-X 2-in-1 NES/Genesis clone on the Genesis 3's audio amp, the PSG was just barely audible and needed to be amplified through 2 transistors to be properly balanced with the FM Synthesis. Even when using the raw FM audio out of the RetroN3's GOAC, the audio signal was so strong, it was VERY badly distorted. I didn't try the PSG straight out of the GOAC, but I presume it would be much weaker than the FM Synthesis.
So if the PSG is too quiet, I would need to replace the 1.5Kohm resistors with some weaker ones, correct?
Okay, I've given the Audio Mod a shot on my friend's RetroN3 Version 2 with 100Kohm resistors at the inputs instead of 22Kohm resistors, 0.1uF ceramic capacitors instead of 0.022uF capacitors as my friend wants to get deep bass on the RetroN3 Version 2, and two 2.2Kohm resistors in parallel instead of 1Kohm resistors as I can't find my 1Kohm resistors anymore(the two 2.2Kohm resistors in parallel give 1.1Kohms) and... success! Well, somewhat. The FM Synthesis sounds marvelous, but the PSG, while it is audible, is VERY badly distorted. I guess it's time to change those 1.5Kohm resistors and see what's the best resistor value to use instead.
those resistors won't do anything regarding distortion, they only control loudness. distortion comes before the transistors, so you got to deal with the 3.3 and 100K resistors.
Well, I've replaced the 1.5Kohm resistors with 3.3Kohm resistors, and while the distortion is gone, the PSG is too silent. In this situation, what should I do to the 3.3Kohm and 100Kohm resistors? Replace them with something stronger?
EDIT: I've figured out what needs to be done to get the PSG at the proper volume level on the RetroN3 Version 2! Along with replacing the 1.5Kohm resistors by 3.3Kohm resistors, the 3.3Kohm resistor would need to be replaced with a 7.5Kohm resistor. Wow... this thing sounds beautiful. TmEE, try selling your Audio Mod to Geniclone manufacturers. It's quite clear Hyperkin's engineers can't get things done right with Geniclones.
Excellent news! Maybe one day Geniclones can be a viable option for new people. :D
We'll see. I have a semi-functional GN Twin I could apply the Audio Mod to with the alterations I've done to get the Audio Mod working on the RetroN3 Version 2(it will work on a RetroN3 Version 1 as well as uses the exact same GOAC). The GN Twin and many other Genesis/NES 2-in-1s have TERRIBLE audio on the Genesis side(it still beats the hell out of AtGames' emulated garbage), so those could really benefit from TmEE's Audio Mod.
And now, I will record some audio samples of the Audio Modded RetroN3 Version 2 from Sonic the Hedgehog, Thunder Force II MD, Thunder Force IV and G-LOC Air Battle.
Good news for those who have VA7 Genesis Model 1s: the new Audio Mod is strong enough to output audio through the headphone jack. However, the sound does distort if you're using headphones with a volume knob and you set the volume knob to maximum volume. A direct connection from the headphone jack to your TV with a 3.5mm headphone to 2x RCA cable does not have any distortion whatsoever.
In case you're wondering, I bypassed the Genesis' volume slider as I'm not sure how to wire it up to the headphone jack. The sound could change with the volume slider added to the mix.
The circuit itself needed a few tweaks so the audio output matches that of a pre-VA7 Genesis Model 1:
-The 0.022uF capacitors were replaced with 0.1uF capacitors
-The 22Kohm resistors at the FM inputs were replaced with 25.5Kohm resistors as the FM was too loud
-The 750ohm resistors at the 32X PWM inputs were replaced with 300ohm resistors as the PWM was barely audible with 750ohm resistors
-The 1Kohm resistors were replaced with two 2.2Kohm resistors in parallel due to having no 1Kohm resistors(this gives 1.1Kohms, just like I did with the Audio Mod I built for the RetroN3 Version 2)
-One of the outputs has a 220uF capacitor instead of a 100uF capacitor(I ran out of 100uF and the closest I could find in my box of parts was 220uF - doesn't change anything in the audio output)
-The 4558 op-amp was replaced with a TL072 op-amp
-Sega CD audio inputs have been omitted(for now)
Needless to say, the audio the VA7 Genesis Model 1 puts out with the Audio Mod blows a stock VA7 Genesis Model 1 out of the water. And I shall now record some samples for y'all to listen to(I unfortunately had to remove the Audio Mod from my friend's RetroN3 Version 2 as the circuit is too big to fit in the RetroN3's case, so I need to rebuild it on a smaller board).
EDIT: And here are the samples - http://www.mediafire.com/?c3kcheauebttyux
The games used are:
-Thunder Force II MD
-Thunder Force III
-Thunder Force IV
-After Burner Complete
-G-LOC Air Battle
-Sonic the Hedgehog
The audio clips are encoded in FLAC audio, so make sure you have something which can play back FLAC files to listen to the clips.
Now to apply the mod on a VA1.8 Genesis Model 2 I've kept on the side for experiments.
Though I must add: I don't know if the fact I changed the 1Kohm resistors with two 2.2Kohm resistors in parallel would affect the amount of audible quantization noise, but compared to the stock audio amp on any Genesis with an ASIC-based YM3438, I can hear a lot more quantization noise with the Audio Mod than with Sega's stock amps(same with GOAC-based Geniclones). It's not as pronounced as on the YM2612, but I can hear it, especially in After Burner II, where the lack of quantization noise is most noticeable.
with 0.1µF caps the sound should be really dark and dull, the samples sound like you have 0.01 instead
In that case, I think all of my ceramic capacitors on the circuit are 0.01uF. I don't even know what the actual value of those capacitors is as I just found them in a failed video encoder circuit.
Though I think I did something wrong with the circuit, and it could either be caused by the capacitor values being wrong or something else. Basically, I can hear quantization noise(less than on a Genesis with a YM2612, but still audible) whereas with Sega's stock amplifier, I could hardly hear it. And when I plug in a 32X, the quantization noise becomes about as quiet as it should be on the system's modified YM3438. What's the problem here?
What exactly is that you describe as quantization noise ?
Basically some notes in certain games which are inaudible on a Genesis with a YM3438. After Burner II in particular comes to mind as that game has numerous notes which are practically inaudible on a Genesis with a YM3438. I'm gonna get my modified VA7 Genesis Model 1 and my VA3 Genesis Model 2 and record some audio clips from both(and I'll throw in my VA3 Genesis Model 1 as well for comparison purposes with what After Burner II sounds like on a real YM2612).
On a YM2612, notes tend to linger around longer than they do with the YM3438-based ASICs. I'm thinking maybe the YM2612 and YM3438 have different envelope generators?Quote:
Originally Posted by TmEE
This is caused by a bug in the DAC in YM2612. YM3438 is identical to YM2612 except DAC is fixed and it no longer makes one value near silence to be same as one value step above.
On YM3438 you get 0, 1, 2, 1, 0, 1, 2, 1 ....
On YM2612 you get 0, 2, 2, 2, 0, 2, 2, 2 ....
This is what causes the excessive noise, but it makes quiet notes more distinguishable in form of "trailing noise"
Is the actual discrete YM3438 the same as this? Or does it match the YM2612?Quote:
Originally Posted by TmEE
As for the samples, I'm currently uploading them and will post the link shortly.