Help MD1 audio w/ mCD & 32x buffer~amp circuit design (digital audio)

[humbug]

Basicaly I am trying to obtain digital audio from the MD1 w/ mCD & 32x.

Basicaly I have come to the issue that is I need the output impedance values for ym2612, Ricoh RF5C164 and the sound processing chip found on the 32x (name unknown to me) and/or the value in dB (decibels)

I've posted it up on spritesmind but it doesn't seem to have many members... where here there should be more... maybe. Here is the link for reference plus heaps more info.

http://gendev.spritesmind.net/forum/viewtopic.php?t=730

(best to read that thread backwards)

anyhelp would be much appreciated ^^

Thanks

Brendan

[Jorge Nuno]

The output impedance of the YM2612 is 1kOhm or more IIRC, you really should use a voltage follower/or a non-inverting amplifier configuration.
I don't remember it's output level, I could try to measure that if you want, but I need some program to set it's channels to the maximum level, or write the full scale to the DAC (I don't know much about the YM, software side, so don't count on me).

The PSG, on the other hand, is around 6dB (you wanted dBs, you get dBs ), and has like 1kOhm too.
Looking at the MD schematics, you'll see that the audio is mixed together with relations of 0.0431 for the PSG, 1 for the YM and 0.0468 for the external audio sources (both 32x and CD)

Assuming these figures are correct, (I hope they are ) you get that the output of the YM is around -27dB below PSG.

What I did with my audio amp was tapping all signals before going through the electrolithic capacitors (and removing those caps, and everything after them, resistors, the sony amp, everything), Amp'ed the YM outputs by 28, and mixed them with the straight PSG and Ext audio, "without" amplification (actually there's a trimmer to adjust the PSG level ), then added an output stage with a voltage follower, then removed the DC offset with 1500µF capacitors to the headphone jack.

Also you either need a polarization voltage of Vcc/2 as a gnd reference or use symetrical supplies for the opamps.


About the ADC filtering, a first-order low pass filter at 61kHz should be enough. It will give you 40dB attenuation for the 6.1Mhz replica. This means that the spectral components around 6.1Mhz (+/- 61kHz) will have a maximum of 7mV (ignoring noises and crap)

[humbug]

Thats exactly the information I was looking for! Thankyou so much

I'm very interested in the mod you did... as it seems like that was along the lines of what my last idea was lol.

Would you have any documentation do how you mixed them and what amp chip you used?

If you have any links to photos etc for what you did w/ discussion etc... I would be very interested in doing the same mod as you!

Thanks a million

Brendan

[humbug]

I just drew up a quick schematic... but I need someone to check over it.

here is what I came up with. Note: the total output voltage is not allowed to exceed 6.7v otherwise the signal will dump to the ADC overflow to gnd.



please post a markup

[Joe Redifer]

I don't understand why you would want to do this. The audio is all analog to begin with. Converting that analog to digital won't make it sound any better. The only thing you'd be able to extract true digital audio from is the CD... and only 44.1Khz redbook audio since most receivers do not support lower bitrates for digital audio input. It's kind of like converting composite video to HDMI... why??

[humbug]

The reason why is ground loops. Toslink will elimnate all ground loops connecting through my amp.

I know it will not make it "sound" any better, however, it will eleminate any RF interferance and any crappy impedance though RCA cables. It wont be noticable on crappy mono crt speakers, but i definetly hear it though my hi-fi unit at high volume.

i don't know if i can get direct digital audio from the CD unit, but I don't think the difference would be much. Besides, I already have a high quality CD player with toslink output.

Another reason is its just better to have everything in the one format...

anyways... those are my motives.

back to talking electronics?

*edit* its also like converting RGBs into 1080p. kinda like xrgb3 with the hdmi output???

[Jorge Nuno]

I used an OPA134PA (single opamp) for the PSG buffering and an OPA4134 (same as the 134 but it's 4 opamps-in-one) for the YM amping and final stage.
I have to build a new version, because this one's done in proto-board, it's too messy with crossing wires and huge resistor leads , and was also done on the fly without schematics following more or less these procedures:
-Mount this sub-circuit;
-Check if it works;
-Move-on to the next section, or fix errors.


Click-me

They were powered by a linear 6Volt regulator (7806), completely separated from the MegaDrive logic and analog supplies, and draws about 30-40mA of current.

Now I would do something different regarding the PSG: I would use separated opamps to fead each channel separately, even though their input would be the same: this reduces cross-signaling between audio channels.


I've read this CS5361's datasheet and what are you going to do with the digital audio stream that comes out?
Don't you have to do some protocol stuff or... I don't know anything about spdif, but I doubt you could connect this directly.
BTW Nice to know the chip it uses a Delta-Sigma typology to oversample the audio.

Oh yes there's more stuff like control bits and bit inversion crap in this protocol. This means you have to do some signal processing between the ADC output and the SPDIF connector.

[humbug]

probably just a silly question, but I'm abit sleepy at the moment, so I will think about it again tomorrow. But I'm going to ask it anyway

Now, I am a little confused about how you managed to set the opamp to 28dB gain. I brushed over the datasheet and noticed it was 120dB gain closed loop, but you didn't say how you wired it up. so that remains a mystery to me at the moment.

another thing, is that you said you used the same dip to do the final stage amping. what voltages where you getting from that? was it anything over 6v typ? I wouldn't want to amp too much then drain to ground before it reaches CS5361. I wanna try to keep the circuit as simple as possible for obvious reasons.

You also said you would use a seperate op amp to seperate the L and R streams which sounds like a good idea, but then wont the signal become too strong? how would I lower it enough to enter into the mixing circuit? Also, could someone please tell me where the PSG/32x/mCD is on the MD1 PCBA? I cant seem to find it >_<

Anyway...

Now, as far as CS5361 goes, I plan to couple that directly with a CS8406 then using the txp output directly to drive a toslink transmitter.

if you see any issues with that... please let me know.

anyway, im gonna sleep on it now.

Thanks so much for your input

cheers

Brendan

[Jorge Nuno]

Ahh so you have another chip, ok I somehow missed that part.
The output swing in my circuit is around 0.6 Vpp.

No, the final stage is just a buffer, It's a circuit with a gain of 0dB.
This is the same thing I did for the PSG. It's just to provide some isolation between the PSG output and the stuff that comes after it, and between each channel.

120dB closed loop gain is the maximum amplification you can acheive with this chip.
To control the opamp gain you have to do somekind of negative feedback, or else you'll end up with an open-loop circuit.

This is one type of the negative feedbacks you can make, and the one I did:
Vin is one of the audio ouputs of the YM2612, Vout is then mixed with "the rest"


In my case:

Rf = 100kOhm, Rg = 3.6kOhm
gain (linear, V/V) = 1 + Rf / Rg = 28.78
Gain (dB) = 20 x Log10(gain) = 29.18 dB

Also that ground is a "virtual ground", it's actually 3V in my setup. This is because, as I said before, my circuit is powered by a single +6V supply.

[humbug]

So how did you obtain 3v for your virtual ground? just a 6v reg with a resistor? how many buffers and virtual grounds could I power with 1x 6v reg?

Also, how would wone work out the final voltage output with all these op amps and resistors in the way? I added a resistor and a virtual ground to the "cirrus" circuit to allow for amplifying the signal up to 6volts max... would that work do you think?

So, here is the latest markup. please let me know if i should change anything.



Thanks so much for your help

cheers

[Jorge Nuno]

Each virtual ground is speparated from each other and is created by resistor dividers of the same value (1Kohm IIRC), between +6V and Gnd.

Your schematic has a few hard-to-tell-by-text errors, I'll try to make one later. And the second part of the circuit presented on the chip's datasheet is... weird.

[humbug]

So something like this.



is there a way I can simplify the ic chip in buffer circuit? it looks like there is too much crap in there.

could I connect annologue ground from ym2612 to ADC AIN- circuit input terminals?

what voltages are present after the amping/buffering and the summing of the signals? I want to know if it needs to have a second stage amp.

thanks

[humbug]

also what the heck is VQ :S

[humbug]

http://www.datasheetarchive.com/pdf-...DSA-136130.pdf



to be coupled with the previously selected spdif transmitter or CS8402 A coupled with 74HC74, 74HC00, LM 7805 and a 12.288MHz oscillator.

I am still unsure of which opamp i should use. I might consider the MC33202

please mark it up plus comments

*EDIT* I also noticed I should give the ADC its own voltage regulator for supply, eliminating the need for those 2 other capacitors.

[humbug]

so these are my limits. My circuit doesn't have a final stage amp, so one needs to be installed, or the last buffer to be modified as an amplifier? (i am not sure if that is possible?)

So I need to know the total voltage/impedance after summing the signals together. Jorge, you said the outswing of yours was 0.6v, where abouts in the circuit is that?

lets just say that the output voltage of the summing junction is 0.6v, is it just a case of adding an op amp of 6.67dB gain?

cheers ^^