Quantcast

Page 1 of 2 12 LastLast
Results 1 to 15 of 26

Thread: Fixing the sound of a VA2.3 Model II Genesis without a Mega Amp

  1. #1
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Sega CD Fixing the sound of a VA2.3 Model II Genesis without a Mega Amp

    Update: Cleaned up the instructions and uploaded updated pictures. (And apparently I futzed those pics up too... bloody hell. I'll fix them AGAIN once I get more free time.)
    Hi all,

    If you're hankering for better sound out of your Model II VA2/VA2.3 console but are unwilling or too broke to splurge on a Mega Amp (or just prefer the challenge of changing the built-in audio circuit), you can modify the on-board mixing/amplifier circuit and achieve a noticeable improvement in sound quality. All capacitors and resistors indicated are 2.00mm x 1.25mm size / size code 0805, 2012 metric. I haven't tested this with a 32X or Sega CD so I don't know how they'll sound after the fixes, but by itself it sounds surprisingly clear and crisp, much better than it did before at any rate. The VA2 follows the same guidelines, but the physical locations of the pieces to change is different. Just match up the part numbers and you'll be fine (they're the same part numbers on both the VA2 and VA2.3).

    A Mega Amp 2.0/2.1 still has considerable advantages over the built-in audio circuit, such as potentially even higher sound quality depending on the op-amp(s) you use, more flexible filter options, possibility of trivially adding a headphone output (Pro version) and adding a dedicated 7805 to further reduce noise, but this should work in a pinch. Hope this helps someone out!

    Big thanks to Ace and JamesF for a lot of insight and help on this, the advice on op-amps, and for having better ears than me.

    1. Replace the 27K resistors at R23 and R25 with 24K (this reduces the YM2612 pre-amp gain from 8.18 to 7.27)
    2. Replace R31 with a bridge or 0-ohm jumper
    3. (Only if needed) Replace R2 with a 750 ohm resistor if the PSG is too loud afterward (not always needed; seems to vary by console)
    4. Replace the BA10358 at IC3 with a 4558 (cheapest), TL072 or M33078 (best), and remove the 4.7K resistors at R12 and R13 (these are used to bias the original '358 into class-A, but they are pull-ups and thus induce DC offset in the sound; the 4558 and better have class-A or AB outputs so they don't need these pull-ups anyway)
    5. Remove the capacitors at C72 and C73, and add 120pF caps between pins 1-2 and 6-7 of IC3 (this replaces the old passive filter with a much more effective active bandwidth limiter with a 55 kHz corner, preventing it from adding a fourth filter stage on the FM!)
    6. Replace the 1800pF caps at C21 and C23 with 1600pF (4.9 kHz output filter corner) or 1200pF (6.63 kHz output filter corner).

    Now depending on your preference, if you don't mind getting rid of the Sallen-Key filter on the FM to get more frequency response:
    6. Replace the 1300pF caps at C56 and C58 with 1000pF caps
    7. Remove the 5600pF caps at C20 and C22
    This removes the Sallen-Key stage, leaving the FM effectively second-order filtered with a rather high 7.9 kHz first-stage corner frequency, sounding considerably clearer and sharper afterward. The rest of the sound (PSG, 32X, Sega CD, etc.) is first-order filtered, and will be filtered at the 4.9~6.63 kHz corner dictated by the caps at C21 and C23. *Don't* remove C56 and C58, these caps are needed internally by the 315-5684.

    But if you prefer the sound of a 3rd-order filter (possible noise reduction benefits) and want to preserve the Sallen-Key FM filter, but simply want a bit clearer FM response:
    6. Replace the 1300pF caps at C56 and C58 with 1000pF
    7. Replace the 5600pF caps at C20 and C22 with 4700pF
    These two cap pairs are important, as they will keep the Q factor, phase margin, etc. close to the original setup, but bumps the corner frequency up from 5.8 kHz to 7.3 kHz making things sound noticeably less degraded.

    Afterward the FM and PSG seem to be about right in volume balance, and the rest of the sound is also slightly clearer as well.
    Attached Images Attached Images
    Last edited by Eep386; 12-13-2017 at 06:04 AM. Reason: MOAR corrections... also added Mega Amp's 750 Ohm PSG pulldown (credits to Ace)

  2. #2
    State Laced Master of Shinobi ComradeOj's Avatar
    Join Date
    Dec 2012
    Location
    New Mexico, USA
    Age
    23
    Posts
    1,146
    Rep Power
    47

    Default

    Nice! Got any sound samples? I would be curious to hear how it compares to a mega amp.

    I have a VA2.3, and fitted it with an unfiltered mega amp. I might have been tempted to try this solution if I knew of it at the time.
    Modded consoles:
    Master System (v7040) with s-video & direct AV out
    Model 1 with 10mhz overclock & halt switches
    Model 1 with 10mhz 68010
    Model 2 VA2.3 with unfiltered Mega Amp, & s-video
    Model 3 VA1 with compatibility fixes & s-video
    32X with s-video
    Visit my web site at www.mode5.net
    Or my collection of homebrew Genesis games, programs, and music on SEGA-16!

  3. #3
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Default

    I'll try to record some samples, though I can't say it'd be fair to expect very good quality recordings (at least, none without a ton of noise) due to my poor equipment...

  4. #4
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Default

    Mother of derp! I meant to say change R31, not R55, to a 0-ohm jumper. I've corrected the instructions accordingly.

    Do NOT change R55!!!! If you did, put the original resistor that was there, back!!!

  5. #5
    Creator of the Mega Amp Raging in the Streets Ace's Avatar
    Join Date
    Jun 2008
    Location
    Montreal, QC, Canada
    Age
    27
    Posts
    3,713
    Rep Power
    41

    Default

    You might want to consider replacing the 2nd-order low-pass filter capacitors by 270pF capacitors. From messing with the 315-5684 and Sega's awful VA0-VA1.8 Genesis Model 2 amp (with TL074s replacing the LM324s, etc.), it seems the ideal way to keep that 2nd-order low-pass filter and make it much harder to perceive is to set the filter to a cutoff frequency close to the YM2612's sample rate. With 270pF capacitors in the 2nd-order filter, this will give a cutoff frequency just past the chip's sample rate of 53KHz, that being around 58.95KHz, which the 315-5684 can handle pretty well (a bit on the limit as, at least with the YM3438, replacing the 560ohm resistors on the power pins with a direct connection to 5V and Ground on a VA3 board will cause crackling, so maybe 330pF capacitors would be better for higher gain). The point is to keep the filter as far outside the audible audio spectrum as possible without exceeding the bandwidth of the amplifier.
    For the final output stage, I would replace the 1800pF capacitors with 1600pF capacitors. This will give a cutoff frequency matching the basic Mega Amp 2.0 design. Also, that 10358 op-amp has to go; put a TL072 there at the very least and adjust the VRef to around 3V.
    HATES ATGAMES WITH A PASSION


    Mega Amp: An all-new audio circuit for your Sega Genesis/MegaDrive and clones.

    Note: If you want to contact me on Skype, identify yourself or your contact request will be rejected.

  6. #6
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Smile

    OK I finally have some samples of the mod up. My Goodwill-special recording equipment really doesn't do it justice (the improvement is MUCH more obvious over the TV), and you'll probably notice an evil amount of noise on the recording (that's my $6.20 firewire A/V recording pod) in addition to the usual buzzing etc. you get from the A/V output of a typical Model 2. I did remove the DC offset though in Audacity beforehand though.

    I modded a VA2 in the same manner (the part designations are 100% the same, only the physical location of the parts are different on the board). You'll notice it sounds considerably less muddy after the mod, and the PSG is now clear and undistorted, and its filtering is much closer to the FM's (though I wonder if 180pF is a little too heavy, perhaps 82pF would be better after the rest of the BA6166 is modded). I was short a few components so I had to substitute a 510 ohm for the 560/680 ohm pull-down for the PSG (that's probably why the PSG is just a touch too quiet here), and i used 4700pFs for the caps at C72 and C73.

    Before:https://www.dropbox.com/s/wsk04ggs87...ilter.ogg?dl=0
    After:https://www.dropbox.com/s/la3zszph93...ilter.ogg?dl=0

    @Ace: On a different VA2.3 I tried replacing the BA10358 with its direct drop-in upgrade the BA3404, but that didn't help anything. If anything it sounded slightly worse; I think it might have needed a slightly higher VRef (like 3v). I do have some LM4558 clones on hand though from a different project; I could try bumping VRef up a bit and giving those a shot to see if it's really worth the trouble to replace the BA10358.

    What is funny though, is a straight-up LM358 will sound better than a BA10358, even though they're supposedly the same part! The BA10358 has even poorer slew rate and GBW.
    Last edited by Eep386; 09-18-2017 at 07:45 PM.

  7. #7
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Default

    Nvm, 1200pF on C72 and C73 begins to introduce distortion. 4700pF seems to be as low as we can go there before things begin to distort noticeably.

    @ComradeOj: For reference, this would be most comparable to the 'filtered' option of the Mega Amp.
    Last edited by Eep386; 09-18-2017 at 07:17 PM.

  8. #8
    Creator of the Mega Amp Raging in the Streets Ace's Avatar
    Join Date
    Jun 2008
    Location
    Montreal, QC, Canada
    Age
    27
    Posts
    3,713
    Rep Power
    41

    Default

    Why are you using a filter capacitor on PSG? Leave that out, the filter on the 315-5684's final stage will take care of that.

    After looking a little more closely at the filter setup on the 10358, it looks as if Sega is using a passive RC filter setup using the 3.3Kohm series resistors on the YM2612's outputs and directly connects that to the inverting op-amp. I'm a bit iffy on that setup; mind testing what would happen if you use an active low-pass filter instead by soldering a capacitor in parallel to each of the feedback resistors? Use the best op-amp you can get your hands on and keep the cutoff frequency as high as possible.
    HATES ATGAMES WITH A PASSION


    Mega Amp: An all-new audio circuit for your Sega Genesis/MegaDrive and clones.

    Note: If you want to contact me on Skype, identify yourself or your contact request will be rejected.

  9. #9
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Default

    Sure, I'll try adding some active filters there. I don't imagine the result will be much different, but you never know. I could slap some 82pFs there (since I got a bunch at a reasonable price) and see how that sounds.

    At least Sega correctly wired the YM2612's MOL/MOR outputs to have a pull-down and a capacitor *before* those 3.3K voltage dividers. XD Should I replace those with a jumper before trying the active filter? What about the 10nF/4700pF capacitors at C72 and C73, should I leave them in or pull them out?

    I did however discover that the BA10358 is getting a 2.5 volt VRef straight from the ASIC, it's not deriving VRef from the RGB encoder (thank goodness!).

    Regarding the PSG filter cap, I heard that systems using the 5684/BA6166 tend to have somewhat uneven PSG filtering. 180pF was indeed extremely muffled, 82pF sounds much less so. I made the PSG filter optional anyway, but for the sake of testing I included it. Going by my Alfred E. Neuman ears I don't hear any weird PSG filtering issues with the capacitor fitted. Or is this problem really applicable to the VA3/4?
    Last edited by Eep386; 09-18-2017 at 10:22 PM.

  10. #10
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Default

    OK, I tried your idea of using active low-pass filters instead. On the stock BA10358 it sounded absolutely *nasty* with 36pF low-pass active filters, like it's extremely smeared and over-amplified. On the BA3404, however, it sounds much better, though I had to drop the gain resistors to 22K to fix some clipping problems. I might have to adjust it further, though - it sounds a little distorted on some sounds. It's funny how much different the BA3404 is from the BA10358, even though the BA3404 is really just a 50% less shitty LM358.

    Not sure if this is an improvement over the passive filters on the board though yet. Or do I need a better op-amp still, or am I just putzing something up in my usual fashion?
    Last edited by Eep386; 09-19-2017 at 12:01 AM.

  11. #11
    Creator of the Mega Amp Raging in the Streets Ace's Avatar
    Join Date
    Jun 2008
    Location
    Montreal, QC, Canada
    Age
    27
    Posts
    3,713
    Rep Power
    41

    Default

    Quote Originally Posted by Eep386 View Post
    At least Sega correctly wired the YM2612's MOL/MOR outputs to have a pull-down and a capacitor *before* those 3.3K voltage dividers. XD Should I replace those with a jumper before trying the active filter? What about the 10nF/4700pF capacitors at C72 and C73, should I leave them in or pull them out?
    Slight correction here: those 3.3Kohm resistors are not part of a voltage divider, they're part of the inverting amplifier made with the 10358. If you're going for an active filter, remove C72 and C73, but DO NOT remove the 3.3Kohm series resistors. This will mess up your gain.

    Quote Originally Posted by Eep386 View Post
    I did however discover that the BA10358 is getting a 2.5 volt VRef straight from the ASIC, it's not deriving VRef from the RGB encoder (thank goodness!).
    Good to know. Shouldn't be too difficult to set up a voltage divider and filter capacitor to get a higher/lower VRef.

    Quote Originally Posted by Eep386 View Post
    Regarding the PSG filter cap, I heard that systems using the 5684/BA6166 tend to have somewhat uneven PSG filtering. 180pF was indeed extremely muffled, 82pF sounds much less so. I made the PSG filter optional anyway, but for the sake of testing I included it. Going by my Alfred E. Neuman ears I don't hear any weird PSG filtering issues with the capacitor fitted. Or is this problem really applicable to the VA3/4?
    That is not was is meant by "uneven filtering." It's caused by the overly aggressive second-order low-pass filter on the FM inputs to the 315-5684, and you will notice the same thing if you simply replace the 10324s/C324Gs/LM324s on the VA7 Genesis Model 1 to VA1.8 Genesis Model 2. Take out that capacitor.

    Quote Originally Posted by Eep386 View Post
    OK, I tried your idea of using active low-pass filters instead. On the stock BA10358 it sounded absolutely *nasty* with 36pF low-pass active filters, like it's extremely smeared and over-amplified. On the BA3404, however, it sounds much better, though I had to drop the gain resistors to 22K to fix some clipping problems. I might have to adjust it further, though - it sounds a little distorted on some sounds. It's funny how much different the BA3404 is from the BA10358, even though the BA3404 is really just a 50% less shitty LM358.
    Dude, come on, 36pF is FAR too little capacitance for such a small feedback resistor. Your low-pass filter has a cutoff frequency of around 200.953KHz, that is far too high. Use a 120pF capacitor and 24Kohm resistor in the feedback, that will give a cutoff frequency of around 55.263KHz. It's just around the cutoff of the second-order low-pass filter if you use 270pF capacitors on both filters.

    Keep this mathematical formula in mind when calculating low-pass filter cutoff frequencies:
    Fc = 1/(2πRC)

    Insert your resistor and capacitor values into R and C to get your cutoff frequency, but make sure your units are in Ohm and Farad. uF is 10-6 Farad, 10-9 for nF and 10-12 for pF. Taking the example of a 24Kohm resistor and 120pF capacitor, that would be 24000ohm and 120x10-12 Farad. If you run this through a calculator, you will get your frequency in Hz, which would be 55263.133Hz (there are more decimal numbers, but I generally prefer rounding to 3 significant digits - it's up to you how precise you want your calculations to be).

    Quote Originally Posted by Eep386 View Post
    Not sure if this is an improvement over the passive filters on the board though yet. Or do I need a better op-amp still, or am I just putzing something up in my usual fashion?
    If the BA3404 can't handle a 55.263KHz low-pass filter, solder in a TL072 and give it a 3V VRef.
    HATES ATGAMES WITH A PASSION


    Mega Amp: An all-new audio circuit for your Sega Genesis/MegaDrive and clones.

    Note: If you want to contact me on Skype, identify yourself or your contact request will be rejected.

  12. #12
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Default

    Quote Originally Posted by Ace View Post
    Dude, come on, 36pF is FAR too little capacitance for such a small feedback resistor. Your low-pass filter has a cutoff frequency of around 200.953KHz, that is far too high. Use a 120pF capacitor and 24Kohm resistor in the feedback, that will give a cutoff frequency of around 55.263KHz. It's just around the cutoff of the second-order low-pass filter if you use 270pF capacitors on both filters.
    Oh, okay. I guess that means I'll be ordering some 120pF capacitors, as I don't have any on hand. Interestingly even with the 36pF caps I could tell there was less noise on the output, so the active filter is still an improvement in some ways, even though the filter cutoff was way outta whack.

    Oh yeah, just for fun I tried disconnecting the VRef and feeding the BA3404 2.98v from a custom VRef, and it just made everything distorted. But then, I was still using those janky 36pF capacitors, so obviously that wasn't going to help much of anything.

    As you can probably tell I'm still pretty new to this, but I'm trying my damnedest to learn.

    Quote Originally Posted by Ace View Post
    That is not was is meant by "uneven filtering." It's caused by the overly aggressive second-order low-pass filter on the FM inputs to the 315-5684, and you will notice the same thing if you simply replace the 10324s/C324Gs/LM324s on the VA7 Genesis Model 1 to VA1.8 Genesis Model 2. Take out that capacitor.
    Alright. I wasn't quite sure there if it was a fault of the 5684/BA6166 or something else. At what location are the second-order filter caps, are they the caps I replaced with 1000pFs, C56 and C58?

    Oh yeah, I don't think I'll be ordering any more BA3404s in the future. They're more expensive than RC4558s and TL072s, though they are indeed much easier to adapt to existing LM358 circuits without changing a lot of external components.
    Last edited by Eep386; 09-19-2017 at 01:06 PM.

  13. #13
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Default

    Well I finally got off my lazy arse and swapped the BA3404 with one of my RC4558s. I left the passive filter in there for now until my 120pF caps arrive. It works well for the most part, but some sounds are now clipping a bit. Poring over the datasheets I discovered the RC4558's input offset current is WAY higher than the BA10358 and TL072's so that's obviously not going to work as a 'drop-in' replacement!

    Fortunately I have some TL072's on order, so I'll give those a shot (and raise VRef up to 3v) when they arrive.
    Last edited by Eep386; 09-19-2017 at 04:08 PM.

  14. #14
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Default

    On the RC4558-fitted system I replaced C56 and C58 with 560pF caps (I don't have any 270pF caps on hand, 560pF was the closest value I have) and now I get this sound from it: https://www.dropbox.com/s/2nabahk51k...lters.ogg?dl=0

    You might hear a little bit of distortion on some of the highs. (Gain or bias issue? Some other derpy thing I overlooked?) I'm still using 1200pF output mix caps - I think I hear why you recommend 1600pF (aka 1.6nF) output mix filter caps, Ace... I'll try to get some from an electronics store in Sacramento today.

    My horrible capture equipment doesn't want to go over 32kHz sadly so you're not really hearing how much it opens up.

  15. #15
    Road Rasher
    Join Date
    Sep 2014
    Location
    Conservative enclave in California
    Posts
    252
    Rep Power
    8

    Default

    Ace! My 120pFs came in today, so I removed the 4.7nF caps on C72 and C73, and soldered the 120pFs across the -IN and OUT terminals on both sides of the RC4558. FM sounds mighty good actually, very clean and clear (well, as clear as the 560pF pre-mix filter stage and 4700pF FM filters allow). The crackling/distortion on the unadjusted RC4558 output seems almost completely eliminated too; I guess the '4558 is okay for this after all? (I'd put in one of my new TL072s, but the pads are starting to get pretty brittle from me futzing with them. Fortunately I still have extra VA2.3s to test on.) I hear very little noise too. I'll have to make _another_ order to get some 270pF caps for the pre-mix filter.

    Lemme get some samples up for you to listen to (bear in mind my equipment hard-limits to 32kHz so it will sound more distorted and muffled than it really is)...

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •