I'm documenting a fix for missing / dodgy audio on a R-Type Leo M92 board, mainly because I could not find any schematics for the audio section.
Background
M92 board running in a new Astro city, recently fixed issues with missing sprites by replacing the 3-pin header and jumper J3, which had intermittent contact. After then running the game for a few hours started to have audio problems, with volume jumping around or being distorted. Audio tended to be missing when the board was powered from cold but would creep back in as the board warmed up, but would sound very clipped.
Investigation
Could not find any connection problems or suspicious solder joints on connectors.
Other boards have good audio in the cabinet so ruled out the volume pot in the Astro city power supply.
Wiped volume pot on game board back and forth after applying servisol, no improvement
Board volume pot gives good readings in circuit with a multimeter
Traced out the schematic for the audio circuit back to the DAC, see attached (main reason for this post)
Using a scope saw good audio at VR1 wiper but distorted / missing audio at the Jamma edge
Decided the audio amp IC201 (MB3730A) probably at fault, desoldered it
Attempted to test the amp on a breadboard using the 'typical circuit' given in the datasheet and feeding in a sine wave from a signal generator. Saw a lot of instability and oscillation even with quiescent input. Not 100% proof it was dodgy (because the datasheet demands some capacitor types be used that I didn't have) but enough.
Fix
Was slightly nervous about the replacement amp that arrived because it had some kind of double-printing and I've received a few fake ICs lately.
Fitted new amp.
Replaced all electrolytic caps because of warnings on other forums of leakage from these Elna caps. All the removed caps actually tested ok (capacitance and ESR), no obvious signs of leakage but 2 had legs that seemed a bit corroded and would not take fresh solder well.
A warning to anybody planning to re-cap their board: I'm ok at soldering but some of the caps were difficult to replace. Several were hard where a leg was connected to the ground plane and that often happens. C209 though was a real swine, the +ve leg connects to a 5V power plane. I was eventually able to remove the +ve leg but could not get enough heat into the area to be able to clear the hole (with a soldering iron, flux, pre-warming with hot air and a good solder sucker pen). I had to solder the replacement cap into the top half of the original solder and confirm continuity to both sides of the board with a meter.
All works ok with the new amp now.
I've attached a sketched schematic of the audio section. I've done my best but please use it with caution. I think it is correct although I don't know how the speaker connects on the Astro city side, i.e. is it driven in bridge mode (like it comes off the board) or is only one amplifier output used that returns to 0V. Hopefully it may help somebody if they have a similar audio problem.
Background
M92 board running in a new Astro city, recently fixed issues with missing sprites by replacing the 3-pin header and jumper J3, which had intermittent contact. After then running the game for a few hours started to have audio problems, with volume jumping around or being distorted. Audio tended to be missing when the board was powered from cold but would creep back in as the board warmed up, but would sound very clipped.
Investigation
Could not find any connection problems or suspicious solder joints on connectors.
Other boards have good audio in the cabinet so ruled out the volume pot in the Astro city power supply.
Wiped volume pot on game board back and forth after applying servisol, no improvement
Board volume pot gives good readings in circuit with a multimeter
Traced out the schematic for the audio circuit back to the DAC, see attached (main reason for this post)
Using a scope saw good audio at VR1 wiper but distorted / missing audio at the Jamma edge
Decided the audio amp IC201 (MB3730A) probably at fault, desoldered it
Attempted to test the amp on a breadboard using the 'typical circuit' given in the datasheet and feeding in a sine wave from a signal generator. Saw a lot of instability and oscillation even with quiescent input. Not 100% proof it was dodgy (because the datasheet demands some capacitor types be used that I didn't have) but enough.
Fix
Was slightly nervous about the replacement amp that arrived because it had some kind of double-printing and I've received a few fake ICs lately.
Fitted new amp.
Replaced all electrolytic caps because of warnings on other forums of leakage from these Elna caps. All the removed caps actually tested ok (capacitance and ESR), no obvious signs of leakage but 2 had legs that seemed a bit corroded and would not take fresh solder well.
A warning to anybody planning to re-cap their board: I'm ok at soldering but some of the caps were difficult to replace. Several were hard where a leg was connected to the ground plane and that often happens. C209 though was a real swine, the +ve leg connects to a 5V power plane. I was eventually able to remove the +ve leg but could not get enough heat into the area to be able to clear the hole (with a soldering iron, flux, pre-warming with hot air and a good solder sucker pen). I had to solder the replacement cap into the top half of the original solder and confirm continuity to both sides of the board with a meter.
All works ok with the new amp now.
I've attached a sketched schematic of the audio section. I've done my best but please use it with caution. I think it is correct although I don't know how the speaker connects on the Astro city side, i.e. is it driven in bridge mode (like it comes off the board) or is only one amplifier output used that returns to 0V. Hopefully it may help somebody if they have a similar audio problem.
