John Bennett
Senior Member
I've had a Double Dragon for many, many years that had a graphics fault. Another Double Dragon came up for sale with an audio fault, so I acquired that - see if I can make 2 working boards from 2 broken boards
. Plus I thought I'd try and do it a bit more scientifically than I used to decades ago (i.e. randomly jabbing wires onto things).
So, the audio issue.... PCM sounds OK, but FM not working and making a horrid noise. Fingers over the op-amps before the audio amp revealed they were stinking hot. Confusingly, the board had recently been repaired and there was flux, suggesting these chips had been done (they weren't originals).
So some new op-amps and all was fixed!
I did think it was the YM3012 at first as the outputs were stuck, but it might've been because it uses a pair of op-amps as part of a local feedback circuit..
I'll change the highpass capacitors as they're a bit tatty, but not quite sure what killed the old op-amps. Perhaps something odd happened with the -5V rail, perhaps certain op-amps are more delicate... not sure (TL082s were fitted, which don't seem to have any obvious shortfalls vs the original 4358s). Something to keep an eye on anyway.
The other board has always had a problem drawing one of the object layers. However when I booted it up (been years in storage), all was blue.
Probing the signals, red and green were off. Looking at the schematics, red and green share common circuitry at the pallet RAM stage - 74LS245 and a TMM2018 RAM chip. So a fault will take out both those colours, but leave blue. After much pain, I socketed the RAM on both games and did a swap - it was the RAM, so that's on the shopping list.
So just the object issue remaining:
I had noticed that touching the RAM and EPROM in the top corner of the main PCB adjusted the screen, so a starting point. These were part of the MCOL circuit, which is the layer at fault.
A bit of a probe around revealed one of the VPOS lines was permanently low to the EEPROM.
Oddly, it was fine at the inter-board ribbon cable.
Following the track revealed a small break (I've scraped it back to copper here, but the dull indent in the centre is the break).
Joining the track and all was good:
Rather satisfying
As an aside (as I mentioned in the other thread), the CGA-VGA converter struggled with both boards. Turned out to be due to DD's noisy sync signal upsetting the 3.3V input to the converter. An inlike 470 ohm resistor worked as a temp bodge (although the converter is only used for my test harness)
John Bennett2017-12-14 17:41:38
So, the audio issue.... PCM sounds OK, but FM not working and making a horrid noise. Fingers over the op-amps before the audio amp revealed they were stinking hot. Confusingly, the board had recently been repaired and there was flux, suggesting these chips had been done (they weren't originals).
So some new op-amps and all was fixed!
I did think it was the YM3012 at first as the outputs were stuck, but it might've been because it uses a pair of op-amps as part of a local feedback circuit..
I'll change the highpass capacitors as they're a bit tatty, but not quite sure what killed the old op-amps. Perhaps something odd happened with the -5V rail, perhaps certain op-amps are more delicate... not sure (TL082s were fitted, which don't seem to have any obvious shortfalls vs the original 4358s). Something to keep an eye on anyway.
The other board has always had a problem drawing one of the object layers. However when I booted it up (been years in storage), all was blue.
Probing the signals, red and green were off. Looking at the schematics, red and green share common circuitry at the pallet RAM stage - 74LS245 and a TMM2018 RAM chip. So a fault will take out both those colours, but leave blue. After much pain, I socketed the RAM on both games and did a swap - it was the RAM, so that's on the shopping list.
So just the object issue remaining:
I had noticed that touching the RAM and EPROM in the top corner of the main PCB adjusted the screen, so a starting point. These were part of the MCOL circuit, which is the layer at fault.
A bit of a probe around revealed one of the VPOS lines was permanently low to the EEPROM.
Oddly, it was fine at the inter-board ribbon cable.
Following the track revealed a small break (I've scraped it back to copper here, but the dull indent in the centre is the break).
Joining the track and all was good:
Rather satisfying
As an aside (as I mentioned in the other thread), the CGA-VGA converter struggled with both boards. Turned out to be due to DD's noisy sync signal upsetting the 3.3V input to the converter. An inlike 470 ohm resistor worked as a temp bodge (although the converter is only used for my test harness)
John Bennett2017-12-14 17:41:38
