After fixing JB's Space Invaders 91 PCB (1 dead eprom) and Moon Cresta (dead 555's etc.) I've found motivation to start repairing some boards and actually writing them up, I'm just using a probe and piggybacking, plus pulling some usual suspects as the desolder gun seems to be working really well. IC testing is being done where possible with a wellon VP-280 eprom programmer that also tests most rams and 74 and 4000 series logic. Ive got a scope but its buried in the loft at the moment.
I started off with a Juno First;
Firstly replaced visibly bad 1000uF capacitor C30
Powered board up - played but with following problems;
No Sound
Horizontal lines on screen 'jailbars'
Sprites were blocks of colour
Jailbars - I knew that, like williams boards, Juno First uses banks of 4116 ram for the screen ram (NEC D416 in this case) and that these are a common failure. As they only have one output bit it is fairly easy to probe each one on pin 14 to check their output. 2 Of the rams had dead outputs. Replacing them fixed the jailbar issue. Since found with another Juno board the 4116's can partially fail giving a working active low but not being able to drive the line to active high. Should probably dig out the scope to probe these in future but was lucky this time round.
Sprites - The sprite data is contained in the bank of three roms on the top board. Probing these I found the chip select and output enable lines were all stuck, I'd expect them to be pulsing. Looking at the schematics showed these lines were controlled by a 74LS139 at 7H. The inputs looked good and the chip was a rusty looking fujitsu so I pulled and tested it. It failed. Replacing it fixed the sprites.
Onto Sound - Finger test brushing the amp showed that was working, the z80 was active so something was happening. I noticed a stuck pin on the 374 at 3H which failed once pulled, but replacing it didnt change the lack of sound. I then noticed some address lines of the Z80 were stuck low rather than pulsing. For the higher address lines that could be expected given the small program code it is runnning, but A8 was suspect, especially as A9 was pulsing.
I pondered this for a bit, probing related IC's and finding nothing suspicious, then decided to pull the two 2114 RAMs at 3N/M - just because I was running out of ideas and they have a fairly high failure rate, and could explain the z80's behaviour.
With the RAM swapped sound reappeared, although some sounds were missing, most noticeably the fire sound. Looking at the schematics again showed sound was generated by an ay3910 and an 8039 microcontroller. It made sense that the AY was generating the synthy noises and that the fire was a more noise based sound that the 8039 was generating.
The 8039 was linked the eprom at 4P, and the eproms output enable signal was generated by pin 9 of the 8039, which was dead. I'd verified the eprom so decided to pull the 8039. Swapping it with a working one off a Gyruss brought back the missing fire sound, but it was very muffled. The only chip between the 8039 and the amp was a NEC uPC451 op amp. Didnt have any of these but found an equivalent LM324. Swapping this brought the fire sound back to proper volume.
Bit of a marathon but nice to have it up and running.
I started off with a Juno First;
Firstly replaced visibly bad 1000uF capacitor C30
Powered board up - played but with following problems;
No Sound
Horizontal lines on screen 'jailbars'
Sprites were blocks of colour
Jailbars - I knew that, like williams boards, Juno First uses banks of 4116 ram for the screen ram (NEC D416 in this case) and that these are a common failure. As they only have one output bit it is fairly easy to probe each one on pin 14 to check their output. 2 Of the rams had dead outputs. Replacing them fixed the jailbar issue. Since found with another Juno board the 4116's can partially fail giving a working active low but not being able to drive the line to active high. Should probably dig out the scope to probe these in future but was lucky this time round.
Sprites - The sprite data is contained in the bank of three roms on the top board. Probing these I found the chip select and output enable lines were all stuck, I'd expect them to be pulsing. Looking at the schematics showed these lines were controlled by a 74LS139 at 7H. The inputs looked good and the chip was a rusty looking fujitsu so I pulled and tested it. It failed. Replacing it fixed the sprites.
Onto Sound - Finger test brushing the amp showed that was working, the z80 was active so something was happening. I noticed a stuck pin on the 374 at 3H which failed once pulled, but replacing it didnt change the lack of sound. I then noticed some address lines of the Z80 were stuck low rather than pulsing. For the higher address lines that could be expected given the small program code it is runnning, but A8 was suspect, especially as A9 was pulsing.
I pondered this for a bit, probing related IC's and finding nothing suspicious, then decided to pull the two 2114 RAMs at 3N/M - just because I was running out of ideas and they have a fairly high failure rate, and could explain the z80's behaviour.
With the RAM swapped sound reappeared, although some sounds were missing, most noticeably the fire sound. Looking at the schematics again showed sound was generated by an ay3910 and an 8039 microcontroller. It made sense that the AY was generating the synthy noises and that the fire was a more noise based sound that the 8039 was generating.
The 8039 was linked the eprom at 4P, and the eproms output enable signal was generated by pin 9 of the 8039, which was dead. I'd verified the eprom so decided to pull the 8039. Swapping it with a working one off a Gyruss brought back the missing fire sound, but it was very muffled. The only chip between the 8039 and the amp was a NEC uPC451 op amp. Didnt have any of these but found an equivalent LM324. Swapping this brought the fire sound back to proper volume.
Bit of a marathon but nice to have it up and running.