Dragon 32 board restoration : 32k MkII PC10087 Issue 6 (1983)
Incoming fault: powers up then goes to blank screen : Visual inspection & repair
|
Resistor between IC10(4) and IC12(10) ? This is the 'Sony Colour Banding' mod, though it helps anyway so was left in place... sort of! The mod at the IC10 end had solder bridged between the pin and the pass through track so the resistor was repositioned to use a nearby via instead of the IC pin, then the bridge was cleared. (No bearing on the known power up issue) |
|
|
IC20 (MC14529BCP) : the underside of the board was a bit mangled by previous solder work, again not going to be part of the main issue but I didn't like it, so I de-soldered the IC, cleaned the board, installed new via rivets then fitted a socket.
|
|
IC15 pin 1 was bent underneath the chip and not making contact with the socket, the pin was too sharply bent to allow straightening so a new pin was overlaid. (This was A11 on the Address bus, we will see this again later))
|
IC12 pin 20 was actually missing, sheared off at the chip but also not even in the solder pad below, there was no evidence that this had been reworked so my hypothesis was that it left the factory this way. In a surprising way this pin was also A11 on the Address bus, what are the chances?
I removed the IC, ground back the chip to allow a new pin to be installed and tested, a new 6847 was to be fitted anyway but I was interested to see what affect the missing pin had, apparently none, it still booted to a blank screen! |
|
Checked voltage across the board, the power supply I used was a known good one anyway so barring a shorted tantalum I didn't expect any issues.
Dragon booted up with normal message then it went to an orange screen then a blank green screen...
Dragon booted up with normal message then it went to an orange screen then a blank green screen...
|
I used a scope and logic analyser on CPU, ROMs, RAM, in fact all the CPU chain showed nothing out of the ordinary, in actually looked like the boot sequence was looping at some point in the initialisation. Looking at a disassembly of the ROM, after the sign-on message is display the next part of the process is initialise the PIAs (6821s), the keyboard PIA (IC26) being first, the scope traces on this looked 'odd', so, de-soldered, socketed and a new 6821 fitted and as if by magic all was now well with the world. |
|
I then followed through with the full restoration:
Added the missing decouplers, I am not a fan of cost saving measures by leaving parts out of a design, and OK leaving them out may have also been a design decision it was possible to see a small but noticeable noise level difference on the power rails close to the missing components.
Cleaned all connectors and pots with De-Oxit
Full Electrolytic recap with Nichicon capacitors (bar one)
Full Tantalum capacitor recap
tested with a cartridge OK.
Still to do: test:
joystick port
cassette port
Added the missing decouplers, I am not a fan of cost saving measures by leaving parts out of a design, and OK leaving them out may have also been a design decision it was possible to see a small but noticeable noise level difference on the power rails close to the missing components.
Cleaned all connectors and pots with De-Oxit
Full Electrolytic recap with Nichicon capacitors (bar one)
Full Tantalum capacitor recap
tested with a cartridge OK.
Still to do: test:
joystick port
cassette port