DivIDE and rule...

Got my DivIDE interface, complete with 1Gb CF card from the Netherlands (thanks Ben) and almost straight away (after playing Jet Set Willy, of course) decided that I wanted the more advanced functions and read/write capability of ESXDOS, rather than the standard FATWare.

Having spent a couple of hours fiddling around with it and FATWare with the help of Velesoft's Firmware Selector menu which should help me along 'till ESXDOS matures and gains more features.

Some of the things I'd like to see implemented in ESXDOS would be some better documentation, some kind of AUTOEXEC feature and, of course LFN support.

Nope, I was doing it right...

Seems that I was doing it right all along, and that my mini-monitor just doesn't like my +2B. Someone gave me an old 20" Samsung CRT TV and the modded composite output on that is crisp with just a little ghosting but flicker free.

This is what happens when you assume things I suppose...

DSCN5864, a photo by Rick Dickinson on Flickr.

You're doing it wrong...

Looks like I went about my +2a composite mod all the wrong way - a bit of reading later and I realised that the composite signal has to come directly from the TEA2000.

I soldered a length of solid core wire to the outgoing side of D31, thinking that the diode would still afford the TEA2000 some protection with the other end going to the socket inside the modulator case. At first I tried leaving the 100uf capacitor soldered in place from my earlier exploits, this gave me a distorted but recognisable signal for the first time, the image is still 'jumpy' and there is a fair bit of ghosting in evidence, perhaps this is because I chose not to bypass/remove D31 but I'll mess about with that later.

Found my 29 year old copy of Manic Miner, dropped into 48k mode and typed the immortal 'LOAD ""' followed by Enter and pressed play...

Spectrum +2 Has Arrived

Got my +2 today, need to composite mod it before I can test it further than plugging it in and seeing the power light come on. I did it quickly but something is amiss as there is no picture, maybe something is touching the metal of the keyboard shield. I'll spend a bit more time with it when the kids have gone to bed...

The continuity continues...

Half way through checking every pin of the CPU socket for bad joints, broken tracks etc. as per the Issue 4a/b Schematic at Retroisle - no problems so far apart from the following...

Either the schematic contains 2 glaring errors so far, or my brains do. CPU Pin 2 goes pin 12 of IC3 on the schematic but is connected to pin 2 on my board, as it does in the Issue 3b Schematic. Also, pin 5 of the CPU socket is shown as connected to pin 27 of the ULA in the 4a/b schematic, same in the 3b schematic, but on my board it connects to pin 37 on the ULA like in the Issue 5 Schematic.

I'm not going to mention it elsewhere yet in case I'm wrong, or have the short from hell as a result of my dodgy soldering, or both... :).

So, 20 pins (1 to 20) later and half my optimism gone, everything is as it should be so lets hope the next 20 reveal the problem - otherwise I'm stuck.

EDIT: It also looks like I've found another error, if I am right then this is the same for the 2b, 3b and 4a schematics - pins 4 & 5 of IC24 are reversed and only seems to be correct on the Swedish Issue 3 schematic I found.

Oops...

After the desoldering of the z80 went without incident, I expected the socketed replacement to go without incident too. I soldered the socket quickly and cleanly into place, replaced the CPU and hey presto...  a choice of either red, magenta or black border with flashing attributes.

Just in case I destroyed the original CPU during the desoldering process, I had already purchased a spare and fitting this gives me additional black and white bars. The internal speaker is buzzing and nothing is getting unduly hot, I have also checked fairly thoroughly for shorts and solder bridges.

I suspect the problem is due to a poor solder joint on the top side of the board, I'll check them thoroughly before crying for help. There is no response to the reset button and sometimes the pattern changes after being left powered on for a minute or two, hopefully I haven't damaged anything else while working.

EDIT: Information regarding this specific fault is contained in the Spectrum Repair Guide, looks like I need to check each and every individual PCB track going to the socket for continuity.