Xpantastic!

Hey there,

My xpander - which I got last year - has started doing odd behaviour, and I just want to float it past you guys to see if it is what I hope it is; or if not, help to narrow it down.

I was using my xpander yesterday, generating kick-ass sounds as normal. Then after about 20 minutes, suddenly the patch name display just filled up with asterisks. The sound engine still worked - and I was able to continue editing my sound. I then went to change the patch at which point the whole thing hung. No sound being produced and the buttons did nothing. I then rebooted it. The thing sprung to life as per normal but no sound was coming out.

I then turned it off and unplugged from mains and left it a while. A couple of hours later I replugged it and turned it on. Booted up but still no sound. After a hunch, I went into the voices page only to find all the voices were set to off -- I turned them on and the xpander was working as it should.

Again, I played with it for about 40 minutes (really addictive synth that it is), and just as I was about to turn it all off, the displays just went blank and the leds went on solidly. I power cycled it, and this time it didn't boot - just stuck with the leds on and no display.

I turned it off and left it about a minute; rebooted and it displayed the message of Resetting the Memory (patches won't be effected). I pressed yes, only for it to again lock with all the LEDs on and no display.

I again unplugged and left it a couple of hours. Coming back, it again showed the Reset the memory display, but this time it then booted. After that, I turned it off and called it a night.

I'm hoping this is a battery issue -- it looks like it's using the original battery and surely that thing has to be dying by now. I opened it up and the caps all look fine. Checking the battery voltage (albeit in circuit), it's at 3.2v - which seems fine, though that is not under load. Reading online, it looks like when the battery dies, it can do all kinds of odd things to the synth; though checking the schematics I'm not totally sure of how this would affect the system once it's booted (and has been running successfully for half an hour or so).

I've ordered a new battery and will check the voltages with the system on later to see what's happening there (I might as well scope it, so I can also check for ripple).

So does this sound like the likely cause? The only other thing I can think of (beyond the diodes around the battery and 5v line dying) is the psu giving spikes or drop outs -- which I'm really hoping it isn't. Mainly as the caps in those things can be pretty nasty. However, it is a Japanese unit and they seem to be known for having shitty PSUs. If it does end up being that (ugh), would just replacing the electrolytics be fine, or should I look at replacing the power supply entirely with a more modern design?

Cheers,

deano

Hopefully it is. Though looking at the schematics, I'm not totally sure why the battery failing would quite cause this.

I could see that with it failing, some of the RAM bits would become corrupted; and so I would expect either the patches to go crazy, or the system just not booting up properly. However, it is mostly booting up fine and only failing after about half an hour. In the last case, it failed whilst I was just playing some midi notes on a custom sound (ie, not saved to patch memory). So that sounds not totally like it would be the fault of the RAM bits becoming corrupted whilst the machine is off. My only guess looking at the schematics, are the failing of battery causing the system to pull a bit too much on the +5v power line and get an undercurrent situation which then affects all the digital chips. Or maybe the RAM bits which are being corrupted are the settings memory but they're corrupting in a way that only manifests itself in rare occasions?

Regardless, I'll replace the battery when that arrives tomorrow since there's nothing to lose by doing so (particularly being a thirty year old battery...); but I will also try and check the following:

1 - The diodes in the battery to +5m line are working correctly (if the one by the battery is dead, then the +5v power line would be running through there - increasing the current usage which can't be good for it)
2 - Check the voltage level of the +5m line both with the system on and off
3 - Scope out the +5v line, and make sure it's relatively clean

My main reluctance to do these checks is just paranoia of the system dying from whatever is up with it (a minor paranoia), and zapping myself on the incredibly exposed psu board that is right below the processor board.

I'm pondering whether it might also be worthwhile to just replace the psu board with something more modern. However, since it requires a whole gamut of voltages (9v ac, +/-5v, +/-12v and that annoying +55v) it doesn9;t look like there's an off-the-shelf unit I can grab; but instead would have to get a basic switched mode power supply and then recreate the output transformer circuit from the xpander to get all those voltages. A lot of work, and really would need me to create a proper circuit board --- massively ambitious to the point of being impractical.

Of course, if there's an easier way or someone else has done the footwork for this, that would be amazing.


But enough rambling - let's hope the battery is the blame and the psu is still going strong.