Well my player has just died on me. I was driving along listening to the radio when all of a sudden the display corrupted and dissapeared and the audio stopped. A reseat made no difference all I'm left with is a pulsing led.
Once I got home I connected up the serial port and got the initial message which gets as far as the 'let fly the penguins of linux' and then nothing. The real bummer is my toolkit is currently locked in a room I have no access to until Monday, all I can do it take the front and the top off.
Could a disconnected RAS wire stop the unit from booting?

Plus my myth box is playing up too. All in all a rather [censored] day.

OK, here I go again

Not sure whether a floating RAS connection on the chips can cause problems or not. When debugging the >32MB kernel issue on a 48MB player it seemed as though I needed to tie unused banks high to avoid them asserting data lines when they shouldn't. When I did my 64MB upgrade I stupidly forgot to upgrade my kernel first, resulting in my needing to deactivate two banks to get below 32MB so that I could upgrade, and in that case I didn't need to tie those two banks high. Go figure.

This non-boot problem can be caused by address bus errors preventing the flash from being read, so that's somewhere else to check should the problem be non-obvious.

Good luck.

And I thought that I had problems, commiserations!

Just got my spare toolkit from work so I've finally taken the thing apart and what greeted me? Disconnected RAS wire. Sadly I won't have access to my iron until tomorrow. But hopefully crisis averted. Genixia, any good tips for a solid connection to the RAS pin on the strongarm 'cos obviously my method wasn't good enough.

Also, I fixed my Mythbox too, bloody ITV changing their PIDs.

hot melt glue is your friend!

Vibration is your killer here - so you want to be able to prevent the wire vibrating, and your prevention needs to be as close to the joint as possible. I prefer to solder to the top of the pin and glue on top of the CPU. Either hot-glue or silicone glue will work - if you go for silicone, make sure you use an acid-free silicone such as GE Silicone II (not sure if it's available in the UK, check Homebase etc).

As for the soldering itself, don't forget to tin the wire first or you'll lose solder from the joint as it wicks up the wire. Make sure you also use some no-clean flux to ensure that the surfaces are clean. I _really_ like the no clean flux pens for this small work.

Glad to hear that it's not a head-scratcher.

Glad to hear that it's not a head-scratcher.

Me too, almost as soon as it happened I thought it would be the RAS wire, in my opinion it's always been the weakest point. I would assume the wire simply disconnecting and not touching anything is unlikely to have damaged anything else in the process. I hope, I'll have to pop down to our test room and have a go under the microscope again.

So do you think Servisol Silicone Adhesive Sealant will do the trick? It's apparently neutral cure and non-corrosive so it sounds like the stuff to use.

He's booting again, i ended up pulling the RAS pin up and out and soldering onto that. After testing I've covered all three wire joints in adhesive (nasty sticky stuff), I have to say the chip is a bit of a mess, but he seems to be booting fine.

Chip covered in snot....

neutral cure and non-corrosive

Sounds good. Now you see why I glue on the top of the chip!