I'm trying to fix a problem with a Rio Central player. The player is stuck at boot with the Rio Central logo in the display. After a few seconds the player opens the CD tray and that's it - all further actions just close the tray and then reopen it.

My guess is that it is unable to bootstrap from the harddisk in the player (the original 40GB one). I've taken the harddisk out and mounted it in one of my machines - I was able to mount 2 of the 4 partitions, one cannot be mounted because it's swap and the remaining one isn't recognized by Knoppix.

I've also tried to boot from the 1.10 version of firmware CD, but the CD player also ejects this disc after some time.

I've managed to get the player 'working' once in the last couple of days (sometimes it works for a little while), but on inserting a keyboard and hitting Ctrl-Alt-Del it just said that libreadline.so.2 has an invalid ELF header.

I've checked this file on the Knoppix system, and it looks as if the file has a valid ELF header.

All symptoms lead me to believe that the problem is either a faulty harddisk or some problem on the IDE channel. Another point that puzzles me is that the massively cooled - possibly power regulator - on the board stays cool, which suprizes me because of the massive cooling body applied.

I hope somebody can give me some advice about this issue - I'd really like to get this machine working again!

Cheers,
Jeroen

Another thought hit me: is the 1.10 version on the internet a 'full' firmware replacement or should the 1.02 original firmware be installed first? Maybe the bottstrap is corrupted and the CD ejects kinda asking for the firmware, and that the system cannot start from the 1.10 disc?

PS: just tried with a fresh 80GB disc - same symptoms.

Jeroen

Okay - just checked the site and it appears that the version without CDDB can only be installed on a system that is already loaded with a previous version of the firmware. The complete version however does (still) not unzip. To make matters even worse I cannot find my original CD...

Would somebody be so kind as to create an ISO image of the original CD and put it up for download (or perhaps point me to a location where it is available)?

Tnaks in advance,
Jeroen

Sure, Randy (TheCentralGuy) just sent it to me, I'll rip an iso and put it up somewhere for you to download.

Matthew

Thanks Matthew - please let me know when it's ready.

Cheers,
Jeroen

It probably won't be till friday. I'll be in the office on friday and can just copy it over to the web server. Randy sent me both the orginal and the new CDs, I'll put both up for you to try.

Matthew

Okay - friday is fine - tnx!

Jeroen

Hey Matthew! How's it going?

So did the CDs I sent help to fix your issue with your Central? I hope they did...

Thanks, Randy

Shoot. Monday. For Sure.

Matthew

Hah - I thought you had a night shift

Cheers,
Jeroen

Pretty please...

Better late than never, I guess.

Anyone who wants it is welcome to download it here:
http://ih-ws1.berkeley.edu/~mkocher/hsx109/HSX109.iso
It's a good 600+ megs.

Has anyone looked into deciphering the cddb format? It'd be nice to substitute the current freedb file.

Matthew

I have briefly taken a look into the CDDB files that are on disc, but it isn't immediately obvious what the format is (at least, not to me). I've been having the same requirement since the player isn't hooked up to the inet.

It may be easier to just map the 'original' gracenote server to localhost via /etc/hosts and create a daemon that handles CDDB lookups via a local FreeDB database. If I have time I will put some effort into building this and share it here.

Jeroen

The freedb idea sounds like a good one. My parents central is online at the moment, so it's not much of a problem.

Did you get a chance to try out the image?

Matthew

Yeah - just tried the image and the Central just sits at its initial 'Rio Central' screen and spits out the CD after some time. I'll try and swap the CD-ROM drive to see if that is the cause of the problem and then probably try to work out how the serial connector on the board works so that I can hook up a terminal to see what happens at boot. I'm not very hopeful though...

Jeroen

Well, still no progress... The Central still only displays the 'Rio Digital Audio' screen and refuses to boot. Since I have got it working (although only briefly) I assume it might be a bad solder contact somewhere on the board. I have reapplied solder to both the IDE and front-panel connector but to no avail.

When I power on the Central I see some LED's flickering - D1 remains red and D3 seems to indicate activity on the IDE channel. Can somebody please confirm that this is how it works on a working Central - perhaps D1 is a fault indication...

Thanks in advance,
Jeroen

Hugo told me when i was visiting the guys in Cambridge some weeks ago that the serial connector is a "standard 3 V serial interface". You just need a buffer and check the wiring to connect to a Terminal. Output should be like on the empeg serial port.

About the change of the CD-drive he mentioned, that the drivers were especially written for that model to get the best ripping results. There are some workarounds for firmware-bugs of the drive included. But just for testing purposes it should work with any CD-drive, i think...

Ok, I've applied the upgrade to the central here, and it's pretty broken. I wouldn't reccomend anyone using the CD image I posted. It'll boot, and occasionally work for a while, but eventually the player software gets into a restarting loop and nothing seems to work besides flipping the power switch in back. My mother says it did the same thing with the version without CDDB loaded on it. Both installs of 1.10 have been fresh installs, the first a blank hard drive, the seccond a restore as new. Maybe an upgrade would be better?

Randy, have you used your 1.10 CD to upgrade sucessfully to a working player?

Unforunatly, I forgot the CDs Randy sent me (lucky I posted it online, eh?), so I can't check 1.02 CD unless someone can post the iso of it.

Matthew

I've used that image that I sent you on 4 Centrals, 3 are mine and 1 is my dad's...I have upgraded already running machines and I've also initialized hard drives from scratch since I have removable drive caddies on my 3 machines. There shouldn't be any issues with the image...

If your parent's Central acts the same way that it did before my image was installed, it doesn't sound like the image is the issue...

You can always try to install the software from the original CD I sent you and see if that makes any difference....

Randy

The connector marked 'debug' is probably the one you mean. It has 5 pins, so 1 too many for a standard GND/+/Rx/Tx connector. I've already measured that it supplies 3.3V, so conversion to TTL is necessary.

I still assume that the problem lies in the IDE bus. When fiddling with my fingers around the XTAL near the ARM CPU (yes, I know I shouldn't do that) I can get the screen to blank and the player to stop responding to anything. It at least tells me that the CPU is running and executing commands (well, ejecting the CD every time, but that's better than nothing).

I've already taken a more detailed look at the board and there are 3 3.3V/TTL buffers/line drivers between the CPU and the actual IDE bus. Maybe I'm lucky and it's just one of them that is behaving strange. I'll hoop up my scope and see what I can find.

Problem of course is how to replace them - I've done this sort of work before using a normal soldering iron, and although I manage I can tell you it's not a job for the fainthearted...

I suppose there's nobody out there that has the actual schematics of the board? They would really really be helpful right now...

Jeroen

Just as a reminder: You don't necessarily need the TTL serial connector if all you want is a shell prompt to poke aorund on. If you've got a USB keyboard, plug it in, and Ctrl-alt-del will get you a shell prompt right there on the Central's screen.

I don't know what stuff you could type at the shell prompt to help diagnose the issue, but at least you can get one.

I have tried using a USB keyboard as well, but I think that the player chokes in its POST. I think it has a problem finding a boot device and spits out the CD in order to be able to lboot from that. Hitting CAD on the USB keyboard doesn't do anything...

Jeroen

BTW, have you tried swapping out the IDE cable? I'm constantly killing IDE cables by pressing the IDE ribbon cables on sharp MB pins, etc.

http://ih-ws1.berkeley.edu/~mkocher/hsx109/HSX109-1.02.iso

The orginal CD. I havn't tried it in my parents' central yet. Hopefully things will improve.

Yep - tried swapping IDE cable as well - still no luck. Downloading 1.02 now and will try that...

Welll, burnt the 1.02 image using cdrecord and k3b (in case I somehow screwed up the cmd line parms). Same symptoms alas...

I must therefore conclude there's a hardware fault. But since I got the player to boot some time ago, and hooking up a keyb and hitting CAD gave me a 'invalid ELF header' on some shared object file (which if you hook the HD up to another machine gives you a perfectly valid ARM ELF binary) I think the problem is in the IDE bus (which also explains the system being unable to boot from the CD).

I think I will try and find somebody here in the Netherlands who has access to a logical analyzer - I'm pretty convinced that one of the 3.3/5V buffers is behaving awkward, but finding this using a plain scope will prove quite tricky.

Any advice is of course still appreciated.

And of course again, Matthew - thanks for putting the images online. Unfortunately it didn't help but it will help me further once the HW fault is isolated and fixed.

Jeroen

Well, I hope for everybody here that they do not have to rely on Rio Customer Support. I've sent them an email asking for assistance and they told me to press the RESET button on the player. I've asked them where it's located on a Rio Central and they are probably still looking - haven't heard from them since...

Meanwhile I've been doing some more investigation - I've drawn a schematic of the IDE bus part of the player and hooked up my scope. What I found is that one of the 3.3/5V transceivers on the board (LVX3245's) that provide the bridge between the 5V peripheral section and the 3.3V internal logic (which the ARM processor uses) is not behaving properly. Pin 4 of the middle transceiver (of which the B-side is hooked to DATA1 on the IDE bus) shows no signal when the device is transferring data. On any other DATA channel on the A-side I can see the bits streaming past. I can also see the bitstream on the B-side at DATA1 of the chip (pin 20).

On pin 4 I DO see data when the IDE interface is initializing and also see this data back a pin 20. Apparently the chip's B1->A1 logic is fried.

I'll get a new one and replace it later this week - I'll let you all know what the results are.

Jeroen