Ok, my original 30GB drive crapped out and I have since put in a 60GB. I am running linux as my OS. This is what I get when trying to use upgclient:

./upgclient -p /dev/ttyUSB0 ../300662-car2-builder-20011022.upgrade
empeg-car Upgrade client.
empegupgrade: Using upgrade file '../300662-car2-builder-20011022.upgrade' to device '/dev/ttyUSB0' (115200 baud)
Checking upgrade file integrity [100%]
Pumping player software
Finding unit [ 4%]
Upgrading flash (bootloader)
Programming flash [ 0%]
ERROR: 9 Flash program failed
Upgrade aborted due to errors [ 0%]
Upgrade failed :-(

Upgrade failed with error 0x80040057


Then I tried the regular emplode software under XP in vmware. It takes about 45 minutes for it to finish flashing the kernel, reboots, then I get:

could'nt find pump" occured during stage 0x20

What am I doing wrong???

I know at this point I need to flash the builder kernel, pop in a HD, it formats (and populates the filesystem?) and then I might be able to copy over the files I was able to recover from my old drive. Then flash the latest empeg kernel (Beta 3?) and then hijack.

Is this the builder from http://www.empeg.com/support.html

I don't remember seeing the number 30062 in front of the builder file.

Also, stupid question, you're using a serial cable and not trying to do this through USB right?

If you used the hard disk builder file from the empeg site that is your problem, it is corrupt. Use Mark's bid disk builder that should solve the problem. Did anyone find out who needs to change the file on the empeg site?

I have used the building from the original site, the one someone posted here in the forums and Marks big disk builder. The whole time the HD is plugged in. These aren't hot-swappable HD's

All of that is fine. The problem doesn't seem to be the HD. I have plugged it into an adapter and into my pc and was able to manually set the partitions and format it but it still didn't like that for some reason. The problem it seems is not being able to flash it properly with any of the builder kernels. Flashing using windows takes about 45 minutes and with linux it's not even working(see above). Flashing using the latest empeg software or hijack works fine.

This is what I get after 45 minutes of it flashing the big disk kernel and then failing to "find pump" I doesn't format the drive at all.

Copying kernel...
Calling linux kernel...
Uncompressing Linux..................................... done, booting the kerne
l.
99) 16MB DRAM
Command line: mem=16m
Calibrating delay loop... 99) 16MB DRAM
Command line: mem=16m
Calibrating delay loop... 99) 16MB DRAM
Command line: mem=16m
Calibrating delay loop... 207.67 BogoMIPS
er cache hash table entriable entries: 2048 (order 2, 16k)
Buffrsity Computer Society NET3. TCP
TCP: Hash tables configured (ehash 16384tocols: ICMP, UDP, (irq = 16) is a SA110
0 UART
SA1100 UART
ttyS02 at 0xf8030000 (irq = 16) is a SA1100 UART
S97f50
e is 636f6972 'rioc'
Found custom animation at offset 0x97f50
Tuner: losed.
=0, ID=-1
Scheduling custom logo.
empeg display initialised.
empeg dsp audio initialised
empeg dsp mixer initialised
eminitialised, PDIUSBD12 id 1012
empeg state support initialised e089/88c1 (save to d0005100).
empeg RDS driver initialised
empitialized: 16 RAM disks alised (first boot
RAM disk driver in
Probing primary interface...
K size
empeg single channel IDE
Probing primary interface...
K size
empeg single channel IDE
Probing primary interface...
K size
empeg single channel IDE
Probing primary interface...
ide_data_ wrote 0x5555 read 0x5580
te 0xaaaa read 0xaa80
ide_data_test: wrote 0x5555 read 0x5580
te 0xaaaa read 0xaa80
ide_data_test: wrote 0x5555 read 0x5580
ide_data_test: wrote 0x0000 read 0xff80
idea read 0xaa80
ide_data_test: wrff80
ide_data_test: wrote 0xaaaa read 0xaa80
ide_data_test: wrote 0x5555 read 0x5580
ide_data_t0xff80
ide_data_test: wrote 0xaaaa read 0xaa80
ide_data_test: wrote 0x5555 read 0x5580
te 0xaaaa read 0xaa80
ide_data_test: wrote 0x5555 read 0x5580
ide_dread 0xff80
ide_data_test: wrof80
ide_data_test: wrote 0xffff est: wrote 0x5555 read 0st: wrote 0xaaaa read 0xaa80

ide_data_test: wrote 0x5555 read 0x5580
ide_data_test: wrote 0x0000 read 0xff80
ide_data_test: wrote 0xffff read 0xff80
ide_data_test: wrote 0xaaaa read 0xaa80
ide_data_test: wrote 0x5555 read 0x5580
ide_data_terote 0x5555 read 0x5555
te 0xaaaa read 0xbeaa
ide_data_test: wrote 0x5555 read 0x5555
te 0xaaaa read 0xbeaa
ide_data_test: wrote 0x5555 read 0x5555
hda: FUJITSU MHK2060AT, ATA DISK drive
ide0 at 0x00ache, CHS=12416/15/63
empeg-flash driver initialized
w/512kB Cid/revision 0x3349
smc9194.c:v0.12 03/06/96 by Erik Stahlman (eSK: ext2 filAC 00:02:d7:26:0b:b7
Partition check:
hda:
RAMDIs [1 disk] into ram disk... RAMDISK: Loading 320 blockdone.
EXT2-ram disk... ot fs on 03:05
uperblock
Kernel panic: VFS: Unable to mount root fs on 03:05
uperblock
Kernel panic: VFS: Unable to mount root fs on 03:05
uperblock
Kernel panic: VFS: Unable to mount root fs on 03:05

Mmmm.. the disk hasn't "been built" is what that is telling you.

This could fail for a number of reasons, mostly user-error, but sometimes a flakey USB-serial converter, or just the wrong version of the upgrader (windows only, the Linux version has been broken almost since the beginning), or the wrong disk builder (use the BIG DISK builder, even on small disks).

Cheers

USB-serial has always worked for all sorts of things. I'm able to use minicom and hyperterminal just fine as well as flash other kernels including hijack just fine.

I'm using the upgrader that comes with emplode 2.0. I just downloaded the emplode for v3 firmware but it is only a zip file with the emplode.exe and emplode.map. I have copied these to the rio/emplode directory and am running the upgrade again but I think it only uses the empegupgrade.exe and doesn't care what version emploded.exe you have.

I am using the big disk builder from http://rtr.ca/bigdisk/builder_bigdisk.upgrade

Same problem:

ide_data_test: wrote 0x5555 read 0x5d55
hda: FUJITSU MHK2060AT, ATA DISK drive
ide0:
RAMDISK: ext2 filAC 00:02:d7:26:0b:b7
Partition check:
hda20 blocks [1 disk] into ram disk... ck 0
RAMDISK: Loading 320 blocks [1 disk] into ram disk... done.
EXT2-fs warning: checktime reached, running e2fsck is recommended
VFS: Mounted root05: rw=0, want=2, limit=0
dev 03:0ess beyond end of device
03:el panic: VFS: Unable to mount root fs on 03:05
uperblock
Kernel panic: VFS: Unable to mount root fs on 03:05
uperblock
Kernel panic: VFS: Unable to mount root fs on 03:05
----------------------------------

I have been able to partiton and format this drive just fine as well as run an fsck on it with no errors.

By the way, unrelated to your issue, your cut-and-pastes from Hypeterminal are being mangled by the hyperterminal clipboard bug. Use the capture text feature instead. Described here.

I'll just use minicom next time

Tired of having to do all this through windows anyway.

correct. There are no partitions on it. Am I supposed to make them manually? I tried doing that following these directions but it didn't seem to work out right.

http://www.differentpla.net/content/2004/03/partition-empeg


Am I supposed to make the partitions manually then use the builder kernel to populate the filesystem?

I just spent the last hour writing a detailed post about every version of software I'm using and every little detail of every test i've performed. Unfortunately, I went to post it and expired and I lost it all. I don't have the patience to go through this again. Just know that I do this sort of thing for a living and I know enough to read through all posts on this forum regarding these issues as well as all the related empeg sites. I am using the latest version of everything including the epmeg software, Marks big disk kernel, i've tried the latest developer kernel as well as hijack. All of these flash fine except for the big disk kernel. Sorry, it flashes fine, but does not format or setup the hard drive like it is supposed to (pump error).

Is there anyone around the northern Boston area who might be able to help? I have tried everything. I just want the tunes back in my car.

Better yet .. .does anyone have a disk image or maybe a copy of the filesytem of the empeg? Not sure if a disk image will work for a different size HD (mine is a 60GB). If not, I could partition it manually and copy over the filesystem.

You can extract it yourself from the .upgrade file with UpgradeSplitter.

Edit: fixed URL. Left off the "http://" and the BBS munged it.

http://www.incase.de/files/empeg/myprogs/UpgradeSplitterBuilder.zip

Or did you mean to obfuscate that link?

Nope. Fixed. Thanks.

I agree with Stu's post about checking the EIDE header on the empeg. I've fixed several with cold solder joints after going through the same frustration you're seeing. Drive goes bad, replace drive, only to find out it's a bad cable or cold solder joints on the header that started the problems in the first place. If you have have solder skills, this is a quick job.

Another thing that has caused problems is using a USB to serial adapter. The driver tends to put the thing at COM 21 or something crazy, manually changing it to a lower COM port helps.

How do I work with these "chunk" files? I don't see any way of writing them to a partition or being able to mount them as a disk image. dd maybe?

Quote:
I agree with Stu's post about checking the EIDE header on the empeg. I've fixed several with cold solder joints after going through the same frustration you're seeing. Drive goes bad, replace drive, only to find out it's a bad cable or cold solder joints on the header that started the problems in the first place. If you have have solder skills, this is a quick job.

The original HD is definitely bad. Plugged into my pc I was getting IO errors and have the files/partitions were completely gone. The solder points on the board look fine as does the cable. I certainly won't rule this out and might reheat the points after a few more things to try out.

Quote:
Another thing that has caused problems is using a USB to serial adapter. The driver tends to put the thing at COM 21 or something crazy, manually changing it to a lower COM port helps.

The serial adapter was working fine for both minicom, cutecom and hyperterminal under vmware. It shows up as ttyUSB0 which I map to com1 under windows. I have been able to flash dev kernels and hijack with it just fine. The problem i've been seeing is the builder kernel not building the HD.