Tale of Woe - Failed mem upgrade | General

You are not logged in. [Log In] empegbbs.com » Forums » empeg-car » General » Tale of Woe - Failed mem upgrade

Topic Options

#273574 - 06/01/2006 12:50

Tale of Woe - Failed mem upgrade

bbowman
enthusiast

Registered: 12/05/2002
Posts: 205
Loc: Virginia, USA

Well, all, I regret to say that after attempting the mem upgrade, I have failed and also killed my empeg in the proces. I've been soldering for more than half my life, so I thought I would be up to the task - but I suppose that I wasn't. Here is my taile of woe and horror.

I followed the instructions to dissaemble the empeg and smoothed down the extra solder on the test pads where the memory board would go. I also ground down the tip of my soldering iron so that I would fit nicely into the holes.

After adding some matching everything up, I taped the board into place and began filling the wholes. Then I added the two leads onto the capacitors. I decided to test it on just the 48MB before doing the challengin 64 MB line and after flipping the dip switchs 1 and 4 to off, I plugged it in and booted.

It booted fine, but no extra DRAM:

Code:


empeg-car bootstrap v1.02 20001106 (hugo@empeg.com)

If there is anyone present who wants to upgrade the flash, let them 

speak now,

or forever hold their peace...it seems not. Let fly the Penguins of Linux!



e000 v1.04

Copying kernel...

Calling linux kernel...

Uncompressing Linux..................................... done, booting 

the ker.Linux version 2.2.17-rmk5-np17-empeg52-hijack-v444 

(hijack@rtr.ca) (gcc versio5

Processor: Intel StrongARM-1100 revision 11

Checking for extra DRAM:

c1000000: wrote ffffffff, read e91ba9f0

NetWinder Floating Point Emulator V0.94.1 (c) 1998 Corel Computer Corp.

empeg-car player (hardware revision 9, serial number 10101307) 16MB DRAM

Command line: mem=16m



etc...

So I rechecked the connections - made sure that I had continuity and heated up the poads under the cpu again.

This time:
Code:


Checking for extra DRAM:

c1000000: wrote ffffffff, read efffffff

NetWinder Floating Point Emulator V0.94.1 (c) 1998 Corel Computer Corp.

empeg-car player (hardware revision 9, serial number 10101307) 16MB DRAM

Command line: mem=16m

Calibrating delay loop... 207.67 BogoMIPS

the wrote/read was changing which lead me to believe that I just had a bad connection some where.
So I when through the connections again. to no avail.

I went through this scenario several times without success and began doing one contact at a time between reboots.

It go to the point that after one resoldering of a contact, the empeg would not boot up. The red LED would not light up either, but the blue led in the ethernet connection would light up.

So I believed that I had bridged something under the board, so I used the braid to clean out the holes and lifted the board. To my horror, the test pads came up with it. I disconnected everything and still the empeg will not even post.

I have included the horrific pictures of the aftermath and on one is an arrow to the last contact that I attempted to solder. It was after this one that the empeg stopped working.

Is there anyway I can get this fixed? maybe a jumper wire?

Here is the last full boot message I got out of it:

Code:

 

empeg-car bootstrap v1.02 20001106 (hugo@empeg.com)

If there is anyone present who wants to upgrade the flash, let them 

speak now,

or forever hold their peace...it seems not. Let fly the Penguins of Linux!



e000 v1.04

Copying kernel...

Calling linux kernel...

Uncompressing Linux..................................... done, booting 

the ker.Linux version 2.2.17-rmk5-np17-empeg52-hijack-v444 

(hijack@rtr.ca) (gcc versio5Processor: Intel StrongARM-1100 revision 11

Checking for extra DRAM:

c1000000: wrote ffffffff, read efffffff

NetWinder Floating Point Emulator V0.94.1 (c) 1998 Corel Computer Corp.

empeg-car player (hardware revision 9, serial number 10101307) 16MB DRAM

Command line: mem=16m

Calibrating delay loop... 207.67 BogoMIPS

Memory: 15008k/16M available (984k code, 20k reserved, 368k data, 4k init)

Dentry hash table entries: 2048 (order 2, 16k)

Buffer cache hash table entries: 16384 (order 4, 64k)

Page cache hash table entries: 4096 (order 2, 16k)

POSIX conformance testing by UNIFIX

Linux NET4.0 for Linux 2.2

Based upon Swansea University Computer Society NET3.039

NET4: Linux TCP/IP 1.0 for NET4.0

IP Protocols: ICMP, UDP, TCP

TCP: Hash tables configured (ehash 16384 bhash 16384)

IrDA (tm) Protocols for Linux-2.2 (Dag Brattli)

Starting kswapd v 1.5

SA1100 serial driver version 4.27 with no serial options enabled

ttyS00 at 0xf8010000 (irq = 15) is a SA1100 UART

ttyS01 at 0xf8050000 (irq = 17) is a SA1100 UART

ttyS02 at 0xf8030000 (irq = 16) is a SA1100 UART

Signature is 636f6972 'rioc'

Found custom animation at offset 0x9bb84

Tuner: loopback=0, ID=-1

Scheduling custom logo.

empeg display initialised.

empeg dsp audio initialised

empeg dsp mixer initialised

empeg dsp initialised

empeg audio-in initialised, CS4231A revision a0

empeg remote control/panel button initialised.

empeg usb initialised, PDIUSBD12 id 1012

empeg state support initialised 0089/88c1 (save to d0005000).

empeg RDS driver initialised

empeg power-pic driver initialised (first boot)

RAM disk driver initialized:  16 RAM disks of 4096K size

empeg single channel IDE

Probing primary interface.

edit : Fixed title
--l0ser

Attachments
273638-dsc02200_a.jpg (263 downloads)

Edited by l0ser (09/01/2006 20:25)

Top

#273575 - 06/01/2006 20:03 Re: Tale of Woe - Faiiled mem upgrade [Re: bbowman]
mlord carpal tunnel Registered: 29/08/2000 Posts: 14491 Loc: Canada	Don't Panic! This is fixable. I've done similar repairs to units in the past. But do see if you can get RobS to have a look at it -- he also can fix stuff like this, and is far more likely (odd as this may sound) to get 'round to it before I ever would. But it IS fixable. EDIT: A better (more clear, and brighter) photograph might help here Cheers Edited by mlord (06/01/2006 20:07)
Top

#273576 - 07/01/2006 13:44 Re: Tale of Woe - Faiiled mem upgrade [Re: mlord]
schofiel carpal tunnel Registered: 25/06/1999 Posts: 2993 Loc: Wareham, Dorset, UK	Agreed. you need a better macro photo of this. I am beginning to wonder if there is now going to be a flood of damaged empegs as this expansion board begins to arrive. Be aware people that this is a real precision job that may take more than a whole day to install. It is MOST DEFINITELY NOT for the inexperienced. I am reconsidering whether I want to offer a service to install this board. I will make a decision and repost soon. _________________________ One of the few remaining Mk1 owners... #00015
Top

#273577 - 07/01/2006 13:53 Re: Tale of Woe - Faiiled mem upgrade [Re: mlord]
bbowman enthusiast Registered: 12/05/2002 Posts: 205 Loc: Virginia, USA	Thank you for your comforting words. I have placed some more pictures on an external site: Pictures of board I realize that the memory upgrade is very unlikely at this point. I just want to get this thing working again. Thank you. _________________________ Brent RioCar MK][a 20GB+80GB '96 Saab 900s (Not any more) Still looking for a good way to install in a 2010 BMW 3 series with iDrive/NAV
Top

#273578 - 07/01/2006 14:01 Re: Tale of Woe - Faiiled mem upgrade [Re: schofiel]
maczrool pooh-bah Registered: 13/01/2002 Posts: 1649 Loc: Louisiana, USA	I agree; these boards are very tempermental and difficult to install. We install them, but it's not a particularly enjoyable task. Brent, your pictures are still pretty unclear, but it looks like you damaged an inner layer trace. Stu _________________________ If you want it to break, buy Sony!
Top

#273579 - 07/01/2006 18:02 Re: Tale of Woe - Faiiled mem upgrade [Re: bbowman]
mlord carpal tunnel Registered: 29/08/2000 Posts: 14491 Loc: Canada	The only photo there that is even close to being non-blurry is dsc02206.jpg Might as well delete the rest of them. I would suggest you try removing the two largish solder blobs from the bottom row of traces, and then look for broken/damaged traces from under where the memboard was placed. There's at least one broken trace visible in the least blurry photo. If that still doesn't help, then PUT THE SOLDERING GEAR AWAY!! You cannot do any more good without a circuit schematic or layout diagram (only RobS has access to those, and he's not allowed to share them). Or possibly a second empeg board to compare with could help -- preferably an unpopulated board. Again, RobS has those. Cheers
Top

#273580 - 07/01/2006 19:56 Re: Tale of Woe - Faiiled mem upgrade [Re: mlord]
schofiel carpal tunnel Registered: 25/06/1999 Posts: 2993 Loc: Wareham, Dorset, UK	I can see at least two shorts and a broken track. There's also a lot of very small blobs around the pads bottom left that really need to be removed. I really would like to see this under the microscope here to check. I'm sad this has gone wrong for you. It may well be that the board was simply defective and your work was good - I hope not. A lot of care went into assembling these boards. _________________________ One of the few remaining Mk1 owners... #00015
Top

#273581 - 08/01/2006 05:45 Re: Tale of Woe - Faiiled mem upgrade [Re: schofiel]
genixia Carpal Tunnel Registered: 08/02/2002 Posts: 3411	Quote: I can see at least two shorts and a broken track. ... It may well be that the board was simply defective and your work was good - I hope not. A lot of care went into assembling these boards. I don't know how you can see anything in those pics! IMO it's unlikely that the board was bad. The first log contains the classic result indicating bad RAS connections for both the high and low words chips. The later logs show evidence that the RAS connections were fixed. The final log suggested that D12 on the high word wasn't connected. _________________________ Mk2a 60GB Blue. Serial 030102962 sig.mp3: File Format not Valid.
Top

#273582 - 08/01/2006 13:17 Re: Tale of Woe - Faiiled mem upgrade [Re: genixia]
mlord carpal tunnel Registered: 29/08/2000 Posts: 14491 Loc: Canada	Quote: IMO it's unlikely that the board was bad. The first log contains the classic result indicating bad RAS connections for both the high and low words chips. The later logs show evidence that the RAS connections were fixed. The final log suggested that D12 on the high word wasn't connected. I concur. But I have had the odd bad RAM chip mess up an otherwise perfectly decent RAM upgrade.. (hi Taym!). Cheers
Top

#273583 - 08/01/2006 22:59 Re: Tale of Woe - Faiiled mem upgrade [Re: mlord]
bbowman enthusiast Registered: 12/05/2002 Posts: 205 Loc: Virginia, USA	Thanks, I have deleted those and uploaded a new set. I tell you, getting a good clear picture of this thing has been about as challenging as soldering this board! Updated pictures Thank you for that advice. I don't consider myself to be a soldering novice, but I'm not a pro either (even though I've done many soldering projects over the past 20 years). Believe me when I say that I am deeply embarrased that this has happenned, but for the love of my empeg, I am looking for a solution here. Does anybody have a good picture of the traces (particularly of the bottom row) that they could send me? Also, once I find out those traces, Is ther a good way for me to figure out what pins (where) are being connected so I could do a jumper somewhere on the top of the board. I suppose that I am the mercy of Rob on some of that. Rob? Any help would be greatly appreciated at this point. Brent _________________________ Brent RioCar MK][a 20GB+80GB '96 Saab 900s (Not any more) Still looking for a good way to install in a 2010 BMW 3 series with iDrive/NAV
Top

#273584 - 09/01/2006 12:21 Re: Tale of Woe - Faiiled mem upgrade [Re: bbowman]
mlord carpal tunnel Registered: 29/08/2000 Posts: 14491 Loc: Canada	Quote: Does anybody have a good picture of the traces (particularly of the bottom row) that they could send me? Attached -- taken with an old Nikon E990 3mp digicam, and a desk lamp. Cheers Attachments 273911-dscn2176.jpg (210 downloads)
Top

#273585 - 09/01/2006 12:32 Re: Tale of Woe - Faiiled mem upgrade [Re: mlord]
mlord carpal tunnel Registered: 29/08/2000 Posts: 14491 Loc: Canada	And here is a better close-up of the area in question, after I adjusted a few camera settings. Cheers Attachments 273913-segment.jpg (235 downloads)
Top

#273586 - 09/01/2006 19:10 Re: Tale of Woe - Faiiled mem upgrade [Re: bbowman]
gogee2000 new poster Registered: 21/09/2000 Posts: 40	Hi there, I don't have any advice, but I do share your feelings. I started an upgrade myself and stopped at the point where the install of the board would have taken place. Not feeling confident enough, I shipped it off to Eutronix and they were able to do it. There were problems though, but Eutronix did there absolute very best at getting my empeg back to life. I'm so very happy to have my empeg working again, even though I only have 16 of 64meg, but that's better than the 12meg :-) Gregory
Top

#273587 - 10/01/2006 01:15 Re: Tale of Woe - Faiiled mem upgrade [Re: mlord]
bbowman enthusiast Registered: 12/05/2002 Posts: 205 Loc: Virginia, USA	Thank you, Lord. Those are excelent pictures. I'll look through to make sure that I'm not shortng any of those traces. _________________________ Brent RioCar MK][a 20GB+80GB '96 Saab 900s (Not any more) Still looking for a good way to install in a 2010 BMW 3 series with iDrive/NAV
Top

#273588 - 14/01/2006 22:43 Re: Tale of Woe - Faiiled mem upgrade [Re: maczrool]
bbowman enthusiast Registered: 12/05/2002 Posts: 205 Loc: Virginia, USA	Yes, I do believe that an inner trace was damaged. Hopefully it was only one. But suspect that it was just one since at the end, I was doing only one contact at a time between tests before it died. I have repaired eight traces that were broken and with a little luck, I can find two connections somewhere on the board that would short the damaged inner trace. Any idea how I can find those? Since I have a good idea of under which test pad the damaged was done, I'm fairly confident of where it broke. It was bottom row, second from the left. Also, is it reasonable to assume that the small holes next to the test pads go down to at least the next layer down? Maybe I could connect a couple of those? Also, this may seem like a stupid question, but how many internatl layers are there? Is it just one or two? After fixing those other traces, I'm starting to get a little hope that I may be able to fix this thing. Either way, I hope that this thread can serve to reinforce how important it is to be careful when installing this mem board. _________________________ Brent RioCar MK][a 20GB+80GB '96 Saab 900s (Not any more) Still looking for a good way to install in a 2010 BMW 3 series with iDrive/NAV
Top

#273589 - 17/01/2006 12:37 Re: Tale of Woe - Faiiled mem upgrade [Re: bbowman]
bbowman enthusiast Registered: 12/05/2002 Posts: 205 Loc: Virginia, USA	Thinking back, Here is a method I would have used that would have made this install easier: Before putting on the memory board, I would have made a bunch of small, flat coils with a head using very small wire. Kind of like small dancng cobra snakes where the head would stick up just high enough to poke through the holes of the memory board. This way, I could be sure that my connection is good without having to get the heat down to the test pad post board placement. My problem what not being able to see below the board to know if the connection was being made. If I could see that pin (the cobra head) I would be almost certain that the connection was being made. If I can fix that inner layer trace, than I may proceed with that strategy (I beliieve to be my only choice other than to abandon the upgrade alltogether). Of course, the salvation of the empeg relies on finding where that broken inner trace needs to be shorted. _________________________ Brent RioCar MK][a 20GB+80GB '96 Saab 900s (Not any more) Still looking for a good way to install in a 2010 BMW 3 series with iDrive/NAV
Top

#273590 - 17/01/2006 15:56 Re: Tale of Woe - Faiiled mem upgrade [Re: bbowman]
larry818 old hand Registered: 01/10/2002 Posts: 1039 Loc: Fullerton, Calif.	Quote: If I can fix that inner layer trace, than I may proceed with that strategy (I beliieve to be my only choice other than to abandon the upgrade alltogether). Of course, the salvation of the empeg relies on finding where that broken inner trace needs to be shorted. I used to design the conveyor machines that replaced thru-hole plating for innerlayer connection. The stuff the conveyor machines uses is, essentially, carbon powder and corn starch. It forms a very low resistance resistor between layers. Two products I know of that do this are called "shadow" and "black hole". It might be worth a look.
Top

#273591 - 17/01/2006 19:19 Re: Tale of Woe - Faiiled mem upgrade [Re: bbowman]
peakmop journeyman Registered: 02/07/2004 Posts: 95 Loc: 384400 km from the Moon	I did several memory board upgrades without a hitch (knock on wood). A couple of suggestions: 1) Based on your pictures I see that pads adjacent to the CPU ones are not completely flat. That could create problems while soldering CPU pads. I absolutely flattened entire area where the memory board would sit on. It takes a while but is a good way to ensure excellent connections from the get go. 2) In order to make sure you placed your board right the first time is to put small amounts of solder in several spots across the area where the board would be placed. This way you create small semi-circular spheres that will lock your board in place instantly. A better way is to make these solder spheres for all of the memory board holes so you never have to excessively heat the underlying pad.
Top

#273592 - 17/01/2006 19:24 Re: Tale of Woe - Faiiled mem upgrade [Re: peakmop]
wfaulk carpal tunnel Registered: 25/12/2000 Posts: 16706 Loc: Raleigh, NC US	Here's an idea. Probably not a very good one. Put down those bits of solder and have a fine wire stick straight up out of them. Then you could thread the board onto those pieces of wire which would come loose when you soldered it down. Alternately, somehow make the wire itself solder (maybe a higher melting temp than the stuff used to tack it down?) so that you don't have to remove it at all. _________________________ Bitt Faulk
Top

#273593 - 17/01/2006 20:42 Re: Tale of Woe - Faiiled mem upgrade [Re: wfaulk]
schofiel carpal tunnel Registered: 25/06/1999 Posts: 2993 Loc: Wareham, Dorset, UK	I have been looking at snap-off via pins to solder onto the pads BEFORE the board is dropped over them. The pin tips then protrude through the memory board. This would then be a straightforward matter of soldering. However, there is the issue of clearance between the case and pin tips to address. WIP. I am becoming more convinced that the RAS connectivity failures observed are due to the DIP switches not having good connectivity after they were soldered on: as Patrick observes, I suspect you need to snap the switches back and forth a few times to clean them then test for continuity/lack of in both positions to ensure they are OK. _________________________ One of the few remaining Mk1 owners... #00015
Top

Previous Topic

Index

Next Topic