Hot Memory and No Boot

Message

vlyne · #1 Post by **vlyne** » Tue Nov 14, 2006 9:42 pm

I’m in the process of resurrecting a supposedly non-booting T20 (700MHz, 128Mb, DVD…). It turned out to be booting but the HD sounded broken (rattling noise when flipped – broken heads?). I replaced the HD and got the Symantec Recovery Disk going but forgot about the 256Mb minimum RAM requirement so I stopped the process and installed an IBM 128Mb memory chip which resulted in that scary so-called “blink of death”. I quickly stopped the machine and found the IBM memory chip was very warm to the touch – much warmer than the original Toshiba chip (which actually was not even warm). The machine booted OK without the IBM chip. Next I extracted the Toshiba chip and put the IBM chip in its slot. I discharged my fingers of static and lightly touched the IBM chip whilst trying to boot up. Sure enough, the “blink of death”, and then the chip got warm very quickly, particularly the leftmost mounted chip. I quickly stopped the machine. I made sure the contacts were scrupulously clean and carefully inspected all contacts and all mounted chips. Tried again with the same result. Conclusion: The IBM chip is stuffed. One potential cause of the “blink of death” is a stuffed memory chip. For memory, the specs on the stuffed chip were:

Labelled: IBM 128Mb SDRAM FRU: 20L0265
Samsung sticky label: M464S1724BT1-L1L
Chip Marking 464S824CT1 SO-DIMM Rev 1.0

The good chip was a Toshiba THLY6416G1FG-80L

If you are going to try the empirical “touch of the finger” test, make sure you discharge yourself first

So the next task on my list is to try and find out what part of the memory circuit would break if I had left the machine to go well beyond the few seconds.

Cheers

#2 Post by **rkawakami** » Tue Nov 14, 2006 10:37 pm

I'll give this a try on my T22 when I get home tonight, but I don't recall any overheating memory module when it last "blinked" at me. I can't remember if it's even in any shape to be powered up at the moment. I think I pulled both batteries after re-assembling it a couple of weeks ago, but we'll see.

The "normal" blink-of-death symptom usually results in an inert piece of plastic/metal/inorganic mass exhibiting NO electron flow at all. I've held down the power button for many seconds (20-30) without detecting any noise, smell, themal activity or physical movement. No, I haven't licked the machine OR the problem yet

.

P.S. Sounds (?) like there's a short on the memory module... although I can't really hear that far away...

vlyne · #3 Post by **vlyne** » Tue Nov 14, 2006 10:54 pm

rkawakami wrote:I'll give this a try on my T22 when I get home tonight, but I don't recall any overheating memory module when it last "blinked" at me... The "normal" blink-of-death symptom usually results in an inert piece of plastic/metal/inorganic mass exhibiting NO electron flow at all.

I didn't get any dead memory beeps so it either passed the initial boot memory test or was waiting for the return signal. I guess one test which I didn't do was make sure the boot system did beep when there was no memory chips installed. I suspect some other fault in the memory circuit could cause the same "blink".

rkawakami wrote:P.S. Sounds (?) like there's a short on the memory module... although I can't really hear that far away...

I certainly felt it

There's still some resistance in the "short" and the system may have noticed a problem after a while but I wasn't going to let it go on till something broke.

I haven't licked the problem either. The closest I've come is that 9V battery trick to test if the battery is OK or not.

Cheers

#4 Post by **rkawakami** » Tue Nov 14, 2006 11:01 pm

vlyne wrote:I haven't licked the problem either. The closest I've come is that 9V battery trick to test if the battery is OK or not.
Cheers

Hah! 9V batteries are too easy... Try that trick with a "D" cell

.

vlyne · #5 Post by **vlyne** » Tue Nov 14, 2006 11:15 pm

rkawakami wrote:Hah! 9V batteries are too easy... Try that trick with a "D" cell .

I'll concede defeat and admit that trick would take some licking

#6 Post by **rkawakami** » Wed Nov 15, 2006 3:03 am

Okay, I know I should have expected this, but here is what happened when I tried to turn on my "blink-of-death" T22 tonight...

It started up.

Several times now.

About two dozen times, in fact.

It has NOT failed to start up properly. (Yeah, I know, a double-negative...)

No matter how quickly I cycle the power off and on.

Each time it does, the HDD LED flashes briefly (HD has been removed), then the other four LEDs all light up, then three of them go out and leave the power LED remaining lit. Just like it should.

It's running memtest86+ on two 128MB sticks put back in. I have it attached to a port replicator with an external floppy drive containing the memtest86+ boot disk. It has also turned on quite nicely OFF of the port replicator as well.

The system was sitting for at least two weeks with the main battery removed and the CMOS battery still connected. I know this because the RTC was still showing Daylight Savings Time which changed Oct 29th. The hard drive and Ultrabay drives were also removed (and remain so). I originally removed both 128MB memory modules (and forgot about that) so when I turned it on tonight, I was greeted with the 3-3-1 error. I turned it off, then turned it back on again. Same error code. Shut it down, put in the two modules and turned it on again. And again, and again, and again...

I have NOT yet plugged in the AC adapter, or the hard drive, or the Ultrabay drive. I am going to leave it this way for several days and see if it will still continue to start up.

The memory modules are only slightly warming up. Nothing out of the ordinary I would say. The system acts like nothing is wrong. I just had it shut off for a couple of minutes while I was typing this and I turned it on again. It shows the IBM splash screen (for quite some time; about 30 seconds), then gives me the white letters on red background message saying "Hard drive not installed...." for about 15 seconds, then boots the floppy drive. The BIOS boot order has been set to floppy, CD, then hard drive so I don't know why it's looking at the HD before booting the floppy. It also takes the same 30 seconds or so to access the BIOS, all the while displaying the "Entering IBM BIOS Setup Utility" message, then the "Hard drive not installed" message, THEN it finally shows the BIOS setup info. Boot mode is "Quick" and extended memory testing is disabled. BIOS revision is 1.08 (16ET28WW).

More to follow in a couple of days....

vlyne · #7 Post by **vlyne** » Wed Nov 15, 2006 4:20 am

Whoa! How inconvenient of the T22 to start working again!

Let's hope it doesn't get away!
Cheers

#8 Post by **rkawakami** » Thu Nov 16, 2006 3:14 am

Well, the T22 is back to normal again. By that I mean it doesn't start up! Yup, the blink-of-death has resumed. It started up fine tonight, several times in fact, and then I let it run memtest86+ for about 30 minutes (1st mistake). I turned it off (2nd mistake) and tried turning it back on. No go.

Blink.
Blink.
Blink.

<Sigh> I guess I'll leave it with the main battery installed for another two weeks and see what happens. In any event, this story is not over yet...

Oh yeah, the 256MB of memory wasn't that warm even after running memtest for 30 minutes. I'm pretty sure it was operating at 700Mhz. It's a 2648-V4U and since it's running on battery, the 900Mhz CPU is being throttled down by SpeedStep.

vlyne · #9 Post by **vlyne** » Thu Nov 16, 2006 7:35 am

rkawakami wrote:<Sigh>

Sigh indeed! Whatever it is seems to go " bad" once it's warmed up but doesn't affect a running system? It's only when you go to reboot it that it makes a difference. But then, you had it going many times yesterday... Makes no sense! Must be an intermittent contact that's wavering on the edge.

Time for my head to make contact with the bed instead of the brick wall

Cheers

Hot Memory and No Boot

Hot Memory and No Boot

Who is online