Thinkpads Forum

Hello,

I assembled a Frankenpad using a T61 (8896-2FU) motherboard and components in a T60p (2007-93U) UXGA case. The motherboard has an Intel GPU and a 2Ghz Core2Duo.

I have 2 4G RAM simms which I've used in my (unmodified) X61 Tablet for several weeks, and I put those in the Frankenpad. After a day or so I began to experience lockups with the machine.

I returned the simms to the X61 Tablet, and experienced lockups with that machine as well, and began to think the RAM was bad. I have since re-seated the simms, and the X61 is running solidly as before.

I put some other simms (1x1G each) in the Frankenpad and continued to have lockups, even after carefully re-seating the simms.

I'm using a Memtest CD to exercise the machine, and it locked up once so far after less than half an hour.

The latest experiment I'm trying is I've flashed the Frankenpad with the latest MIddleton (2.29-1.08) bios - and since that (over an hour) I've had no problems.

The donor T61 was dropped, and disassembled and (carelessly) re-assembled by the previous owner. I'm thinking that perhaps the bios was damaged (static?) or the bios was not a good match for the parts that went into the assembly?

When I was running the machine before the lockups with lm-sensors in Xubuntu, I noticed the two cores reported temperatures several degrees apart, like 44 & 52C. However, the fan did not run inordinately.

Does anyone have any experience or suggestions about this? I'll update if I get a lockup with the new bios, or if the memory test runs to completion with no errors.

UPDATE: After 104 minutes, no lockup. The only 800Mhz RAM I have is the 4G simms, which are in use in the X61 tablet now. The memory being checked is 667Mhz and I am getting lots of errors - but no lockup on the machine itself.

Thanks in advance,
Bret

bwaldow wrote:The memory being checked is 667Mhz and I am getting lots of errors - but no lockup on the machine itself.

Well, if you are getting even one error during the memory testing, then I wouldn't use that module(s). A lockup essentially means the same thing; there's a memory error critical enough to cause the system to hang, even before it gets reported by memtest.

rkawakami wrote:Well, if you are getting even one error during the memory testing, then I wouldn't use that module(s). A lockup essentially means the same thing; there's a memory error critical enough to cause the system to hang, even before it gets reported by memtest.

In fact, I would expect errors as I am using 667Mhz bus RAM in an 800Mhz bus system. I am more interested in whether the system locks or not, and any ideas anyone has about that.

I have the two 4G 800Mhz simms to try later, but they are in use in the X61 tablet now, so I'm just looking for continuing problems or not with the Frankenpad itself.

Regards,
Bret

1) My understanding is the T61 8896-2FU systems are specified to use PC2-5300, aka 667Mhz modules. This comes from the ltwbook.pdf document. How do you get 800Mhz?
2) Assuming that your system does require PC2-6400 (800Mhz) modules and you are plugging in slower ones, the system shouldn't even boot as the SPD information on the modules will be wrong (the missing or bad 1-3-3-1 beep error code on boot).
3) System lockups can be caused by memory errors (in addition to other things that can also freeze a computer); it just depends on where in the memory map the error(s) occur. Such lockups are therefore random. You could see many errors being reported by memtest86+ and never get a lockup OR you could see the diagnostic freeze (the WallTime clock stops) and no errors are displayed.
4) My guess is that one or both of your PC2-5300 modules are defective. Return them to the X61 (which is also spec'ed at 667Mhz), let it sit un-powered for an hour (to get as close to room temperature as possible) and then start up the memtest86+ diagnostic. Let it run for at least three complete passes through the all of the different patterns. If you get any errors or the system locks up, then test them separately to see which one is defective.
5) If you are overclocking the system, then all bets are off.

My apologies. You may be correct about the required speed.

I have 800Mhz PC2-6400 4Gbyte modules. I am ambiguous above as Crucial's on-line selector suggests T61s use either 667Mhz or 800Mhz modules. I don't know which models would want one or the other, or perhaps it's not that but rather that both can work.

Aside from the need for 800Mhz, do you have any knowledge of it being ok or a mistake to use them on a machine spec'ed for 677Mhz modules?

I did hear that 1-3-3-1 beep error code when I put a single module into memory slot 2 only on the X61. I have also heard a 1-4-1-1 code (several times) but have not found an explanation of what that means - under different circumstance, but with this machine - two modules together. Perhaps that was "poor seating".

I will try your suggestion 4, allowing the "big" modules to run overnight in the X61. In the meantime as well, I will also put my "best" modules (a matched set of 2G Kingston KFJ-FPC218/2G) into the Frankenpad to see if it behaves well with the modules well-seated. That may resolve my fears I do or don't need a new system board.

Most of my simms are mis-matched mongrels that came with the several T60s and T61 I bought preparing for this assembly, although at least one has an actual Lenovo label on it.

Either because these simms are defective, or because the Frankenpad is good and I need more, can you recommend a memory vendor? I don't know anyone that will guarantee this configuration as 8G in a T61 is not a factory recommendation, and a Frankenpad certainly isn't either.

Thanks for your help.

Regards,
Bret

1-3-3-1 beeps means missing or faulty RAM.
Eliminate that module and go on testing with other modules.

I put my "best" (matched Kingston) simms into the Frankenpad and got lockups and video artefacts - vibrating vertical thin lines through the bios screen. I think the Frankenpad has an ex-system board, and I'll look to get another one.

I know how to do it now, anyway.

The X61 is cooling, and I'll put the 4G simms in it and go through the testing regime mentioned above. I'll also put my Kingston simms into my T60p to get the benefit of them in the meantime.

Thanks for the advice.

Regards,
Bret

clean the contacts of the chips with alcohol, and you may want to polish them with a pencil eraser too, then clean with alcohol again. Also look at the board carefully under magnification, see if you can find and corrosion. A tooth brush can be used to clean the simm slots.

If you do find any corrosion on the board, as a last chance before trashing the board you can try some baking soda and a toothbrush to clean it up. You may need some water but it must be distilled. Water is usually something you don't want near electronics, so this is only after you've accepted that it's the boards last chance. Steam is good for cleaning too, but you must let the board dry thoroughly before powering up. I'd recommend a day or two in a hot and dry location.

After cleaning all connectors, reseat the ram, cpu and all connectors and you may get lucky.


Good Luck

Thanks for the suggestions.

I followed rkawakami's advice, and ran the simms through several rounds of the Memory Test with no flaws found at all.

The simms are currently working in the X16 tablet with no apparent problems. I found other obvious problems with the Frankenpad, and I consider the motherboard flawed now. I am looking to source another.

Regards,
Bret

With the replacement motherboard, now the Frankenpad is smooth and runs cool. I pushed the CPU to 100% and the cores went to 92C & 88C and not above after 15 minutes. It idles under 40C in both cores (X-Window but no programs running).

I am considering putting a T500 HSF and perhaps a T7800 Merom CPU. And 8G of RAM, of course. I'll burn it in for a little while longer before I do any of that.

Thanks to the forum for all the good advice.

Regards,
Bret

Have the same problem. Got some BSOD's with 2x4 Gb samsung 333 Mhz PC-2 5300 (frankenpad).
Perfom memtest86 with each stick. One of them found 2 errors during 8 passes later (I succesfully perfom with this stick those operations that initiate BSOD with 2 modules), another over 2500 errors during 2 passes. I have test another 3 modules (1 Gb each) on the same system and found no errors during 5 passes for each stick.
Does this approve that reason of BSOD in RAM?