Quasi-locked 1.4GHz TBird?

[xcourse]

Purchased the above-mentioned cpu, installed it with a MC462A HS and 68cfm Delta on an A7V133. It booted fine and ran OK, although it didn't seem to process SETI@home as fast as my old 1100@1200(9x133). After it had run for a while without problems, I started running it faster. 11x133(1466) was no problem, but that was the highest multiplier it would take, regardless of FSB. If I entered 11.5 or 12, it still came up at 11. I checked the L1 bridges under magnifying glass and they're all connected and solid. So I did what any self-respecting overclocker would do - I went to the L6 bridges (which were all open to give it the 10.5 default multiplier) and closed the 2x and 4x contacts to make the default multiplier 12.5. Evidently that was a huge mistake; the cpu is now dead, near as I can figure.

Has anyone else come across this problem with accessible multipliers on this processor?

 

[xcourse]

bump for discussion. I'm really perplexed about this.

 

[Courtland]

I MIGHT of read this on AT few days ago. Are you sure you didn't crack the core removing the swiftech, thats what I did, lol.

 

[xcourse]

I looked carefully for any signs of damage to the core because that's the first thing I thought of also - all of the core's edges are clean and sharp. And everytime I removed the Swiftech, including several times from the 1100 TBird (which is still running fine, evidently), I backed the mounting screws out 1/2 turn at time, just like I install them.

I think that something inside the core didn't like my closing the 2x and 4x adders at the L6 bridges. It worked fine on this 1100; I did it on this cpu so that it would start up when mounted on the A7V133. But in that case, I lowered the default multiplier instead of trying to raise it. Perhaps there is something that should also have been done to the L3, L4, or L5 bridges in conjunction with modifying the L6 bridges. It seems that this was a case of too little knowledge being a bad thing.

Can anyone else add something - ideas or just advice not to tread where only fools go?

 

[nealh]

How did you close the L6 bridge ..may seem stupid but can you remove it....

 

[Garet Jax]

xcourse,

Did you modify the L3 and L4 bridges as well?

Typically L3, L4 and L6 all have to changed in unison. Refer to Tom's article here for more information.

Looking at the article myself, it looks like your L3 bridge is OK, but your L4 bridge is wrong. To change it, you will need to cut the 2 connected bridges and disconnect the 2 connected ones.

 

[xcourse]

I connected the 2x and 4x adders on the L6 bridges by using Loctite rear window defroster repair kit conductor paint. After applying it, I reinstalled the cpu, immediately realized it wasn't going to even post, pulled it back out and scraped off the paint. I re-reinstalled it, and still nothing at all happened. (That old sinking feeling in the pit of one's stomach...)

I'll read the article you point to, Garet Jax, and get back on after I've digested it for a few minutes. Maybe the 1400 can be resurrected, but I doubt it.

 

[xcourse]

Garet Jax

Thank you. I'll go ahead and try those extra mods since I have nothing to lose - the cpu is acting dead anyway, so if nothing else, it'll be good practice at steadying my hand!

Anyone else have any ideas on this? Especially puzzling is the fact that the L1 Bridges ARE connected - the cpu came that way - yet when running it was multiplier-locked at 10.5, with the gratuitous 11 thrown in just like the 13 was thrown in with some of the AXIA 1200's(200 FSB). Perhaps AMD made more than just surface changes when they produced the cpu's with five stepping digits (e.g. AYHJA, AXIAR, etc.). Maybe the L1 bridges are no more than red herrings now. "They're all connected, so the cpu must be unlocked."

 

[Pabster]

What motherboard?

 

[WarCon]

I unlocked my AXIAR chip just fine. Have you tried cleaning it with acetone to make sure it isn't still making contact in one of those bridges? It is an indent where the laser cut and it might still be full of defogger and just not visible.

Worth a look.

.........edit.........
You probably have done this already, but did you clear the cmos to make sure it just didn't like the setting? If you are using the keyboard reset(holding the insert key and turning it on) you should know that it doesn't always work.

 

[nealh]

I have and seen post where 1.4 200 fsb version is multiplier locked but the 266 are not..I am not sure how true this is

 

[Garet Jax]

xcourse,

Any update? I am interested in the overall outcome.

 

[Assimilator1]

xcourse
I'm also interested ,any news?

Inccidently....
although it didn't seem to process SETI@home as fast as my old 1100@1200(9x133).
FSB is more important than outright cpu MHz for SETI.BTW are you in a team?

 

[xcourse]

Whew! Lots of replies, and lots to reply to. Thanks to all.

Asus A7V133 motherboard, as stated in my first post. And, although I didn't say, I'm using an Enermax 431W power supply. Like it, too.

Garet Jax: I applied the mods to the L3, L4, and L6 bridges, but chickened out and decided to try a multiplier of 11.5 instead of 12.5. That would have been ridiculous of me, and I realized it eventually. (Even if it is an AYHJA 'y' chip.) I did the mods last night but didn't feel sharp enough by the time I finished to go ahead and install it. And today, a shooting match kept me busy all day. So it's still not done. I'll probably install and try it tomorrow, and I apologize for keeping you waiting.

Warcon: Yes, I tried both pressing the insert key and resetting the CMOS by switching off the power and shorting across the reset solder points with a screwdriver. Neither worked, and by that time I was getting REAL WORRIED.

The cpu is, per the retail box it came in, the 266MHz FSB version.

When I said that it seemed as if the 1100, running at 9x133, was processing Seti@home faster than the 1400, the 1400 was running at 11x133, or 1466. Even when I upped the FSB to 136, it still seemed doggie. And by the way, the A7V133 I'm using just plain won't go above 139 FSB. Not with the 1100 (which didn't surprise me that much because it is just a 200 FSB cpu, originally), and not with the 1400 (which at least has more tolerance built in with its 266FSB). And I know the limitation is just the motherboard because the memory is Mushkin Rev3 SDRAM, and the graphics card is a GForce3. It's just not a real OC-friendly motherboard, just like not all cpu's are created equal, even when it says they should be.

Again all, thanks for your interest, and I'll have more to post tomorrow after I've tried the cpu again. Please wish me luck.

 

[WarCon]

Good luck man.....

 

[Garet Jax]

xcourse,

No rush man...

Take your time and have fun doing it.

 

[Assimilator1]

xcourse
And by the way, the A7V133 I'm using just plain won't go above 139 FSB
Presumabley you tried a much lower multiplier to ensure the cpu wasn't the limiting factor?

Re SETI WU times ,if you only tried 1 WU on your 1.4 GHz it's probable that you got a VLAR WU ,these can take up to 50% longer than normal WU's!:Q.
Btw you didn't say if you were in a SETI team or not.

Good luck with your cpu & don't give up on it yet!

 

[xcourse]

To all: thanks for the "best wishes," but it now appears that my worst fears are true. Reinstalled the cpu, cleaned the base of the MC462A with acetone, applied a fresh coat of AS, mounted the HS, etc etc. End result was the same as before - nothing. I'll clean everything off again with my little diamond-tipped Dremel scraper, then use acetone to make sure, and give it one last try. I think something burped inside it though. If it is indeed a dead player, I'll hang it up as a monument to my hubris. Alas!

Re: Seti. I guess it's possible that the system was working on one of those more difficult WU's - I didn't realize there were different types. No, I'm not part of a team; there's just the computer I built for my wife (running a 700E@933 purchased from Compu-wiz), and the AMD-based machine I built for myself (that is now defunct, for a while). They are jointly sharing the same user name, "Get Sirius," and have completed 153 units so far since joining two months ago.

If I do get this processor going, however long it takes, I'll post that fact here, and what finally worked to get it going. Thanks again to all for your suggestions and comments.

 

[Assimilator1]

Xcourse
Sorry to hear that your cpu is dead ,that's really bad news

BTW you'd be welcomed to join the SETI Team Anandtech ,if you're interested check out my sig & have a scout around the Distributed Computing Forum section.